EPA-600/3-77-001b
  January 1977                            Ecological Research Series
                           OF
                       OF
                   METEOROLOGY
                  INTERNATIONAL CONFERENCE
                  ON PHOTOCHEMICAL OXIDANT
                  POLLUTION AND ITS  CONTROL
                         Proceedings:  Volume II
                              Environmental Sciences Research Laboratory
                                  Office of Research and Development
                                  U.S. Environmental Protection Agency
                             Research Triangle Park, North Carolina 27711
:.*«

-------
                 RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection  Agency, have  been grouped into five series. These live broad
 categories  were established to facilitate further development and application of
 environmental technology.  Elimination of traditional grouping was consciously
 planned to  foster technology transfer and a maximum interface in related fields.
 The five series are:

     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies

 This report has been assigned to the ECOLOGICAL RESEARCH series. This series
 describes  research on the effects of pollution on humans, plant and animal
 species, and materials. Problems  are assessed for their long-  and short-term
 influences.  Investigations include formation, transport, and pathway studies to
 determine the fate of pollutants and their effects. This work provides the technical
 basis for setting standards to minimize undesirable changes in living organisms
 in the aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                   EPA-600/3-77~001b
                                                   January 1977
                 INTERNATIONAL CONFERENCE
                            ON
              PHOTOCHEMICAL OXIDANT POLLUTION
                      AND ITS CONTROL
                  Proceedings:  Volume II
Hosted by the United States Environmental Protection Agency
                   September 12-17, 1976
                  Raleigh, North Carolina

         Coordinated by the Triangle Universities
                Consortium on Air Pollution

                   with the patronage of
   Organization for Economic Cooperation and Development
                         Edited by

                     Basil Dimitriades
        Environmental Sciences Research Laboratory
           Research Triangle Park, N. C.   27711
        Environmental Sciences Research Laboratory
            Office of Research and Development
           U. S. Environmental Protection Agency
      Research Triangle Park, North Carolina   27711

-------
                                   PREFACE

     Since the late 1940's when the first studies of atmospheric pollution
began, photochemical smog has been evolving into a full-fledged problem with
an enormously strong and complex impact upon man's lifestyle.  Smog is no
longer a subject of mainly academic interest.  Its presence and consequences
are now fully felt by the layman and perhaps most painfully by the automobile
user.  The use of automobiles is being restricted in the large urban centers
in the U. S., and even more restrictive measures including gasoline rationing
have been considered.  The merits of the economic incentives traditionally
associated with industrial growth are now being seriously questioned, and
pressures grow stronger for a review and a more realistic appraisal of the
environmental, energy, and industrial growth priorities.  Such a strong and
multifaceted impact of the photochemical pollution problem makes it imperative
that all judgment regarding the causes, occurrence, and solution or alleviation
of the problem be made responsibly and with the highest degree of confidence.
Thus, while it is extremely important that the health and welfare of the people
be protected, it is equally important to ascertain that such protection is
really achieved and that the health danger is not traded for other equally
bad or worse problems.

     Intensive studies conducted in the past 3-4 years have resulted in an
abundance of suggestive evidence that in part supported and in part refuted
the earlier understanding, but did not resolve all existing issues.  Thus, the
international scientific community is still divided on the issue of the justi-
fication of the 0.08-ppm ambient air quality standard for oxidant, one objection
arising from the questionable achievability of such a standard.  A newly revived
issue of major importance pertains to the relative roles of the hydrocarbon
and nitrogen oxide precursors in the urban and rural oxidant formation processes.
The viewpoint of the U.S. Environmental Protection Agency supporting maximum
control of the hydrocarbon and limited control of the N0x emissions is challenged
in several Conference papers.  More specific issues raised by the new evidence,
and debated in the Conference, are the achievability of the ambient oxidant
standard, the role of stratospheric ozone, the role of the natural ozone
precursors, the utility of current air quality simulation models, and the
significance of short- and long-range photochemical pollution transport.

     Aside from the debate on the above issues, the Conference will gather and
bring into focus the latest developments in the areas of physical-chemical
research methodology, biological effects of oxidants, and emission control
methods.  It was felt that inclusion of such widely diverse subject areas
would be essential for examining the balance, rationality, and effectiveness
of the entire oxidant control effort.

-------
     The Conference is expected to be extremely valuable to the U.S.  Environ-
mental Protection Agency in that it will provide a forum for presenting the
Agency's evolving viewpoint on oxidant control strategies, and for sounding out
scientific receptivity to this viewpoint.  The Conference is of value also to
the international community of scientists and government administrators in
that it will provide an opportunity for comparing the problems, experiences,
and control policies of one country with those of others.  Considering the now
established international range of pollution transport, and the implications
of oxidant-related control upon international trade, such interaction among
countries is more than justified.   For these reasons the organization of the
Conference was undertaken by the U.S. Environmental Protection Agency, with
the patronage of the Organization for Economic Cooperation and Development.
                                        B. Dimitriades
                                      IV

-------
                                   CONTENTS


PREFACE [[[     iii

ACKNOWLEDGMENT [[[     xv
SESSION 1 - ANALYTICAL METHODS FOR OXIDANTS AND .......................       1
            PRECURSORS - I
                       K.H. Becker
 1-1      METHODOLOGY FOR STANDARDIZATION OF ATMOSPHERIC ..............       3
          OZONE MEASUREMENTS
          J.A. Hodgeson, E.E. Hughes, W.P. Schmidt, and
          A.M. Bass

 1 -2      ULTRAVIOLET PHOTOMETER FOR OZONE CALIBRATION ................      13
          A.M. Bass, A.E. Ledford, Jr., and J.K. Whittaker

 1 -3      HYDROCARBON AND HALOCARBON MEASUREMENTS :  ....................      19
          SAMPLING AND ANALYSIS PROCEDURES
          R.B. Denyszyn, L.T. Hackworth, P.M.  Grohse, and
          D.E. Wagoner


SESSION 2 - ANALYTICAL METHODS FOR OXIDANTS AND .......................      29
            PRECURSORS - II
            ChcuAman:  K.H. Becker

 2-1      A NEW CHEMILUMINESCENT OLEFIN DETECTOR FOR ..................      31
          AMBIENT AIR
          K.H. Becker, U. Schurath, and A. Wiese

 2-2      GC-CHEMILUMINESCENCE METHOD FOR THE ANALYSIS OF .............      41
          AMBIENT TERPENES
          R.L. Seila

 2-3      MEASUREMENTS OF SULFATE, INORGANIC GASEOUS NITRATE ..........      51
          AND OTHER CONSTITUENTS IN THE ATMOSPHERE
          T. Okita


-------
 2-5      STATUS OF CALIBRATION METHODS FOR OZONE MONITORS ............     67
          R.J. Paur, R.K. Stevens, and D.L. Flamm


SESSION 3 - SOURCES OF TROPOSPHERIC OZONE - I ..........................     73
                       R.A. Rasmus sen
 3-1      METEOROLOGICAL CONDITIONS CONDUCIVE TO HIGH LEVELS ... ........     75
          OF OZONE
          T.R. Karl and G.A. DeMarrais

 3-2      OZONE IN RURAL AND URBAN AREAS OF NEW YORK STATE - ..........     89
          PART I
          P.Coffey, W. Stasiuk, and V. Mohnen

 3-3      OZONE MEASUREMENT AND METEOROLOGICAL ANALYSIS OF ............     97
          TROPOPAUSE FOLDING
          V.A. Mohnen, A. Hogan, E. Daniel sen, and P. Coffey

 3-4      METEOROLOGICAL FACTORS CONTROLLING PHOTOCHEMICAL ............    109
          POLLUTANTS IN SOUTHEASTERN NEW ENGLAND
          R.A. Dobbins, J.L. Nolan, J.P. Okolowicz, and
          A.J. Gilbert
SESSION 4 - SOURCES OF TROPOSPHERIC OZONE - II ............... ..........    119
                       R.A. Rasmussen
 4-1      AN ASSESSMENT OF THE CONTINENTAL LOWER TROPOSPHERIC .........    121
          OZONE BUDGET
          R. Chatfield and R.A. Rasmussen

 4-2      URBAN KINETIC CHEMISTRY UNDER ALTERED SOURCE CONDITIONS .....    137
          L.A. Farrow, I.E. Graedel , and T.A. Weber

 4-4      THE EFFECT OF OZONE LAYERS ALOFT ON SURFACE .................    145
          CONCENTRATIONS
          T.N. Jerskey, T.B. Smith, and W.H. White


SESSION 5 - SOURCES OF TROPOSPHERIC OZONE - III .................. ......    155
            ChcuAman:  R.A. Rasmussen

 5-1      OZONE CONCENTRATIONS IN POWER PLANT PLUMES : .................    1 57
          COMPARISON OF MODELS AND SAMPLING DATA
          T.W. Tesche, J.A. Ogren, and D.L. Blumenthal

 5-2      OZONE AND NITROGEN OXIDES IN POWER PLANT PLUMES .............    173
          D. Hegg, P.V. Hobbs, L. Radke, and H. Harrison

-------
 5-3      THE ANALYSIS OF GROUND-LEVEL OZONE DATA FROM ................     185
          NEW JERSEY, NEW YORK, CONNECTICUT, AND MASSACHUSETTS:
          DATA QUALITY ASSESSMENT AND TEMPORAL AND
          GEOGRAPHICAL PROPERTIES
          W.S. Cleveland, B. Kleiner, J.E. McRae, and R.E. Pasceri

 5-4      CHEMICAL AND METEOROLOGICAL ANALYSIS OF THE MESOSCALE .......     197
          VARIABILITY OF OZONE CONCENTRATIONS OVER A SIX-DAY
          PERIOD
          W.D. Bach, Jr., J.E. Sickles, II, R. Denyszyn, and
          IW.C. Eaton

 5-5      OZONE AND HYDROCARBON MEASUREMENTS IN RECENT OXIDANT ........     211
          TRANSPORT STUDIES
          W.A. Lonneman


SESSION 6 - OZONE/OXIDANT TRANSPORT - I ...............................     225
            ChaMman:  A. P. Altshuller

 6-1      TRANSPORT OF OZONE BY UPPER-LEVEL LAND BREEZE - AN ..........     227
          EXAMPLE OF A CITY'S POLLUTED WAKE UPWIND FROM ITS CENTER
          E.K. Kauper and B.L. Niemann

 6-2      OZONE FORMATION IN THE ST. LOUIS PLUME ......................     237
          W.H. White, D.L. Blumenthal, J.A. Anderson, R.B. Husar,
          and W.E. Wilson, Jr.

 6-3      LONG RANGE AIRBORNE MEASUREMENTS OF OZONE OFF THE ...........     249
          COAST OF THE NORTHEASTERN UNITED STATES
          G.W. Siple, C.K. Fitzsimmons, K.F. Zeller, and
          R.B. Evans

 6-4      AIRBORNE MEASUREMENTS OF PRIMARY AND SECONDARY ..............     259
          POLLUTANT CONCENTRATIONS IN THE ST. LOUIS URBAN PLUME
          N.E. Hester, R.B. Evans, F.G. Johnson, and
          E.L. Martinez

 6-5      OZONE IN HAZY AIR MASSES ....................................     275
          R.B, Husar, D.E. Patterson, C.C. Paley and
          N.V. Gillani


SESSION 7 - OZONE/OXIDANT TRANSPORT - II ..............................     283
                       A. P. Altshuller
 7-1      THE TRANSPORT OF PHOTOCHEMICAL SMOG ACROSS THE ..............    285
          SYDNEY BASIN
          R. Hyde and G.S. Hawke

 7-2      OXIDANT LEVELS IN ALBERTA AIRSHEDS ..........................    299
          H.S. Sandhu

                                      vi i

-------
 7-3      AN INVESTIGATION OF LONG-RANGE TRANSPORT OF 	    307
          OZONE ACROSS THE MIDWESTERN AND EASTERN
          UNITED STATES
          G.T. Wolff, P.J. Lioy, G.D. Wright, R.E. Meyers,
          and R.T. Cederwall

 7-4      OXIDANT AND PRECURSOR TRANSPORT SIMULATION IN 	    319
          THE RESEARCH TRIANGLE INSTITUTE SMOG CHAMBERS
          J.E. Sickles, II, L.A. Ripperton, and W.C. Eaton

 7-5      OZONE EPISODES ON THE SWEDISH WEST COAST 	    329
          P. Grennfelt
SESSION 8 - IMPACT OF STRATOSPHERIC OZONE 	    339
            CkcUxman:  R. Guicherit

 8-1      OZONE OBSERVATIONS IN AND AROUND A MIDWESTERN 	    341
          METROPOLITAN AREA.
          G. Huffman, G. Haering, R. Bourke, P. Cook, and M. Sillars

 8-2      A "TEXAS SIZE" OZONE EPISODE TRACKED TO ITS SOURCE 	    353
          J.W.  Hathorn, III and H.M. Walker

 8-3      APPLICATION OF 1960's OZONE SOUNDING INFORMATION 	    381
          TO 1970's SURFACE OZONE STUDIES
          P.R.  Sticksel

 8-4      THE ROLE OF STRATOSPHERIC IMPORT ON TROPOSPrfERIC	    393
          OZONE CONCENTRATIONS
          E.R.  Reiter
SESSION 9 - THEORIES ON RURAL OZONE/OXIDATES 	    411
            ChcuAmzm  B. Dimitriades

 9-1      RESEARCH TRIANGLE INSTITUTE STUDIES OF HIGH OZONE 	    413
          CONCENTRATIONS IN NONURBAN AREAS
          L.A. Ripperton, J.J.B. Worth, F.M. Vukovich,
          and C.E. Decker

 9-2      IMPORTANT FACTORS AFFECTING RURAL OZONE CONCENTRATION 	    425
          F.L. Ludwig, W.B. Johnson, R.E. Ruff, and
          H.B. Singh

 9-3      A MECHANISM ACCOUNTING FOR THE PRODUCTION OF OZONE IN 	    439
          RURAL POLLUTED ATMOSPHERES
          M. Antell

 9-4      NET OZONE FORMATION IN RURAL ATMOSPHERES 	    451
          T.Y. Chang and B. Weinstock
                                     vm

-------
 9-5      THE KINETIC OZONE PHOTOCHEMISTRY OF NATURAL 	    467
          AND PERTURBED NONURBAN TROPOSPHERES
          T.E. Graedel and D.L. Allara
SESSION 10 - PHYSIOLOGICAL EFFECTS OF OXIDANTS - I 	    475
             Ckotixman:  J. Knelson

10-1      ON THE RELATIONSHIP OF SUBJECTIVE SYMPTOMS TO 	    477
          PHOTOCHEMICAL OXIDANTS
          I. Mizoguchi, K. Makino, S. Kudou and R. Mikami

10-2      EFFECTS OF OZONE PLUS MODERATE EXERCISE ON PULMONARY 	    495
          FUNCTION IN HEALTHY YOUNG MEN
          B. Ketcham, S. Lassiter, E. Haak, and J.H. Knelson

10-3      EFFECTS OF OZONE AND NITROGEN DIOXIDE EXPOSURE OF 	    505
          RABBITS ON THE BINDING OF AUTOLOGOUS RED CELLS TO
          ALVEOLAR MACROPHAGES
          J.G. Hadley, D.E. Gardner, D.L. Coffin, and D.B. Menzel

10-4      RELATIONSHIPS BETWEEN NITROGEN DIOXIDE CONCENTRATION	    513
          TIME, AND LEVEL OF EFFECT USING AN ANIMAL INFECTIVITY
          MODEL
          D.E. Gardner, F.J. Miller, E.J. Blommer, and
          D.L. Coffin

10-5      DEVELOPMENT OF OZONE TOLERANCE IN MAN 	    527
          M. Hazucha, C. Parent, and D.V. Bates


SESSION 11 - PHYSIOLOGICAL EFFECTS OF OXIDANTS - II 	    543
             CkcuAman:  J. Knelson

11-1      TOXIC INHALATION OF NITROGEN DIOXIDE IN CANINES 	    545
          T.L. Guidotti and A.A. Liebow

11-2      THE EFFECT OF OZONE ON THE VISUAL EVOKED POTENTIAL 	    555
          OF THE RAT SUPERIOR COLLICULUS AND VISUAL CORTEX
          B.W. Berney, R.S. Dyer, and Z. Annau

11-3      HEALTH EFFECTS OF SHORT-TERM EXPOSURES TO N02-03 	    565
          MIXTURES
          E. Ehrlich, J.C. Findlay, J.D. Fenters, and
          D.E. Gardner

11-4      BIOCHEMICAL INDICES OF NITROGEN DIOXIDE INTOXICATION 	    577
          OF GUINEA PIGS FOLLOWING LOW LEVEL-LONG TERM EXPOSURE
          B. Menzel, M.B. Abou-Donia, C.R. Roe, R. Ehrlich,
          D.E. Gardner, and D.L. Coffin
                                      IX

-------
1 1 -5      BENEFIT EFFECTIVE OXIDANT CONTROL ...........................     589
          R.A. Bradley, M. Dole, Jr., W. Schink, and S.  Storelli


SESSION 12 - EFFECTS OF OXIDANTS ON VEGETATION - I .....................     599
                        W.W. Heck
12-1      OZONE INDUCED ALTERATIONS IN PLANT GROWTH AND ...............    601
          METABOLISM
          D.T. Tingey

12-2      OXIDANT LEVELS ON REMOTE MOUNTAINOUS AREAS OF ..... ...........    611
          SOUTHWESTERN VIRGINIA AND THEIR EFFECTS ON
          NATIVE WHITE PINE (PINUS STROBUS L.)
          E.M. Hayes, J.M. Skelly, and C.F. Croghan

12-3      THE EFFECTS OF OZONE ON PLANT-PARASITIC NEMOTODES ...........    621
          AND CERTAIN PLANT MICROORGANISM INTERACTIONS
          D.E. Weber
SESSION 1 3 - EFFECTS OF OXIDANTS ON VEGETATION - II ...................    633
             Ckcuiftman:  W.W. Heck

13-1      GROWTH RESPONSE OF CONIFER SEEDLINGS TO LOW OZONE ..... .......    635
          CONCENTRATIONS
          R.G. Wilhour and G.E. Neely

13-2      MACROSCOPIC RESPONSE OF THREE PLANT "SPECIES" TO . . ..........    647
          OZONE, PAN, OR OZONE + PAN
          D.D. Davis and R.J. Kohut

13-3      RELATIVE SENSITIVITY OF EIGHTEEN HYBRID COMBINATIONS ........    655
          OF PINUS TAEDA L. TO OZONE
          L.W. Kress and J.M. Skelly

13-4      EFFECTS OF OZONE AND SULFUR DIOXIDE SINGLY AND IN ... ........    663
          COMBINATION ON YIELD, QUALITY, AND N-FIXATION OF
          ALFALFA
          G.E. Neely, D.T. Tingey and R.G. Wilhour


SESSION 14 - REACTIVITY AND ITS USE IN OXIDANT-RELATED CONTROL ........    675
                        J.G. Calvert
14-1      MULTIDAY IRRADIATION OF NOX ORGANIC MIXTURES ......... ........    677
          W.A. Glasson and P.H. Wendschuh

14-2      HYDROCARBON REACTIVITY AND THE ROLE OF HYDROCARBONS .........    687
          OXIDES OF NITROGEN AND AGED SMOG IN THE PRODUCTION
          OF PHOTOCHEMICAL OXIDANTS
          J.N. Pitts, A.M. Winer, G.J. Doyle, K.R. Darnall and A.C. Lloyd

-------
14-3      APPLICATION OF REACTIVITY CRITERIA IN OXIDANT- 	     705
          RELATED EMISSION CONTROL IN THE USA
          B. Dimitriades and S.B. Joshi

14-4      PHOTOCHEMICAL REACTIVITY CLASSIFICATION OF 	     713
          HYDROCARBONS AND OTHER ORGANIC COMPOUNDS
          F.F. Farley


SESSION 15 - ATMOSPHERIC CHEMISTRY AND PHYSICS 	     727
             Chairman:  J.G. Calvert

15-1      DECOMPOSITION OF CHLORINATED HYDROCARBONS UNDER 	     729
          SIMULATED ATMOSPHERIC CONDITIONS
          F. Korte and H. Parlar

15-2      PHOTOOXIDATION OF THE TOLUENE-N02-02-N2 SYSTEM IN A 	     737
          SMALL SMOG CHAMBER
          H. Akimoto, M. Hoshino, G. Inoue, M. Okuda , and
          N. Washida

15-3      THE CHEMISTRY OF NATURALLY EMITTED HYDROCARBONS 	     745
          B.W. Gay, Jr. and R.R. Arnts

15-4      MEASUREMENT OF PHOTONS INVOLVED IN PHOTOCHEMICAL 	     753
          OXIDANT FORMATION
          D.H. Stedman, R.B. Harvey,  and  R.R.  Dickerson

15-5      ACTIVE SOLAR FLUX AND PHOTOLYTIC RATE IN THE 	     763
          TROPOSPHERE
          J.T. Peterson, K.L. Demerjian, and K.L. Schere


SESSION 16 - MATHEMATICAL MODELS OF OZONE/OXIDANT AIR QUALITY - I ....     775
             ChjcuAman:  K. Demerjian

16-1      PHOTOCHEMICAL AIR QUALITY SIMULATION MODELLING: 	     777
          CURRENT STATUS AND FUTURE PROSPECTS
          K.L. Demerjian

16-2      THE SYSTEMS APPLICATIONS, INCORPORATED URBAN AIRSHED 	     795
          MODEL:  AN OVERVIEW OF RECENT DEVELOPMENTAL WORK
          S.D. Reynolds


SESSION 17 - MATHEMATICAL MODELS OF OZONE/OXIDANT AIR QUALITY - II ...     803
             CfwuArnan:  K. Demerjian

17-1      A SURVEY OF APPLICATIONS OF PHOTOCHEMICAL MODELS 	     805
          J.E. Summerhays

-------
17-2      TESTS OF THE DIFKIN PHOTOCHEMICAL DIFFUSION MODEL ...........     817
          USING LOS ANGELES REACTIVE POLLUTANT PROGRAM DATA
          G.H. Taylor and A.Q. Eschenroeder

17-3      DEVELOPMENT OF A MARKET CHAIN MODEL FOR PHOTOCHEMICAL ......     827
          OXIDANT PREDICTION
          J.R. Martinez

17-4      A PRELIMINARY INVESTIGATION OF THE EFFECTIVENESS OF AIR ....     837
          POLLUTION EMERGENCY PLANS
          W.F. Dabberdt and H.B. Singh


SESSION 18 - OXIDANT-PRECURSOR RELATIONSHIPS AND THEIR INTERPRETATION .    849
             IN TERMS OF OPTIMUM STRATEGY FOR OXIDANT CONTROL - I
                        J.N. Pitts
18-1      A "J" RELATIONSHIP FOR TEXAS ................................    851
          H.M. Walker

1 8-2      AN ALTERNATIVE TO THE APPENDIX- J METHOD .....................    871
          FOR CALCULATING OXIDANT- AND N02- RELATED CONTROL
          REQUIREMENTS
          B. Dimitriades

18-3      COMBINED USE OF MODELING TECHNIQUES AND SMOG CHAMBER .........    881
          DATA TO DERIVE OZONE-PRECURSOR RELATIONSHIPS
          M.C. Dodge

1 8-4      OUTDOOR SMOG CHAMBER STUDIES :   EFFECT OF DIURNAL ............    891
          LIGHT DILUTION AND CONTINUOUS  EMISSION ON OXIDANT
          PRECURSOR RELATIONSHIPS
          H.E. Jeffries, R. Kamens, D.L.  Fox, and B.  Dimitriades

18-5      USE OF TRAJECTORY ANALYSIS FOR DETERMINING EMPIRICAL ........    903
          RELATIONSHIPS AMONG AMBIENT OZONE LEVELS AND METEOROLOGICAL
          AND EMISSION VARIABLES
          E.L. Meyer, Jr., W.D. Freas, III, J.E. Summerhays .,  and
          P.L. Youngblood


SESSION 19 - OXIDANT-PRECURSOR RELATIONSHIPS AND THEIR .................    915
             INTERPRETATION IN TERMS OF  OPTIMUM STRATEGY FOR
             OXIDANT CONTROL - II
             Chcuxman:  E.L. Meyer

19-1      REPORT ON OXIDANTS AND THEIR PRECURSORS IN CANADA ...........    917
          L. Shenfeld

19-2      PRECURSOR CONCENTRATION AND OXIDANT FORMATION IN SYDNEY .....    927
          G.H. Allen, K. Post, B.S. Haynes, and R.W.  Bilger

-------
19-3      SMOG POTENTIAL OF AMBIENT AIR, SAMPLED AT DELFT	      943
          NETHERLANDS:  THE EFFECT OF INCREASING NOX CONCENTRATION
          0. van Ham, and H. Nieboer

19-4      SUMMARY:  A PRELIMINARY INVESTIGATION OF EXPECTED 	      955
          VISIBILITY IMPROVEMENTS IN THE LOS ANGELES BASIN FROM
          OXIDANT PRECURSOR GASES AND PARTICULATE EMISSION CONTROLS
          C.S. Burton, T.N. Jerskey, and S.D. Reynolds


SESSION 20 - CONTROL OF OXIDANT PRECURSOR EMISSIONS - I 	      969
             Chavman'-  R.W. Bilger

20-1      TRAFFIC MANAGEMENT AS A MEANS OF OXIDANT PRECURSOR 	      971
          CONTROL FROM LOW-SPEED SATURATED TRAFFIC
          R.B. Hamilton and H.C. Watson

20-3      CONTROL OF VEHICLE REFUELING EMISSIONS 	      989
          A.M. Hochhauser and L.S. Bernstein
SESSION 21 - CONTROL OF OXIDANT PRECURSOR EMISSIONS - II 	     1001
             Ctau/uncm:  R.W. Bilger

21-1      NOX CONTROL TECHNOLOGY FOR STATIONARY SOURCES 	     1003
          G.B. Martin and J.S. Bowen

21-2      EMISSION ESTIMATES OF N02 AND ORGANIC COMPOUNDS FROM 	     1015
          COAL-FIRED BED COMBUSTION
          P.E. Fennelly

21-3      EMISSIONS ASSESSMENT OF THE CHEMICALLY ACTIVE FLUID 	     1025
          BED (CAFB) PROCESS
          A.S. Werner, R.M. Bradway, D.F. Durocher, S.L. Rakes,
          and R.M. Statnick

21-4      NOX CONTROL BY ABSORPTION 	     1035
          G. Sakash

21-5      DEVELOPMENT OF A LOW EMISSIONS PROCESS FOR ETHYLENE 	     1039
          DICHLORIDE PRODUCTION
          W.S. Amato, B. Bandyopadhyay, B.E.  Kurtz, and R.H.  Fitch


SESSION 22 - AIR QUALITY AND EMISSION TRENDS  	     1051
             ChdLnmani  R.E. Neligan

22-1      THE IMPACT OF EMISSIONS CONTROL TECHNOLOGY ON 	     1053
          PASSENGER CAR HYDROCARBON EMISSION  RATES AND PATTERNS
          F. Black
                                     xm

-------
22-2      CONTROL OF OXIDANTS IN SYDNEY 	     1069
          D. Iverach and M.G. Mowle

22-3      OXIDANT AND PRECURSOR TRENDS IN THE METROPOLITAN 	     1077
          LOS ANGELES REGION
          J. Trijonis, T. Peng, G.  McRae, and L.  Lees

22-4      COMPARISON OF PAST AND PROJECTED TRENDS IN OXIDANT 	     1095
          CONCENTRATIONS AND HYDROCARBON EMISSIONS
          R.M. Angus, E.W. Finke, and J.H. Wilson

22-5      TRENDS IN AMBIENT LEVELS OF OXIDANT AND THEIR POSSIBLE 	     1103
          UNDERLYING EXPLANATIONS
          E.L. Martinez, N.C. Possiel, E.L. Meyer, L.G. Wayne,
          K.W. Wilson, and C.L. Boyd


SESSION 23 - ON THE OZONE/OXIDANT CONTROL STRATEGY IN U.S	     1113
             Chapman:  A.B. Bromley

23-1      TRENDS IN PHOTOCHEMICAL OXIDANT CONTROL STRATEGY 	     1115
          J. Padgett

23-2      PROBLEMS WITH CONVERTING STATE-OF-THE-ART PHOTOCHEMISTRY ...     1123
          TO STATE LEVEL CONTROL STRATEGIES
          W. Bonta and J. Paisie

23-3      OXIDANT CONTROL UNDER SECTION 110 OF THE CLEAN AIR ACT 	     1135
          J.L. Pearson

23-4      OXIDANT CONTROL STRATEGY:  RECENT DEVELOPMENTS	     1143
          B. Dimitriades

23-5      CONTROL REGULATIONS FOR STATIONARY SOURCES OF HYDROCARBONS .     1155
          IN THE UNITED STATES
          R.T. Walsh

          PREPARED COMMENTS ON JAPANESE PHOTOCHEMICAL AIR QUALITY ....     1167
          STANDARDS AND CONTROL STRATEGIES
          Professor R. Kiyoura
                                     xiv

-------
                                ACKNOWLEDGMENT

     The assistance of the Conference Program Committee members in organizing
the technical program of the Conference is gratefully acknowledged.  Program
Committee members were:  Dr. A.P. Altshuller, EPA, USA (Chairman); Dr.  K.H.
Becker, U. Bonn, Germany; Dr. R.W. Bilger, U. Sydney, Australia; Dr.  J.C.
Calvert, Ohio State U., USA; Dr. B. Dimitriades, EPA, USA; Mr.  M.  Feldstein,
San Francisco, Calif., USA; Mr. D.R. Goodwin, EPA, USA; Dr. R.  Guicherit,
IGTNO, Netherlands; Dr. A.B. Bromley, OECD, France; Dr. M. Hashimoto, Environ-
ment Agency, Japan; Dr. J. Knelson, EPA, USA; Mr. R.E. Neligan, EPA,  USA;
Mr. J. Padgett, EPA, USA; Dr. J.M. Pitts, U. California, USA; Dr.  R.A.  Rasmussen,
Washington State U., USA; and Mr. J.O. Smith, EPA, USA.

     Conference arrangements were made by the Triangle Universities Consortium
on Air Pollution, under the direction of Dr. Laurence Kornreich, through
Research Grant #800916-0451.
                                      xv

-------
                                                                          11-4
           BIOCHEMICAL INDICES OF NITROGEN DIOXIDE INTOXICATION OF
             GUINEA PIGS FOLLOWING LOW LEVEL-LONG TERM EXPOSURE

           D. B. Menzel, M. B. Abou-Donia, C. R. Roe, R. Ehrlich,
                      D. E. Gardner, and D.  L.  Coffin*
ABSTRACT

     In on e^ort to determine po**ibte enzymatic indicator* o& nitrogen di-
oxide.  (N02) damage, and to evatuate protective mechanism*, lung and red btood
ceti*  (RBC) gtutatkione. peroxida*e (GSHP), £ang and pta*ma ty*ozyme  (LVZ),
tung and pta*ma acid pho*phata*e  (AP) and RBC acetytchotine*teraAe (CHE)
tevelA Mere determined in guinea pig* a^ter 7 day* or 4 month* expo*ure to 0.5
ppm W02.  Short-teAm expo*ure depre**ed RBC GSHP and etevated lung LVZ and RBC
CHE.   Long-teAm expoAure depreAAed RBC GSHP, pta*ma LVZ, AP and CHE but ete-
      tung AP.  Ve.prt**ion o^ RBC GSHP and tack ofi induction o{, tung GSHP *ug-
          enzyme. i* not protective. againAt NO2-  Vamage re*pon*e* dominate
           ixpo*uAc* OA Ahown by the reteaAe o& creatine pko*pkokina*e iAo-
            long-term expo*ure* may represent atteration* in celt population*.
Long-term e^ect* on tung LVZ and RBC GSHP may repreAent di^erence* ^rom
ozone  (03) intoxication.  Evaluation cfa protective agent* and indicator* ofa
intoxication require chronic expo*ureA a* *hort-term expo*ure* are not
equivalent.


                                INTRODUCTION

     As evidenced by this Symposium, exposure to oxidant gases continues to be
a pressing human health problem.  The development of tests for early damage
from inhaled pollutants remains an elusive and critical aspect of this problem.
Discovery of sensitive indicators of pollutant damage could result in better
estimates of the health hazards of exposure and potentially indicate the
underlying basic mechanism of intoxication.

     Both nitrogen dioxide (N02) and ozone (03) catalyze the peroxidation of
unsaturated fatty acids found in biological membranes (1).  Peroxidation is
highly damaging to the function of cells and is a potential common mechanism
of intoxication for oxidant gases.  Glutathione peroxidase (GSHP) may repre-
sent a detoxification mechanism for lipid peroxides (2), and is elevated in
*D.  B. Menzel, M. B. Abou-Donia, C. R. Rose, Duke Medical Center, Durham,
 North Carolina.  R. Ehrlich, IIT Research Institute, Chicago, Illinois.
 D.  E. Gardner, D.  L. Coffin, U.S. Environmental Protection Agency, Research
 Triangle Park, North Carolina.
                                     577

-------
rats on short term (7 day) exposure to ozone,  but not nitrogen dioxide (3).
Creatine phosphokinase (CPK), lactate dehydrogenase (LDH), serum glutamic
oxalic transaminase (SGOT), and serum glutamic' pyruvic transaminase (SGPT) are
released from heart,  liver, and muscle on damage and may indicate damage to
lung and other tissues following N02 inhalation.  Lysozyme is likely to be
released into the serum from damaged alveolar macrophages rich in this enzyme
(4,5), while acid phosphatase is an enzyme found in lung lysosomal fractions
capable of tissue digestion (4).  Another index of potential  distal injury on
N02 inhalation is plasma cholinesterase (6).

     Since most of these indicator enzymes have been evaluated only on short-
term exposures of 7 days or less to oxidant gases, the response to a near am-
bient level of N02 (0.5 ppm) was examined in  both short-term (7 days) and
long-term (4 months)  exposure.


                            METHODS AND MATERIALS

     Guinea pigs of the Hartley strain (250-300g, Murphy Breeding Labs) were
exposed to 0.50 +. 0.04 ppm of N02 for 8 hours  per day for 7 days or for 4
months as previously described (7).  At the end of the appropriate time period
the animals were exsanguinated by cardiac puncture, and the organs removed and
frozen on dry ice.  The organs were kept frozen until analysis.  Blood was
collected with 10"4M EDTA to prevent coagulation, and red blood cells (RBC)
and plasma were obtained by centrifugation.  A comparable control group was
exposed to charcoal filtered air and sacrificed in an identical manner.   All
of the animals were randomized from the same  group of animals obtained from
the supplier.

     Lungs were homogenized in 9 ml of isotonic phosphate buffer for each gram
of tissue.  The homogenate was centrifuged at 100,000 g for 1 hour and the
supernatant, soluble fraction was used for the measurement of GSH peroxidase,
acid phosphatase, lysozyme, and protein (8).   GSH peroxidase was determined
using cumene hydroperoxide as the substrate and coupled with added GSH reduc-
tase at pH 7.6 (2).  Acid phosphatase was determined by the method of Bessey
et al. using jD-nitrophenyl phosphate as a substrate in 300 rnM acetate buffer
at pH 5.0 (9).  Lysozyme was determined in the soluble fraction and plasma by
the Worthington (Freehold, New Jersey) assay  kit procedure, which utilized a
turbidometric determination of the digestion  of Micrococcus lysodeikticus at
pH 6.2.  Results are reported as equivalent amounts of crystalline egg while
lysozyme (10).

     Plasma cholinesterase was determined as  described before (11), using di-
butyryl thiocholine as the substrate at pH 7.5.   RBC GSHP levels were deter-
mined using hemolysates and the conditions described above.  Plasma enzymes,
lactic dehydrogenase (LDH), creatine phosphokinase (CPK), glutamic oxalic
transaminase (SGOT),  and glutamic pyruvate transaminase (SGPT) were determined
as previously described (12).  Gel electrophoresis was used to separate the
isomeric forms of LDH and CPK.

     Student's "t" test was used to compare the means calculated from the
individual values.  The numbers of animals examined in each assay are shown in
the figures.  Each value represented an average of three assays per animal.

                                     578

-------
                           RESULTS AND DISCUSSION

     Many of the studies of the biological effects of oxidants have focused on
short-term exposure on the assumption that little or no change in the response
of the animal occured on continued exposure.   It is now clear that exposure to
oxidants results in a complex, time-dependent response on the part of the
animal.  The short-term effects observed in our studies are contrasted to
those observed after prolonged exposure.

     The exposure schedule selected for these experiments was an 8-hour expo-
sure per day for either 7 days or for 4 months.   Guinea pigs were randomized
and selected from the same group that was ultimately exposed for 4 months, and
the 7-day exposed animals represented a subsample of the original group of
animals.

     Short-term exposure is characterized by indices of tissue damage.  An
elevation of the serum enzymes LDH, CPK, SGOT, and SGPT was observed (Figure
1).  The elevations of enzymic activity were variable with only the total CPK
levels being marginally significant statistically.  Because of the large vari-
ability of the enzyme levels in the N02-exposed group, these enzymes are of
little value as specific indicators of N02 damage.  The elevation of the CPK
and LDH activity was examined in detail by gel electrophoresis.   No specific
tissue isoenzyme pattern could be discerned and, therefore, probably repre-
sents a non-specific type of cell damage.

     Short-term exposure to N02 also resulted in a marked elevation of the
plasma cholinesterase (Figure 2), while long-term exposure resulted in a de-
pression of this enzyme.  CHE levels are most indicative of hepatic and myo-
cardial disease.  CHE is elevated in hemochromatosis, but depression of plasma
CHE is most often associated with active heptocellular disease (6,13).  In-
creased plasma levels of CHE are seen with cardiac surgery but depressed
values  are  consistently  associated with  myocardial  infarction  (13).   Aside
from secondary metastatic carcinoma of the liver, pulmonary disease, including
pneumonia, is not associated with changes in CHE levels.  Most likely the
alterations observed, especially the depression of CHE levels on long-term
exposure, are related to hepatic injury.

     Both ozone and N02 catalize the peroxidation of unsaturated fatty acids:

                            OH

         RCH = CHR + 03 -* RCHOOH + RCHCHR

                                     OOH

                              N02

         RCH - CHR + N0? -> R-CHCH2R + RCHCHR

                                        NOH


                                    579

-------
o_
IJ3
oo
                                                        H
                                                    'ON
                                                                    r-o,
           CD
           CD
     CD
     CD
                                                QJ
                                                E
                                                >,
                                                M
                                                Zl
                                                O)

                                                E
                                     OJ
                                     (/I
1
1
IPrvu
^ON
                               |--|
                                        LPi
                                        CD

                                        CD

                                         A

                                         Q.

                                        *
            i
            cr>
            CD
 I
CD
CD
CD
CD
CD
CD
CNJ
CD
CD
                                                                                  O

                                                                                  CD
                                                'ON
                                                                                  (/I
                                                                                  O
                                                                                  CL
                                                                                  X
                                                                                  
-------
        12-1
        10-
         8-
         6-
     UJ
     _l
     o

     3-
         2-
7 DAYS
MONTHS
                              «••**
         CM
        O
                              (33)
                                                      tt:
                                                                  #**
                                                                   CM
                                                                  O
                             (34)
       (33)
              ***
                  P  > 0,001
           Figure 2.  Effect of N0? exposure on plasma cholinesterase.
We have suggested that peroxidation  of unsaturated  lipids is a common mechanism
of toxic action of both of these gases (1).   Another indicator of RBC damage,
Heinz body formation,  has been observed with  fatty  acid ozonides and ozone
exposure (14).   Ingestion of vitamin E by  both mice and men results in protec-
tion of circulating RBCs against ozonide initiated  Heinz body formation (15)
and the lethal  effects of N02 and ozone in rats  (16).  Vitamin E as a membrane-
bound antioxidant in cells may act to prevent the propagation of peroxidative
damage within the cell membranes (17).

     Glutathione peroxidase (GSHP) has been proposed to act as a detoxifica-
tion mechanism for lipid peroxides formed  during ozone inhalation (3).
                                     581

-------
                             GSHP
                2GSH + ROOH 	> GSSG + ROM + H20
                                 GSH

                              Reductase
GSSG + NADPH 	> 2GSH +  NADP+
GSHP is dependent upon glutathione reductase and NADP-reductase for full  activ-
ity.  In the RBC glucose serves as the source of electrons for reduction  of
NADP via glucose-6-phosphate dehydrogenase.   Ozonide-initiated Heinz body for-
mation is not inhibited by glucose in vitro  suggesting that glucose dependent
pathways of detoxification are not as important as vitamin E reserves in  the
prevention of Heinz bodies.  A single patient with glucose-6-phosphate dehydro-
genase deficiency was observed to have the same propensity to form Heinz
bodies as the normal person having eaten an  average U.S.  dietary intake of
vitamin E (D. B. Menzel, unpublished results).

     On short-term N02 exposure RBC GSHP  levels fell, but were not depressed
on  long-term exposure  (Figure 3).  RBC and lung GSHP levels decreased with
age, independent of N02 exposure.  Chow et al.(3) have shown that GSHP levels
are increased by ozone exposure of rats, but no increase in GSHP was found
with N02.  Similarly, no increases in GSHP levels were observed here.  Pre-
sumably increases in GSHP on ozone exposure  were to detoxify peroxides formed
by  ozone-initiated peroxidation.   If GSHP does function as a detoxification
mechanism for organic peroxides,  the exposures used here to near ambient
levels of N02 do not appear sufficient to induce higher levels of the enzyme
in  guinea pigs.  While ozone is a more toxic gas, N02 produces more peroxides
than ozone in model systems of polyunsaturated fatty acids (1).  N02 exposure
should induce GSHP more easily, but does not in both rats and guinea pigs.
Though both ozone and N02 evoke strikingly similar anatomical responses and
have similar chemical properties  to initiate lipid peroxidation, the failure
of N02 to induce higher levels of GSHP suggests basic differences in the mech-
anisms of detoxification of ozone and N02.

      Cellular  damage  in the lung may result in increased phagocytosis.  Lyso-
zyme and acid phosphatase are two lysosomal  marker enzymes associated with
phagocytosis and repair in the lung (4).  There were no immediate effects of
N02 on the levels of the plasma acid phosphatase, but long-term exposure re-
sulted in a significant depression (Figure 4).  No effects could be detected
in the lung on  short-term exposure,  while acid phosphatase levels were elevated
on continuing exposure.  The increase in lung acid phosphatase levels in  the
presence of depressed plasma levels  suggests  a continuing elevation of phago-
cytosis in  the   lung due to NO.  exposure.   Lysozyme levels were elevated in both
lung and plasma on  short-term exposure (Figure 5).   Long-term exposure resulted
in a significant depression of the levels of the enzyme in both tissues.   The
depression  of lysozyme correlates with the increased susceptibility to airborne
pathogens of N07-cxposed animals.   N02 exposure is likely to interfere with the
defense against pathogenic bacteria.

                                     582

-------
                                   H
ZfU:
LT>
                                   H
     co
                                  aw
               en

               'oJ

               CD
                                                                        O)
                                                                        CO
                                                                        03
                                                                        X
                                                                        O
               CD
               Q-
                                                                        O)
                                                                        c
                                                                        o
                                                                        fO
     co
     X
h
^ 2oN
                                                         CXI
                                                                        cn

                                                                        c
                                                                        O

                                                                        CD
                                                                       to
                                                                       O
                                                                       Q_
                                                                       X
    JO

    cr
h
Hy T W
Gin
                                                                       cu
PQ
ct:
     co
l_,,
r
i i
CD CD
LA CT
:i-
1 HI
1 dl
•1 3ON

V

i i
CD CD
N^ CXI
V
CD
• — 1
CT r— 1
K> CD
— ' CD
CD
A
cr D_
^— * j*.
*
*
                 9W/NIW/Q3ZIQIXO Hd\/M 310Wrt
                                                                       oo

                                                                       O)

                                                                       en
                                   583

-------
     co
     X
           H
                •ON
                                              LT\
                                                                                     C/l

                                                                                     O)

                                                                                     O>
    CD
    UD
  I


LPi
 I

CD
 I

CD
                      CD
                      CNI
                                                           I

                                                           CD
                             9W/NIW/310WI1
                                                                                     01
                                                                                     en
          H
                •ON
          H
                                                         CNJ
                                              CXJ
                                                 CL

                                                 O
                                                 _c
                                                 Q.

                                                 -o

                                                 o
                                                 O3
                                                                                     c
                                                                                     03
                                                                                     oo
                                                                                     03
     00
     X
                                                     H
                                                     CD
                                                     CD

                                                     CD

                                                      A

                                                      Q_
                                                                              #
                                                                              *
                                                                       OJ
                                                                       S-
                                                                       3
                                                                       (/I
                                                                       o
                                                                       Q.
                                                                       X
                                                                       O)
                                                                                     o
                                                            M-
                                                            o
oo
•en

Q_
     00
                                                     LPl
                                                     CD

                                                     CD
                                                                        CD
                                                            o
                                                            o>
                                                                                     •3-

                                                                                     
-------
     to
     X
                              H
                                        in
                                            yi\/
oo
v

CD
S- CD
x_x ^
Q.
i#
*
           00
                                                 'ON
                                                                         LO
                                                                         CD

                                                                         CD

                                                                         A

                                                                         CL

                                                                         *
                                                                              O)
                                                                             4-
                                                                             UJ
                                                                             LD

                                                                             CD
                 CNJ
                                  (9/9rt)  3WAZOSA~|
                                      585

-------
     In summary, of the indicators of damage studied, elevation of plasma acid
phosphatase in the presence of depressed plasma lysozyme and cholinesterase
appears most useful for future studies of indicators in man.  Since glutathione
peroxidase has not been observed to be elevated in rats or  in guinea pigs on
N02 exposure, glutathione peroxidase may not function in N02 intoxication as a
detoxification mechanism.
                               ACKNOWLEDGEMENT

     This work was supported by a contract from the United States Environmental
Protection Agency.  The technical work of Mrs. Jacqueline Carothers and the
secretarial assistance of Ms. Elaine Smolko are acknowledged.
                                  REFERENCES
       1.   Roehm,  J. N., J. G. Hadley and 1L B. Menzel.  Oxidation of
           Unsaturated  Fatty Acids by Ozone and Nitrogen Dioxide:  A
           Common  Mechanism of Action.  Arch. Environ. Health  23;
           142-148 (1971).

       2.   Little,  C.,  R. Olinescu,  K. G. Reid and P. J. O'Brien.
           Properties and Regulation of Glutathione Peroxidase.  J.
           Biol. Chem.  245: 3632-3636 (1970).

       3.   Chow, C.  K., C. J. Oil lard and A. L. Tappel.  Glutathione
           Peroxidase System and  Lysozyme in Rats Exposed to Ozone
           and  Nitrogen Dioxide.  Environ. Res.  7_: 311-319 (1974).

       4.   De Lumen, B. 0., S. Taylor, N. Urribarri and A. L. Tappel.
           Subcellular  Localization  of Acid Hydrolases in Rat Lungs.
           Biochim.  Biophys. Acta  268: 597-600 (1972).

       5.   Myrvik,  Q. N., E. S. Leake and B. Fariss.  Lysozyme Content
           of Alveolar  and Peritoneal Macrophages from the Rabbit.
           J. Immunol.  86; 133-136  (1961).

       6.   MacQueen, J. and D. Plant.  A Review of the Clinical
           Applications and Methods  for Cholinesterase.  Am. J. Med.
           Technol.  39: 279-287  (1973).

       7.   Ehrlich,  R.  and M. C.  Henry.  Chronic Toxicity of Nitrogen
           Dioxide.  Arch. Environ.  Health  17_: 860-865 (1968).

       8.   Lowry,  0. H., N. J. Rosebrough, A. L. Farr and R. J.
           Randall.  Protein Measurement with the Folin Phenol
           Reagent.  J. Biol. Chem.  193: 265-275 (1951).
                                     586

-------
 9.   Bessey, 0.  A., 0. H. Lowry and M.  J.  Brock.   A Method
     for the Rapid Determination of Alkaline Phosphatase with
     Five Cubic Millimeters of Serum.  J.  Biol. Chem.   164:
     321-329 (1946).

10.   Prockop, D. J. and W. D. Davidson.  A Study of Urinary
     and Serum Lysozyme in Patients with Renal Disease.  New
     Engl.  J. Med.  270:  269-274 (1964).

11.   Abou-Donia, M. B. and D. B. Menzel.  Fish Brain Cholines-
     terase:  Its Inhibition by Carbamates and Automatic Assay.
     J. Comp. Biochem. Physio!.   2_: 99-108 (1967).

12.   Roe, C. R., L. E. Limbird,  G.  S. Wagner and S. T. Nerenberg,
     Combined Isoenzyme Analysis in the Diagnosis of Myocardial
     Injury:  Application of Electrophoretic Methods for the
     Detection of the Creatine Phosphokinase MB Isoenzyme.
     J. Lab. Clin. Med.  80: 577-590 (1972).

13.   Moore, C. B., R. Birchall,  H.  M. Horack and H. M. Batson.
     Changes in Serum Pseudo-Choiinesterase Levels in Patients
     with Diseases of the Heart, Liver or Musculo-Skeletal
     System.  Amer. J. Med. Sci.  214:  538-548 (1957).

14.   Menzel, D.  B., R. J. Slaughter, A. M. Bryant and H. 0.
     Jauregui.  Heinz Bodies Formed in Erythrocytes by Fatty
     Acid Ozonides and Ozone.  Arch. Environ. Health  30:
     296-301 (1975).

15.   Menzel, D.  B., R. J. Slaughter, A. M. Bryant and H. 0.
     Jauregui.  Prevention of Ozonide-Induced Heinz Bodies in
     Human  Erythrocytes by Vitamin  E.  Arch.  Environ.  Health
     30: 234-236 (1975).

16.   Menzel, D.  B., J. N. Roehm and S.  D.  Lee.  Vitamin E:
     Environmental and Biological Antioxidant.  J. Agri. Food
     Chem.   2£:  481-486 (1972).

17.   Menzel, D.  B.  "The  Role of Free Radicals in the Toxicity
     of Air Pollutants (Nitrogen Oxides and Ozone)", in Free
     B^li§i]A lH Molecular Biology  and Pathology, W. A. Pryor,
     Ed., Vol. Iff  Academic Press  TT976T
                              587

-------
                                                                            11-5
                       BENEFIT EFFECTIVE OXIDANT CONTROL

          R. A.  Bradley,  M.  Dole, Jr., W.  Schink, and S.  Storelli*
ABSTRACT
     A method AA pn.zAe.nt, jjo-t estimating oxA.da.nt health damage  baAe.d on avait-
a.b£e health  e^ect -oi^oAma^con.   Tktb method wa6 aied to analyze  the. bpatiat
and tejnpoAal dAAtAibution o^ oiu.da.nt htalth damage. in CaLi&oinia.   The. spatial
and tejnpofiat analyA-ib  t>kow& that ma/oA poAtionA o^ oxA.da.nt he.alth damage. ocam.
during a Lunit^d nuwbeA ofa houAA that have, high oxA.da.nt te.ve.lA.   This nuuJtt
AaggestA that A-igni^-icant Ae.duiCtA.onA -tn oxA.da.nt he.atth damage.  doaJid be accomp--
tu>he.d Lk oxA,dant  £eveZi on ep^iode day* and at the. Mout autoA couJLd be De-
duced.  Although tonga anceAtaAnt^.eJ) u)eAe fiound -in ab-6o£u£e  damage. utAmate.*,
the. tieJLative. di^eAenceA in damage oueA ipace and tuna. w)eAe  ^ound fie.aAonably
conAi^te.nt whe.n appeA  and toweA. bound e^>timatej> weAe. made..   Commonly aied OAA.
qaaLity i.ndicatou weAe &ound to be mx^£eadxlng pti&dictionA fiox. damage..   A
change in the. conceptual. axA pollution conVioi mode.1 ^OA analyzing control
.{>tSLate.gie.4 iA pAOpo^ed to -inclu.de. diA.o,ct con^ideAatA.on o^ damage..


                                  INTRODUCTION

     The mandate for air pollution control is based upon the public's desire
to obtain protection against the detrimental  effects of air  pollution.   To
carry out this mandate,  air quality standards have been set  to identify levels
of air quality desirable for protection of the public health and  welfare.
Air pollution control  efforts are focused on the attainment  and maintenance
of air quality at  or cleaner than the standards.

     Decisions as  to which strategies should be implemented  for meeting or
surpassing the standards are usually made considering factors  such as cost,
expected emissions  reduction, expected improvement in air quality, and whether
air quality  standards  would be met.   Benefits of control efforts  are considered
in terms of  improved air quality, not in terms of reduced air-pollution damage.

     We believe that in  strategy evaluation a major focus needs to be on bene-
fits in terms of reduced air-pollution damage.  With this belief  in mind,  we
analyzed oxidant health  damage in California using available dose-response
relationships to determine whether air quality is a suitable indicator for
oxidant health damage,  and whether there is a need to consider the spatial and
temporal distribution  of damage  in air pollution decision-making.
^California Air Resources  Board,  Sacramento, California.

                                      589

-------
     This paper presents a summary of our analysis.   First, the method for
estimating oxidant health effects is presented.   This is followed by an anal-
ysis of the spatial and temporal differences in damage and then by a comparison
of air-quality and health-damage estimates.   Then a new conceptual model  for
analyzing air pollution is proposed, and benefit-effectiveness  analysis is
discussed.
                 METHOD OF ESTIMATING OXIDANT HEALTH DAMAGE

DOSE-RESPONSE RELATIONSHIPS

     We based our estimates of health damage on dose-response relationships
developed by Leung, Goldstein, and Dal key (1).   We considered these relation-
ships the best available for characterizing the human health effects of oxi-
dant in view of the paucity of available information.

     Leung, Goldstein, and Dalkey conducted a delphi survey of 14 medical  pro-
fessionals familiar with air pollution health effects.   Each professional  was
asked to estimate the oxidant level at which 0, 10, 50, and 90% of 14 differ-
ent population subgroups would experience three levels  of health impairment:
discomfort, disability, and incapacity.   The 14 population subgroups considered
were:  normal, children, old age, heart condition (mild), heart condition
(severe), hay fever (sinusitis), influenza, upper respiratory infection,
asthma, acute viral bronchitis, actue bacterial pneumonia,  chronic respiratory
disease, mild chronic obstructive lung disease, and severe chronic lung
disease.  A logistic equation was then used to process  the information on  oxi-
dant concentrations to assess impairments for each population category:


         In ~-p = A + B ln(0x)

where    F = percentage of population affected

        Ox - oxidant concentration to which the population is exposed for
             1 hour (in ppm)

         A = calculated intercept parameter

         B = calculated slope parameter

This equation may be stated in the equivalent form
             i  x  A n B
             1  + e  Ox

     What resulted was 42 dose-response functions  giving percentage of popula-
tion affected by hourly oxidant concentrations with three kinds  of estimates:
low, best, and  high.   The low and high estimates were at the 5 and 95°^ confi-
dence levels.
                                     590

-------
 COMPOSITE  DAMAGE  FUNCTIONS

     We  took  the  functions  developed by  Leung, Goldstein, and Dalkey and pro-
 duced  three per-capita composite damage  functions by combining the 42 dose-
 response functions  for the  best, low, and high health-damage estimates.  To
 develop  each  function, we weighted each  subpopulation curve by our estimate of
 the  relative  size of each subpopulation  group in California.  In addition, we
 weighted the  three  levels of  health impairment as discomfort = 1, disability =
 6.6, and incapacity = 54.5.   These weighting factors were based on estimated
 medical  treatment costs  and productivity loss for typical symptoms for each
 severity level.

 POPULATION-WEIGHTED DAMAGE  ESTIMATES

     Next, we took  the hourly oxidant air-quality data at 76 oxidant air-
 monitoring locations in  California for calendar year 1974 and combined this
 with the per-capita composite health-damage functions shown in Figure 1 to
 obtain per-capita oxidant health damage.  We then combined population data
 with the per-capita estimates to obtain  population-weighted damage estimates.
 The  overall method  for estimating health damage resulting from oxidants is
 discussed  in  detail by Schink and Dole (2).
          EVALUATION OF HEALTH DAMAGE ESTIMATES OVER SPACE AND TIME

     Using the method delineated in the previous subsection, we estimated
population-weighted oxidant health damages for the eight California air basins
for which ambient oxidant data were available.  These estimates (low, best,
and high) are listed in columns 1-3 of Table 1.  The results illustrate the
uncertainty surrounding absolute health-damage estimates.   The low (5%) and
high (95%) limits are from 10" to 2600%, respectively, of the best estimate.
Columns 4-6 of Table 1  show the estimated percentage share of damage for each
air basin using the high, low, and best estimates.

     Although the absolute damage estimates show wide variability, the percent-
age shares, or relative health damage estimates, for the various air basins are
reasonably consistent and similar, with from 88 to 96% of the damage estimated
occurring in the South  Coast Air Basin.   These results indicate that by far the
greatest amount of oxidant health damage in California is in the South Coast
Air Basin, suggesting that a major portion of the resources devoted to oxidant
control in California is appropriately devoted to improving the air quality in
this basin.

     We  believe the similarity and consistency  of the relative  damage estimates
are largely a function  of the shape of  the per-capita composite damage func-
tion,  not  its scaling factor.  Thus we  expect  relative damage estimates devel-
oped from  different damage functions to yield  similar results if  the  functions
exhibit  similar shape characteristics.  We consider the important shape char-
acteristics to be:

     t   Zero or minimal damage at low  oxidant  concentrations corresponding to


                                     591

-------
                                                                    HI GH
                                                                    BEST
                                                                  (Median
                                                                    LOW
  Figure 1
     0.10      0.20      0.30      0.40      0.50

      ONE  HOUR  OXIDANT  EXPOSURE  LEVEL (ppm)


Composite,  severity-weighted, per-capita,  oxidant
functions.
                                                                 0.60
health-damage
         low or negligible risk to people at oxidant levels at or below the
         air quality standard.

     •   Increasing health damage as oxidant levels increase.

     0   Increasing rate of change in health damage as oxidant levels increase.

     These characteristics result in the greatest amount of damage occurring
at the highest oxidant levels.   This result can be seen in Table 2,  which
shows the percent of per-capita oxidant health damage at or above various air-
quality levels at Azusa, California, one of the more heavily impacted areas in
the South Coast Air Basin.   As  indicated in the table, from 11 to 22% of the
damage is estimated to result from 14 hours having oxidant levels 0.30 ppm and
above during the year.   From 61 to 79% of the health damage is estimated to
occur during 223 hours (2.5% of the year) having levels 0.20 ppm and above.
                                     592

-------
        TABLE  1.   CALIFORNIA OXIDANT HEALTH DAMAGE FOR CALENDAR YEAR 1974
	IN  WEIGHTED HEALTH  DAMAGE UNITS AND PERCENT OF TOTAL DAMAGE


                      Total damage  (x  10  )               % Share of total

Air  Basin            Low        Best        High          Low     Best    High
South Coast
San Diego
San Francisco
Bay Area
San Joaquin
Valley
Southeast
Desert
Sacramento
Va 1 1 ey
North Central
Coast
5,955-3
52.3
62.3
59-8
30.2
21.9
1.2
57,330
892
688
683
332
254
16
1,469,263
39,947
49,696
53,672
22,199
21,581
2,204
96.30
0.84
1.01
0.97
0.49
0.35
0.02
95-27.
1.48
1.14
1-13
0.55
0.42
0.03
88.50
2.41
2.99
3.23
1.34
1.30
0.13
 South  Central
 Coast                 0.9         12         1,579        0.01    0.02    0.09

 Total             6,183-9    60,207     1,660,141       99-99   99-99   99-99
This  table  indicates  that  major portions  of  the oxidant health damage occur
during a  limited  number  of hours  that  have high oxidant levels during the
year, suggesting  that significant  reductions  in oxidant health damage could be
accomplished  if oxidant  levels on  episode days could be reduced.

            COMPARISON OF AIR QUALITY AMD HEALTH DAMAGE ESTIMATES

     We compared health-damage estimates and commonly used oxidant air-quality
parameters, highest hourly reading for the year,  annual average,  annual  aver-
age of the daily maximum hour, and annual hours above the 0.08 ppm standard,
to determine how well air quality parameters represent health damage.   Per-
capita health-damage and air-quality parameters for 1974 for two locations,
San Jose and Upland, are presented in rows 1  and 2 of Table 3.

     San Jose and Upland are  the two most severely impacted locations in the
San Francisco Bay Area Air  Basin and South Coast Air Basin respectively.  The
estimated relative damage  at  San Jose as compared to Upland is 1.1% as shown

                                     593

-------
        TABLE 2.   PERCENT OF PER CAPITA OXIDANT HEALTH DAMAGE OCCURRING
             AT OR ABOVE SELECTED 0X1 PANT LEVELS AT AZUSA IN 197*4

Hours Percent of oxidant damage at
Oxidant
level
0.10
0.15
0.20
0.25
0.30
0.35
at or above
level
98?
505
223
71
}k
3
Low
99
9A
79
51
22
8
Best
99
93
78
^9
20
7
or above level
High
96
8*4
61
32
11
3








	
TABLE 3- PER-CAPITA 0X1
OXIDANT INDICATOR VALUES
DANT HEALTH
FOR UPLAND
DAMAGE AND SELECTED
AND SAN JOSE IN 197*t

Annual
Location
Annual 197^4
per capita maximum
health hour
damage (ppm)
Annual
average
(ppm)
average
da i 1 y
maximum
(ppm)
of
hou r
Annual
hours
above
0.08 ppm
 San  Jose
 Upland
 0.68



60.^3
 0.22
 O.V4
0.020
0.050
0.052


0.1M+
 162


1360
 San  Jose
 as % of
 Upland
 1.1!
50.0%
             36.1!
               11.9*
in row 3 of the table.  A comparison of this percentage with those for the air
quality parameters indicates the air quality parameters fail to capture the
large difference that exists in health damage between the two locations.   Thus
air quality seems to be a misleading indicator of health damage.

     Similarly, changes in health damage are much greater than changes in air
quality.   This can be seen more clearly in Table 4.   This table has estimates
                                     594

-------
        TABLE 4.   ESTIMATED PERCENTAGE CHANGE  IN  OX IDANT  HEALTH  DAMAGE
       	FOR CHANGES IN OX IDANT  AIR QUALITY	


                                       Estimated  change  in
Change in
oxidant
ai r qual i ty
20% improvement
!0% improvement
10% deterioration
20% deterioration
ox
low (%)
-68
-42
+63
+ 158
idant health
best (%)
-67
-41
+60
+ 146
damage
high (%)
-56
-32
+42
+95

of expected changes in health damage associated with a 10 and 201 reduction
or a 10 and 20% increase in all hourly oxidant readings during the year.   High
and low estimates are included as well as the best estimate.   The table indi-
cates a 10% improvement in oxidant air quality is expected to yield a 32 to
42% reduction in oxidant health damage.  This indicates the reduced damage
expected with small improvements in air quality is much greater than one would
expect when just looking at the change in air quality.


                   CONCEPTUAL AIR POLLUTION CONTROL MODEL

PAST MODEL AND PROPOSED MODEL

     Based on the preceding comparison of health damage and air quality, we
propose a change in the standard conceptual air-pollution control model illus-
trated in Figure 2.  As shown in Figure 2, air quality, the result of the
interaction of emissions, meteorology, and atmospheric chemistry, is improved
through emission control.  Currently, we analyze various possible emission-
control strategies by means of feedback loop A, i.e., by comparing the result-
ant changes in air quality produced by control strategies.   The goal is to
improve air quality to meet air quality standards.   Although  air quality
varies over both space and time, a single air-quality parameter (e.g., the
maximum pollutant concentration as measured at the worst location in an air
basin) is usually used in the analysis by means of feedback loop A.

     We propose a change in the feedback loop to consider damage (loop B).
Due to the large uncertainties in the absolute damage estimates, we recommend
damage analysis be made on a relative basis by looking at relative differences
in damage estimates associated with various strategies until  more accurate
absolute damage estimates can be made.   Furthermore, as techniques for analyz-
ing other aspects of damage (e.g.  vegetation damage) are developed,  their
results can be included in the analysis.


                                     595

-------
    r
                                   — B —
   EMISSION
   CONTROL
                     •EMISSIONS
  AIR
"QUALITY
 DAMAGE
ESTIMATION
                            METEOROLOGY,

                        ATMOSPHERIC CHEMISTRY
              Figure 2.   Conceptual  air-pollution control model.


     Consideration  of damage  represents  an  improvement  in  two ways.   The first
is that damage estimates place  the  focus  more  clearly on the  parameter of
greatest concern to the public.   As was  noted  earlier in this paper,  commonly
used air quality parameters appear  to  be  misleading  indicators of oxidant
health damage.  The second improvement involves  consideration of the  spatial
and temporal variations in damage.  This  consideration  is  accomplished by
using damage functions (dose-response  relationships) and receptor population
weighting factors to provide  a  weighting  scheme  for  totaling  damage  over space
and time.

BENEFIT-EFFECTIVENESS EVALUATION

     With the consideration of  damage  by  means of  loop  B,  benefit-effectiveness
evaluations can be  made.  Benefit-effectiveness  evaluation  involves  the analy-
sis of benefits (reduced damage)  associated with various options or  strategies
to determine how to maximize  the  benefits for  a  given allocation of  resources.
These evaluations would consider  the spatial and temporal  distribution of dam-
age and would identify control  options providing the greatest reduction in
damage for a given  abatement  cost.  Though  beyond  the scope of this  paper, we
believe alternative control strategies can  be  evaluated within a stochastic
nonlinear mathematical programming  framework using weighted damage as an
objective function.  This function  would  be minimized subject to input and/or
expenditure level abatement cost  constraints.
                                  CONCLUSIONS

     Relative health damage estimates  are  reasonably  consistent over space and
     time, even though there are  wide  variations  in absolute  estimates.

     Air quality appears to be a  misleading indicator of health damage.   Com-
     monly used air-quality parameters  do  not  capture the large variation in
     the spatial and temporal differences  in health damage.

3.   Significant reductions in per-capita  oxidant  health damage can be achieved
1.


2.
                                      596

-------
     if reductions are made at the worst locations at the time of the worst
     oxidant damage.   Thus, oxidant health damage could be significantly re-
     duced if oxidant levels on episode days could be reduced.

4.   Air pollution decision-making on which strategies to use needs to con-
     sider oenefits in terms of reduced damage.   This consideration may be
     accomplished by using relative damage estimates in benefit-effectiveness
     evaluation.
                                 REFERENCES

     Leung, S.,  Goldstein, E.,  and Dalkey, N. ,  Human Health Damages  from
     Mobile Source Air Pollution, California Air Resources  Board,  Sacramento,
     California, March 1975.

     Schnik,  W.  0., Jr.,  and  Dole, M.,  Jr., The Development and Use  of Rela-
     tive Health Damage Estimates in Air Pollution Decision Making,  (Presented
     at the 51st Annual  Conference of the Western Economic  Association,  San
     Francisco,  California,  June 24-27, 1976),  California Air Resources  Board,
     Sacramento, California,  June 1976.
                                    597

-------
              SESSION 12
  EFFECTS OF OXIDANTS ON VEGETATION - I
                  :  W.W.  Heck
North Carolina State University at Raleigh
                  599

-------
                                                                            12-1
           OZONE  INDUCED ALTERATIONS IN PLANT GROWTH AND  METABOLISM

                                 D. T.  Tingey*
ABSTRACT
     Plant  foliage i*  a primary receptor o& ozone, and  other  ox^dant*  that
cau*e biochemical and  phy '* io logical, change* -in the.  foliage and the*e  foliar
change* OSLO. reflected  in the. development o& other plant  organ*.   The.  biochemi-
cal changes that are induced by ozone, are the, cau*e o& visual, injury,  growth
fie.ddctA.on.,  lo**  oft plant vigor,  and ultimately determine, plant community and
broader eco*y*tem level* ofi re*pon*e.  Changes -in plant  phy*iology and bio-
c.hemLt>tH.y can pnovlde.  4 e.n^^uU.\j^Lty mank pA.odu.ae-d by the. plant in ne.&pon&e.
to ozone., and a  clo&e. fie.lati.on-t> hip e,xJJ>tt> between the. amount  o£ injiviy and the.
amount o& &tneAt> &thyle.ne. produced.  In addition, &thyle.ne. can  incA^a&e. Izafa
ab&cAAbion and phe.nol bi.o£yntheAiA and induce. bioAyntheAiA o& e,nzymeJ> involve.d
in legion
     When plants  one. exposed to ch/ioni.c. le,ve£i> o& ozone., leAA AugaA it  i>ton.e.d
in the. nootA.   Chsionic. expo^uAed alt>o caa6e an incAe.a&e. in the.  phenol content
o& the. foliage., which occuu ptiiofi to the. appearance o& visual  injury.   The.
phenol increase. iA  aJU>o related to the ultimate &oHaA injury and  the Atipple
symptom Aeen in injured tibAue u> frequently the result o& phenol  accumulation.
     The ei&ectA  o&  ozone are not limited to the foliage.   In  general,  the
        o& the  ozone occur {,irAt on root growth followed by top  growth.   When
the production  o^ photo^ynthate -ci limited, plants restrict the  &low o& carbo
hydrate* to the root*.

     Chronic e^ect* 0$ ozone are not limited to *ymptomology  and reduced
growth, a* ozone  can al&o alter quality (e.g., by increasing the nitrogen (M)
content o^ the  foliage).   Even though there li> higher M in  the leave*  ofa for-
age plant*, they  provide le** protein/ unit area o^ crop cover.   Thi* ha* the
net e^ect o& reducing  the forage protein available to animal*.  Torage crop*
al*o exhibit a  decrease in $- carotene content.

     The*e phy*iological change* help explain why alteration*  in growth and
plant vigor can occur.   The retention o& carbohydrate in the leave*  reduce*
the growth ofa the root* and re*ult* in le** energy *tored.  Thi* reduction in
*U.S.  Environmental  Protection  Agency, Corvallis, Oregon.


                                       601

-------
                   u&timatzly i>tow plant giowth and tended the. ptant mote
          to otheA t>tA..  Phy^-iolog-ical change px.ov
oft monit-Otiing the, health and v^goti o^ the. p&ant wJLh oh. uiithout ux^6iia£ injury.


                                INTRODUCTION

     Plants exist in two different environments:  the foliage and stems are
immersed in a gaseous environment while the roots in the soil have only limited
gas exchange.  Even though the plant survives in two different environments,
plant growth is the summation of the interaction of events that occur in the
shoot and root.  As a whole, the plant is autotropic while the individual
organs are heterotrophic depending upon one another for specific assimilated
or elaborated substances (1).  The roots supply water, minerals, nitrogen com-
pounds, and phytohormones.  The foliage supplies carbohydrates, vitamins, co-
factors, nitrogen compounds and phytohormones (1).   These elaborated substances
from the root and shoot are translocated throughout the plant and are required
for normal growth.

     Plant growth and yield is the end product of the biochemical processes
related to uptake,  assimilation, biosynthesis, and trans location.  Growth re-
quires that the plant photosynthetically trap solar energy and extract the
necessary minerals  and water from the soil.   The various plant organs convert
these raw materials into the required compounds for plant growth.  These
materials are frequently produced in excess at the site of production and are
translocated throughout the plant.   If the distribution of assimilated or
elaborated substances among the plant's organs is disrupted, a reduction in
growth and/or yield will result.

     The foliage is the primary route by which ozone enters the plant and is
also the primary site where it exerts its effect.  To penetrate to the site of
action, ozone diffuses across the leaf boundary layer, through the stomatal
pores, which in part control the amount of ozone entering the leaf; passes
through the intercellular spaces; and dissolves in the hydrated cell surfaces.
When the ozone dissolves in the water, it decomposes into free radicals, which
exert the ozone effect (2).

     The effects of ozone on the plant can be divided into several levels, de-
pending on the time period after exposure that the effects are observed.  The
earliest effects are an increase in membrane permeability, an increase in the
leakage of ions such as potassium from the cell, a decrease in carbon dioxide
fixation, and a stimulation of stress ethylene production (2,3).  These early
changes are followed by both enzyme inactivation and/or activation of other
specific enzymes, alteration in metabolite pools, and modified translocation
(4,5).   These biochemical changes are ultimately expressed in increased leaf
drop, visual foliar injury, and a reduction in plant growth and yield (6,7,8,
9).   Thus plant changes that are initially biochemical in origin are ultimately
expressed as visual injury, losses in plant growth and yield, and reduced
plant vigor; they ultimately determine plant community and broader ecosystem
levels of response.

     Photochemical  oxidants can depress plant growth and yield by one or a
combination of several possible mechanisms, such as, reducing the carbon

                                     602

-------
dioxide uptake, decreasing the amount of photosynthate formed, reducing mineral
uptake from the soil, altering the amount of types of compounds elaborated by
the different organs, altering the rate of trans location, or changing the
source-sink relations for biosynthesized compounds among the various plant
organs, or reducing the life of photosynthetic organs.


                           RESULTS AND DISCUSSION

     Plants produce low levels of hormonal ethylene under normal conditions,
but when subjected to environmental stresses, ethylene production increases.
This increased ethylene production following stress is frequently referred to
as wound or stress-induced ethylene.   Injured, but not dead tissue is the site
of the stress-induced ethylene production (10).  Craker (11) suggested that
the production of stress ethylene is the cause of many of the plant responses
to ozone.   Stress ethylene is liberated from plants by as little as 15-minutes
exposure to ozone and is increased by both the amount and duration of ozone
(3).  The stress-induced ethylene is liberated by the plant prior to any
apparent visual changes and frequently in the absence of visual foliar injury
(3).  Also, a close relation exists between the stress-induced ethylene and
the ultimate foliar injury, indicating that stress-ethylene evolution is a
good measure of ozone stress to plants (3).   The maximum rate of ozone-induced
ethylene production occurred within 2 to 4 hours following the exposure, and
the amount of stress ethylene produced was in a physiologically active
concentration (11).

     Increased leaf abscission is frequently associated with oxidant injury
and can result from increased stress ethylene production (6,12, Tingey unpub-
lished data).   The increased leaf abscission could result in reduced plant
growth by decreasing the available leaf area for photosynthesis.  Also,
reduced plant growth and dwarfness in plants is associated with elevated
ethylene levels (10,13).  The activity of several enzymes involved with phenol
biosynthesis and phenol oxidation can be increased by elevated ethylene levels
(10).

     Howell and Kremer (14) identified pigments formed in ozone-exposed leaves
exhibiting visual injury as polymerized phenols, proteins, and amino acids.
These researchers suggest that the ozone induces changes that either allow
more oxygen to enter the cells or allow oxidative enzymes to react with phenols,
forming quinones that can polymerize with proteins and amino acids to form the
pigments seen in injured tissue.   In ponderosa pine foliage, phenols increase
prior to the appearance of visual injury (5).   In alfalfa plants experiencing
chronic ozone exposure, there is  an 8% increase in total phenols (see Table 1).
There is also a positive correlation between the amount of phenols that
accumulate in the tissue and the  amount of foliar injury (15).  The activities
of the enzymes of phenol biosynthesis (phenylalanine ammonia lyase) and phenol
oxidation (polyphenol oxidase, and peroxidase) associated with lesion forma-
tion are stimulated prior to the  appearance of visual injury (4).   The in-
creases in enzyme activities can  all  be stimulated by ozone-induced stress
ethylene,  and ethylene is apparently the trigger for lesion formation (4,10,
11).  Visual injury is associated with reduced growth.  Studies with radishes
have shown a linear relationship  between the foliar injury observed at the


                                     603

-------
            TABLE 1.   THE INFLUENCE OF CHRONIC OZONE EXPOSURES ON
               THE TOTAL GROWTH AND SOLUBLE RESERVES OF ALFALFA*

Ozone Concentration
0
50
Total Dry
Matter (g)
49.6
27.1
Total Nonstructural
Carbohydrates (g)
9.3
4.2
% TNC
Total Dry Matter
19
15

*For experimental  details, see Reference 18.

final harvest and the percent growth reduction measured (Y = 1.6 + 0.9X, where
Y = % growth reduction and X = % foliar injury to the plant; for experimental
details see reference 16).

     The incorporation of carbon dioxide is required for plant growth.  A
reduction in photosynthesis is ultimately expressed as decreased dry matter
production.  When plants are exposed to either acute or chronic ozone doses,
there is a temporary reduction in photosynthesis.  Alfalfa was exposed to a
low concentration of ozone (50 nl/1, 6 hrs/day, for 10 wks) the total dry
matter produced for the season was reduced 45% (Table 1).   In addition, the
total nonstructural carbohydrates (TNC) were reduced 55%.   The exposed plants
also contained lower levels of TNC as a percentage of the total dry matter
than did the control plants.  TNC reserves are used by the plant to drive its
metabolic and growth processes in both shoots and roots.

     In addition to a reduction in photosynthesis and the TNC reserves in the
plant, there is also a change in the partitioning of photosynthate in plants
exposed to ozone stress.  In ponderosa pine exposed to ozone (100 nl/1,6 hrs/
day, for 20 wks) there was a retention of sugars and starch in the top of the
plant and a decrease in these metabolites in the roots (5).  The reduction of
soluble reserves could decrease the ability of the plants to initiate growth
the following spring.

    When  shoot-synthesized  carbohydrates  bacame  too  deficient  to maintain  the
normal growth rate, sinks such as roots and fruits are more severely affected
than the foliage.  This is consistent with the observation that organs remote
from the site of synthesis or absorption are the most severely penalized under
stress conditions that affect the process (1).  Hence, when either storage
roots, fruits, or seeds have a high demand for photosynthate, they will exper-
ience a reduction in translocated photosynthate, and growth will be reduced if
an ozone stress has reduced photosynthesis during this period.  This reduction
in photosynthate available to the sinks during a period of high metabolic de-
mand explains why yieled reductions are greater during specific phenological
stages (16).

     The root and shoot are in constant competition for energy and resources
for their growth and development (1).  The shoot-root ratio measures the re-
sultant pattern of the differential growth of the two organs.  The shoot-root


                                     604

-------
 ratio can be used to measure  the effect  of environmental stresses on plant
 growth patterns.  The  effect  of chronic  ozone exposures on  the shoot-root
 ratios is shown  in Table  2.   Each  species exhibited  its own  characteristic
 ratio.   However, all species  showed  an increased shoot-root  ratio even  at the
 50  nl/1  ozone  treatment.   This increase  could result from a  stimulation  of
 foliage  growth or a reduction in root growth.   But the  levels of ozone  used
 suppressed both  foliage and root growth  (7,8,9).  Hence, the increase in the
 shoot-root ^atio could only result from  a greater suppression of root growth
 in  relation to foliage growth.  The  greater  reduction in root growth would be
 predicted from the reduction  of photosynthesis, trans location, and  accumula-
 tion of  photosynthate  in  the  roots reported  in  the above paragraphs.  The re-
 duction  in root  growth and stored  reserves has  been  reported in the field
 (19).  Studies have shown that trees experiencing air pollution stresses are
 more susceptible to other environmental  stress  (20,21).

   TABLE  2.   THE INFLUENCE OF CHRONIC OZONE EXPOSURES ON THE SHOOT-ROOT RATIO*
 Ozone Concentration
        nl/1
                   Plant Species
                           Alfalfa
                Soybean
   Radish
Tobacco
0
50
TOO
1.1
1.3
1.7
8.9
9.3
10.0
2.2
4.0

7.3
11.1


 *For experimental  details, see References 7,8 and 9.

     The reduction in root growth is also reflected in reduced mineral uptake
(Table 3).  The uptake of phosphorus (P) and potassium (K) was reduced about
30%, while the uptake of calcium (Ca) and magnesium (Mg) was reduced approxi-
mately 45%.  The reduction would be expected to result from the lowered meta-
bolic reserves in the roots from the ozone stress.  Mineral uptake is depend-
ent on photosynthesis and metabolic reserves (22).


      TABLE 3.   MINERAL  UPTAKE OF ALFALFA EXPOSED  TO CHRONIC OZONE LEVELS*
     Ozone
 Concentration
    (nl/1)
 Total Plant
Material,  Dry
     (g)
  Elemental  Composition
(nig/Total  Plant Material)
                                                      K
                                        Ca
                         Mg
0
50
100
1.82
1.51
0.99
11
10
8
67
61
46
20
15
11
7
6
4

 *For experimental  details,  see Reference 9:   Elemental  Analyses conducted by
  Soil  Science Department,  North Carolina State University,  Raleigh,  N.C.
                                      605

-------
      Symbiotic nitrogen  fixation  by  legumes  is  controlled  in  part  by  photo-
 synthesis  and  photosynthate  distribution  (23).  The  photosynthate  is  required
 in  the  roots to provide  the  energy to  reduce  and  incorporate  atmospheric
 nitrogen  into  organic  forms.   Nitrogen  fixation was  measured  in  alfalfa plants
 exposed to either  0  or 50  nl/1  ozone  (Table  4).   The  total  nitrogen fixed  by
 the  plants throughout  the  growth  period was  depressed  40%.  However,  the
 efficiency of  nitrogen fixation was  not affected  by  the ozone treatment.   This
 illustrates the dependency of  nitrogen  fixation on photosynthesis  and growth.
TABLE 4. EFFECT OF OZONE ON NITROGEN FIXATION BY ALFALFA*

Ozone
Concentration
(nl/1)
0
50
Total Plant
Material
(TPM) (g)
49.6
27.1
Nitrogen
Fixed
(mg/TPM)
1350
793
Efficiency
(mg N/g Plant
Material )
27
29
.
           For experimental  details,  see Reference 18.

      Forage quality  is  also altered  by  chronic ozone exposure  (Table  5).  The
 ozone treatment  increased the  content of protein,  ami no acid.,  and phenols,  and
 decreased  the  level  of  e-carotene  in the foliage.

         TABLE  5.   EFFECTS OF  CHRONIC OZONE  EXPOSURES ON THE  QUALITY  OF
                                 ALFALFA FORAGE*
            Chemical  Constituent                    %  of  Control  Level


            Protein                                        128

            Ami no  Acids                                    119

            Total  Nonstructural Carbohydrates               98

            Phenols                                        108

            B-Carotene                                      76


         *  For  experimental  details/ see  reference  18.

     The protein increase is not of significance because total dry matter pro-
duction was reduced and therefore protein production/unit ground area was also
reduced (18).  As a result, the amount of protein available for feed is de-
creased.  Even though there are significant changes in the organic content of
alfalfa foliage, there do not seem to be any significant changes in the miner-
al composition (Table 6).

                                     606

-------
                TABLE 6.   THE EFFECT OF CHRONIC OZONE EXPOSURES
                  ON THE  MINERAL CONTENT OF ALFALFA FOLIAGE*


            Ozone Concentration      Elemental  Composition (%)
            (nl/1)                    P -     K      Ca      Mg
0
50
100
0.6
0.6
0.7
5.0
5.0
5.1
1.4
1.3
1.3
0.5
0.4
0.4

               For experimental  details,  see Reference 9.
                                   SUMMARY

     Ozone concentrations near and below the current standards have been shown
to have significant effects on plant qrowth and metabolism.  The majority of
effects observed, such as reduced root growth, mineral uptake, and nitrogen
fixation, seem to result from reduced photosynthesis and photosynthate distri-
bution.  This reduction in metabolic reserves ultimately slows plant growth
and renders the plant more sensitive to other stresses.  Physiological changes
can provide a sensitive means of monitoring the health and vigor of the plant
with or without visual injury.
                                 REFERENCES

 1.  Aung, L. H., Root-Shoot Relationships.In:  The Plant Root and its Environ-
     ment, E. W. Carson, Ed., University Press of Virginia, Charlottesville,
     Virginia, 1974, pp. 29-61.

 2.  Heath, R. L. , Ozone IN:  Response of Plants to Air Pollution, J. B.  Hudd
     and T. T. Kozlowski, eds.,  Academic Press, New York, N.Y., 1975. pp. 23-
     55.

 3.  Tingey,  D.  T., Standley and R. W. Field, Stress Ethylene Evolution:   A
     Measure  of Ozone Effects on Plants, Atmos. Environ. 10 (in press), 1976.

 4.  Tingey,  D.  T. , R.  C. Fites, and C. Wickliff, Activity Changes in Selected
     Enzymes  from Soybean Leaves Following Ozone Exposure, Physiol. Plant. 33:
     316-320, 1975.

 5.  Tingey,  D.  T., R.  G. Wilhour, and C.  Standley, The Effect of Chronic
     Ozone Exoosures on the Metabolite Content of Ponderosa Pine Seedlings,
     Forest Science (In press),  1976.
                                     607

-------
 6.   Menser, H. A., H. E. Heggestad, and J. 0. Grosso, Carbon Filter Prevents
     Ozone Fleck and Premature Senescence of Tobacco Leaves, Phytopathology
     56:466-467, 1966.

 7.   Tingey, D. T., W. W. Heck, and R. A. Reinert, Effect of Low Concentrations
     of Ozone and Sulfur Dioxide on Foliage, Growth and Yield of Radish, J. Am.
     Soc.  Hort. Sci., 96:369-371, 1971.

 8.   Tingey, D. T., R. A. Reinert, C.  Wickliff, and W.  W.  Heck, Chronic Ozone
     and Sulfur Dioxide Exposures, or Both, Affect the Early Vegetative Growth
     of Soybean, Can.  J.  Plant Sci.  53:875-879, 1973.

 9.   Tingey, D. T.  and R. A. Reinert,  The Effect of Ozone and Sulfur Dioxide
     Singly and in  Combination on Plant Growth, Environ.  Pollut. 9:117-125,
     1975.

10.   Abeles, F. B., Ethylene in Plant Biology, Academic Press, New York, NY,
     1973.

11.   Cracker, L. E., Ethylene Production from Ozone Injured Plants, Environ,
     Pollut., 1:299-304,  1971.

12.   Thompson, C.  R. and  0.  C. Taylor, Effects of Air Pollutants on Growth,
     Leaf Drop, Fruit Drop,  and Yield of Citrus Trees, Environ. Sci. Technol.,
     3:934-940, 1969.

13.   Aharoni, Y.,  C. T.  Phan, and M. Spencer, Ethylene Production as Related
     to Dwarfness  and Earliness in Plants, Can. J. Bot. , 51:2243-2246, 1973.

14.   Howell, R. K. , and D.  F. Kremer,  The Chemistry and Physiology of Pigment-
     ation in Leaves Injured by Air Pollution, J. Environ.  Qua!., 2:434-438
     1973.

15.   Howell, R. K., Phenols, Ozone and Their Involvement in the Physiology of
     Plant Injury  In:   Air Pollution Related to Plant Growth, M. Dugger, ed.,
     American Chemical Society Symposium Series, 3:94-105,  1974.

16.   Tingey, D. T., J. A. Dunning, and G. M. Jividen, Radish Root Growth Re-
     duced by Acute Ozone Exposures, Proc. Third Internat.  Clean Air Congress,
     A154-A156, 1973.

17.   Hill  A. C. and N. Littlefield, Ozone Effect on Apparent Photosynthesis,
     Rate of Transpiration and Stomatal Closure in Plants,  Environ. Sci. Tech.
     3:52-56, 1969.

18.   Neeley, G. E., D. T. Tingey and R. G. Nil hour, Effects of Ozone and
     Sulfur Dioxide Singly and in Combination on Yield, Quality and N-Fixation
     of Alfalfa, Proc. Internat.  Conf. Photochem. Oxidant Pollut. (In Press),
     1976.

19.   Parmeter, J.  R.,  Jr. and P.  R.  Miller, Studies Relating to the Cause of
     Decline and Death of Ponderosa Pine in Southern California, Plant Disease

                                      608

-------
     Reporter, 52:707-711, 1968.

20.   Lux, H.  Ergenisse von Zuwachsuntersuchungen (Bohrspanalysen)  im Rauch-
     schadensgebiet Dubener Heide, Arch.  Forstw.,  14:1103-1121,  1965.

21.   Materna, J. Kriterien zur Kennzeichung einer Imnrisionseinwirking auf
     Waldebestands, Proc.  Third Internal.  Clean Air Cong.,  A121-A123,  1973.

22.   Epstein, E., Mineral  Nutrition of Plants:  Principles  and Perspectives,
     John Wiley and Sens,  Inc., New York,  NY, 414 pp.

23.   Gibson,  A. H., Limitation to Dinitrogen Fixation  by Legumes.   In:
     Nitrogen Fixation.   W. E. Newton and C. 0. Nyman, eds.,  Washington State
     University Press, Pullman, Washington, pp. 400-428, 1976.
                                     609

-------
                                                                           12-2
                 OXIDANT LEVELS IN REMOTE MOUNTAINOUS AREAS OF
                  SOUTHWESTERN VIRGINIA AND THEIR EFFECTS ON
                     NATIVE WHITE PINE (PINUS STROBUS L.)

                 J.  M.  Skelly, C.  F.  Croghan and E. M. Hayes*
ABSTRACT
             £eve£i  wete mon/ttoAed ^on 15 months in 4outfiwe6teAn
Symptom development  in leAponAe. to ox-idant -£eve£4 wcw Aec.oA.ded on. whte pint
f$i.e£d tn&&> uAtng two  m&thocU,:   (a)  indlvJjduaJt j$a-6xLa€e A.e-6pon-6e, and  (b)  to tat
tnji.ii. Ae4pon4e.   Pollution ep-iiode-6 we/te noted in July 1975 and June  1976 and
both Ae4a£ted  ($Aom w.ind4 out o& the. WE.   Symptom development wo* -tepoAted .in
OAAocsiation uuJth. tk.   T^e 1975 symptom e.x.pti&>t>
-------
 estimated a 3% mortality of eastern white pine (Pinus strobus L.]  in the
 affected area.

      Recently, high levels of ozone in excess of the National Ambient Air
 Quality Standard (NAAQS) for photochemical oxidants have been reported; 160
 microgram/m3 or 8 pphm maximum concentration oxidant not to be exceeded more
 than one hour once per year.

      Johnston et al.  (2) monitoring ozone in rural  Garrett County, Maryland,
 from August 29 to September 1972,  found the NAAQS to be exceeded  11.26% of the
 time; one period lasted 26 consecutive hours.   The  mean for the period was 5.7
 pphm.  Richter et al. (1) found similar high ozone concentrations  in the rural
 Mount Storm area of West Virginia.   Coffey and Stasiuk (9) reported ozone
 concentrations in excess of the NAAQS common in the rural  areas of New York
 State and suggested transport of ozone and natural  ozone precursors from the
 stratosphere.   Miller et al.  (13)  attributed high rural oxidant levels to the
 transport of oxidant and man-made  precursors from the population  centers in
 the San Joaquin Valley of California.   Cleveland and Kleiner (7)  found higher
 ozone concentrations when air was  flowing from the  Camden-Philadephia urban
 complex to remote sites 27 to 49 km distant.   Nitigan et al.  (3),  working in
 the Ohio Valley in the summer of 1974, found the NMQS to  be exceeded twice as
 frequently at the rural sites as at the urban sites.   They and others (11,18)
 also support that transport of ozone precursors from urban areas  to rural
 areas occurs under appropriate meteorological  conditions.

      The present study was initiated to identify oxidant symptom  patterns on
 white pine in the mountainous areas of southwestern Virigina.  Study of weather
 patterns associated with various oxidant levels was also conducted.


                             MATERIALS AND METHODS

      From April 1975 through  April  1976 continuous  monitoring took place at
 three rural  sites in Virginia:   (1) Salt Pond Mt. ,  Giles County;  (2) Rocky
 Knob Mt., Floyd County.; and  (3) Shenandoah Valley, Rockingham County.  In
 May of 1976, Station 3 was relocated to Apple Orchard Mt.  in Bedford County.
 Total oxidant has been monitored with 1% buffered potassium iodide (KI) cali-
 brated Mast  Meters (Model 724-2) located within the U.S. Weather  Bureau sta-
 tion structures.

      At the  Salt Pond Mt. site 21  trees within a 15-year old white pine stand
 were grouped into one of three symptom classes (sensitive, intermediate,
 tolerant) with seven trees per class.   Sensitive trees were characterized by
 stunted tree growth and short needles with tip burn and chlorotic  mottle
 symptoms; only current-year needles were present.   Intermediate trees had
 longer needles with chlorotic spots and mottling symptoms; only current-year
 needles were present.   Tolerant trees had long green needles that  were symptom-
. less; two year old needles were present.   In May of 1976 six additional trees
 located near Site 2 were added to  the study.

      After candle elongation  and fasicle emergence  in late May 1975 and 1976,
 one branch was randomly selected on each tree and 12 faciles on the branch


                                      612

-------
terminal and one lateral were selected and marked.  Four fasicles were marked
on the bottom, middle, and top of each candle in sequence.

     All needles on all fasicles were rated for seven symptom types at approx-
imately 11-day intervals during the growing season.  Symptom expression was
rated according to the following scale:  (1) chlorotic spot; (2) chlorotic
mottle; (3) chlorotic band; (4) chlorotic tip; (5) necrotic spot; (6) necrotic
band; (7) tip burn.  The severity index was computed by multiplying the number
of the symptom by the percent injury.  All symptoms were observed during the
study period.

     During the 1976 field season a system for evaluating total tree response
to oxidant exposure was introduced.   This symptom was modeled after that used
by Miller (14) in the San Bernardino Mountains.  Trees were rated quantita-
tively in both upper and lower crown for needle retention, condition, and
length.  Possible scores ranged from 0 to 9 for sensitive, 10 to 14 for inter-
mediate, and 15 to 18 for tolerant trees.  The total tree evaluation system
was used to rate the aforementioned 21 field trees at Salt Pond Mt. along
with 28 new trees on the Blue Ridge Parkway between Rocky Knob and Apple
Orchard Mt., a distance of approximately 145 km.


                                   RESULTS

     1975:  For a portion of the study period, flay to August 1975, the hourly
oxidant concentration was 4.27, 4.96, and 3.15 pphm for Sites 1,2, and 3,
respectively.  On July 3, 1975 all three sites recorded oxidant levels above
the NAAQS (Figure 1).  Throughout this period increases in oxidants were
associated with winds from the NE and stationary high pressure systems.  De-
creases in oxidants were associated with passage of cold fronts from the NW,
warm fronts from the S, and high winds from the SW.
                        i'. .'• » IT 'e .'« ' I 11 1 4 n a It g. a x	i ' i i i 6 T . I .0 ' I \ . .» «

                                     DAT OF MONTM
 Figure 1.  Oxidant in pphm (24 hr average) recorded at three monitoring
            sites in southwestern Virginia for the period June 14 to July 16,
            1975.
                                     613

-------
     The July 1-5 episode was  experienced because of a low pressure  system
(hurricane Amy) being off the  Atlantic Coast, adjacent to New York.   A high
pressure system had moved down from Canada and was stationary over  the Great
Lakes states.  The air circulation around a low pressure system is  counter-
clockwise and clockwise around a high pressure system.   On July 2  this pattern
remained the same, forcing air from the NE south into Virginia (Figure 2).
By July 3 the low pressure system had dissipated; the high pressure  system had
moved to Virginia, allowing for further oxidant synthesis as is characterized
under these conditions.  On July 4 a cold front originating over Canada passed
over Virginia forcing the air of July 2 and 3 off the Atlantic Seaboard.
  Figure 2.   Weather map of meteorological  conditions leading to the oxidant
             episode of July 3 and 4, 1975  in southwestern Virginia.
     From July 1  to July 10 the mean severity index of needle injury of the
Site 1  field trees increased from 105.59 to 197.83 for the sensitive,  2.27 to
32.92 for the intermediate and from 0.41 to 0.64 for the tolerant.   The high
levels  of oxidant of July 3 resulted in a postemergence acute tipburn  to white
pine at all three sites (Table 1).


     1976:  Site averages of oxidant for the period April through July 1976
were 5.56 and 5.42 for Sites 1 and 2 respectively (Site 3 was not relocated
until late May).   The occurrence of increases and decreases of oxidant levels
in response to weather conditions corresponded well with the 1975 findings.

     June 4th marked the onset of an extended episode and from June 5 through
13 oxidant levels above the NAAQS were observed at all three stations for 110,
186, and 122 consecutive hours at Sites 1,2, and 3 respectively.  The same
sites recorded 189, 186, and 134 consecutive hours above 6.0 pphm oxidant
during this period (Figure 3a).
                                     614

-------
      TABLE 1.  SYMPTOM EXPRESSION OF WHITE PINE DUE TO OXIDANT LEVEL


    Tree Class     June 20     July 1     July 10     July 21
1975
Tolerant
Intermediate
Sensitive
.21
1.61
65.63
.41
2.02
105.59
.64
32.82
197.83
.91
34.24
227.16
                   June 22     July 2     July 12     July 22     July 31
1976
Tolerant
Intermediate
Sensitive
0.0
1.08
25.21
0.0
1.28
40.80
0.0
13.71
102.64
0.0
19.60
105.74
0.0
24.45
109.58

     The June 5th through 12th episode resulted from winds out of the NE on
the 5th and 6th and a stationary high pressure system that continued through
the 12th.  The 8th through the 10th had little to no air movement with in-
creasingly clearer skies.  On the llth and 12th, winds were again out of the
NE and the high pressure system persisted.  The 13th and 14th brought the
episode to a close; winds were out of the S and the high pressure system
dissipated (Figures 3b-3j).

     Oxidant injury to the field trees as recorded during the summer of 1976
at Site 1 is recorded in Table 1.  From July 2 to July 12 the mean severity
index of needle injury increased from 40.80 to 102.64 for the sensitive, 1.28
to 13.71 for the intermediate.  Similar increases in oxidant injury were not
observed on the tolerant trees at Site 2.

     At Site 1, the total tree evaluation system showed a marked decrease in
total scores (increased symptoms) for the sensitive and intermediate trees.
There was an average decrease of 11% for sensitive and 12'^ for intermediate
trees.   The tolerant trees showed less than 1" decrease in total score for
the period June 10 through July 31.   For the same period, the white pines
                                    615

-------
Figure 3.   (a) Oxidant in pphm (24 hr average) recorded at three monitoring
           sites in southwestern Virginia for the period of June 1-31, 1976.
           (b-j) Weather maps of meteorological  conditions during the oxidant
           epis'ode of June 5-12, 1976.
                                      616

-------
along the Blue Ridge Parkway  also  showed  a decrease in total  tree score.
There was an average decrease of 37.5%  for sensitive,  22% for intermediate,
and 10% for tolerant trees.
                                  DISCUSSION

     Data before the mid  1960's  indicate  that such  high levels of oxidant were
not previously observed in  rural  environments (5,12).   While working in south-
western West Virginia  in  1962, Berry  (5)  recorded levels 2 to 3 times less
than those recorded in 1975  (Figures  4a-4b);  note the  similarity in diurnal
oxidant patterns.  Berry  (personal  communication, 1976) indicated that his
Mast Ozone Meters were corrected  using  2% buffered  KI  along with air flow
calibration techniques.   Therefore, the observed increased level of oxidant
should be considered as reasonable  estimates  while  taking into account site
location and year to year variation.  The high levels  of oxidant reported in
this study confirm the findings  of  other  researchers  (1,2,3,7,8,9,13).

       	SITED, I9JUNE-IOJULY, 1975
       	SITE I, I9JUNE-IOJU-Y, 1973
       	BERRY, 1964         ,	— ^v
      1 1  1 1 1 1 1
                             I I I I I
      I 2 3456789101112,1 2 3456789IOIII2
            AM      I       PM
                 TIME OF MY
                                                      aTEB, 60UNE-7JULY, 1975
                                                      BERRY, 1964
                                                  I 23456789K>III2|I 2345678910118
                                                        AM      I       PM
                                                             TIME OF CWY
  Figure  4.   (a)  Comparison of 1962 oxidant levels as recorded by  Berry  (5)
             and  1975 oxidant levels as recorded at two moutain  locations  in
             southwestern Virginia,  (b) A similar comparison valley  sites
             1964 vs. 1975.  Locations were different, monitoring  dates
             were identical.

     The data collected  in our study  further  indicate transport of oxidant and
oxidant precursors from  the  urban  environment to the  rural  environment under
appropriate meteorological conditions.  Moreover,  under centaln conditions
high levels of pollutants  can be transported  several  hundred miles as is the
case reported here.  Under the influence  of the determinate meteorological
conditions proceeding July 3, 1975, it is  probable that air pollutants were
transported into southwestern Virginia from areas  as  far away as New York
City.  This is not a new phenomenon.   On  July 2,  1974,  the  air was moving the
opposite direction,  to the NE and  high levels of ozone
Connecticut and Massachusetts from New York City (8).
tions, the oxidant levels  in southwestern  Virginia are
environment as was shown by  the association of oxidant
from the northeast.
                                                        were transported into
                                                        Under adverse condi-
                                                        influenced by the urban
                                                        increases and winds
     Increases in oxidants  during  the  investigation were also observed under
high pressure systems.  The weather  conditions  occurring during the episode
                                     617

-------
of June 1976, were winds out of the NE and an associated high pressure system.
High pressure systems were characterized by disorganized wind patterns and
clear to partly cloudy skies.   Under these conditions, there is no movement of
air from a given geographical  region and photochemical generation of oxidants
is increased.

     It has been hypothesized that the type of reactions generating oxidants
in the rural atmosphere are similar to those generating oxidants in the urban
atmosphere (16,17).  However, natural  synthesis alone in the rural environment
could not singly account for the levels of oxidants being reported.   The urban
reaction system has a far greater emission intensity of oxidant precursors
that are responsible for the undesirable effect of photochemical smog.

     The urban reaction system is more likely the contributor of stable inter-
mediate precursors that would contribute to oxidant synthesis in the rural
environment.  Photochemically, oxidants are generated by the reaction of oxides
of nitrogen  (NOX) and hydrocarbon (HC) (18).  High levels of oxidants are
generated when NO is depleted in the presence of sunlight.   As air moves
from the urban to the rural environment, the HC to NOX ratio increases.  As
this ratio increases, the oxidant dosage increases (11).  This appears to be
the case in the rural environment.   If these premises are correct, then we
are dealing with a photochemical oxidant generation system in southwestern
Virginia that is highly dependent on the sun's energy and should show system-
atic increases and decreases over large geographical regions.  This is the case
in southwestern Virginia.

     The increase in symptom severity index observed on July 10, 1975 was six-
teen fold for the intermediate trees.   This appeared within one week of the
episode peak.  This would seem to indicate a direct response to the high levels
of oxidant occurring July 3 and 4.

     The increase in symptom severity index recorded on July 12, 1976 was
elevenfold for the intermediate trees.  This was recorded approximately four
weeks after the June episode.   Symptom development has not been observed to
require as long as 4 weeks following acute exposure, and this leads us to
believe that this symptom expression may be an accumulative oxidant response,
indicating a low dose over time.

     Both the 1975 and 1976 increases  in symptom expression were observed
during the 2nd week in July.  Berry (6) noted tissue age as a factor in symp-
tom expression.  Assuming needle extension to occur at about the same time in
1975 and 1976, one could correlate this increased expression to the age of the
needle tissue at the time of dosage.

     The total tree evaluation system indicated a decrease in average tree
scores (increase in symptoms)  at both  Site 1 and along the Blue Ridge Parkway.
The trees along the Parkway showed a greater increase in symptoms than did the
others.   This could indicate the presence of higher oxidant levels in the Blue
Ridge versus the southern Appalachian  Mountains.  Age could also be a factor;
the trees along the Parkway were mature while those rated at Site 1  were part
of a 16-year-old stand.  The sample along the Parkway included only one sensi-
tive tree.   This may explain the particularly large decrease '•(37. B%) in the
average sensitive tree score there.

                                    618

-------
     Severe damage to white pine has been observed in this study and statisti-
cal analysis has significantly correlated the 1975 damage with increased oxi-
dant concentration during specific time intervals.


                                 CONCLUSION

     Oxidant levels were monitored for 15 months and distinct weather patterns
were correlated with high levels in the mountainous areas of southwestern
Virginia.  Winds out of the NE were involved in all pollution episodes.   While
symptom development on native and planted white pines continued throughout the
entire growing season, definite increases in severity were associated with the
occurrence of an episode of oxidant increase.


                                 REFERENCES

 1.   Anonymous.   Mount Storm,  West Virginia- Gorman,  Maryland and  Luke,
      Maryland-  Keyser, West Virginia.   Air  Pollution  Abatement Activity.
      Envir.  Prot.  Agency  Rep.  Air  Pollution Control  Office  Pub!. No.  APTD-0656,
      1971.
 2.   Anonymous.   Investigation  of  High  Ozone Concentrations  in  the Vicinity
      of  Garrett  County, Maryland,  and Preston County, West  Virginia.  Envir.
      Prot. Agency  Rep.  R4-73-019,  1973.

 3.  Anonymous,  Investigation of Rural Oxidant Levels as Related to Urban
     Hydrocarbon Control Strategies, Envir.  Prot. Agency Rep. 450/3-75-036,
     1975.

 4.  Berry, C. R., and L.  A. Ripperton, Ozone, a Possible Cause of White Pine
     Emergence Tipburn, Phytopathology, 53:552-557, 1963.

 5.  Berry, C. R., Differences  in  Concentrations of Surface Oxidant Between
     Valley and Mountaintop Conditions in the Southern Appalachians, J.  of the
     Air Poll. Control Assoc.,  14(6) 238-239, 1964.

 6.  Berry, C. R., Age of Pind  Seedlings with Primary Needles Affects Sensi-
     tivity to Ozone and Sulfur Dioxide, Phytopathology 64:207-209, 1974.

 7.  Cleveland, W. S., and B. Kleiner, Transport of Photochemical Air Pollu-
     tion from Camden-Philadelphia Urban Complex, Envir. Sci. and Tech., 9:
     869-872, 1975.

 8.  Cleveland, W. S. and B. Kleiner, Photochemical Air Pollution:  Transport
     from the New York City Area into Connecticut and Massachusetts, Science
     191:179-181, 1976.

 9.  Coffey, P.  E. and W.  N. Stasiuk, Evidence of Atmospheric Transport of Ozone
     into Urban Areas, Envir. Sci. and Tech. 9:59-62, 1975.

10.  Costonis, A. C. and W. A.  Sinclair, Relationships of Atmospheric Ozone
     to Needle Blight of Eastern White Pine, Phytopathology, 59:1566-1574,
     1969.

                                     619

-------
11    Heicklen,  J.,  Photochemical  Smog:   Its Cause and Cure,  Perm.  State CAES
     Pub.  No.  299-73,  1973.

12.   Junge, C.  E. ,  Air Chemistry  and Radioactivity,  Intern.  Geophys.  Ser.  4,
     New York,  Academic Press, 1963.

13.   Miller, P.  R., M. H.  McCulchan, and H. P.  Milligan, Oxidant Air Pollution
     in the Central Valley,  Sierra Nevada Foothills, and Mineral King Valley
     of California, Atmos.  Envir.  6:623-633, 1972.

14.   Miller, P.O.,  Oxidant-induced Community Change in a Mixed Conifer Forest,
     Advances in Chemistry Series 122, American Chem. Soc.,  Wash., D. C., 101-117,
     1973.

15.   Morris, C.  L., Ozone Damage  to Eastern White Pine in Western  Virginia,
     Virginia Division of Forestry,  Forest Pest Survey,  1973.

16.   Photochemical  Smog and  Ozone Reactions, Advances in Chemistry Series  113,
     American Chemical Society, Washington, D.C., 1972.

17.   Ripperton,  L.  A., H.  Jeffries,  and J.  B.  Worth, Natural  Synthesis of
     Ozone in the Trophsphere, Environ.  Sci. and Tech.,  5(3):246-248, 1971.

18.   Stephens,  E.  R.,  Chemistry of Atmospheric Oxidants, J.  of the Air Poll.
     Contr. Assoc.  19(3):181-185, 1969.
                                     620

-------
                                                                              12-3
               THE EFFECTS  OF OZONE ON PLANT-PARASITIC NEMATODES
                 AND CERTAIN  PLANT-MICROORGANISM  INTERACTIONS

                                  D. E. Weber*
ABSTRACT
     The. i.n^Lue.nc.e. o& CUA  pollutant* on paAa*itL£**ing popu-
lation*  o£ both j$emo£e and mote, nvmatodu, and  the. total number, ofa zgg* and
          Tfie.atmwt ofi *oybe.an with 03 and the. 03-S02 mtxtuAe,, gwatly *up-
         the. fie.pfioducti.on o& *tubby fioot
Re.pfioducti.on o^ *ting nmatode^  (Be.lonola4jmu*_ longi-caudatu* )  wa* u*uatly not
afafie.cte.d by the. pollutant*.   The. QfiOMth oft mmatode.~tn^e.*te.d *oybe.an*  wo*
i,nhib  ^na.(\ajiMH\ .  Plants weAe, e.xpo*e.d to the. pollu-
tant* be^ofie. and a&teA. nematode.  inoculation*.  FolioA. injuny fie^ulting fjfiom
(}3 and the. 03-S02 mi.xtuAe. pfu.oA  to inoculation wa* a**oc^iate.d with  a pronounced
*upptie.**i.on o£ nejnatode, fie.pfio auction.


                                  INTRODUCTION

      Oxidants and other air  pollutants injure plant foliage and inhibit the
growth and productivity of plants.   Air pollutants also affect interactions
involving vegetation and microorganisms (1,2).  Field and laboratory investi-
gations  have shown that ozone (03) influences interactions involving fungi,
bacteria,  and viruses associated with a large number of plants (Table  1).
These effects range from an  increase in the infection of plants by  such patho-
gens  as  Botrytis cinerea (3)  and Fomes annosus (4) to an inhibition in the
''U.S. Environmental Protection  Agency, Corvallis, Oregon.


                                       621

-------
            TABLE 1.   THE EFFECT OF OZONE  ON  INTERACTIONS  INVOLVING
                           PLANTS AND MICROORGANISMS

Organisms
FUNGI
Fungi on:
leaves of
pinto bean
roots of
pinto bean
Botrytis on:
potato
potato
geranium
broad bean
Fomes annosus
Experimental
conditions


greenhouse
greenhouse
greenhouse
field
greenhouse
greenhouse
on
Ozone
concentration


.06 yl/1, 8hr/
day, 28 days
.1-.15 yl/l,8hr/
day, 4-28 days
.15-. 25 yl/1
6-8 hr
ambient air
.07-.! yl/1
10 hr/day
15-30 days
.20 yl/1 , 8 hr
ambient air
ponderosa pine field
Fusarium in
tomato
Puccini a on:
oats
wheat

wheat
Uromyces on
ointo bean
greenhouse
greenhouse
growth
chamber

growth
chamber
greenhouse
.06-. 08 yl/1
6 hr/day, 5 days/
wk, 45 days
.1 yl/1, 6 hr/
day, 10 days
.06 yl/1 ,6 hr/
day, 3-17 days
.06 yl/1, 6 hr
.30 yl/1, 5 hr
ozonated olefins
.4-. 5 yl/1, 3 hr
Effect Reference


Increased number of
fungi
Same
Increased severity &
deveopment of disease
Same
Increased infection
on injured leaves
Less foliar injury
by Oo on leaves
Colonization by
fungus increased
Suppressed plant nrowth
and delayed wilt
symotoms
Inhibited fungal qrowth
Inhibited funqal growth
& spore oroduction
Infection inhibited
Protected areas of
leaves from 0^ injury
Same


14
8
3
3
15
16
4
17
5
6
6
13
11
(continued)
                                    622

-------
Table 1 (continued)
Organisms Conditions
Erysiphe on:
barley


Aureobasidium
on white pine
Lophodermium
on white pine
BACTERIA
Pseudomonas in:
soybean
kidney bean
Rhizobium
nodules on:
pinto bean
Lee soybean
soybean
VIRUSES
Bean Common
mosaic in bean
Tobacco etch
in tobacco
Tobacco mosaic
virus in:
tobacco
pinto bean

growth
chamber


growth
chamber
growth
chamber

growth
chamber (?)
greenhouse

greenhouse
greenhouse
greenhouse

growth
chamber
field
growth
chamber


growth
chamber
growth
chamber
Concentration

.05 yl/1 ,6 hr/
day, 3 days
6 hr/day
for 2 days
6 hr/day, more
than 2 days
.07 yl/1
4.5 hr
.07 y/1
4.5 hr

.25 yl/1,
4 hr
.60 yl/1,
1 hr

.1-.15 yl/1, 8
hr/day, 4-28 days
.75 yl/1,
1 hr
.12 yl/1
5 days/wk, 3 wk

.25 yl/1, 4 hr
ambient air
.25 yl/1, 4 hr


.30 yl/1, 3-6 hr
.10 yl/1, 3 hr
Effect Reference

Leaf infection by
condia inhibited
Fungal growth
inhibited
Fungal growth
enhanced
Fungus on leaves
injured by 0^ only
No effect on colon-
ization of host

Fewer bacterial
lesions
Areas near lesions pro-
tected from 0~ injury

Nodules absent
Lower numbers and
weights of nodules
Fewer nodules

Inhibited development
of 03 injury of leaves
Same
Same


Same
Increase in number
of viral lesions

19
19
19
20
20

7
12

8
9
10

21
21
22


13
23
623

-------
 development of rust fungi  (5,6), and a bacterial pathogen (7).   The results of
 several studies have indicated that root nodule formation by nitrogen-fixing
 bacteria is inhibited when leguminous plants are exposed to 03  (8,9,10).   In
 other research, microorganisms have been implicated in the protection of leaf
 tissues from 03 injury (11,12,13).

     Research involving the effects of 03 on plant and microfauna associations
 has not been found.  Plant-parasitic nematodes were utilized to study the
 effects of 03 on one aspect of plant and microfauna interactions.

     Nematodes comprise a group of about 500,000 species.  They are present in
 soil and aquatic environments in large numbers.  Some are important parasites
 of plant and animals.   Many feed on microorganisms and decaying materials.
 They transport bacteria and fungi within the rhizosphere and are involved in
 the dissemination of plant pathogens.  Nematodes provide a source of nutrients
 for organisms and may be involved in the mineralization of organic matter (24,
 25).  This paper reports the results of one study to determine  the influence
 of 03, singly and in a mixture with sulfur dioxide (S02), on soil and foliar
 pi ant-parasitic nematodes.


                                 PROCEDURES

 SOIL NEMATODES

     The parasitism of soybean [Glycine max L. (Merr.) 'Dare']  by soil nema-
 todes was investigated in greenhouse facilities.  Seedlings of  soybean were
 grown from seed in sand.   Roots were inoculated with a commercial inoculum of
 nitrogen-fixing bacteria CRhizobium japonicum).  Seedlings were then exposed
 to 0.25 microliter/liter 03 and S02, singly and in mixture, and to charcoal-
 filtered air in 12.5 m3 exposure chambers.   Four-hour exposures were made
within a 1-week period prior to nematode inoculations and on three alternate
 days/week after inoculation until experiments were terminated.   Three to five
 plants were utilized in each treatment.

     Fifteen days after planting, soybean seedlings were inoculated with nema-
 todes.   The data summarized in this paper are from single experiments, involv-
 ing both endo- and ectoparasitic species.  Similar results were observed in
 replications of these experiments.  In experiments involving a  sedentary endo-
 parasite, roots of test plants were inoculated with 9000 eggs per plant of the
 soybean cyst nematode (Heterodera glycines).  Two ectoparasites were studied in
 related experiments.  In these experiments, roots of soybean were inoculated
with 400 sting (Belonolaimus lonqicaudatus) and 100 stubby root (Trichodorus
 christiei) nematodes per plant.

     Soybean cyst nematode experiments were terminated 36 and 67 days after
 inoculation.  Female cysts were washed from roots.  Twenty cysts per sample
were crushed to evaluate the production of eggs and larvae.  Males were ex-
 tracted from samples of sand (26).  Nematodes and eggs were counted and values
converted to per plant totals.
                                     624

-------
     Experiments with sting and stubby root nematodes were terminated 48 and
79 days after inoculation.  Population counts per plant were determined from
sand samples.

FOLIAR NEMATODES

     Schwabenland Red begonia was used in the evaluation of the effects of in-
jury by 03 and S02 on the foliar nematode, Aphelenchoides fraqariae.   Three
begonia plants were exposed to each pollutant at 0.25 microliter/liter concen-
trations and to filtered air as described for one, 4-hour exposure period.
Utilizing two groups of plants, pollutant exposures were conducted 3 days be-
fore leaves were inoculated with nematodes or 3 days after inoculation.  Four
leaves per plant were inoculated with 150 nematodes per leaf.

     Thirty days after inoculation, nematodes were extracted from leaves and
counted from diluted samples.  The amount of injury on each inoculated leaf
was estimated on a 0 to 100% scale 3 days after exposure of each plant group
and 30 days after inoculation.


                           RESULTS AND DISCUSSION

     The influence of 03 and the 03-S02 mixture on the soybean cyst,  stubby
root, and foliar nematodes was inhibitory.  This inhibition was probably
caused by a suppression in the growth of the host, foliar injury, and various
metabolic changes associated with the host-parasite-pollutant interaction.
The effect of S02  alone on nematode populations, plant weights, and the forma-
tion of nodules was similar to the control.

EFFECTS ON SOIL NEMATODES

     The inhibitory effects of 03 and the 03-S02 mixture were first observed
36 days after inoculation, where the number of female cysts/plant was slightly
lower than the control.   At 67 days, the inhibitory effect was much larger and
amounted to a 73 to 77% suppression in the number of cysts/plant as compared
with the control.  Treatment of soybean with 03 and the 03-S02 mixture also
inhibited the production of male nematodes and the total number of eggs and
larvae (Table 2).

     Damage of soybean by 03 and the 03-S02 mixture may have depressed the
quantity and quality of organic and inorganic substances.  Since the growth
of shoots and roots was inhibited by these pollutants (Figure 1), the chemical
constituents of soybean were probably altered.  These substances are utilized
by nematodes in the establishment of a feeding relationship with the host.
If plant chemistry is modified, the growth and reporduction of nematode popu-
lations may be impaired.  Tingey (27) discussed various mechanisms that may be
involved in ozone-induced modifications of the chemistry of roots and foliage.
In addition, the production of toxins in roots of plants damaged by 03 and the
03-S02 mixture may have influenced the growth of nematode populations.

     Formation of nodules in roots of soybean by nitrogen-fixing bacteria was
not affected by the pollutants.  The effects of the pollutants on nodulation


                                     625

-------
           TABLE  2.   INFLUENCES  OF 03  AND  S02  ON  THE  SOYBEAN  CYST NEMATODE

          	AT  67 DAYS  AFTER  INOCULATION	

                                          Numbers  per plant       ______
Treatments
Control
so2
°3
Mixture
Females (Cysts)
1,955
1,738
525*
451*
Males
2,656
2,239
356*
555*
Eggs and larvae
146,013
112,220
43,573*
46,321*

  *Significantly  different  from the  control  at  P=0.05.
                    16
in
I
(T
e>
 i
i-
e>
LU
                  x
                  en
                  LU
                    '2
                          '
       P
       I
       iF
                                   1
                                  i!
                                                 LSD 0.05
     JBS
     i
     &:<
     £&
                              SHOOT  ROOT

                         CONTROL
                SULFUR
                DIOXIDE
OZONE
MIXTURE
                                    TREATMENTS
   Figure 1.  The influence of 0. and S02 on growth of soybean infected with
                        soybean cyst nematodes.

were probably masked by the severe inhibition  of this process  by  soybean  cyst
nematodes (28).

     Since 03 is highly reactive, it is doubtful that this  pollutant would
have a direct effect on soil organisms.  This  has been confirmed  in experi-
mentation which demonstrated that 03 does not  pass  through  sand,  peat, and
gravel soil mixes.

     The influence of 03 and the 03-S02 mixture on  the two  ectoparasites  was
different.  Population counts of sting nematodes were usually  not  significantly
affected by exposure of soybean to the pollutants.  However," the  exposure of
soybean to 03 and the 03-S02 mixture greatly inhibited the  stubby  root nematode.
                                      626

-------
This inhibition was observed  at 48  and  79  days  after inoculation and ranged
from 72 to 91% of the control  (Table  3).

     The growth of soybean  infested with  the ectoparasites  and the formation
of root nodules was suppressed by 03  and  the 03-S02  mixture (Figure 2).   These
effects of the pollutants had little  influence  on  the size  of the sting nema-
tode populations.  The  pollutants,  however may  have  influenced reproduction of

   TABLE 3.  INFLUENCES  OF 03 AND  SOQ  ON REPRODUCTION OF COMBINED POPULATIONS

                  OF  STING AND STUBBY  ROOT NEMATODES ON SOYBEAN
Treatments
  Days after
  inoculation
                            Nematodes per plant
                              jtinjj
                                          Stubby root
Control

SO
2

°3

48
79
48
79
48
79
2,794
14,029
2,428
11,009
2,802
11,276
19,748
25,547
16,190
23,604
3,589*
7,051*
Mixture


48
79
2,130
9,244*
1,791*
4,010*

* Significantly  different from the control  at P=0.05.
         70
CO


oc

 I
t-
X
(5
LU
         60


         50


         40


         30
      CO on
      LU 2O
          10


          0
•:&•:
1
        m
                                                        LSD 0.05
                              :i-:*;
                              •:•:•':
                              :';;.-»

if
'&$'
                     m
                                     SHOOT ROOT NOOULES -
                                 70


                                 60


                                 50
                                                        CO
                                                        ID
                                                        o
                                                      409
                  CONTROL    SULFUR DIOXIDE     OZONE

                                  TREATMENTS
                                         MIXTURE
                                                                30


                                                                20


                                                                10


                                                                0
                                                        o

                                                        o:
                                                        Lul
                                                        00
 Figure 2.   The influence of 0- and S02  on  growth  and nodulation of soybean
             infected with populations of sting  and stubby root nematodes.
                                      627

-------
 the stubby  root nematode by inhibiting the development of root tips.  Since
 root tips are more suitable feeding sites for this nematode than other root
 regions  (29), a decline in the number of apical areas could limit nematode
 feeding  and reproduction.  The sting nematode is less specific with regard to
 feeding  sites (30).

 EFFECTS  ON  THE FOLIAR NEMATODE

     Injury of begonia leaves by 03 and the 03-S02 mixture appeared to inhibit
 the reproduction of the foliar nematode.  This inhibition was associated with
 leaf injury induced by 03 and the 03-S02 mixture within 3 days after the expo-
 sure of  plants to the pollutants.  Injury resulting from the infection of
 leaves by nematodes during the 30-day period following inoculation also prob-
 ably contributed to the inhibitory response.   The inhibition of nematode repro-
 duction was greater when plants were exposed to 03 and the 03-S02 mixture before
 inoculation with nematodes than when plants were exposed to the pollutants
 after inoculation (Table 4).   The deviation in the amount of injury within the
 two inoculation treatments was probably the result of differences in environ-
 mental conditions (light, temperature, humidity) associated with the exposure
 periods, rather than a protective effect by the presence of nematodes.


     TABLE 4.  RESPONSES OF BEGONIA AND THE FOLIAR NEMATODE TO 03 AND S02


 Treatments              Leaf injury (% leaf)        Number of nematodes/leaf
 	3 days* 30 days#	

 Before inoculation

    Control                  0      48                     43,700

    S02                      0      50                     38,950
    03                      77      86                      9,000

    Mixture                 68      75                     13,400
 After inoculation
Control
S02
03
Mixture
Standard Error
0
0
58
38
4
48
58
72
61
6
45,100
43,300
34 ,500
27,900
5,200
* Injury was estimated on a scale of 0-100% at 3 days after exposure.
# Injury estimated on a scale of 0-100% when nematodes were extracted from
  leaves, at 30 days after inoculation.
     The inhibitory effects on nematode reproduction were probably related to
the destruction of nematode-feeding sites and conditions that made begonia
leaves less suitable as a host for this nematode.

                                     628

-------
                                 CONCLUSIONS

     These studies have provided an evaluation of the effects  of 03,  alone  and
combined with SO^, on the plant-parasitic association of both  soil  and foliar
nematodes.  Previous investigators demonstrated that many types  of  plant-
parasitic associations, involving fungi, bacteria, and viruses,  are influenced
by 03.   These investigations have shown that an additional  type  of  plant-
parasitic association, involving nematodes, is also influenced by 03.   In para-
sitic associations, 03 has been found to inhibit or enhance the  development of
microorganisms on plants.   Most results indicate that the effect probably in-
volves  an ozone-induced modification in the susceptibility of  the host or,  in
some cases, a change in the pathogenicity of the microorganism.

     There is a need to examine the broader aspects of these associations.
Microorganisms are required in the decomposition of organic matter.   The asso-
ciation of saprophytic fungi with foliage and roots have been  examined in only
a few studies.  In these studies, the number of fungi was increased by expo-
sure of plants to 03.  The increase appeared to be related to  the presence  of
injured tissues.   However, in laboratory studies, 03 was found to inhibit
several saprohpytic microorganisms.   Oxidants and other pollutants  may directly
interfere with the growth of microbial populations associated  with  the process
of decomposition.   In other instances, the inhibitory effects  of oxidants on
photosynthesis, plant growth, and other physiological processes  may influence
decomposition by suppressing the availability of organic materials.   By inter-
fering  with decomposition, 03 could alter the normal cycling of  nutrients with-
in ecosystems.  Other microbial processes are important components  of ecosys-
tems.   The suppressive effects of 03 on the formation of root  nodules by
nitrogen-fixing bacteria may reduce usable forms of nitrogen.  Also a disrup-
tion in the mycorrhizal association of fungi and plant roots could  inhibit
plant nutrition.
                                REFERENCES

        Heagle,  A.  S.   Interactions  Between Air  Pollutants and Plant
        Parasites.   Ann.  Rev.  Phytopath., 11:365-388,  1973.

        Babich,  H.,  and G.  Stotzky.  Air Pollution and Microbial Ecology.
        Critical  Reviews  in Environ. Cont., 4:353-388, 1973.

        Manning,  W.  0., W.  A.  Feder, I. Perkins, and M. Glickman.  Ozone
        Injury and  Infection of  Potato  Leaves by Botrytis cinerea.  Plant
        Dis.  Reptr.,  53:691-93,  1969.

        James R.  L.,  F. W.  Cobb, and J. R. Parmeter.  Effect of Photo-
        chemical  Air  Pollution on Susceptibility of Ponderosa and Jeffery
        Pine  to  Forties annosus.   In:  Proceed, of the Am. Phytopatological
        Society,  1976.  p.  108.

        Heagle,  A. S.   Effect of Low-level Ozone Fumigation on Crown Rust
        of Oats.  Phytopathology, 60:252-54, 1970.
                                     629

-------
 6.  Heagle, A. S. and L.  W.  Key.   Effect  of  Ozone on the Wheat Stem
     Rust Fungus.  Phytopathology,  63:397-400,  1973.

 7.  Laurence, J. A., and  F.  A.  Wood.   Ozone  Exposure Protects Soybean
     from Pseudomonas glycinea.   In:   Proceed,  of The Am. Phytopath-
     ological Society, 1976.   p.  105.

 8.  Manning, W. J., W. A. Feder,  P. M.  Papia,  and I. Perkins.  Influence
     of Foliar Ozone Injury on Root Development and Root Surface Fungi
     of Pinto Bean Plants.  Environ. Pollut., 1:305-312, 1971.

 9.  Tingey, D. T., and U. Blum.   Effects  of  Ozone on Soybean Nodules.
     J. Environ. Qua!., 2:341-342,  1973.

10.  Reinert, R. A., D. T. Tingey,  and C.  E.  Koons.  The Early Growth of
     Soybean as Influenced by Ozone Stress.   Agronomy Abstr., 63:148, 1971.

11.  Yarwood, C. E., and J. T.  Middleton.  Smog Injury and Rust Infection.
     Plant Physio!., 29:393-395,  1954.

12.  Kerr, D. D., and R. A. Reinert.   The  Response of Bean to Ozone as
     Related to Infection  by Pseudomonas phaseolicola.  Phytopathology,
     58:1055, 1968.

13.  Brennan, E., and I. A. Leone.   Suppression of Ozone Toxicity
     Symptoms in Virus-Infected Tobacco.   Phytopathology, 59:263-264, 1968.

14.  Manning, W. J., and P. M.  Papia.   Influence of Long-Term Low Levels
     of Ozone on the Leaf  Surface Mycoflora of  Pinto Bean Plants.
     Phytopathology, 62:497.   1972.

15.  Manning, W. J., W. A., Feder,  and I.  Perkins.  Ozone Injury Increases
     Infection of Geranium Leaves  by Bortrytis  cinerea. Phytopathology,
     60:669-670, 1970.

16.  Magdycz, W. P., and W. J.  Manning.  Botrytis cinerea Protects Broad
     Beans Against Visible Ozone  Injury.   Phytopathology, 63:204, 1973.

17.  Manning, W. J. and P. M.  Vardago.   Ozone and Fusarium:  Effects on
     the Growth and Development of  a Wilt-Susceptible Tomato and a Wilt-
     Resistant Tomato.  In:  Proceed,  of The  Am. Phytopathalogical
     Society, 1976.  p. 107.

18.  Heagle, A. S., and L. W.  Key.   Effect of Puccinia graminis f. sp.
     tritici on Ozone Injury  in Wheat.   Phytopathology, 63:609-613, 1973.

19.  Heagle, A. S., and A. Strickland.   Reaction of Erysiphe graminis f.
     sp. hordei to Low Levels  of Ozone.  Phytopathology 62:1144-1148, 1972.

20.  Costonis, A. C., and  W. A. Sinclair.  Susceptibility of Healthy and
     Ozone-injured Needles of  Pinus  strobus to  Invasion by Lophodermium
     pinastri and Aureobasidium pullulans. Eur. J. Forest Pathol. 2:65-
     73, 1972.

                                  630

-------
21.  Davis, D. D., and S.  H.  Smith.   Bean  Common  Mosaic Virus Reduces
     Ozone Sensitivity of Pinto Bean.   Environ. Pollut., 9:97-101, 1975.

22.  Moyer, J. W., and S.  H.  Smith.   Oxidant Injury  Reduction on Tobacco
     Induced by Tobacco Etch  Virus Infection.   Environ. Pollut., 9:103-
     106, 1975.

23.  Brennan E., I. A. Leone.  Interaction of Tobacco Mosaic Virus and
     Ozone in Nicotiana sylvestris.  J.   Air Pollut.  Cont. Assoc., 20:470,
     1970.

24.  Jenkins, W. R., and D. P.  Taylor.   Plant Nematology.  Reinhold
     Publ. Corp., New York, 1967.   270  pp.

25.  Nicholas, W. L.  The Biology  of Free-Living  Nematodes.  Clarendon
     Press, Oxford, 1975.   219  pp.

26.  Byrd, 0. W., Jr., C.  J.  Nusbaum, and  K.  R. Barker.  A Rapid Floata-
     tion-Seiving Technique for Extraction of Nematodes From Soil.  Plant
     Dis. Rep., 50:954:957, 1966.

27.  Tingey, D. T.  Ozone Induced  Alteration  in the  Metabolite Pools and
     Enzyme Activities of Plants,   pp.  40-57.   In:   M. Dugger, ed.; Air
     Pollution Effects on Plant Growth.  American Chemical Society
     Symposium Series 3, 1974.   130 pp.

28.  Lehman, P. S., D. Huisingh, and K.  R.  Barker.   The Influence of
     Races of Heterodera glycines  on Nodulation and  Nitrogen-Fixing
     Capacity of Soybean.  Phytopathology,  56:357-360, 1966.

29.  Russell, C. C., and V. G.  Perry.   Parasitic  Habit of Trichodorus
     christiei on Wheat.  Phytopathology,  56:357-360, 1966.

30.  Standifer, M. S., and V. G. Perry.  Some Effects of the Sting and
     Stubby Root Nematodes on Grapefruit Roots.   Phytopathology, 50:152-
     156, 1960.
                                 631

-------
                SESSION 13
  EFFECTS OF OXIDANTS ON VEGETATION - II
                   :  W.W. Heck
North Carolina State University at Raleigh
                    633

-------
                                                                            13-1
          GROWTH REPONSE OF  CONIFER SEEDLINGS TO  LOW  OZONE CONCENTRATIONS

                            R. G. Wilhour and G. E. Neely*

ABSTRACT
     Ju.ve.nile. &e.e.dling& o{,  Je^Aet/ p-tne (P-inuA j e.{,fire.yi. GAev. and
Monte Aet/ p-ine. (P.  radiata V.  Von] ,  there, p-tne  (P.  contoAta van., contorta
Vougi.  ex Load}', lodge.pole.  p-ine. (P.  c.ontorta VGA.  muMjq.ya.na. LBalfi. ] Critch{].
•6ugoA ptne ( P .  JLa.mbeJvtia.na  t?oug£.T7 wuteJtn white.  p-ine (P."~mon£ico£a Voagi. }
pondeAota pi.ne.  (P. pondeAota  Law*.}, Sitka 4pAuce  (Pieea &itcho,n&-U> '[Bong.]
Carr.}  and Voaglat 1-ai  ( PA zudotMLga me.nz>i&>-ll  [M^ibTJT^anco . )  w)2Ae exposed
io a AtandaJidiz&d dotty ozone, (03)  do^e (an aueAage ojj 10 pphm  (200 pg/m3),
           ciat/)  4eueia cfcu/4 peA wee.fe /en oatdooA  e.xpoAuAe chambe/Li.  COH^AO^S
     tne,at&d with c.hajicoaJt- ^UJiojuLd ambient OAA.   Pot& weAe placed ancfeA  the.
               'Le.atmen^s i,mme.diateJty faolloui^ng  ieedcng and AemowLneci tkane.  up
to 22
     Gnowtk ti&spon&o, o& tke.  nine, aoni-fieA. 4pe(U.e^> wai  euaXaated faottowi.ng  IS
continuous  wnnkA ofa expo^uAe to  03.   Signifi.ic.ant gAowtfi Aecfu.ctxon4 Ae/ated
to t^ie  03 tAe.atme.nt u)eAe ^.e£ati.vely con^-i-ittint oniy |$OA pondeAoia p^.ne  and
we^teAn white, p-ine,,  Stgnifilcant growth K.zductionA  (g-iv&n 04 I Aedactton  fjAom
control) weAe ob^eAued x'n Aoot length (12.11}*, dn.y weight  (PW) , 4tem  (21.31)*,
and W  Aoot (26.31)** Of, pondeAo^a ptne and W Cottage (12.91)** and W item
(9.?|)* tn  W2xiteAn white p-cne.   Tne proposition o^ totaJL potion. ^uA^ace  tn/ated
on pondeAo^a ptne exposed to 03  wai  Ae£ative£y ntg/t (201)  bat ptiactic.all.y
ab&mt  on wuteAn white, p-oie (71).   No constant tieJLationi>hip between ^o-tiai
i.njuAy  and  growth Ae^pottie to 03 woci ob^eAved between ^peaie^.  The 'ie£ati.ue
^o^taA  tnj'uAt/ Ae^ponie o^ tne c.ont|$eA4 u)O4 pondeAo^a  ptne (201), £odgepo£e
p-uie.  (16%},  4hoAe ptne  (731),  Je^Aey ptne (21), VouglaA fiiA  (71), 4u.gaA  ptne
(II), we^teAn white, pine.  (1%} , Sitka 4pAu.ce  (71), and MonteAey ptne  (01).
Ttp necAo4t4 OA chlofioAiA weAe the mo^t common &ymptam& o{^ 0^ i.njuA.y to
primary need£e^; typical chlorotic. mottitng wai ob-ieAued on a ^e
need£e4 o^  ponderoAa ptne.
    The  e^ect o^ (?3 on pondeAo^a ptne -deeding  emeAgence i-ua^ examined  b(/
counting  the numbeA o^ emeAged  4eed£tng4 tn (?3 and  contAo^ treatments on  a
b>uoeeftt(/  ichedu£e beginning  two weekt, a^te.r ieeddig.   S-igcu^-ccant** dx,j$-
          -en  ^eedfoig emergence  befween f*;c'.atmenti  ^ir^t appealed faour
      a{)ter  tending.  Ai'  f/ie  fc'Amaiatani o^ fhe  emergence expct-cwicnf, a
greater number u{] 4 
-------
daily 03 dose, fax. 0,  1,  2,  3,  4,  5,  6,  OA  1  day*  peA weefe fati IS continuous
o>eefe4.  Only W 4tem  (16%}  woi significantly**  Aeduced friom the. control by
03; t/tos dxLJ^eAence woi  obseAve.d  at  the. 6  {19%} and  7 (771)  cta^s peA weefe
03 tAeotmenta.  Fo-tcoA Injury Mas slight between  0,  I,  2,  and 3 dai/4 ofi 03
expo^uAe peA week (aveAaged 2.41) bat tncAeaied moAke.dly at e.xposu)tcs lotting
4 day* peA week OA longeA  (aveAaged  771).
              hanve^ts o& pondeAo-ia ptne se.e.dlings throughout a 10 week te^t
peAlod showed no statlAtlcalty A-ignifi-icant growth AzductionA  in plant height
OA fioot langtk, and no di^&ie.nc.eA weAe expected ^Aom  ^uA^he^t expo^uAe. to
    03 do4e ctied -en tn^6 expeAtment.  Highly significant tfinatm^nt dl^oji-
           not detected tn PW top; howjeueA, a  distinct fiend  o
-------
     This research was designed to investigate the relative growth response
of several forest tree species to low 03 doses applied over a long time and
to intensively investigate the growth of a selected tree species growing
under such a pollutant schedule.

                              METHODS AND MATERIALS

     Nine western conifer species (Jeffrey pine,  Monterey pine,  shore pine,
lodgepole pine, sugar pine, western white pine, ponderosa pine,  Sitka spruce,
and Douglas fir) were used in the study.  Seeds were stratified  four weeks at
5°C.  and planted in Jiffy-Plus potting medium contained in 10 x  30 cm high
pots.  Four seeds were arranged at each of four predetermined locations in
each pot and covered with 3-4 mm potting medium.   A total of 496 pots of pon-
derosa pine and 36 each of the remaining eight species were evenly assigned
to six outdoor exposure chambers (12).  Species and treatments were randomly
located in six blocks within each test chamber.  The experimental treatments
commenced immediately after seeding in mid-June and continued into November.

     Ozone produced by irradiating charcoal-filtered ambient air with high
intensity ultraviolet light was maintained in three test chambers an average of
10 pphm (200 ug/m3), six hrs. per day (9:30 am-3:30 pm) and seven days per week
for 22 weeks.  The 03 concentrations were allowed to follow a pattern of low
(8±1  pphm) in the morning hours and gradually increasing to 12±1 pphm later
in the day.   Ozone concentrations were continuously monitored with Mast meters
(Mast Development Co., Ames, Iowa) calibrated weekly with a MEC  1000 ozone
calibrator (McMillan Electronics Co., Houston, Texas).  The MEC  calibrator
was periodically compared to the standard 1% neutral-buffered potassium iodide
method.   Charcoal-filtered ambient air was passed into three remaining chambers
that housed control  plants.

SEEDLING EMERGENCE

     The effects of the 03 treatment on the emergence of ponderosa pine
seedlings were evaluated by counting the number of emerged seedlings on a
biweekly schedule for six weeks.  All treatments  were analyzed that included
exposure of ponderosa pine to the standardized daily 03 dose (10 pphm, six
hrs.  per day) seven days per week for six weeks;  appropriate controls were
also observed.  The first observation was two weeks after seeding.  The few
seedlings that emerged after the six week observation period were not included
in the emergence data.

     Seedlings were thinned on the same biweekly schedule referred to above.
The four most uniform seedlings per container were retained, one at each of
the predetermined positions.

SPECIES SENSITIVITY

     Growth of the nine species to 03 was evaluated in mid-November following
20 continuous weeks of exposure to 03.  The seedlings were carefully removed
from the potting medium and roots were washed clean with tap water.  Root
length was measured as the distance between the root collar at soil level and
the primary root tip.  Plant height was measured as the distance between the

                                      637

-------
cotyledons and the tip of the apical bud.  Foliar injury response was approxi-
mated by estimating the percent of the foliar surface area injured.

     Stems, roots, and foliage were bagged separately by four plant (pot)
samples and dried at 70°C for 72 hrs; dry weights (DW) were recorded.  All
response parameters were recorded as a four plant (pot) average.

VARIABLE 03 TREATMENT

     Effects of variable 03 exposure durations on ponderosa pine were tested
by rotating 18 pots per treatment weekly between treatments.   Seedlings were
effectively exposed to the standardized daily 03 dose for 0,  1, 2, 3, 4, 5,
6, or 7 days per week for 18 continuous weeks.  The response  parameters
referred to earlier were measured at the termination of the study.  The
statistical treatment of the data in this study and the previous study in-
cluded an analysis of variance and an F-test to detect treatment differences
at the 90, 95, or 99% confidence level.

PONDEROSA PINE GROWTH

     The growth response of ponderosa pine seedlings over time to the daily
03 schedule referred to earlier administered seven days per week was estimated
by biweekly harvests beginning two weeks following seedling emergence and
ending 18 weeks later.  The parameters measured were plant height, root length,
DW top (stem + foliage), and DW root.  The data were plotted  and a non-linear
least squares method was used to fit the curves.  The growth  curves of outside
diameter (OD) top and OD root showed a divergence between treatments with
successive harvests indicating that treatment differences would most likely
appear at later harvests.  The last four harvests, representing 14,-16, 18,
and 20 weeks of exposure to 03, were chosen for regression analysis.  The 20
week treatment to 03 in this experiment is identical to the 03 treatment of
ponderosa pine in the species sensitivity test.

                             RESULTS AND DISCUSSION

SEEDLING EMERGENCE

     Results of the experiment on the influence of 03 on seedling emergence
demonstrated that rate of seedling emergence was affected by  03 exposure.
The number of seedlings emerged at the three observation periods was always
greater in the 03 treatment (Table 1).  Significant differences in seedling
emergence between treatments first appeared at four weeks after seeding (95%
confidence level).  At six weeks, seedling emergence in the 03 treatment was
higher than controls at the 90% confidence level.

     How 0-j treatment caused an increase in both rate of ponderosa pine
seedling emergence and total number of seedlings emerged cannot be inferred
from this experiment.  Blum and Tingey (4) demonstrated that  03 did not pass
through sand, peat, and gravel soil mixes; this suggests an indirect effect
of 03 on ponderosa pine seedling emergence.  The effect may be related to
altered soil chemical or microbiological components mediated  by 03 reacting
with the surface layer of the potting medium.

                                     638

-------
SPECIES SENSITIVITY

     Response of the nine test species to 20 weeks of treatment to the
standardized daily 03 dose (10 pphm for 6 hr) for 20 continuous weeks was
evaluated in detail.  Average growth and injury values are presented in
Table 2.  Ozone seriously injured foliage of ponderosa pine, lodgepole pine,
and shore pine.  The relative foliar response based on differences in % foliar
injury between treatment and control plants was ponderosa pine (20%), lodge-
pole pine (16%), shore pine (13%), Jeffrey pine (2%), Douglas fir (1%),
sugar pine (1%), western white pine (1%), Sitka spruce (1%), and Monterey
pine (0%).  Necrosis of needle tips was the predominant symptom of 03 injury.
Cotyledons and oldest primary needles were most severely affected.  Chlorotic
mottling occasionally occurred on secondary needles of ponderosa pine.  The
tip necrosis symptom in this study resembles that described by Davis and
Wood (9) as occurring on primary needles of several conifers treated with
acute 03 exposures (25 pphm 03 for eight hrs.); they also observed chlorotic  .
mottle.  A comparison of the results of the studies indicates that tip necrosis
or chlorotic mottle may follow chronic or acute exposure of conifers to 03.
Neither symptom is, therefore, necessarily specifically related to chronic
or acute 03 exposures.
        TABLE 1.  EFFECT OF CHRONIC OZONE EXPOSURE ON SEEDLING EMERGENCE
                             OF Pinus ponderosa Laws

Treatment
Control


Ozone


Time after
seeding
(weeks)
2
4
6
2
4
6
Emerged
seedlings
(number)
1279
1457
1516
1299
1545**
1589*
Germination
(%)
37.0
42.2
43.9
37.6
44.7
46.0

         * Significantly different  from  control at  90",  confidence  level
           (Chi Square test)


        ** Significantly different  from  control at  95%  confidence  level
           (Chi Square test)
                                     639

-------
     Significant growth reductions occurred in ponderosa pine and western
white pine.  Significant growth reductions in ponderosa pine due to 03 treat-
ment were root length (12%)*, DW stem (21%)*, and DW root (26%)** (Table 2).
In western white pine, DW foliage (13%)** and DW stem (9%)* were significantly
reduced by the 03 treatment.  Interestingly, one of the two species showing
significant growth reductions would be ranked very sensitive (ponderosa pine)
based on foliar injury and the other (western white pine), resistant.  Growth
of other species showing a relatively high amount of foliar injury (lodgepole
pine and shore pine) was not significantly affected by the 03 treatment.
Evidently visible injury is not a reliable predictor of growth response of
conifer species to chronic 03 exposures.

VARIABLE 03 TREATMENTS

     Because of the importance of ponderosa pine as a timber species  and the
knowledge already available on its 03 sensitivity (18,  19), it was selected
for detailed examination of response to 03.  Only stem DW was significantly**
reduced from the control by the 03 treatment and then only by exposures
lasting six (19% reduction) or seven (17% reduction) days per week for the
18 week exposure period.  In contrast,  foliar injury response varied  dramati-
cally among 03 treatments.   Foliar injury remained constantly low (average
2.4%) at 03 exposures of 0, 1, 2, and 3 days per week (Figure 1A).  A pronounc-
ed increase in injury occurred when seedlings were exposed to 03 for  4 to 7
days per week (average 16.5%).  Significant growth reductions were not consis-
tent with elevated foliar injury.  Either foliar injury is a more sensitive
response indicator of ponderosa pine to 03 exposure or no consistent  relation-
ship exists between foliar injury (at the 03 doses tested) and growth response.
A similar disparity in 03 sensitivity evaluations between growth and  foliar
sensitivity was also noted by Jensen (14) on deciduous  tree species.

PONDEROSA PINE GROWTH

     Sequential, biweekly harvests were taken throughout a 20-week growth
period to examine growth response of ponderosa pine seedlings to the  standard-
ized daily 03 dose for variable biweekly exposure increments of up to 20
weeks.  The data were plotted and a non-linear least squares fit was  used
to fit the curve to the data.  Each point on the growth curves (Figures 1B-
1F) is the average value for 18 pots (72 seedlings).  The growth patterns
of plant height and root length were not obviously affected by the 03 treat-
ments (Figures IB and 1C).   Maximum growth was attained at 10 to 12 weeks
following seedling emergence, regardless of treatment.   Regression analysis
of the last four harvests (14, 16, 18,  and 20 weeks of exposure to 03) showed
no significant differences between treatments.  Maximum plant height  and
root length were attained prior to the termination of the experiment; there-
fore, no further growth differences were expected between treatments  if 03
treatments had continued.

     A regression analysis of DW top (stem + foliage) for the last four
* Significantly different at 90% confidence level.
**Significantly different at 95% confidence level.
                                     640

-------
-3

Z
tr
u
o
o
IT
25




20




 15




 10




 5




 0






70



60



50



40



30



20



1.0



 0
         8      12       16      20

           TIME SCALE, weeks
                  T
                  J_
-L
            5      10     15

           TIME SCALE, weeks
                            20
                                     E 7.0
                                     o


                                     ^6.0
                                    8
                                    y
              5.0



              4.0



              3.0
                                    S2 2.0
                                    UJ


                                    Hl.O

                                    <

                                    al  o
              .80


              .70
| 40


^.30

Q
o .20


  .10


   0
                 0      5      10     15

                       TIME  SCALE, weeks
                              20
                        5      10     15

                       TIME SCALE, weeks
                              20
           5      10     15     20

          TIME SCALE, weeks


 Figure 1 (A to  E).   Response of Fiji us ponderosa  Laws,  to  chronic 03

 exposure: (A)  foliar injury, (B) height growth,  (C)  root  growth,

 (D) top growth  (DW),  and  (E) root growth (DW).
                                 641

-------
        TABLE 2.  RESPONSE OF CONIFER SEEDLINGS TO CHRONIC OZONE EXPOSURES

Species
Douglas fir
Jeffrey pine
Lodgepole pine
Monterey pine
Ponderosa pine
Shore pine
Sugar pine
Western white pine
Sitka spruce
Root
length9
7.5
0
13.0
1.6
12.1*
2.8
0
5.7
14.3
Plant
height3
6.1
2.2
7.7
0
11.3
2.2
0
0
0
DW
foliage3
12.1
5.6
2.4
0
12.4
0
0
12.9**
7.1
DW
stem3
14.8
1.6
8.4
0
21.3*
6.3
0
9.1*
13.6
DW
root6
9.4
5.4
14.7
0
26.3**
10.1
0
12.2
17.9
Foliar
injuryb
1
2
16
0
20
13
0
1
1

  a  Percent reduction from control

  b  Percent injury on 03 treatment  minus injury on control

  *  Significant at 90% confidence level

  ** Significant at 95% confidence level

observations showed significant treatment differences only at a relatively
low confidence  level  (80'/>).  A similar regression analysis for DW root showed
that treatment  differences were highly significant  (98%).  Of the growth
parameters  analyzed,  root growth  (measured by DW) was most sensitive to the
chronic 03  exposures  and differences increased with increased exposure time.
The average reduction  in root growth from the control at the last observa-
tion period (20 weeks  of 03 exposure) was approximately 26%.

     The curves for growth of ponderosa pine seedlings by DW top and DW root
illustrate that the 03 treatment affected the growth pattern (Figures ID and
IE).  Growth curves for the 03 treated plants more closely reached an asymptote
than did the corresponding curves for the controls.  The 03 treatment appa-
rently prompted an early growth senescence in ponderosa pine seedlings.  The
form of the growth curves show that maximum seasonal growth was not reached
during the course of this experiment and maximum seasonal  growth losses were
probably not represented by this experiment.

     The direct effects of 03 on growth of ponderosa pine seedlings observed
in this experiment relate to biochemical  response of the same test seedlings
as reported elsewhere (22).   The 03  exposure increased the Kjeldahl  nitrogen
and amino acid pools in the roots.  The levels of soluble sugars,  starch,

                                     642

-------
and phenols tended to increase in the tops (stem + foliage)  and decrease in
the roots of the 03-exposed plants.   Altered metabolic responses were observed
early in the 03 treatments prior to  the appearance of noticeable growth
differences.  For example, differences in Kjeldahl nitrogen  content of the
roots of ponderosa pine exposed to 03 were significantly greater than the
controls throughout the experiment;  the first observation followed only two
weeks of exposure to 03.  In the tops, the starch level in the 03-treated
plants exceeded the control level at all harvests except the first where
the amounts were similar.  The combined effect of reduced root and shoot
growth, stimulation of early senescence, and altered metabolic pools could
affect the ability of seedlings to survive harsh environmental elements and
reduce competitive ability of 03-sensitive individuals.

     These observations concerning reduced growth and altered metabolic pools
in conifer seedlings exposed to low  03 levels demonstrate the need to consider
the response of forest tree vegetation in development and assessment of se-
condary air quality standards.  A great need exists for detailed investi-
gations concerning the physiological and growth response of pollutants on
trees and subsequent consequences in surviving harsh environmental elements
and competition from associates.

                                REFERENCES

 1.   Barnes. R.L.  1971.  Effects of chronic exposure to ozone on soluble
      sugar and ascorbic acid contents of pine seedlings.  Can. J. Bot. 50:
      215-219.

 2.   Barnes,  R.  L.  1972.  Effects of chronic  exposure  to  ozone on  photo-
      synthesis  and  respiration of pines.   Environ.  Pollut.  3:133-138.

 3.   Berry,  C.R.   1971.   Relative sensitivity of red,  Jack,  and white pine
      seedlings  to ozone and  sulfur  dioxide.   Phytopathology  61:213-232.

 4.   Blum,  U.  and D.  Tingey.   1977.   A study of the potential  ways  by which
      ozone  could  reduce root  growth and nodulation of soybean.   Atm.  Env.
      (in  press).

 5.   Botkin,  D.B.,W.H.  Smith,  and  R.W.  Carlson.   1971.   Ozone suppression
      of white pine  net  photosynthesis.   J.  Air Pollut.  Control  Assoc.  21:
      778-780.

 6.   Cole,  A.F.W.  and M.  Katz.   1966.   Summer ozone concentrations  in southern
      Ontario  in  relation to  photochemical  aspects and  vegetation  damage.
      J. Air Pollut.  Control  Assoc.   16:201-206.

 7.    Davis,  D.D.   1974.   Relationship  between age and  ozone  sensitivity of
      current  needles  of ponderosa pine.   Plant Dis.  Reptr.  58:660-663.

 8.    Davis,  D.D.   1976.   Interaction  of acute doses of ozone  and  PAN  on
      young  ponderosa  pine  seedlings.   Plant  Dis.  Reptr.  (submitted).
                                     643

-------
 9.   Davis, D.D. and F.A. Wood.  1972.  The relative sensitivity of eighteen
      coniferous species to ozone.  Phytopathology  62:14-19.

10.   Davis. D.D. and F.A. Wood.  1973.  fho influence! of environmental
      factors on the sensitivity of Virginia pine to ozone.  Phytopathology
      63:371-376.

11.   Davis, D.D. and F.A. Wood.  1973.  The influence of plant age on the
      sensitivity of Virginia pine to ozone.  Phytopathology  63:381-378.

12.   Heagle, A.S., D.E, Body, and E.K. Pounds.   1972.  Effect of ozone on
      yield of sweet corn.  Phytopathology  62:683-687.

"t3.   Houston,  D.B.  1974.  Response of selected Pinus strobus L. clones to
      fumigations with sulfur dioxide and ozone.  Can. J.  For.  Res.  4:65-68.

14.   Jensen, K.F.   1973.  Response of nine forest tree  species to chronic
      ozone fumigation.   Plant Dis.  Reptr.   57:914-917.

15.   Jensen, K.F.  and L.S.  Dochinger.  1974.   Responses of hybrid poplar
      cuttings  to chronic and acute levels  of ozone.   Environ.  Pollut.
      6:289-295.

16.   Kress, L.W.  1971.  Effect of age,  time,  and pollutant concentration on
      the response  of three  conifer species to  ozone.   Phytopathology 61:130
      (abstr.)

17    Miller, P.R., M.H. McCutchan,  and H.P. Milligan.  1972.  Oxidant air
      pollution in  the central valley, Sierra  Nevada  foothills, and Mineral
      King Valley of California.  Atmos.  Environ.   6:623-633.

18.   Miller, P.R., J.R. Parmeter, Jr., B.H. Flick,  and  C.W. Martinez.   1969.
      Ozone dosage response  of ponderosa pine seedlings.  J. Air Pollut.
      Control Assoc.  19:435-438.

19.   Miller, P.R., J.R. Parmeter, Jr., O.C. Taylor,  and E.A. Cardiff.   1963.
      Ozone injury to the foliage of Pinus ponderosa.   Phytopathology  53:1072-
      1076.

20    Research Triangle  Institute.  1973.   Investigation of  high ozone con-
      centration in the  vicinity of Garrett County, Maryland and Preston
      County, West Virginia.  EPA-R4-73-019.  106 pp.

21.  Santamour,  F. S.,  Jr.1969.  Air  pollution  studies  on  Platanus and
     American elm  seedlings.  Plant Dis. Reptr.   53:482-484.

22.  Tingey, D.T., R.G.  Wilhour, and  Carol  Standley.  1976.  The effect  of
     chronic ozone exposures  on  the metabolite  content  of  ponderosa  pine
     seedlings.  Forest  Sci.  (in press).

23.  Treshow, M. and D.  Stewart.  1973.  Ozone  sensitivity  of  plants in
     natural communities.  Biol. Conserv.   5:209-214.


                                     644

-------
24.   Wood, F.A.   1970.  The relative sensitivity of sixteen deciduous tree
     species to ozone.  Phytopathology  60:579-580 (abstr.).
                                   645

-------
                                                                           13-2
                 MACROSCOPIC  RESPONSE  OF  THREE  PLANT "SPECIES"
                     TO  OZONE,  PAN,  OR OZONE  +  PAN

                       D.  D. Davis  and  R.  J.  Kohut*
ABSTRACT

      ThAee plant species o& vaA.yi.nQ ta.wnom.Lc  classification weAe expo6ed to
ozone., peAoxyacetyl nitAate  (PAW), on a combination  o{^  the.  tmo pa£.Cafrw.fi.
Species &ele.cte.d uieAe pondeAosa pine  (Pinus  pondenosa vaA.  pondeto^a],  hybrid
poplaA (PopuliLS tAichocaApa  x P. ma.ximov04c.zii done  3$%}, and pinto bean
{Phaseollis' ~vuZQ_aAiA "Pintu~l 71W].  PondcAof>a pine,  and hybrid  popCai ate vein
susceptible to ozone, bat WAiAtant to PAW.  Pin-to bean <.A  itticcptc'bCc  to
both oxidantA, AeJ>pondi.ng with -two (ici.ttnctlt/ cit^c/teuf and ipafcat'C// i
symptom typ&&.  Ex.po&uA$, ufa  ponddAo^a pine,  wkite  in an ozone-ieiiictu'e
4tage, to the combined pollutant* AeAiitted in significantly let>t>  damage than
that induced by ozone alone.  In contAa&t, injuAy  on hybAid poplaA induced by
combined pollata.nti> MU generally QAeateA than  that  produced  by ozone alcne.
ExpoAuAe o& pinto bean to combined ozidantb  completely  inhibited  the PAW Aympton
on the abaxi.al leafi AuAfiace  and had a AyneAgi&tic  OA additive e^fiect on the
ozone symptom ex-pnuAion in  the adaxj,al leafa buA&ace.

                                  INTRODUCTION

     Air quality standards designed to protect  vegetation are often based on
research involving the exposure of plants  to singular pollutants.  However,
plants growing in nature are often exposed  to ambient air containing multiple
pollutants.  Two common, phytotoxic photochemical  oxidants  that may co-exist
in polluted air are ozone and peroxyacetyl  nitrate (PAN) (10).   Although the
phytotoxic nature of each oxidant has  been  clearly established, the response
of plants to the combined pollutants  has  not been  reported.  Thus, this project
was initiated.

     The research described  herein was conducted over a two-year period from
1974 to 1975.  Results from  three separate  experiments, involving three
diverse plant types, have been  combined and  summarized  for  publication in this
paper.
*D. D. Davis, Pennsylvania State University,  University Park,  Pennsylvania.
 R. J. Kohut, University of Minnesota,  St.  Paul,  Minnesota.

                                      647

-------
                                HYBRID POPLAR

MATERIALS AND METHODS

     Cuttings of hybrid poplar (Populus maximowiczii  x trichocarpa),  grown
out-of-doors in pots, were exposed for four hours in  groups of 12 to  either
0.18 ppm ozone, 0.18 ppm PAN, or the two pollutants combined at these concen-
trations.  The exposures began when the plants were three weeks old from bud-
break and were repeated six times at two week intervals throughout the summer.

     Plants were exposed in a specially modified growth chamber (13)  at environ-
mental conditions of 23°, 75% relative humidity (RH),  and 25 Klux light inten-
sity, hereafter described as "standard conditions".  Exposure and monitoring
techniques for each pollutant were as previously described for ozone (5) and
PAN  (1,  5, 9,  12).  A similar growth chamber housed the control plants at the
above environmental conditions.  The ozone and PAN monitors (chemiluminescent
and  electron capture GC, respectively) were pollutant specific and not subject
to interference during  the combined pollutant exposures.  The same exposure
and  control procedures  were also used for poderosa pine and pinto  bean.

     On the third day after exposure,  symptoms were evaluated  on the  15 most
distal leaves on each plant.   The percentage  of the leaf surface affected was
estimated using previously described reference charts  (5).   The  injury  estimates
for the 15 leaves on each of the 12 plants  were averaged to produce a treatment
injury level.

RESULTS

     None of the plants exposed to PAN developed visible symptoms.  The plants
exposed to either ozone or the ozone + PAN  mixture developed areas of dark
brown to black  bifacial  necrosis.   There was  no difference in  the time of
symptom initiation with either treatment or in the distribution of symptoms on
the individual  leaves.   The recently matured  leaves on the plant developed the
most severe symptoms in both treatments.   Visible symptoms were not observed
below the 15 most distal leaves on the plant.

     The injury level  produced by the ozone + PAN exposure was significantly
greater than the sum of that produced by the  separate  ozone or PAN exposures
in four of the  six experiments (Table 1); in  two experiments the injury levels
were not significantly different.

                                PONDEROSA PINE

MATERIALS AND METHODS

     Ponderosa  pine seedlings in the juvenile needle  stage were raised from
seed in pots in the greenhouse until exposure to ozone, PAN, or the combined
pollutants.  A constant level  of 0.05 ppm PAN and 0.40 ppm o/one was  maintained
in the exposure chamber at standard conditions.   On two consecutive days, }?0
pine seedlings  were exposed to this combined  dosage for 2, 4,  6, or 8 hours
beginning at 0900 hours.  Following exposure, plants  remained in qrowth cham-
bers for five days at standard conditions during symptom development.

                                     648

-------
 TABLE 1.  VISUAL ESTIMATES OF THE PERCENTAGE OF ADAXIAL INJURY PRODUCED ON
 HYBRID POPLAR BY EXPOSURE FOR 4 HOURS TO EITHER 0.18 ppm OZONE, 0.18 ppm
 PAN OR THE TWO POLLUTANTS COMBINED AT THESE CONCENTRATIONS.




Exposure
Dates
(1975)
June 20-22
July 5-7
July 19-21
Aug. 3-5
Aug. 16-18
Aug. 29-31





Ozone

0.8
0.0
0.4
5.4
28.8
0.0


% Injury a


PAN

0.0
0.0
0.0
0.0
0.0
0.0





Ozone + PAN

3.7
0.6
2.1
5.8
52.6
2.8
Confidence
level of the
synergistic
interaction
(F-test)

95%
N.S.b
90'-:
N.S.b
99%
99%

 a  The mean percentage of foliar injury on the 12 plants in the treatment.

 b  No significant pollutant interaction at the 90% confidence level.
RESULTS

     The symptoms induced by ozone or ozone + PAN were identical and were
similar to typical ozone injury on ponderosa pine.  Seedlings were tolerant
of PAN and did not develop symptoms.  The percentage injury induced by ozone
+ PAN at 4, 6, and 8 hours was significantly (0.01) less (antagonistic
response) than that produced by ozone alone (Figure 1).  Injuries from the
two-hour exposures were statistically (0.05) similar.
                                 PINTO BEAN
MATERIALS AND METHODS
     Ten-day old pinto bean plants in the trifoliate leaf stage were exposed
for four hours to either 0.30 ppm ozone, 0.05 ppm PAN, or the two combined
at these concentrations.  Twenty plants maintained in growth chambers were
used in each exposure.  A series of three exposures was conducted on three
consecutive days and replicated five times.   Control plants were iraintained
in another chamber at'the same environmental conditions durinq the exposure.
After the exposure, all plants were returned to their original growth chambers
during symptom development.

                                     649

-------
            50  -
            40  -
        V-i

        *r~)
        c
        cfl
        •H
        c
        0)
        O
        (-1
        0)
        P-I
30  -
20  -
            10  -
                                 HOURS OF EXPOSURE


    Figure 1.   Percent foliar injury on primary needles of ponderosa pine
               exposed to 0.40 ppm ozone, or 0.40 ppm ozone + 0.05 ppm PAN
               for 2,4,6 or 8 hrs.   Each dot. represents the average value
               determined for 30 seedlings in two replications.   PAN alone
               induced no injury.   Ozone alone induced significantly (.01)
               greater foliar injury than did ozone + PAN at the 4,6, and
               8 hour exposures (T-test).  Injury was not induced by two
               hours of exposure.
     The first trifoliate leaf on each plant was evaluated for injury three
days after exposure.  The adaxial and abaxial surfaces of leaflets were
evaluated separately.  The symptom types and the percentage of the leaflet
surface affected bv each type were recorded using the two factor system pre-
viously described (5).  An average percent plant injury and treatment injury
was calculated for each leaf surface.  Percent adaxial and abaxial injury
were also averaged to calculate injury on the combined leaf surface.
                                     650

-------
RESULTS

      The ozone exposures produced a typical  light tan fleck on the adaxial
leaf surface, while PAN produced primarily abaxial bronzing with some light
tan bifacial necrosis.  The symptoms produced by the combined pollutant ex-
posures were mainly restricted to the adaxial leaf surface with a nearly
complete suppression of symptoms on the abaxial  surface.  The adaxial symptom
produced by the combined pollutants was fleck similar in appearance to that
produced by ozone but slightly more yellow.  In  the one replication where
abaxial symptoms were produced in response to the combined pollutant exposure,
the symptom was a bronzing similar to that produced by PAN.

     Three of five replications indicated a synergistic response on the ad-
axial surface in the combined pollutant exposures, while the response was
additive in the other two replications (Table 2).  Injury on the abaxial
leaf surface indicated an antagonistic response  in all five replications.
When the combined leaf surface response was evaluated, all of the replica-
tions indicated an antagonistic response.

                                 DISCUSSION

     The possibility that ozone and PAN might interact chemically in the ex-
posure chamber atmosphere was investigated to ensure that the observed inter-
actions were not the result of chemical reactions within the chamber.  A
stable ozone concentration was established in the exposure chamber and PAN
was then introduced.  The chemiluminescent monitor provided a continuous
reading of ozone concentration and indicated that PAN had no effect on the
ozone concentration.  The procedure was then reversed by establishing a
stable PAN concentration and introducing ozone.   No change in PAN concentra-
tion was detected using the electron capture GC.  Therefore, it was possible
to discount an atmospheric reaction between the two pollutants as an explana-
tion for any changes in injury levels produced by the combined pollutants.
This indicated that the diverse response of the three taxonomically different
species was physiological in nature within the host.  The results of these
exposures provide the first evidence that the photochemical oxidants ozone
and PAN can interact when producing injury to vegetation.

      The symptom types induced by the combined  pollutants was typical  of that
induced by ozone alone.  The foliage of hybrid poplar produced brown to black
areas of bifacial necrosis, typical of ozone injury on this clone (5).   The
needles of ponderosa pine showed a chlorotic mottle, similar to that induced
by ozone alone (3,7,8).  This was as expected, since these two species were
resistant to PAN (2,5), as are many woody plants (4).   However, pinto bean,
which is sensitive to PAN, also responded with an adaxial fleck similar in
appearance to that caused by ozone alone.  Thus, in each case the combi-
nation of ozone and PAN induced symptoms on all  three species similar to
those induced by ozone.  This is somewhat analagous to the symptom response
of various species to mixtures of ozone and sulfur dioxide, in which the
ozone type symptom was augmented (6,11).   The recently matured foliage of
a plant is generally recognized as being most sensitive to ozone.  The
symptoms observed in both the ozone and ozone/PAN treatments were most severe
on this recently matured tissue with all  three plant types.


                                     651

-------
  TABLE 2.   PERCENTAGE VISIBLE FOLIAR INNURY ON THE PRIMARY LEAVES OF PINTO
  BEAN PRODUCED BY 4-HOUR EXPOSURES TO EITHER 0.30 ppm OZONE, 0.05 ppm PAN, OR
  THE TWO POLLUTANTS COMBINED AT THESE CONCENTRATIONS.

Rep.
1


2


3


4


5


Leaf
Surface
Adaxial
Adaxial
Combined
Adaxial
Adaxial
Combined
Adaxial
Adaxial
Combined
Adaxial
Adaxial
Combined
Adaxial
Adaxial
Combined

Ozone
20
0
10
12
0
6
25
0
12
21
0
10
28
0
14
% Injury3
PAN
4
66
35
13
57
35
3
35
18
5
53
29
4
23
13

Ozone/PAN
25
0
13
24
4
14
42
0
21
45
0
22
45
0
22
Response^
NS
ANT
ANT
NS
ANT
ANT
SYN
ANT
ANT
SYN
ANT
ANT
SYN
ANT
ANT

 a  The mean percentage of visible foliar injury for the 20 plants in each
    treatment.
    Response was evaluated for each leaf surface

      NS = non-significant at 95% level of confidence;
     ANT = antagonistic response;
     SYN = synergistic response (/-test)
     These results indicate that a plant does not have to be susceptible to
both pollutants for them to interact and produce injury.   Such findings should
be considered when evaluating the importance of an air pollutant in the field.
It must be remembered that in addition to the pollutant of interest, a plant
in the field may also be subjected to other pollutants.  While these pollu-
tants may be considered innocuous on their own, their potential  to enter into
interactions must be considered.  Also, plant response to simultaneous expo-
sure to ozone and PAN will likely vary, depending upon the species, dosacie and
ratio of the combined pollutants, age of plant or leaf, leaf surface, environ-
mental conditions, and method of evaluation.  This makes  it very difficult to
                                     652

-------
consider results from interaction studies when formulating air quality stan-
dards, unless constant findings are achieved over a wide range of species and
conditions.
                                   KEFERENCES

 1.   Darley, E. F., K. A. Kettner, and E. R. Stephens.  1963.  Analysis of
      peroxyacetyl nitrates by gas chromatography with electron capture
      detection.  Anal. Chem.  35:589-591.

 2.   Davis, D.D.  1975.  Resistance of young ponderosa pine seedlings to
      acute doses of PAN.  Plant Dis. Reptr.  59:183-184.

 3.   Davis, D.D., and J.B. Coppolino.  1974.  Relationship between age
      and ozone sensitivity of current needles of ponderosa pine.  Plant
      Dis. Reptr.  58:660-663.

 4.   Drummond, D. B.  1971.  Influence of high concentrations of peroxy-
      acetyl nitrate on woody plants.  Phytopathology  61:128 (Abstr.)

 5.   Kohut, R.  1975.  The interaction of 0^ and PAN on hybrid poplar and
      pinto bean.  The Penna. State Univ. Cen. Air Environ. Studies
      Pub. No. 428-76, 60 pp.

 6.   McDowell, F. D. H. and A.  F. W. Cole.  1971.  Threshold and synergistic
      damage to tobacco by ozone and sulfur dioxide.  Atmos. Environ.
      5:553-559.

 7.   Miller, P. R., J. R. Parmeter, Jr., 0. C. Taylor, and E. A. Cardiff.
      1963.  Ozone injury to the foliage of Pinus ponderosa.  Phytopathology
      53:1072-1076.

 8.   Richards, B. L., Sr., 0. C. Taylor, and G. F. Edmunds, Jr.  1968.
      Ozone needle mottle of pine of southern California.  J. Air Pollut.
      Control. Assoc.  18:73-77.

 9.   Stephens, E. R., F. R. Burleson and E. A. Cardiff.  1965.   The
      production of pure peroxyacetyl nitrates.  J. Air Pollut.  Control
      Assoc.  15:87-89.

10.   Taylor, 0. C.  1969.  Importance of peroxyacetyl nitrate (PAN) as a
      phytotoxic air pollutant.   J. Air Pollut. Control Assoc.  19:347-351.

11.   Tingey, D. T., R. A. Reinert, J. A. Dunning and W. W. Heck.  1973.
      Foliar injury responses of eleven plant species to ozone/sulfur
      dioxide mixtures.  Atmos.  Environ.   7:201-208.

12.   Wood, F. A., and D. B. Drummond.  1974.  Response of eight cultivars
      of chrysanthemum to peroxyacetyl nitrate.  Phytopathology  64:897-898.


                                     653

-------
13.    Wood,  F.  A., D.  B. Drummond, R. G. Wilhour, and D. D. Davis.  1973.
      An exposure chamber for studying the effects of air pollutants on
      plants.   Penn.  State Univ., Agric. Exp. Stn. Prog. Rep. 355.  7 pp.
                                     654

-------
                                                                             13-3
                      RELATIVE SENSITIVITY  OF  EIGHTEEN HYBRID
                      COMBINATIONS OF  PINUS TAEDA L.  TO OZONE

                           L. W. Kress  and  J.  M.  Skelly*

ABSTRACT

     LobtoLty pine. se.e.dtings weAe. e,x.pose,d  to  ozone. (03) to seJLe.ct a se.nsiti.ve.
and a toteAant CAOSS to be. aied ai> bio indicators o& ambie.nt aiA potiution
Ie.ve2s.   Eighte,e.n hybrid combinations  o& JLobiotty pine. weAe. scAe,e.ne.d  &OK.
se.nsiti.vity to 0.10-0.25 vtiteA ozone./l^iteA aiA.  CioSA 6-13 x 2-8, the. most
se.nsitive. in  &ive, ofa zight e.x.posuA&s,  was  among  the, ^ive. most se.nsiti.ve. -in
^euen o&  &ight expoiuA^s.  In both ne.ptic.ationt>  o^ 0.15 \ititoA 0-^/tit^i  exposed to 0.10 vLit&i  03/titeA oJJi, and Lt  was in
one. cfa those.  fie.ptications that 6-13 x. 2-8  was not among the. fcve. most se.nst-
tive..  C^oss  6-13 x 2-8 was ne.vest among the. tun  most tolerant.  Cno&s 540 x 504
was among the, stx. most toi.eAa.nt -in s the. e^ight e.xpoSusizs
and had no -injuAy in two o{^ those, e^ight expo^uAei.  The. ftzsults 'indic.ate.d
that tt was fie.as-ib£e. to use. se.ns-itive. and  toieAant CAOS&ZS o& LobtoLty pine,
as bioi,ndi,c,atou oft amb
-------
production rate at this installation was used to estimate the amount of air
pollution emitted.

     Past studies at the RAAP have demonstrated significant negative cor-
relations between cyclic industrial production levels and the annual! radial
increment growth of planted loblolly pines (5); other studies reported simi-
lar data with several other forest species (4, 6).   Research elsewhere has
demonstrated that loblolly pine was relatively sensitive to 03 or S02 (1,
2).  The current investigators felt that loblolly pine would therefore be a
good candidate for use as a bioindicator of the reduced RAAP air emissions
as abatement facilities become operational.

     Loblolly pine is the most important species to the forest industry in
the South.  The Commonwealth of Virginia planted 83 million loblolly pine
seedlings in 1973, and the Southern Forest Free Improvement Program has an
active breeding program based on gaining a 5% improvement in tree growth
form, growth rate, fiber content, disease resistance, and several other im-
portant parameters.  The importance of the species  generated the breeding
program from which we have been able to obtain seed of controlled crosses
of loblolly pine.  Controlled crosses minimize variation in the plant popula-
tion much the same, although not as effectively, as clonal  selection.

     Part of a continuing research program to biologically monitor ambient
air pollution levels has involved the use of a tolerant and a sensitive line
of loblolly pine planted in open-top chambers,.  Eight trees of each line
have been planted within chambers drawing ambient air or charcoal-fil-
tered air; an open plot was also established.  Height growth, needle length,
and foliar symptoms on these trees are to be measured over the naxt three to
six years.

     The objective of the study reported within this paper was to screen
several hybrid crosses of loblolly pine for sensitivity to 03 and to choose
the most sensitive and most tolerant cross for use  in the field studies utiliz-
ing open-top chambers.  A secondary objective was to collect preliminary data
as to whether or not breeding for tolerance or sensitivity of loblolly pine
to air pollution would be feasible.

                              MATERIALS AND METHODS

     Seed of 18 hybrid lines of loblolly pine was obtained from the Southern
Forest Tree Improvement Program and was greenhouse  grown in a 2:2:1 weblite:
vermiculiterpeat moss soil mix in four-inch plastic pots with four trees per
pot.  Each pot received 1 gram of osmocote fertilizer and was watered rou-
tinely.  The greenhouse was provided with charcoal-filtered air and supple-
mental lighting of 21,000 lux was provided on a 14-hour photoperiod by high
pressure sodium lamps.

     During July 1975 when the trees were 2 1/2 to  4 weeks old, they were
exposed to ozone for eight hours at corrected Mast  readings of 0.10, 0.15,
0.20, and 0.25 pliter/liter air (10, 15, 20, and 25 pphm).   The exposure chambers
employed a one-pass-through air system, drawing in  charcoal-filtered air hav-
ing a chamber residence time of two minutes.  The pollutants were injected

                                     656

-------
 into the  intake lines just prior to their entry into each chamber.  Pollutant
 concentrations were maintained at the desired levels by monitoring at plant
 height and adjusting the flow of pollutant into the chamber.  The 03 levels
 were monitored continuously with Mast oxidant meters and spot checked regu-
 larly with the 1% neutral buffered potassium iodide (KI) method.  Lighting
 for the chambers was 30,000 lux provided by high-pressure sodium lamps.
 Environmental conditions were 25-30°C and 70-85% RH.

     Foliar symptoms were evaluated prior to exposure, immediately follow-
 ing exposure, and 1, 7, 14, and 21 days following exposure.   Symptom expres-
 sion was  assigned values of 1-7 in an attempt to place an importance on each
 symptom type (Table 1).
             TABLE 1.   SYMPTOM CLASSES USED TO EVALUATE  LOBLOLLY  PINE
                HYBRIDS EXPOSED TO VARIOUS LEVELS  OF 03  FOR 8  HOURS
                                SYMPTOM CATEGORIES
                                1     Mottle
                                2     Chlorotic  spot
                                3     Chlorotic  band
                                4     Chlorotic  tip
                                5     Necrotic spot
                                6     Necrotic band
                                7     Necrotic tip
     The severity value was then multiplied by the percent of the needle
affected.  This yielded a severity index for each needle.  The severity indices
were summed for all needles on a tree to yield a T (total injury) value.  The
T value was then divided by the total number of injured need"!es to yield an
I (average injured needle) value; an indication of how severely each needle
was injured.  The T value was also divided by the total number of needles
on the tree to yield an A (average of all needles) value; an indication of
how severely each tree was injured.  To provide smaller numbers to work with
statistically, a value of .5 was added (to eliminate 0 values) to each of the
T, I, and A values, and then the square root of each value was calculated.
Each of the respective values was then summed and averaged for the four
trees of each cross.  This procedure yielded six values (T, St, I, SI, A, SA)
denoting the same severity index for each cross.  An ANOVA was calculated
using each value, and the means for each cross were tested by Duncan's Multiple
Range Test.

                                     RESULTS

     There was substantial variation in the symptom expression response to
03 between the loblolly pine crosses tested.  There also was substantial

                                     657

-------
variation in the symptom expression of the  individual  trees  within  a  given
cross.   In ranking the means,  however, in  terms  of  sensitivity  and  tolerance
to 03 there was virtually no difference between  the six  (T,  ST,  I,  SI,  A,  SA)
severity index values, and therefore the ST value was  chosen to  work  with
for the rest of the analyses.

     The mean symptom severity indices for cross 6-13  x  2-8  in  replication  1
of 15 pphm 03/8 hours have been listed in  Table  2.

  TABLE  2.   SYMPTOM SEVERITY  INDICES  OF SENSITIVE LOBLOLLY PINE  CROSS 6-13  x
    2-8  EXPOSED TO 15 PPHM 03  FOR  EIGHT HOURS  FOR USE  IN TREE BIOINDICATOR
                                SYSTEMS AT  RAAP
                                INJURY INDICES'
                T - 6114                                ST - 74
                1-306                                SI - 17
                A -  175                                SA - 12


*A11 indices are an average of four values; T is the average total injury value,
 I is the average of total value divided by the number of injured needles, and
 A is the average of total value divided by the total number of needles.  The
 ST, SI, and SA values are, in effect, the square roots (before averaging) of
 T,I, and A values.
     The variation between the four individual values that were averaged to
yield the T, I, and A values accounted for their respective square root val-
ues (which were taken before averaging) not being exact square roots.

     For example, with no variation between the four values that make up
the T value in Table 2, the ST value would be the square root of the T value.

     The symptom progression of four of the crosses following exposure to
15 pphm 03/8 hours have been listed in Table 3.

     Lines 6-13 x 2-8 and 2-40 x 2-8 were the two most sensitive and 540 x 504
and 501 x 504 were the two most tolerant.  There were no symptoms immediately
following the exposure, but symptoms developed very rapidly on the most
sensitive line.  If all the needles were 100% necrotic, the injury index
would have been approximately 150.  A value of 74 indicated relatively severe
injury; all needles 50? necrotic, half of the needles 100% necrotic  and the
rest asymptomatic, or somewhere in between.

     Table 4 shows the two most sensitive loblolly pine crosses and  the num-
ber of other crosses that each of these two lines was more sensitive than to
03 (two replications each of 25 and 15 pphm 03/8 hours).   In both replica-
tions of 15 pphm 03/8 hours, cross 6-13 x 2-8 was significantly more sensi-

                                     658

-------
  TABLE 3.  SYMPTOM EXPRESSION (ST SEVERITY INDEX VALUE)  OF FOUR LOBLOLLY  PINE
     CROSSES EXPOSED TO 15 PPHM 03 FOR 8 HOURS IN THE DEVELOPMENT OF A TREE
                         BIOINDICATOR SYSTEM AT THE RAAP
                               DAYS AFTER EXPOSURE
                           1                7              14              21
6-13 x 2-8*
2-40 x 2-8
540 x 504
501 x 504
58
2
0
0
66
9
0
0
71
13
3
0
74
14
4
0

 ^Crosses 6-13 x 2-8 and 2-40 x 2-8 were relatively sensitive and crosses 540
  x 504 and 501 x 504 were relatively tolerant.

   TABLE 4.  THE TWO MOST SENSITIVE LOBLOLLY PINE CROSSES AND THE NUMBER OF
       CROSSES THAT EACH WAS SIGNIFICANTLY MORE  SENSITIVE TO 03 THAN IN
           THE DEVELOPMENT OF A TREE BIOINDICATOR SYSTEM AT THE RAAP
                                     TREATMENTS*

                     25 pphm/8 hr                  15 pphm/8 hr
     Cross           12                  12
6-13 x 2-8
2-40 x 2-8
14
7
13
11
17
0
17
0

   lVTwo replications each of 25 pphm 03/8 hours and 15 pphm 03/8 hours.

tive than all of the remaining 17 crosses.  At 25 pphm/8 hours, cross 2-40 x
2-8 was significantly more sensitive than 7 and 11 of the remaining 17 lines/

     The symptom severity indices of the two most and two least sensitive
crosses over two replications each of the four ozone levels have been listed
in Table 5.  Any value underlined denotes it was ranked among the five most
sensitive crosses.  A value with a line over it signifies the cross was
among the six most tolerant.

     Cross 6-13 x 2-8, the most sensitive in 5 of 8 of the exposures, was
among the five most sensitive in 7 of 8 of the exposures.  It was also signifi-
cantly more sensitive than all remaining crosses in both replications of
15 pphm 0, for eight hours.  Cross 540 x 504 was among the six most tolerant in 6
of 8 of the exposures, and in the one exposure in which it was not it sustained
little injury.  Cross 501 x 504 was among the six most tolerant in 7 of 8
exposures, but in the remaining exposure, it was among the five most sensitive.

     Table 6 shows the effect of time of year on the sensitivity of loblolly

                                     659

-------
  [ABLE  5.  SUMMARY OF DATA USED IN SELECTION OF 6-13 x 2-8 AND 540 x 504 FOR
          USE AS BIOINDICATOR LINES IN FURTHER STUDIES AT THE RAAP
                                       OZONE TREATMENT
                  25 pphm/8 hr   15 pphm/8 hr   20 pphm/8 hr   10 pphm/8 hr
                      Rep            Rep            Rep            Rep
     Cross         12       12       12       12
6-13 x 2-8
2-40 x 2-8
540 x 504
501 x 504
Highest value+
90*
76_
40
8
90
105
2i
30"
5
105
74
11
4
2"
74
97
9_
2"
]1_
97
18
2!
I
w
39
5
9_
T
0"
16
3
0"
0
0"
3
1
1
T
0
2

  bar  under  -  ranked among five most sensitive of eighteen
  bar  over - ranked among six most tolerant of eighteen

  *  -  Symptom  severity of the two most and two least sensitive loblolly pine
      crosses  in two replications each of four levels of ozone.

  +  -  The highest value of any cross for each replication is listed at the
      bottom.
    TABLE 6.   SYMPTOM SEVERITY INDEX (T VALUE)  OF THE  MOST SEVERELY INJURED
LOBLOLLY PINE CROSS OVER A VARIETY OF EXPOSURE  CONDITIONS  AT DIFFERENT TIMES OF
      THE YEAR.   TREE AGE WAS HELD RELATIVELY CONSTANT FOR EACH  EXPOSURE
Exposure Date                     pphm 0,,/hr                    Severity Index


    4/22                             50/8                            1191
    5/5                              25/8                             163
    6/9                              25/8                              40
    7/2                              15/8                            6114
    7/8                              15/8                            9935
    7/14                             20/8                            1734
    7/22                             20/8                             450
                                     660

-------
pine to ozone.

     The injury (T) values increased from 163 and 40 in May and June to 6114
and 9935 in the first week of July, and lower 03 concentrations were used in
the latter exposures.  Then in the middle of July, the values drop again to
1734 and 450.  Tree age at time of exposure was held relatively constant.

                                   DISCUSSION

     The preceding data, along with preliminary N02 and S02 screening data,
indicate that it is feasible to use sensitive and tolerant crosses of lob-
lolly pine as bioindicators of pollution abatement programs at the RAAP.
Based on foliar symptoms in response to 03, there is a good breakdown be-
tween the most sensitive and the most tolerant crosses.  Preliminary screen-
ing indicates that those lines most sensitive to 03 are also the most sensi-
tive to N02 or S02.

     Crosses 6-13 x 2-8 and 540 x 504 were selected for use as the bioindica-
tors to be used in the open-top chamber field studies at the RAAP.  Cross
6-13 x 2-8 was significantly more sensitive than 540 x 504 in five of the
eight replications.  In the remaining three, neither cross sustained signi-
ficant injury.  The difference between the crosses was based solely on foliar
symptoms.  A goal of the open-top chamber study will be to demonstrate effects
on growth as well as foliar symptoms.

     The results indicate that sensitivity to 03 may be heritable enough to
warrant further study into the feasibility of breeding programs.   Besides
the 18 crosses included in this data, a limited number of seedlings of four
other crosses were also screened.   One of these four (2-8 x 523)  was also a
very sensitive cross, demonstrating that the three most sensitive crosses
all involved line 2-8.   Information on how the primary needles of each tree
responded is now recorded, and information on how the secondary needles re-
spond will  be accumulated in future studies.   The responses must  remain con-
sistent for breeding programs to be feasible, and for the different crosses
to be useful as bioindicators to be planted at the RAAP to demonstrate the
success or failure of the abatement programs.

     The time of year effect was an interesting but puzzling problem.   It
has been noted by several other researchers on several  other species of
plants.  Peak sensitivity, holding age constant, usually occurred at the end
of June and the first part of July.   It may be tied partly to age, day length,
and RH in a field situation, but even controlling those variables the difference
still  occurred.   The dramatic difference noted in this  study may  be due in large
part to RH.   During the winter and spring months it was difficult to maintain
high RH levels in the exposure chambers.   This effect may be very significant
at the RAAP.  At times  when plants were not in a sensitive condition,  the
pollutant levels could possibly be higher without causing significant adverse
effects to the plants.   This data  also demonstrates that we need  the results
of combined pollutant,  long-term ambient studies to be  able to evaluate more
fully the growth impact of the pollution regime.
                                     661

-------
                         CONCLUSIONS AND RECOMMENDATIONS

     Based on foliar symptoms, there appeared to be enough variation among
lines of loblolly pine to allow their use as bioindicators of ambient air
pollution.  The results also indicated that sensitivity or tolerance may be
heritable enough to warrant further study into the feasibility of breeding
programs.
     Future study must examine the effects on loblolly pine of 03, N0x, and
S02 singly and in combination.  Preliminary data indicates that. thosexlines
sensitive to 03 are also sensitive to N02 or S02.   Growth impact due to
long-term low-level exposures to combinations of all  three pollutants must
also be examined in field studies and in controlled environment studies.

                                     REFERENCES

1.   Berry, C.  R.  Age of Pine  Seedlings  with Primary  Needles Affects
     Sensitivity to Ozone and Sulfur Dioxide.   Phytopathology 64:207-209,
     1974.

2.   Ham,  D.  L. The Biological  Interactions  of  Sulfur  Dioxide and  Scirrhia
     acicola  in Loblolly Pine.   Ph. D.  Dissertation, Duke  Univ.,  Durham,
     N.C.,  1971.   75  pp.

3.   Hayes, E.  M.  and J. M.  Skelly.   Oxidant Levels in  Southwestern
     Virginia  and  Their Effects  on Eastern White Pine.   Proceedings of
     the Amer.  Phytopathological  Soc.,  Vol.  3,  (Abstr.),  (In Press), 1976.

4.   Phillips,  S.  0., J. M.  Skelly, and  H. E. Burkhart.   Inhibition of
     Growth in  Asymptomatic  White Pine  Exposed  to  Fluctuating Levels of
     Air Pollution.   Phytopathology, (In  Press), 1976.

5.   Phillips,  S.  0., J. M.  Skelly, and  H. E. Burkhart.   Growth  Fluctuations
     of Loblolly Pine Proximal  to a Periodic Source of Air  Pollution.  Phyto-
     pathology,  (In Press),  1976.

6.   Stone, L.  L.  and J. M.  Skelly.  The  Growth of Two Forest Tree  Species
     Adjacent  to a Periodic  Source of Air Pollution.   Phytopathology   64:
     773-778,  1974.
                                     662

-------
                                                                             13-4
           EFFECTS OF OZONE AND SULFUR  DIOXIDE  SINGLY AND IN COMBINATION
                   ON YIELD, QUALITY, AND  N-FIXATION OF ALFALFA

                   G. E. Neely, D. T. Tingey, and R. G. Wilhour*

ABSTRACT

     field eKpoAureA ofa alfaalfaa, Me.dica.go Aativa L. cv.  Meba-Slua to ozone.
(03) (TOO  and 200 yg/m3)  and Aulfaur dioxide  (S02)  (756 yg/m3)  weAe conducted
during two growing AeaAonA to determine growth,  quality, and yield efafaectA.
Two experimentA  weAe conducted.  In the. faiAAt,  03  WOA admlnlAtered Alx. houAA
per day, Aeven day* per week to established  alfaalfaa plantA grown -in peat-
vermicullte.   The. -second experiment Involved Aand-grown alfaalfaa plants with
03  (^euen  hoau  p&i day]  and 4u£^uA ctioyx.de  (24 (IOUA^ pe^. day) applied -b-ingitj
and tn combination uiith. 03, 4 even dcti/4 pe^ week.
                       condact.e.d at the.  U.S.  Environmental Protection Agency' i
 (EPA) Co^fa££^4  EnvtAonmantat Re^eaAcA  Labotuttoiy faafan Atte. i.n CotivaiLLb ,
 OfKLQon.   PasiamzteAA  meaiuAed i.nciude.d dug w&igkt (W)  o^ ^oAage, 4^abfa£e,
 and  ioott>;  total, pnotz-in, amino acidi,,  total nonAtsw.ctuA.al caAbonydAatut,  (TNC),
 phii.no L^ ,  and ft-cafiotane. content*; and nitA.oge.n- Citing capacity.  Plants weAe
 floutingly hanvej>te.d  w/ien con^Ao£4 Aeacned  7/70  fa£oom.   Roo^: and Atubble. data
     obtJcu.ne.d only at the. .second haAveAt ofa  the. &and al^al^a
     FoAage  dAt/ weight, a fa QJ>toJotLkh.,  dAy  weight* ofa all pollutant
tAe.atme.ntf>  (compaAed to contAolb) 4/iowed a gieate^i Aeductum with iucce^-i-cue
naAue^^i.   In both AtudieA ,  the protean  content  pel unit ivc,ight ofa plant
faoiage. w)O4  i.ncAe.cu>e.d in the. potlut.ant-t>tAe.£&e.d plants by the. Co-it h&ive,!>t.
The. net e.fafae.ct ofa the. polXatantA, howe.ve.fi, WOA a Aeductum In totai protein
pAodaced  peA ^.eatmen^; fie^nlting faAom Aeduced pAoduction ofa to tat d^y  matte-A.
The. 03 tAe.atme.nt (200 yg/m3)  al&o i.ncAe.a&e.d  the.  amino acx.d6 content /en eitafa-
ll*>he.d alfaalfaa at. the. laA.t kaAvut.  Howe.veA, due. to Aedaced dAt/ matter pAo-
duction,  the. total amino acid produced peA sample, wca> reduced.  Ozone
decreased Q-caAotene content on both a unit  weight and total weight baix^s
at the laAt harvest ofa ef>tabl^hed alfaalfaa.   The total TWC In faorage,  !>.tubble,
and Aootb  ofa the pollutant tAeatmen.tf> wai Aedaced by approximately 50  percent
at the loAt haAve&t in both At.udi.eA.  The 03> S02 cr 03 + S>0: land alfaalfaa
tAeatmentA  caused appAoyu.mately a 40 to  60 percent dec-reaAe in tctat .5t/mfa-co-
ticaliy failed- nitrogen (W) In the faorage, stubble, and -loots at the last  har-
*U.S. Environmental  Protection Agency,  Corvallis, Oregon.

                                      663

-------
       T/ie. e^ecti Oj$ t/tc 03 and S02 mix.tuA.iL on. alfiaCfaa did not di^eA grotty
     the. e^ee& o& the. tingle. poltutanU.

                                  INTRODUCTION

     Alfalfa, Medicago sativa L., a perennial legume, is widely grown through-
out the world, with approximately 11 million hectares grown in the United
States.  It has been called "Queen of Forage Crops" because of the superior
protein, mineral and vitamin content, yield and total feed value.

     The following effects of ozone (03) on alfalfa have been reported:
Howell et al. (1) noted variations in the visible injury on 14 strains exposed
four hours to 400 yg/iir 03 in the greenhouse; Tingey and Reinert (2) found 12
and 22 percent reductions in forage (stems and leaves) and root dry weights,
respectively, after greenhouse exposure to 98 yg/m3 0^ for 40 hr/week from
seedling to harvest.  Oshima (3) reported that a dose (concentration x time)
of 2500 pphm hours greater than 10 pphm (200 yg/m3) 03 was required for
significant yield reductions to occur in Moapa 69 alfalfa grown in the ambi-
ent air.

     Changes in alfalfa growth and metabolite pools resulting from 03 expo-
sures are important because they affect crop quality and yield.  They also
may (a) reflect pollution-induced changes in plant metabolism, (b) affect
plant resistance to stress, and (c) affect plant regrowth after winter dormancy
or cutting.  Other studies conducted to determine changes in metabolite pools
in plants exposed to 03 have focused on high concentration, short duration
exposures (4, 5, 6).  Little data on long-term, chronic 03 effects are avail-
able.  Ponderosa pine seedlings exposed to 200 yg/m3 03 for 20 weeks showed a
retention of sugars and starch in the top of the plant and a reduction in the
roots (7).

     Metabolites from the stems and leaves of legumes are translocated to the
roots to provide the energy needed to reduce and incorporate atmospheric
nitrogen (N) into the organic forms that can be used for growth (8).  This
suggests that a reduction in metabolites translocated to the alfalfa roots
reduces the amount of atmospheric nitrogen (N) fixed, thus reducing plant
growth and yield.

     The objective of this research was to determine the effects of chronic
levels of 03 and sulfur dioxide (S02) (singly and in combination) on alfalfa
yield, quality and N-fixation as they related to changes in metabolite levels.

                             EXPERIMENTAL PROCEDURES

     Two experiments were conducted with alfalfa, Medicago sativa L. cv.
Mesa-Sirsa at the EPA's Corvallis Environmental Research Laboratory Farm in
Oregon.  In each study, seed was inoculated with an N-fixing bacteria before
planting in 20 cm diameter plastic pots.  For the established study, the pots
were filled with a peat-vermiculite mix (Jiffy Mix-Plus) containing a slow
release fertilizer; for the sand study, 20 mesh silica sand with a modified
Hoagland solution minus N was added daily.  The only nitrogen source available
to the plants in the sand experiment was symbiotically fixed atmospheric N.

                                     664

-------
Two plants were grown in each of the pots allocated per treatment in  the
established (18 pots) and sand (10 pots) alfalfa studies.   Established plants
were obtained by withholding the alfalfa from pollutant treatment until  after
the first harvest.  Treatment of plants in the sand study  began when  they were
approximately 10 cm seedlings.

     In the established study, plants were exposed for 70  days to either
charcoal-filtered air (control), or 200 yg/m3 03 in charcoal-filtered air;
each treatment was replicated three times.  Ozone was added six hours per day
(9 a.m. to 3 p.m.), seven days per week as described by Heagle et al. (9).
All plants received charcoal-filtered air without ozone the remaining 18 hours
each day.

     Plants were exposed for 68 days in the sand alfalfa study.  One  exposure
chamber was allocated to each treatment:  (a) charcoal-filtered air (control),
(b) 100 yg/m3 03, (c) 156 yg/m3 SO?, and (d) 100 yg/m3 03  plus 156 yg/m3  S0:.
Ozone was added seven hours per day (9 a.m. to 4 p.m.), seven days per week
in charcoal-filtered air; sulfur dioxide was added to either charcoal-filtered
air or charcoal-filtered air plus ozone 24 hours per day,  seven days  per week
as described by Heagle et al. (10).  These plants also received charcoal-
filtered air when the pollutants were not present.  The tops (forage) of all
plants were clipped 8 cm above the growing medium surface  when the controls
reached the 1/10 bloom stage of growth.  Dry weight (DW) of forage was obtained
at each harvest in both studies by drying all plant material  from each pot at
70°C. for 48 hours.  Root and stubble dry weights were obtained only  at the
last harvest (second) of the sand alfalfa experiment.

     Forage from two pots were batched in the established  alfalfa before
metabolite analysis, providing 9 samples per treatment compared to 10 in the
sand alfalfa study.  Dried plant material was analyzed for total protein,
amino acids, phenols, total nonstructural carbohydrates (TNC) and ^-carotene
content by methods previously described (7, 11, 12).

     The data are presented as percent of control with associated 95  percent
confidence limit estimates or other indicated levels for statistically sign-
nificant differences.

                                     RESULTS

ESTABLISHED ALFALFA EXPERIMENT

     Forage metabolite content per unit weight of sample at the first harvest
did not significantly vary from the control for any of the metabolites meas-
ured (Figure 1).  At the second harvest, the amino acid and phenol content of
the forage was significantly higher in the 03-treated plants; no significant
differences occurred in the other metabolites.  By the third harvest, amino
acid and total protein content were 119 and 128 percent of the control, re-
spectively.  There were no significant differences from the control in total
nonstructural carbohydrates (TNC), total phenols or lignin contents,  but the
p-carotene content was reduced to 76 percent of control.

     At the first harvest, the total amount of measured metabolites produced


                                     665

-------
   O
   cr
   h-
   2
   O
   O
   LL
   O
   LJ
   O
   cr
   UJ
   Q_
                 PROTEIN
                 AMINO
                 ACIDS
               HARVEST I      HARVEST II      HARVEST HI

   Figure 1.  Metabolite content of established  alfalfa forage (unit weight).
    *Each value is based on nine observations for each mean.   Significant
                       differences  are shown as:  *5% and **!%.

per sample in forage by Ch-exposed  plants did not vary significantly from  the
control (Figure 2).  The only  significant differences occurring  at  the  second
harvest were a reduction (75%  of control) in TNC and dry  weight  (80% of con-
trol) produced per sample of forage stressed with 03.  By the third harvest,
the total metabolites produced per  sample of 03-stressed  plants  were 40 to 60
percent of control for each of the  metabolites investigated.   The dry weight
of forage treated with 03 was  49 percent of control at the third harvest.

SAND ALFALFA EXPERIMENT

     At first harvest, TNC  forage content on a unit weight basis in  plants
treated with 0<  or S0? singly was greater than the  control  (Figure  3);  this
significant difference was  not present at the second harvest.  However, root
or stubble  TNC  content on a unit basis was  80 to 85 percent of control  for
each of three pollutant  treatments at the second harvest.   The protein  con-
tent of the forage on  a  unit weight basis at first  harvest was not  signifi-
cantly different  from  the control in any of three treatments  (Figure  4).  By
                                     666

-------
     140
PROTEIN

AMINO ACIDS
                                                      LIGNIN

                                                       -CAROTENE
                                                   m OVEN DRY WEIGHT
              HARVEST  I
                 HARVEST II
HARVEST III
   Figure 2.  Metabolite content of established  alfalfa  forage (total weight).

        *The metabolite content per  unit weight was multiplied by  the
         total  treatment weight to give total metabolite produced.
         Each percent is based  on  nine observations for each  mean.
         Significant differences are shown as:  *5% and **!%.

the second  harvest,  protein content in the forage was  113,  123, and  123 per-
cent of control  for  the  Q->, S02, and 03 + S02 treatments,  respectively.  How-
ever, the protein  content of the stubble or roots at the second harvest
was not significantly  different from the control  regardless of treatment.
At the first harvest,  the total amount of TNC produced per  treatment in
forage of plants exposed to 03, S02, and 03 + S02 was  68,  57, and 62 percent
of control,  respectively (Figure 5).   By the second harvest, the total amount
of TNC in the forage  per sample was further reduced to 45,  47, and 54 percent
of control,  respectively.  Totals for root TNC were approximately 43 percent
of control  for each  of the pollutant treatments.

     The total amount  of protein present in forage at  the  first harvest was
approximately 63 percent of control in both the S0: and  S0: + Ch treatments;
the total  protein  in  the 0^ treatment was  not significantly different from
the control  (Figure  6).   At the second harvest, the total  protein produced
in forage of plants  exposed to 03,  S02, and 03 +  S02 was 58, 62, and 69 percent
                                    667

-------
          140
        o
        or
        o
        o
        u_
        o
        Ld
        O
        cr
        LU
        Q_
           120
           100
80
           60
           40
           20
             0
       I


            I

                       i
1
                                             FORAGE

                                             STUBBLE
   e°-3 ROOTS


                   H1    H2
                       0,
                       H1     H2
                          SO,
           H1
             0,
H2
SO,
                 Figure 3.   Total  nonstructural carbohydrate
                   content  of sand alfalfa (unit weight).

     *Each value  is  the mean of ten samples per treatment; H1=Harvest 1,
                                   H2=Harvest 2.

of control,  respectively.  Total protein in the  stubble was 64, 79, and 70
percent of control for 03, S02,  and 03 + S02  treatments, respectively.   Total
protein in the  plant roots of each treatment  group was approximately 60 per-
cent of control.

     The amount of nitrogen in a plant is  related to the amount of protein
in a plant.   Therefore, the same differences  shown above for protein are also
truo for the amount of atmospheric N  that  was  symbiotically fixed and present
in the nllalfa  tissue.

     Iho dry weight of forage at the  first harvest was 69,  54, and 61  percent
of control  for Ot, S0:., and 0,, + SO.*  treatments, respectively  (Figure  7); and
at the second harvest, the forage  dry  weights  were further  reduced to  51, 50,
                                    668

-------
      140
      120
   §
   o
   o
   Lu
   O
   0
   cr:
   LJ
      100
      80
      40
      20 -
        0
                  FORAGE   ~1 STUBBLE

                          .-:o
                           1
I
*1
                           n
                                 i
                                       rsx'

                                I
                                               ROOTS
0°0
O o
06
                                  H1     H2
                                     SO.


         H1     H2

             vy-j               ^j v-^o            V^'"X '   ^^ ^~^r)

Figure  4.  Total protein  content of  sand alfalfa (unit  weight).
                   H1
                    0,
                              :*
                                                               °"°:
                                                               O 0
                                                            S0
and 56
and 03
while
      percent of control, respectively.   Stubble dry weights  for the 0-,,  S025
      + S02 treatments  were 61, 72, and  72 percent of control, respectively^
      root dry weights were approximately 55 percent of control.
                                 DISCUSSION

     The.*  pollutant treatments  caused greater reductions (compared  to con-
trols)  in metabolites produced per sample at successive harvests.  This
indicates that the plants  became less vigorous  as the exposure  time increased,
because the stressed plants  had decreasing total metabolite pools  for growth.
                                   669

-------
       100
    O
        80
    o
    O
    u_  60
    o
    h-
    tr
    UJ
       40
         0

                H1     H2
                     0,

    FORAGE

    STUBBLE

     ROOTS

                                        i
HI     H2
    SCL



H1
  0,
H2
SO,
                 Figure 5.  Total nonstructural carbohydrate
                   content of sand alfalfa  (total weight).

The early  large reduction in  yield of the sand alfalfa exposed to pollutants
noted at the  first harvest was  probably caused by these plants being stressed
at an early growth stage before metabolite pools were established for later
growth.  Exposure of alfalfa  and perhaps other crops during  early growth stages
is apparently very damaging to  yields.

     According to Smith and Silva (13), accumulated carbohydrates in alfalfa
roots are  primarily used for  forage rather than root growth.  The greater
than 50  percent loss in total TNC stored in the roots of stressed plants
at the second harvest of the  sand alfalfa experiment indicates that the
plant's  capacity to produce new top growth had been reduced  significantly.
Also, loss  in root dry weight adversely affects the plant's  ability to absorb
nutrients  needed for growth (14).   The reduction in stored carbohydrates
and smaller roots in pollutant-treated plants probably accounts for the large
reduction  noted in both studies.   Research on tree species exposed to low
levels of  ozone also indicated  greater reductions in plant dry weights (com-
pared to controls) as the season progressed (15, 16).
                                    670

-------An error occurred while trying to OCR this image.

-------
      100
   o
   ^80
   o
   o
   u_60
   o
   LU 40
   o
   cr
   UJ
        0


H1      H2
     0
                      FORAGE

                      STUBBLE
                      ROOTS

                       I


                                 H1     H2
                                     S 0,
H 1
  o.
                      3                  2

                  Figure 7.   Dry weight of sand  alfalfa.
H2
 SO,
                                   REFERENCES

1.    Howell, R. K.,  T.  E.  Devine, and C.  H. Hanson,  Resistance of Selected
     Alfalfa Strains to Ozone.  Crop. Sci.  2:114-115,  1971.

2.    Tingey, D. T.,  and R. A. Reinert.  The Effect of Ozone and Sulfur Dioxide
     Singly and in  Combination on Plant Growth.  Environ.  Pollut.  9:117-125,
     1975.

3.    Oshima, R. J.   Development of a System for Evaluating  and Reporting
     Economic Crop  Losses  Caused by Air Pollution in California III and IIIA.
     State of California Department of Food and Agri.   Sacramento, Cal. , 1975.
     103 pp.

4.    Ting,  I.  P., S. K. Mukerji.  Leaf Ontogeny as a Factor in Suscepti-
     bility to Ozone:  Amino  Acid and Carbohydrate Changes During  Leaf
     Expansion.  Amer.  J.  Bot.  58:497-504, 1971.
                                   672

-------
 5.    Tomlinson, H.,  and S. Rich.   Metabolic Changes in Free Amino Acids of
      Bean Leaves Exposed to Ozone.  Phytopathology  57:972-974, 1967.

 6.    Tingey, D. T.,  R. C.  Fites,  and C. Wickliff.  Activity Changes in
      Selected Enzymes from Soybean Leaves Following Ozone Exposure.
      Physiologia Plantarum  33:316-320,  1975.

 7.    Tingey, D. T.,  R. G.  Wilhour, and C. Standley.  The Effect of Chronic
      Ozone Exposures on the Metabolite Content of Ponderosa Pine Seedlings.
      Forest Science  (In Press), 1976.

 8.    Gibson, A. H.   Limitation to Dinitrogen Fixation by Legumes.  In:  Nitro-
      gen Fixation.   W. E.  Newton  and C. J. Nyman, eds.  Washington State Univ.
      Press, Pullman, Wash., 1976.  pp.  400-428.

 9.    Heagle, A. S.,  D. E.  Body, and E. K. Pounds.  Effect of Ozone on Yield
      of Sweet Corn.   Phytopathology  62:683-687, 1972.

10.    Heagle, A. S.,  D. E.  Body, and G. E. Neely.  Injury and Yield Responses
      of Soybean to Chronic Doses  of Ozone and Sulfur Dioxide in the Field.
      Phytopathology  64:132-136,  1974.

11.    Smith, D.   Removing and Analyzing Total Nonstructural Carbohydrates from
      Plant Tissue.   Research Report 41.  University of Wisconsin, Madison,
      Wis., 1969.

12.    Zscheile,  F.  P. and R. A. Whitmore.  Determination of Carotene in Alfalfa.
      Analytical Chemistry,  19:170-172, 1947.

13.    Smith, D.  and J. P. Silva.  Use of Carbohydrate and Nitrogen Root
      Reserves in the Regrowth of Alfalfa From Greenhouse Experiments Under
      Light and Dark  Conditions.  Crop Science  9:464-467, 1969.

14.    Tingey, D. T.   Ozone Induced Alterations in Plant Growth and Metabolism.
      Proc. of Inter. Conf. on Photochem. Oxidant Poll, and Its Control.   (In
      press), 1976.

15.    Wilhour, R. G., and G. E. Neely.   Growth Response of Conifer Seedlings to
      Low Ozone Concentrations.  Proc.  of Inter. Conf. on Photochem.  Oxidant
      Poll, and Its Control.  (In  press), 1976.

16.    Jensen, K. F.,  and R. G.  Masters.  Growth of Six Woody Species Fumigated
      with Ozone.  Plant Disease Reporter  59:760-762, 1975.
                                     673

-------
                   SESSION 14
REACTIVITY AND ITS USE IN OXIDANT-RELATED CONTROL

             CkcuAman:  J.G. Calvert
              Ohio State University
                       675

-------
                                                                            14-1
                 MULTIDAY IRRADIATION OF NO -ORGANIC  MIXTURES
                                           X
                      W.  A.  Glasson and P. H. Wendschuh*
 ABSTRACT

      An  experimental simulation ofi urban pollution transport to rural environ-
 ments  Mas  applied to (I)  the. irradiation oft several  hydrocarbons and organic
 solvent* and (2)  the irradiation o& a typi-cal urban  hydrocarbon mixture in the.
 presence 0(J  vaA.t/4.ng concentrations of, nitrogen oxi.des.   Maximum ozone cotacen-
 tration  and  ozone, dotage, uiere used 04 measures o^
      It  WOA  fiound that the. i.eact-tv-6tt£4 ojf thAe.e. automotive, paint AolventA in
oxidant  fiosunation,  although A^gnifacantly di.66eAe.nt in the.  e.aJily AtageJ* o& a
24-houA  c-ontinuouA iAAadiation,  meAge.d by the. e.nd  o$ the. >uw&.   In the. tsiant-
pont  A-imutation e,x.pesujmentA , it uxu ^ound that tie.acti.vity Lt> onl.y modeAateJLy
a£teAe.d  on the. second and thiAd day* o& iJAadlation.  Thai,  hydrocarbon ie.ac-
tivity  AtiM. an important coni>ideAatLon in contra tting organi-c ejniAAi.om>
e.ve.n  in  AituationA i.nvolvi.ng long- distance, trampcrt.

      In  the.  nitroge.n oxides variation Atudy, the. result* Ahowe.d that ozone.
veAAuA nitrogen oxA.de^> plot* ^or e.ach irradiation  day d*Uplaye.d the. typ-icaJi
nitric oxA.de.-inhibiti.cn Ahape..   An anatytiA o& thif> urban- to -rural transport
simulation i.ndicate.d that re.ducti.on ofa urban ni^troge.n oxi.dti> to control rural
oxi.dant  could result in an adve.M>e. ufafect on urban ozone, and little. e.^e.ct on
rural ozone..
                                 INTRODUCTION

     Transport of oxidant and/or oxidant precursors  has  been  suggested as the
cause of  rural  oxidant exceeding the air quality standard  of  0.08 pptn(l).  A
number of factors, including dilution, tend to decrease  reactivity differences
with increased time.   Presumably for this reason,  it  has been sugaested that
hydrocarbon  reactivity is not an important consideration in controlling oxidant
in rural  environments  in spite of the obvious advantage  in urban environments
(1).  As  a test of this suggestion, reactivities in  oxidant formation were
determined in  24-hour  continuous irradiations for  three  automotive paint sol-
vents and in an experimental simulation of urban-to-rural  transport using
several hydrocarbons  and organic solvents.  The simulation consisted of three
6-hour irradiation periods separated by two 6-hour dark  periods.   Atmospheric
dilution  was simulated by the nautral leakage from the chamber.
*General Motors  Research  Laboratories, Warren, Michigan,


                                      677

-------
     The results of recent studies have been interpreted to mean that rural
oxidant formation is enhanced by increasing nitrogen oxides(2,3).   These re-
sults lead to the suggestion that reduction of urban NOX will result in rural
oxidant reductions (1).  This suggestion was tested by studying the effect of
varying NOX concentration on ozone (03) formation in the photooxidation of a
hydrocarbon mixture typical of polluted urban atmospheres, using the transport
simulation described above in our work, maximum 03 concentration and 03
dosage were used as measures of oxidant.


                                EXPERIMENTAL

     The experiments were carried out in a smog chamber which has  been de-
scribed previously (4).  Losses due to sampling and leaks were compensated by
the addition of dilution air.  This dilution air was filtered, passed over an
oxidation catalyst, and saturated with water.  The dilution rate was 0.0015%
sec-1 as measured using sulfur hexafluoride (SF6) as a tracer, leading to 80%
loss of reactants in 30 hours.   The light intensity corresponded to ki =
0.042 sec-1, and the experiments were run at an average temperature and rela-
tive humidity of 34°C and 23%,  respectively.

     Nitric oxide/nitrogen oxides (NO/NOX) and 03 were determined continuously
by Monitor Labs Model 8440 and McMillan Electronics Corporation Model 1100
chemi luminescent analyzers, respectively.  Individual  hydrocarbons were deter-
mined with a Perkin-Elmer 900 gas chromatograph equipped with dual flame ion-
ization detectors (FID's), using columns described previously (5).  Formalde-
hyde was determined by the chromotropic acid method (6).


                           RESULTS AND DISCUSSION

HYDROCARBON REACTIVITY

     In the first test of long-term irradiation on reactivity, 1.6 ppm of
three automotive paint solvents was irradiated with 0.8 ppm NOX continuously
for 24 hours.   The results are given in Figure 1.  Solvent C (pre-Rule 66)
formed the most 03 in the first six hours of irradiation, folio-wed by solvent
B (solvent A + 10% xylenes) and solvent A (exempt).  By the end of the 24
hours of irradiation, all the solvents were producing  similar amounts of 03.
It is clear, though, that the 03 dosage experienced in the first six hours is
terms of the atmosphere, the urban environment would be exposed! to higher 03
substantially greater for solvent C and, thus, in concentrations in this case.
It might be argued that continuous irradiation experiments are unrealistic
representations of amtospheric transport processed, due to diurnal variations
in light intensity.

     We have studied the effect of long-term irradiation of hydrocarbon reac-
tivity using the experimental simulation described above.  The results of a
typical  multiday irradiation expeiment are given in Figure 2 for the irradia-
tion of 1.6 ppm solvent 3600 (C in Figure 1) and 0.8 ppm NOX.  Succeeding
irradiation periods show decreasing 03 and nitrogen dioxide (N02)  concentra-
tions, while NO is essentially  completely oxidized in  the first irradiation


                                     678

-------
              < CD  O
oo
v-

o
"/
I
\
\
\
/:
' '
n
-\\
CO \ \
—

1
vO
o-


Jill
LT\ *& Cl^l C\J
o' o o o
*•

1
r— 1
0
X' -
^

c

                                                                    C\J
                                                                    CNJ
                                                                    o
                                                                    C\J
                                                                    oo
                                                                                o

                                                                                t- to
                                                                                   Ol
                                                                                C i.
                                                                                o 3
                                                                                •i- +j
                                                                                -»-> x
                                                                                O -r-
                                                                                O   X
                                                                                s- o
                                                                                Q. -Z.

                                                                                O 4-1
                                                                                E E
                                                                                O CU
                                                                                N >
                                                                                O r-
                                                                                   O
                                                                                f- oo
                                                                                o
                                                                                O -r-
                                                                                i/l 00
                                                                                •r- CL

                                                                                ro O>
                                                                                 O)
                                                                                 i.
                                                                                 CTl
                                                                    CSJ
                     ujdd 'uo!}ej}U9Duoo  auozg

                                 679

-------
   E  E  I
   o.  a.
   O.  0.

      00 O O
   ^J O Z Z O
                                                           13
                                                           O
o>
E
                                                                  S-
                                                                  (1)
                                                                  Q.
                                                                  X
                                                                  c
                                                                  o
                                                                  (O
                                                                 •r—
                                                                 "O
                                                                  s-
                                                                  fO
                                                                  -o
                                                                  o
                                                                  a.
                                                                  CM

                                                                  
-------
period.  Nitrogen dioxide and 03 decrease in the dark periods due to a combina-
tion of dilution, chemical reaction, and wall losses.  Ozone concentrations
formed in a similar experiment without added hydrocarbon or NOX can be used as
a measure of chamber reactivity.  In such an experiment, the irradiation of
background concentrations of hydrocarbon and NOX (0.05 ppm and 0.08 ppm, re-
spectively) produced 0.007, 0.011, and 0.012 ppm 03 in the first, second, and
third irradiation periods, respectively.  These values suggest that chamber
reactivity is not a significant factor in our experiments.

     The results of the irradiation of 1.6 ppm of a series of hydrocarbons,
solvents, and auto exhaust with 0.8 ppm NOX are given in Table 1.  In general,
in each succeeding irradiation period less 03 is produced, with the exceptions
of n-butane and solvent 3694 (a largely paraffinic solvent).  Comparisons of
the data of this study with literature data show general agreement within the
ranges found in the literature.

TABLE  1.  MULTIDAY  IRRADIATIONS OF  VARIOUS  ORGANIC COMPOUND-NO   MIXTURES
                                                              /^ "~     "  '

                                                (03)max  (ppm)
Organic  reactant                    1st  Day        2nd Day        3rd Day
Benzene
Toluene
n-Butane
Propylene
cis-2-Butene
Tetramethyl ethyl ene
Trichloroethylene
Solvent 3694
Solvent 3600
Auto exhaust
0.029
0.25
0.10
0.47
0.43
0.51
0.15
0.15
0.34
0.47
0.016
0.046
0.16
0.21
0.21
0.19
0.11
0.17
0.13
0.23
0.018
0.029
0.077
0.10
0.12
0.063
0.041
0.076
0.073
0.13
     If differences in hydrocarbon reactivity diminish with irradiation time,
then the 03 concentrations formed in the second and third irradiation periods
will show a decreased dependence on the 03 formed in the first irradiation
period.  A regression analysis of the data shown in Table 1 displayed correla-
tion coefficients of 0.76 and 0.67 for the second and third irradiation per-
iods, respectively.  Thus, although the differences in hydrocarbon reactivity
do diminish with irradiation time, the effects of hydrocarbon reactivity on 03
formation are still apparent after three irradiation "days."

EFFECT OF NITROGEN OXIDES

     In the second part of this study, the irradiation of a typical  urban
hydrocarbon mixture was studied as a function of the NOX concentration.  The
average composition of the non-methane hydrocarbon (NMHC) mixture was (%,

                                     681

-------
 v/v): ethane, 9.2; ethylene, 11.6; propane, 18.3; propylene, 8.2; n-butane,
 23.0; isopentane, 10.3; benzene, 4.4; toluene, 9.8; and m-xylene, 5,, 3.   The
 average initial NMHC concentration in our experiments was 0.935 ppm,  and the
 reaction mixture also contained an average of 17 ppm carbon monoxide  (CO) and
 2.8 ppm methane.

     The effect of the NOX concentration at the beginning of each irradiation
 period on  (03)max is given in Figure 3.   The results show the typical NQ-
 inhibition curve for each irradiation period with the highest (03)mdx occur-
 ring at lower NOX concentrations for each succeeding irradiation period.  This
 effect is  undoubtedly due to the lower hydrocarbon concentration present at
 the start  of each succeeding irradiation period.  As expected, the olefins are
 destroyed  most rapidly, while the aromatics are destroyed more slowly and the
 paraffins  are destroyed most slowly.   The reactivity of the system is supple-
 mented by  the presence of the product formaldehyde, and undoubtedly,  other
 aldehydes  as well.  The average formaldehyde concentrations at the start of
 the second and third irradiation periods was 0.122 and 0.088 ppm, respectively.

     Figure 3 cannot be used to properly evaluate the effect of urban NOX con-
 trol on rural ozone since the reduction in NOX would occur primarily  in the
 urban area.  Thus, in Figure 4, 03 formation for all three periods is plotted
 against the NOX concentration at the beginning of the experiment.  Again,
 typical NO-inhibition curves are obtained for the three irradiation periods,
 except that the NOX concentration required for maximum 03 formation increases
with each  succeeding irradiation day; the latter observation is related to the
 fact that  the NOX concentration at the beginning of each succeeding irradiation
 period is  not directly proportional to the NOX concentration at the start of
 the experiment.   The curves suggested that reductions in urban NOX concentra-
 tions will increase 03 on the first day (urban), slightly reduce 0\ on the
 second day (rural), and have little effect on 03 on the third day (rural).
 Thus, the  result of reducing urban NOX concentrations may be to trade off
 higher 03  exposures in the urban population centers for small effects in the
 sparsely populated rural environment.

     Current oxidant control strategy involves hydrocarbon reduction, which
should shift the maxima in Figures 3 and 4 to lower NOX concentrations.  It
seems likely that the potential benefits of NOX control will be even  more dif-
 ficult to  obtain as hydrocarbon levels are reduced.

     It is recognized that the simulation used in the present study is simpler
 than the atmosphere.   Further work is planned to refine the simulation.

     Due to space constraints, this report does not include much of the experi-
mental  detail.   This information can be found in References 6 and 7.


                                  REFERENCES

 1.   References in "Control of Photochemical Oxidants  -- Technical Bases
     and Implications of Recent Findings."  U. S.  Environmental  Protection
     Agency, Office of Air Quality Planning and  Standards,  Researchearch
     Triangle Park, NC (1975)

                                     682

-------
    0.25
    0.20
   0.15
E
a.
_O-

x
TO

 E
   0.10
    0.05
     0.0
                       o
                                      1st Day
2nd Day
      0.0    0.1     0.2    0.3     0.4    0.5    0.6    0.7    0.8    0.9


                                (NO ) (ppm)
                                   A
               Figure  3.   Effect of  NO  on Irradiation

                           of Typical  Urban HC Mix.

                                 683

-------
0-°      0.1    0.2    0.3   ~ 0.4    0.5 ~   0.6    0.7     0.8    0.9
                           (NO) (ppm)
                              X
       Figure 4.   Effect of Urban  NO  on Rural  Ozone.
                                      X
                             684

-------
2.    Jeffries, H.  E., J.  E.  Sickles, II, and L.  A.  Ripperton.   "Ozone
     Transport Phenomena:   Observed and Simulated."  Presented to the
     69th Annual  Meeting of the Air Pollution Control  Assoc.,   Portland,
     OR (1976).

3.    Grimsrud, E.  P.  and R.  A.  Rasmussen.   "Photochemical Ozone Production
     from Captured Rural  Air Masses of Ohio and  Idaho."  Report to the EPA,
     Grant No. 800670  (1974).

4.    Tuesday, C.  S.,  B. A.  D'Alleva, J. M.  Heuss, and G. J. Nebel.  "The
     General  Motors Smog Chamber."  Presented to the 58th Annual  Meeting of
     the Air Pollution Control  Asso., Toronto, Ontario, Canada  (1965).

5.    Groblicki, P. J., R.  S. Eisinger, and M. A. Ferman.  "Design of a
     Mobile Atmospheric Research Laboratory."   GM Research Publ. GMR-1914
     (1975).

6.    Wendschuh, P. H.  "The Photochemical  Reactivity of Several Automotive
     Paint Solvents."  GM Research Publ.  GMR-2079  (1976).

7.    Glasson, W.  A. and P.  H. Wendschuh.  "Multiday Irradiation of
     NO -Organic  Mixtures."  GM Research Publ.  GMR-2236  (1976).
       X
                                     685

-------
         HYDROCARBON REACTIVITY AND  THE  ROLE OF HYDROCARBONS,  OXIDES OF
                  NITROGEN, AND AGED SMOG IN THE PRODUCTION  OF
                             PHOTOCHEMICAL OXIDANTS

                 J.  N. Pitts, Jr., A.  M.  Winer, K. R. Darnall,
                         A. C. Lloyd,  and G. J. Doyle*
ABSTRACT
     A fie.activtty claf,^i^i.catA,on  fan. a. bfioad -6pe.cit.um ofi atmo&pkeJiic.aJLly i.m-
pofitant ofiaanic (and Ae.ve.fial tnoftganic]  compounds ka& ban  developed.  It iA
bat>e.d on the,  fiates o^ fie.action o& these. compounds with the,  hydfiozyl fLO.dic.oJi, the.
pfiimafiy chain cafifiieA -in pkotoc.kimi.C-at ait pollution and, indeed,  -in
cke.mii.tny.  The. cloAAi^-ication kat>  been deue£oped in the. conte.Kt o£ a 5-
fie.activtty A cole. i.n which eo.cn clat>A &pam> an ofide.fi oft magnitude. i.n
le-iative. to methane..   An attempt  hot been made in tkit> papeA. to indonpofiate. ail
compounds in  the. x.e.activtty cJLa&Ai&'ic.ation fan. whic.h hydtiOKyl x.adic.aJL note.
conbtantA have, been e.x.p&iAe.nt cJju>&i.^i.c.atA.on contains
mofie. than 110 compound*,, including  aikanu, a£feene6, afiomaticA , cxt/gena£e6,
natuJiai hyd?ioc.afibonA, halo carbon*,,  amines, and Anlfiide^ .  kpptiOKimateJLy halfi o&
the. note, constant* employed weAe.  de^te.fmine.d -in tkif> labofiatony by eAt.ke.fi {jla&h
photolyAAJ>-tieJ>onanc.e. ^£ao/iei eence meaiuAemenAs ofi by relative, fiate. me.akufie.me.ntt>
in two ^6000  lite.fi e.nvisionme.ntal  c.kambe.fii> , oft by both techniques .

                                 INTRODUCTION

     The concept that different hydrocarbons, and their oxygenated and halo-
genated derivatives, have differing  capacities to produce photochemical oxidant
orginated some 20 years ago in the  classic work of Haagen-Smit and Fox (1).
Following contributions by many other workers, this concept has been incor-
porated in local, state, and federal air pollution control  strategies for at
least a decade, beginning with the  pioneering efforts of the Los Angeles County
Air Pollution Control District (LAAPCD).   Thus, in 1966, the LAAPCD implemented
a regulation  known as Rule 66 restricting the composition of organic solvents so
as to reduce  the use of compounds that were observed to be  of the highest re-
activity in a major study by LAAPCD  workers (2).  Since then, while there has
been a range  of opinion within the  scientific community as  to the absolute
accuracy of predictions of reductions of ambient ozone (0^,) levels based on
  J.  N.  Pitts, A. M. Winer,  G.  J.  Doyle, and  K.  R.  Darnall, Statewide  Air  Pol-
  lution Research Center  (SAPRC),  University  of  California, Riverside,  CA.
  A.  C.  Lloyd, Environmental  Research and Technology, Santa Barbara, CA.

                                      687

-------
detailed application of such reactivity scales,  they have nevertheless received
general acceptance as rational and useful  elements of cost-effective oxidant
control strategies.

     Traditionally,  such reactivity classifications have been based on smog
chamber measurements of various phenomena  observed in the photooxidation of
hydrocarbon-oxides of nitrogen mixtures under conditions approximating those of
polluted ambient atmospheres.   These include the rates of hydrocarbon consump-
tion, nitric oxide (NO) to nitrogen dioxide (N02)  conversion and 03- formation,
as well as maximum levels of 03 reached, total  03  dosage, aerosol  formation,
degree of eye irritation, plant damage, etc.  Much of the reactivity data which
had been accumulated through 1969 for each of these manifestations (except plant
damage) was critically reviewed by Altshuller and  Bufalini  in 1971 (3).   In
addition to noting general agreement in reactivity trends obtained in various
studies (some employing different reactivity criteria), they cited a number of
significant discrepancies in the specific  assignments of reactivity to indi-
vidual compounds and even to whole classes of compounds.

     Such differences are not surprising since the observed values of the
secondary smog manifestations cited above  can be substantially affected by a
number of aspects of the smog chamber analytical and experimental  methodologies
employed.  Important factors in this regard are, for example, intensity and
spectral distribution of the ultraviolet (UV) light source, surface to volume
ratio and nature of the walls and windows  of the chamber, specific measurement
methods employed for reactants and products, purity of matric air, hydrocarbon
(HC) to nitrogen oxides (N0x) ratios, length of irradiation, etc..   Clearly,
differences in reactivity scales are not only of concern per se, but also
because they can lead to important differences in  emission inventories for
reactive hydrocarbons and in strategies for their  control (4-6).

                             RECENT DEVELOPMENTS

     In 1974 the Environmental Protection  Agency (EPA) held a major conference
on solvent reactivity.  Existing hydrocarbon reactivity data (including those
from chamber studies by Battelle, SRI, and Shell)  and approaches to formulating
hydrocarbon reactivity scales were discussed and evaluated at this conference.
Consideration was also given to preliminary evidence then being accumulated
concerning the occurrence of pollutant transport and the recognition that such
transport will enhance the formation of oxidant from "low reactivity" organics.
Data gathered in the succeeding two years  has confirmed the necessity for
control officials to consider the problems of (a)  long-range transport of 03 and
its precursors, and (b) the occurrence of  meteorological conditions resulting in
multi-day "stagnant air" episodes (leading to persistent high 03 levels) in
which ambient air contains not only each day's fresh pollutant burden but also
"aged smog" carryover from the previous day(s).   Under each of these conditions,
a number of compounds previously cited in  some classifications as  "nonreactive"
or of "low" reactivity—for example, propane and n-butane--can react to produce
substantial amounts  of 03.

     The oxidant generating capacity, during long  irradiation periods,, of
hydrocarbons (and oxygenates, halogenates, etc.) formerly classified to be
"nonreactive" has posed serious problems for regulatory agencies.   Control

                                     688

-------
officials are faced with balancing today's serious economic and energy con-
straints against the need to protect the public from elevated oxidant levels not
just in urban centers but also tens to hundreds of kilometers "downwind" of the
urban centers where millions of suburbanites reside.  Dispersion of urban plumes
over long distances also complicates the regulation of hydrocarbon emissions
from industries located in suburban and rural areas, since such areas may
experience 03 levels exceeding the Federal Air Quality Standard despite the fact
that local industries may not be the major sources of 03 precursors.


            REACTIVITY SCALE BASED ON REACTION WITH HYDOXYL RADICAL

     The need for new experimental data sets upon which to base contemporary
reactivity classifications has been emphasized by both the EPA and very recently
by the California Air Resources Board (ARB) (8).   An accepted method of gen-
erating the desired information would be a carefully designed environmental
chamber study in which the 03-forming capacity of each of a very large number of
organic compounds would be determined under a standard set of simulated
atmospheric conditions, including long-term irradiation (i.e., 10-12 hours).
Such a program would be extremely time consuming and expensive and would in-
evitably suffer to some degree from the problems noted in previous chamber
studies of this type.

     Recently, we have proposed a supplementary (rather than a substitute)
approach to obtaining the required data, an approach which does not suffer
significantly from the difficulties cited above (although subject to definite
limitations which will be discussed).  Specifically, we have proposed a hydro-
carbon reactivity scale based not on secondary smog manifestation criteria, but
rather on the primary chemical act of hydroxyl radical (OH) attack on organic
species (9, 10).  We have done so on two bases:  first, that the OH radical is
the most important reactive intermediate in photochemical air pollution, and
second, that a reactivity scale based on the depletion of hydrocarbons by
reactions with OH has utility in assessing hydrocarbon chemical behavior in
polluted ambient air,  since only those compounds which participate at "signi-
ficant" rates in atmospheric reactions are of consequence in the chemical
transformations in ambient air, including, of course, the production of photo-
chemical oxidant.  Thus, our working premise has been that the relative re-
activity of organics towards OH is generally a useful and directly measurable
index of their potential importance in the production of secondary pollutants.
As we have previously pointed out, however, if specific reactivity criteria
unrelated in varying degrees to oxidant formation are of concern—for example,
aerosol formation and associated visibility effects, mutagenic activity, etc.--
our proposed reactivity classification may or may not have any validity what-
soever (11).

            BACKGROUND AND APPLICABILITY OF THE PROPOSED HYDROXYL
                      RADICAL REACTIVITY CLASSIFICATION

     The important role of OH in photochemical smog formation is now well
established on the basis of both experimental and computer modeling studies (12-
20), and OH has been directly observed in both the laboratory and in ambient air
(21-23).  This awareness has led to a dramatic increase during the past several


                                     689

-------
years in the number of compounds for which OH rate constants have been measured.
For example in this laboratory, in addition to extensive absolute rate measure-
ments by flash photolysis-resonance fluorescence (24-34), we have employed two
^6000 liter environmental chambers to measure relative rates, with an accuracy
of ±20%, for the reaction of OH with more than 40 compounds under simulated
atmospheric conditions of temperature, pressure, concentrations, and light
intensity.   These relative rate constants were placed on an absolute basis using
the published rate constants for OH + n-butane or isobutene.  The detailed
kinetic data derived from these investigations have been previously reported
(34-40).  This technique has also been employed recently by Niki arid coworkers
(41).

     From the successful correlation of OH rate constants with the rates of
hydrocarbon disappearance observed in chamber simulations at the SAPRC (34-39)
and Ford Motor Co.  (41) laboratories, we conclude that, to a good approximation,
this correlation can be extrapolated to the atmosphere for (a) alkenes in
ambient air parcels during the early morning hours when 03 levels are generally
quite low (<0.05 parts per million (ppm)), and for (b) alkanes, aromatics, and
most of the other classes of compounds discussed here at essentially all times
and locations.   The latter assumption, namely that an OH rate constant is a
good "reactivity index" for alkanes, aromatics, halocarbons, etc. throughout an
irradiation day (or multiple irradiation days), rests upon the fact that the
rates of reaction of these classes of organics with species such as 03, 0(3P)
atoms and the hydroperoxyl radical are several orders of magnitude slower than
with OH (42-46).  For example, even at the highest ozone concentrations ex-
perienced in ambient atmospheres, 03 will not contribute significantly to the
consumption of alkanes and aromatics.  This is in contrast to the case for
alkenes which, although the rate constants for reaction of 03 with alkenes are
not particularly large (46-48), react rapidly with 03 at the average concen-
trations commonly encountered in polluted ambient air (O.1-0.2 ppm).

     Although a reactivity scale based on OH rate constants can, in principle,
be divided  into a large number of classes, it is convenient, particularly for
purposes of comparison with other scales, to use the five-class scale we have
previously  formulated (10).  This scale is based on order of magnitude dif-
ferences in reactivities of organics toward OH relative to methane (=1).  The
scale is shown in Table 1.  Table 1 also gives the range of half-lives of
compounds appearing in each class assuming depletion_solely due to reaction with
the OH radical  at a concentration of 107 radicals cm  .  Our initial resulting
reactivity  classifications, based on available OH rate data in References 9 and
10, contained 80 and 34 organic compounds, respectively.  In this paper we
attempt to  present a reactivity classification which is exhaustive with respect
to the reported room temperature literature for OH rate constant determinations
(as well as recent, as yet unpublished, measurements in this laboratory).  Thus
we have compiled a reactivity classification consisting of 113 compounds as-
signed into the 5 classes of our proposed reactivity scale.  Table 2 gives the
rate constant (with its source),* the reactivity (relative to methane =1), and
proposed classification for each compound, with the compounds arranged by
* The rate constant data given in Table 2 are not to be construed as criti-
  cally evaluated or recommended values.   In general,  absolute rate deter-
  minations were used when available, and average values are often given.

                                     690

-------
     TABLE 1.  REACTIVITY SCALE FOR ORGANICS  BASED ON RATE OF  CONSUMPTION
                    DUE  TO REACTION WITH THE HYDROXYL RADICAL
  Class
Ha If-life3
  (days)
Reactivity Relative
  to Methane  (=1)
     I

    II

   III

    IV

     V
    £10

   1 - 10

 0.1 - 1

0.01 - 0.1

   <0.01
       I10
     10 - 100
    100 - 1000

   1000 - 10,000
     >10,000
  at-)/2 = 0.693/kontOH].   [OH] is assumed  to  be  10? radicals cm-s,
 chemical  class.  In view of the growing attention being given  to  nitrogen- and
 sulfur-containing compounds in the polluted  troposphere, we  have  included
 ammonia,  methyl amine,  hydrogen sulfide, and  methyl  mercaptan in our classi-
 fication  (as well as carbon monoxide as a matter of interest).  For convenience,
 we  also present, in Table 3, the compounds appearing in each of the 5 classes of
 the SAPRC reactivity scale in order of increasing reactivity within each class.

       TABLE 2.  REACTIVITY CLASSIFICATION OF ORGANICS (AND  SELECTED  INORGANICS) BASED
                        ON REACTION WITH THE HYDROXYL RADICAL

kOH+Cpd
Compound (j> mol-1s"1)xlO"9 Reference
Reactivity
Relative to
Methane = 1
Proposed
Class
Alkanes
Methane
Ethane
Propane
n-Butane
Isobutane
Cyclobutane
n-Pentane
Isopentane
Cyclopentane
Neopentane
n-Hexane
Cyclohexane
2-Methylpentane
3-Methylpentane
2,3-Dimethylbutane
0.0048 64
0.17
1.3
1.8 31,
1.3
0.72
3.9
2.0
3.7
0.49
3.6
4.2
3.2
4.3
3.1
(continued)
691
,70,71,74,76
70,72,76
69,76
62,68-70,75,78
41,70
69
41
36
79
70
36,41,63
41,69,70
36
36
39


                                   1
                                  35
                                  270
                                  375
                                  270
                                  150
                                  810
                                  420
                                  765
                                  100
                                  750
                                  875
                                  665
                                  900
                                  645
                    I
                   II
                  III
                  III
                  III
                  III
                  III
                  III
                  III
                  III
                  III
                  III
                  III
                  III
                  III

-------
TABLE 2.   (Continued)

Compound
2 ,2 ,3-Trimethyl butane
2 ,2 ,4-Trimethyl pentane
n-Octane
2 ,2 ,3 ,3-Tetramethyl butane
Alkenes
Ethene
Propene
1-Butene
Isobutene
cis-2-Butene
trans-2-Butene
1-Pentene
cis-2-Pentene
2-Methyl-1-butene
2-Methyl-2-butene
1-Hexene
Cyclohexene
2 ,3-Dimethyl -2-butene
3 ,3-Dimethyl -1 -butene
1 -Methyl cycl ohexene
1-Heptene
Di -Alkenes
Allene
1 ,3-Butadiene
Isoprene
Al kynes
Acetylene
Methyl ancetylene
Aromatic Hydrocarbons
Benzene
Toluene
n-Propyl benzene
I sopropyl benzene
Ethyl benzene
p-Xylene
o-Xylene
m-Xylene
p-Ethyl toluene
o-Ethyltoluene
m-Ethvl toluene
p-Cyiwnp
1 ,2,3-Tritnethylbenzene
i ,2 4-Tnmecnyl benzene
1 ,3,5-1"- imethyl benrene


kOH+Cpd
( mol-^-Oxlcr9
2.3
2.3
5.1
0.67

4.7
15.1
21.3
• 30.5
32.3
42.1
18
39
35
48
19
43
92
17
58
22

2.7
46.4
47

0.099
0.57

0.85
3.6
3.7
3.7
4.8
7.4
8.4
14.1
7.8
8.2
11.7
9.2
14.9
20
29.7
(continued)
692
Reference
39
70
70
70

33
27
27
27
27
27
41
41
41
29
41
39,41
75
41
39
39

61
36
37

66
61

25,65
25,65
36
36
36
25,35
25,35
25,35
36
36
36
37
25,35
25,35
25,35


Reactivity
Relative to
Methane = 1
480
480
1060
140

985
3145
4440
6355
6730
8770
3750
8125
7290
10,000
3960
8960
19,170
3540
12,080
4585

560
9,670
9,790

21
120

180
750
770
770
1000
1540
1750
2940
1625
1710
24 4 n
"l 920
31 00
41 7'J
6190


Proposed
Class
III
III
IV
III

III
IV
IV
IV
IV
IV
IV
IV
IV
V
IV
IV
V
IV
V
IV

III
IV
IV

II
III

III
III
III
III
IV
IV
IV
IV
IV
IV
IV
IV
IV
IV
IV



-------
TABLE 2.   (Continued)

Compound
!' 'V/K.s
" V 1 ethyl ketone
Methyl isobutyl ketone
Diisobutyl ketone
Alcohols
Methanol
Ethanol
n-Propyl Alcohol
Isopropyl Alcohol
n-Butanol
4-Methyl-2-pentanol
Ethers
Diethyl ether
Di-n-propyl ether
Tetrahydrofuran
Esters
n-Propyl acetate
sec-Butyl acetate
Natural Hydrocarbons
^-Pinene
L--Pinene
p-Menthane
3-Carene
'•-Phellandrene
Carvomenthene
d-Limonene
Dihydromyrcene
Myrcene
cis-Ocimene
Halocarbons
Fluoromethane
Chloromethane
Bromomethane
Dif 1 uoromethane
Chlorofl uoromethane
Dichloromethane
Trifluoromethane
Chi orodi f 1 uoromethane
Dichlorofl uoromethane
Trichloromethane
Tetraf 1 uoromethane
Chlorotrifl uoromethane


ka
OH+Cpd
( molds'1 )x!0~9 Reference

2.0
9.0
15

0.57
1.8
2.3
4.3
4.1
4.3

5.6
10.4
8.8

2.7
3.4

35
41
4
52
70
76
90
101
137
192

0.0096
0.025
0.023
0.0047
0.022
0.084
0.00012
0.025
0.016
0.065
<0. 00024
<0. 00042
(continued)
693

37
37
37

63
63
63
38
63
73

38
38
40

40
40

37
37
37
37
37
37
37
37
37
37

71
28,67,71
67,71
71
71
28,67,71
71
26,71
28,71
67,71
71
71


Reactivity
Relative to Proposed
Methane = 1 Class

420
1875
3125

120
375
480
895
855
895

1170
2170
1830

560
710

7290
8540
830
10,830
14,580
15,830
18,750
21 ,040
28,540
40,000

'i
s
i
1
5
IB
0.025
5
3
14
-O.Ob
^0.09



III
IV
IV

III
III
III
III
III
III

IV
IV
IV

III
III

IV
IV
III
V
v
V
V
V
V
\l

1
!
1
]
1
l!
1
1
I
11
I
I



-------
                                TABLE  2.   (Continued)
        Compound
j. mol
                                kcWpd
                                i -1  -1 \
)x!0
                                         -9
Reference
Nitrogen-Containing Compounds

Ammonia                          0.0988
Methyl amine                     13.2

Sulfur-Containing Compounds

Hydrogen sulfide                  3.16
Methyl mercaptan                 20.4

Miscellaneous
                        30
                        77
                        30
                        77
a"Room temperature"  (i.e., within range 292-305  K).
Reactivity
Relative to   Proposed
Methane = 1   Class
Dichlorodifluoromethane
Fluorotrichloromethane
Tetrachloromethane
Chloroethane
1 ,2-Dichloroethane
1 .2-Dibromoethane
1 ,1-Difluoroethane
1 ,1 -Dichloroethane
1-Chloro-l ,1-difluoroethane
1 ,1 ,1-Trichloroethane
1 ,1 ,l-Trifluoro-2-chloroethane
1 ,1 ,1 ,2-Tetrafluoro-2-
chloroethane
1,1 ,l-Trifluoro-2,2-
dichloroethane
1,1 ,2,2-Tetrafluoro-l,2-
dichloroethane
1 ,1 ,2-Trifluoro-l,2,2-
trichloroethane
<0.0006
<0.0006
<0.0024
0.235
0.132
0.15
0.019
0.16
0.0017
0.0090
0.0063
0.0075

0.0171

<0.0003

<0.0002

26,71
26,71
71
72
72
72
72
72
72
72
72
72

72

72

72

                                                                     -.0.13
                                                                     -0.13
                                                                     -.0.5
                                                                     49
                                                                     28
                                                                     31
                                                                      4
                                                                     33
                                                                      0.35
                                                                      2
                                                                      1
                                                                      2

                                                                      4

                                                                     <0.06

                                                                     -,0.04
                                21
                               2750
                               660
                               4250
                                                    I
                                                    I
                                                    I
                                                   II
                                                   II
                                                   II
                                                    I
                                                   II
                                                    I
                                                    I
                                                    I
                                                    I
                                II
                                IV
                               III
                                IV
Carbon monoxide
Methoxybenzene
o-Cresol
0.084
11.8
20.5
46
77
77
18
2460
4280
II
IV
IV

                COMPARISON WITH  OTHER REACTIVITY CLASSIFICATIONS

      In  our proposed reactivity classification, Class  I  contains methane  and
most  halocarbons.   The remaining half-dozen halocarbons  appear in Class  II  along
with  carbon monoxide, ammonia,  acetylene and ethane.   Almost all alkanes  appear
in  Class III while alkenes other than ethene appear  in Classes IV or V.   Aro-
matics and oxygenates appear  in Classes III or IV and  the natural hydrocarbons
(not  specifically identified  in previous reactivity  classifications) are  of high
reactivity and appear in Classes IV and V.  Clearly, our proposed classification
emphasizes that most compounds  react in polluted atmospheres and Classes  I  and
II  contain the relatively few compounds which have half-lives greater  than  10
days  and 24 hours, respectively.   This result parallels  the conclusion by
                                        694

-------
















f — •»
CO

—
t— J

- a:
I —
1 — ' z
> o

1 — 1
h- Q
O LU
J~
•A i








ra
O

'o

> X
t— . CO

I/) ^t"

05 t

ocn
0

X
CO
•—• •



O3

CTi
O
X
>— i CO

— *d-
co i
LO
03 CO
r— O

X
CO

«^f-
V 	 f

03

CO
O

X
CO


t/1 1
CO
rar-.
t— O

X
CO

«3-













l"a


o
LO r—
LO X
03 CO
C_) *3~
V 1














QJ
C.
QJ QJ
C -(->
QJ QJ 3
C X -Q
QJ QJ 1
4-J jn QJ CM QJ
3 O C QJ 1C
_Q r— QJ C i — CJ
1 O S- QJ >> U QJ
CM >>~O Jd QJ -d S- C
1 U d 4-J d 4J >> QJ
i — QJi — 03CQJOJE EE
>> C >>r— QJ C £ O QJ -r-
-C QJ .C ,— E O -i- S-C U
4-J S-4-> QJ O E Q ~O QJ O

21 o s: C- i 	 iro-c s- c/i
1 1 1 1 fO t "••- >)-r-




QJ QJ QJ
d C QJ C
O QJ C QJ
4-J N QJ N
QJ C -M d
Ja^ S- QJ QJ 3 QJ
QJ -Q C -Q -0 C
C r— -C r— O 1 i — T3
QJ QJ 03 >j 4J OJ QJ >,4-> -— >-,4-J
S- CC S-4-> QJCC -CQJ i -CQ-
QJ OJ QJQJ ^13 OJQJ 4->-^i — 4-JO3

QJ 4-> r— •— O O >>r— C C E r— -C E S_
N , 4-J(L) -r-QJ
CQJ QJ4->4->(UT3-r-JDE aJCQJS-Er—aJ
QJCr— C>— •— C>> CS-r->>03Ch-i:QJ£aJCh- OC

r— 4->_C.— -Cjd.— 03>>E ' -d O >,r— CO 0 QJO C X^f >,QJ4-J
>>U4-J >)4-J4-> >,S_jr >jC4-J-d^: >~) "I/I O- 1 QJ QJ --C S- 13
+J 1 -r- I 1 t I QJ QJ 1 '»- 1 QJ QJ I *T- S- ' 1 I - QJ 1 I

QJ
C
4-> QJ
13 C QJ
-O (13 d
i — QJ 4-J OS
>> C d 4-> QJ
-d O QJ 13 C QJ
+-> -M Q__Q QJ rO QJ 4-J QJ
OJQJ QJ r— i — 4-J4-J"OQJo3 QJC
OJfO ^r-d4-JjO4-030J (UN
.— S- r— 4->4JQj, — , — 4->O QJMC
>^4-J OJ >> QJQJU >ir3Cffl CCQJ
QJ 4-JQJd -d QJ •— E E ra JC^QJ n3QJ-Q
d OJI— 03 QJ 4->CO'i-T- 4-J Q-F— 4-)JDr—
03i — UI4-J d QJ QJ03CS-S-I— QJCi — >, d r— >,
-MOfOfOISQJfaaJdr— 4-J 03 f— I— >> £QJ>14-JQJaJ>)O-
d d r — " -Q d 4-J d OJ O ' — d CX 1 1 O. QJ •!— CJl-d 3 d Q. O O

O __c~ _c " i — M jQ O ^ o3 ~d *"* S— ""S— QJii-QJI^ i — S— Q-
O4-J4-JCM LJ C O OCQ-d4-J OQ~CMCMD-i — COTD^ Or— UO- O
SSS~^SJ2ii£SS^«J^i5«J^ Jd X O QJ O QJ
QJ 4-J O S- O S- d
E QJ C o E o as
OEo QJi — o r— -c
s-OE d-dS- _d4->
OS- raOJUJD LJQJ
r— O d -r- r- -r- -,- QJ "r- Q
-d^O C>,QCjdQ S-
U -d _Q O 4-J t f 03 1 O
•r- O S- E QJ CM CM _C r— r—
S-T— O3EO " "4-J " c~
1— Q CJ  QJ QJ
4J QJ d d
OJ O 03 as
OS- -d .C
S- O OJ 4-> 4-J
O 13 d CU QJ
^ i — 03 O O
i — 4- QJ _d S~ S-
4-03 C 4-> O O
•r- S- ra QJ 13 13
S- 4-J _d O i — " —
4-> QJ QJ 4-J S- 4- 4-
l 4-JQJdQJQJ Oo3 -i—
CM IdrOdO ^S-QJQJt- QJ

CM "jd4-JjdO 4-OJ03 031 QJ Oi

QJtrO«— EOE4^03 4-JCMOJ OJ •— rO ^ QJ

raS-4-Jr— S-OS_-O4-Jd i — -— S- Or— 4->QJ O
-dOQJIO=3OlQJ03 ""OQJS-ICDE QJS-
4-J r— E O 3 r- i — • — E-d r— i 	 i — dOOOOQJdO
QJ-dOS-i — 4-.C «O4-J " "_d03~S-S-S-d03^
E U S- O 4- -r- o i — S-QJ i — • — O-di — OOOo3-di —
O '•- O •— -r- -0 -r- 1 O E 1 l-r-4-J4-r— ^3=3JZ4-J4-
S-S-lD-dS-OS-Oi — O OOS-QJO-dr— i — 4-J OJ T-
O r— r— U4-J S_4->S-jdS-OJS- S-h- ES-O4-4-QJ EX3
^j IM--r-OOOOOOdOO 1 OO-i-'f-OEOO
i — CM OSQ S-i — S-i — O3 ^3oSi — i — i — S-i — QO S- O S- S-
4- "S- I O-dO-dS-i — jd-d-d "O-C t 1 OEEOO
• r— , — 4_>CMi — U I3O4-'l4-4-JOOi — 3 OCMi — • < — Oi — i —
S^ "QJ " jC T— ( — l QJ -i— QJ 1 l "i — -r- " "^1S 	 d-d
h- ' — I— i — CJQQ-i — I— Q S: CM CM i — U-QCMr— OCDOCJ
























QJ
OJ
NI

QJ QJ
-Q d
i — QJ
>> 4-> OJ
Jd 3 QJ d QJ !

QJ d 1 OJ 4-J QJ QJ
£ QJ r— 4-> 3 d 'r— '
QJ-r-QJ4~>l d D3QJ"D '
Ct-Crs*1 — QJOJQJ 1 XroQJ
OJ I— QJ CQ >jCQ_ dCM QJ4-J d

Q-LO ^CM4->CCMdt/)OCQS- ;
QJ "-Q 1 QJ-r- l-i— di — 1 D-
"T~ CO O tO 5T Q_ t/) Q_ OS O CO O
1 " CO T- 1 1 -r- 1 S- >, " CO






1 	 1
O
rO
4-J QJ
QJ 03
O_4->
QJ 1 d
QJ C i — CM QJ
QJ d r— ra O 1 Q-
OJCoSOXCi — i —
d03-dCQJra>,>>


QJQjQJDi — S^OJOJOJ
T* o_ y CQ o D- ^* y -d
i i iri «^•

4-
O

QJ
cn
C


O3
695

-------
 Altshuller  and  Bufalini  in  their  1971 review article that  "almost every  hydro-
 carbon  except methane  can produce  some oxidarit when photooxidized in the pre-
 sence of  high enough ratios  of  hydrocarbons to oxides of nitrogen"  (3).

     Although our  proposed  reactivity scale and classification  is based  solely
 on  consumption  rates of  organics,  Altshuller and Bufalini  have  shown that  this
 measure of  reactivity  is very similar to one based on nitric oxide  oxidation
 rates (3).   They showed  that the  ranking of reactivities of hydrocarbons from
 the nitric  oxide photooxidation studies of Altshuller and  Cohen  (49) and Glasson
 and Tuesday (42) was essentially  the same as that obtained from  the studies of
 hydrocarbon consumption  carried out by Schuck and Doyle (50), Stephens and Scott
 (51), and Tuesday  (52).  The ranking of reactivities for the aromatic hydro-
 carbons in  our  classification is  very similar to that obtained  by Altshuller et al
 (53) and  by Kopczynski  (54,  55).   We have previously shown that  our classi-
 fication  (9, 10) is in  general  agreement with that proposed by  Dimitriades (7)
 in  1974 and noted  where  significant differences existed.   Several of these
 differences arose  from  the  greater degree of differentiation in  reactivity
 permitted by the OH reactivity  criterion.

     A  three-class reactivity classification of organic compounds very recently
 adopted by  the  ARB (56)  is  shown  in Table 4, and is similar to  one  now under
 consideration by the EPA (57).  As described by the ARB (56, 57), Class  I
 includes  low reactivity  organic compounds, yielding little if any ozone  under
 urban conditions.  Class II  consists of moderately reactive organic compounds
 which give  an intermediate yield  of 03 within the first day of  solar irradiation,
 while Class III is limited  to highly reactive organic compounds  which give high
 yields  of 03 within a  few hours of irradiation.

     In general, the ARB three-tiered classification is consistent with  the
 reactivity  classification presented here, based on OH rate constants, with only
 minor exceptions.  For  example, the ARB scale shows the primary  and secondary
•C2+ alcohols to be highly reactive in Class [II, while our scale shows them to
 be  of moderate  reactivity.

              ADVANTAGES AND LIMITATIONS OF THE HYDROXYL RADICAL
                           REACTIVITY CLASSIFICATION

     In contrast to previous reactivity scales, where it is commonly recognized
 that uncertainties in the chamber  data upon which they are based limit the
 number  of classes  into which the reactivity scales can be  meaningfully divided
 (7,  57) our OH  reactivity scale could be divided into a large number of classes
 based on the fact that the general accuracy of OH rate constant  measurements is
 better  than ±20%.  Thus, one significant strength of this  approach  is the  very
 high degree of  resolution obtained in ranking compounds in order of their  rate
 of  reaction with OH.  A second advantage of the present classification is  that
 it  can  be readily extended to include additional organic (and inorganic) com-
 pounds  once their rate of reaction with OH is known.  Thus, it is not necessary
 to  conduct  a major new experimental program each time a significant number of
 new compounds become of interest  (as has been felt necessary in  the usual  smog
 chamber approach,  in order to obtain consistent data for all compounds of  in-
 terest  at a given time).  Finally, our proposed classification gives greater
 weight  than most previous classifications to the alkanes and aromatic hydro-


                                     696

-------
   TABLE 4.  CALIFORNIA AIR RESOURCES BOARD (ARB) REACTIVITY CLASSIFICATION
                          OF ORGANIC COMPOUNDST (56)
        Class I
   (Low Reactivity)
         Class II
   (Moderate Reactivity)
           Class III
       (High Reactivity)
C1-C2 Paraffins
Acetylene
Benzene
Benzaldehyde
Acetone
Methanol
Tert-alkyl alcohols
Phenyl acetate
Methyl benzoate
Ethyl amines
Dimethyl formamide
Perhalogenated
Hydrocarbons
Partially halogenated
paraffins
Phthalic anhydride**
Phthalic acids**
Acetonitrile*
Acetic acid
Aromatic amines
Hydroxyl amines
Naphthalene*
Chlorobenzenes*
Nitrobenzenes*
Phenol*
Mono-tert-alkyl-benzenes
Cyclic ketones
Alky! acetates
2-Nitropropane
C,+ Paraffins
Cycloparaffins
n-Alkyl ketones
N-Methyl pyrrolidone
N,N-Dimethyl acetamide
Alkyl phenols*
Methyl phthalates**
All other aromatic hydro-
carbons
All olefinic hydrocarbons
(including partially halo-
genated)
Aliphatic aldehydes
Branched alkyl ketones
Cellosolve acetate
Unsaturated ketones
Primary & secondary
C2+ alcohols
Diacetone alcohol
Ethers
Cellosolves
Glycols*
C2+ alkyl phthalates**
Other esters**
Alcohol amines**
Co+ Organic acids + di acid**
C3+ di acid anhydrides**
Formin**
(Hexa methylene-tetramine)
Terpenic hydrocarbons
Olefin oxides**
t
 This reactivity classification is identical to that suggested by the EPA in
 August, 1975 (see text of References 56 and 57) except that the ARB has
 moved propane from Class I to Class II.  *Reactivity data are either non-
 existent or inconclusive, but conclusive data from similar compounds are
 available; therefore, rating is uncertain but reasonable.  **Reactivity
 data are uncertain.
                                     697

-------
carbons, which as discussed above, require a longer irradiation period but can,
during downwind transport or stagnant air episodes, contribute significantly to
03 formation.

     In closing, we note several  caveats which should be borne in mind in
applying the proposed OH reactivity scale.  First of all, it is not strictly
applicable to compounds which undergo significant photodissociation in the
atmosphere (e.g., aliphatic aldehydes) or which react in the atmosphere at
significant rates with species other than OH (e.g., the alkenes with 03).  In
such cases, the compound will be more reactive than predicted from a scale based
on consumption due solely to OH attack.   It is interesting to note, however,
that with respect to the reaction of 03  with alkenes this limitation is not as
serious as it would appear to be based only on the relative rates of consumption
of alkenes by 03 and OH, respectively.  This follows from the fact that recently
Niki and coworkers (58) have shown that  03-alkene reactions do not lead pre-
dominantly to radical products, and hence this additional alkene consumption
pathway will  contribute significantly less "additional  reactivity" (over that
obtained for OH reaction alone) than if  it led to radical chain propagation.

     A second, and perhaps more significant, limitation in the application of
our reactivity classification concerns the inherent problem arising from un-
certainties in the identity and fates of subsequent products (7).  For example,
two compounds which have the same rate constant for reaction with OH may not
necessarily have the same oxidant-producing potential in undergoing complete
photooxidation.  As another example, we  have recently reported that the higher
alkanes (>CiJ undergo alkoxy radical isomerization (59), and also that reaction
of alkylperoxy radicals (^C^) with NO leads to alky! nitrate formation rather
than to RO + N02 (60).  Thus, estimates  of 03 formation (or other secondary smog
manifestations) for the higher alkanes which are based  soley on their rates of
reaction with OH are likely to be somewhat high.  These limitations notwith-
standing, the reactivity classification  presented here  should be of utility in
assessing the potential contributions of a wide variety of orgariics to the
production of photochemical oxidants.

                                ACKNOWLEDGMENTS

     This work was supported by the California Air Resources Board (ARB Contract
No. 5-385) and the National Science Foundation-Research Applied to National
Needs (NSF-RANN Grant No. ENV73-02904-A03).


                                  DISCLAIMER

     The contents do not necessarily reflect the views and policies of the ARB
or the NSF-RANN, nor does mention of trade names or commercial products con-
stitute endorsement or recommendation for use.

                                  REFERENCES

1.   A. J. Haagen-Smith and M. M. Fox, "Ozone Formation in Photochemical Oxi-
     dation of Organic Substances,"  Ind. Eng. Chem. 48, 1484 (1956).
                                     698

-------
2.   M.  F. Brunelle, J. E. Dickinsen, and W. J,  Hamming, "Effectiveness of
     Organic Solvents in Photochemical Smog Formation," Solvent Project, Final
     Report, Los Angles County Air Pollution Control District, July, 1966.

3.   A.  P. Altshuller and J.  J. Bufalini, "Photochemical Aspects of Air Pol-
     lution:  A Review," Environ.  Sci. Techno!.  5., 39 (1971).

4.   "Organic Compound Reactivity System for Assessing Emission Control Stra-
     ategies," California Air Resources Board Staff Report No. 75-17-4, Sept-
     ember 29, 1975.

5.   B.  F. Goeller, J. H. Bigelow, J. C. DeHaven, W. T. Mikolowsky, R.  L.
     Petruschell,  and B. M.  Woodfill, "Strategy Alternatives for Oxidant Control
     in the Los Angeles Region," The Rand Corporation, Santa Monica, CA, R-1368-
     EPA, December, 1973.

6.   J.  C. Trijonis and K. W. Arledge, "Utility of Reactivity Criteria in
     Organic Emission Control Strategies for Los Angeles," TRW Environmental
     Services, Redondo Beach, CA, December, 1975.

7.   "Proceedings  of the Solvent Reactivity Conference," U.  S. Environmental
     Protection Agency, Research Triangle Park,  NC, EPA-650/3-74-010, November, 1974

8.   S.  W. Benson, J. G. Calvert,  J. -G. Edinger, C. Foote, A. J. Haagen-Smith,
     H.  S. Johnston, and J.  N. Pitts, Jr., "Final Report of the Committee  on
     Photochemical Reactivity to the Air Resources Board of the State of Cali-
     fornia," April, 1976.

9.   J.  N. Pitts,  Jr., A. C.  Lloyd, A. M. Winer, K. R. Darnall, and G.  J.
     Doyle, "Development and Application of a hydrocarbon Reactivity Scale Based
     on Reaction with the Hydroxyl Radical," Paper No. 76-31.1, presented  at the
     69th Annual Meeting of the Air Pollution Control Association, Portland, OR,
     June 27-July  1, 1976.

10.   K.  R. Darnall, A. C. Lloyd, A. M. Winer, and J. N. Pitts, Jr., "Reactivity
     Scale for Atmospheric Hydrocarbons Based on Reaction with the Hydroxyl
     Radical," Environ. Sci.  Techno!. 10., 692 (1976).

11.   J.  N. Pitts,  Jr., Presentation to California Air Resources Board Committee
     on Photochemical Reactivity,  December, 1975.

12.   N.  R. Greiner, "Hydroxyl-Radical Kinetics by Kinetic Spectroscopy.  II.
     Reactions with C2H6, C3H8, and iso-C^o at 300 K," J.  Chem. Phys. 46_, 3389
     (1967).

13.   J.  Heicklen,  K. Westberg, and N. Cohen, "The Conversion of NO to N02  in
     Polluted Atmospheres,"  Publication No. 115-69, Center for Air Environmental
     Studies, University Park, PA, 1969.

14.   K.  Westberg and N. Cohen, "The Chemical Kinetics of Photochemical  Smog as
     Analyzed by Computer,"  The Aerospace Corporation, El Segundo, CA, ATR-
     70(8107)-!, December 30, 1969.

                                     699

-------
15.  B. Weinstock, E. E. Daby, and H. Niki, "Discussion on Paper Presented by E.
     R. Stephens," in Chemical Reactions in Urban Atmospheres, C. S. Tuesday,
     Ed., Elsevier, NY, 1971, pp. 54-55.

16.  H. Niki, E. E. Daby, and B.  Weinstock, "Mechanisms of Smog Reactions," Adv.
     Chem. Series 113, 16 (1972).
                                           *
17.  K. L. Demerjian, J. A.  Kerry, and J. G.  Calvert, "The Mechanism of Photo-
     chemical Smog Formation," Adv.  Environ.  Sci. Technol.  4., 1 (1974).

18.  J. N. Pitts, Jr. and B. J. Finlayson, "Mechanisms of Photochemical Air
     Pollution," Angew. Chem. Int. Ed. 1_4, 1  (1975).

19.  B. J. Finlayson and J.  N. Pitts, Jr., "Photochemistry of the Polluted
     Troposphere," Science 192, 111  (1976).

20.  B. J. Finlayson-Pitts and J. N.  Pitts, Jr., "The Chemical Basis of Air
     Quality: Kinetics and Mechanisms of Photochemical Air Pollution and Appli-
     cation to Control Strategies,"  Adv. Environ. Sci. Technol., in press (1976).

21.  C. C. Wang, L. I. Davis, Jr., C. H. Wu,  S. Japar, H. Niki, and B. Wein-
     stock, "Hydroxyl Radical Concentrations  Measured in Ambient Air," Science
     189, 797 (1975).

22.  D. Perner, D. H. Ehhalt, H.  W.  Patz, U.  Platt, E. P. Roth, and A. Volz, "OH
     Radicals in the Lower Troposphere," 12th International Symposium on Free
     Radicals, Laguna Beach, CA,  January 4-9, 1976.

23.  D. D. Davis, W. Heaps,  and T. McGee, "Direct Measurements of Natural
     Tropospheric Levels of OH Via an Aircraft-Borne Tunable Dye Laser," Geophys.
     Res. Lett. 3_, 331 (1976).

24.  R. Atkinson, D. A. Hansen, and  J. N. Pitts, Jr., "Rate Constants for the
     Reaction of the OH Radical with H2 and NO (M=Ar and N2)," J. Chem. Phys.
     62^, 3284 (1975).

25.  D. A. Hansen, R. Atkinson, and J. N. Pitts, Jr., "Rate Constants for the
     Reaction of OH Radicals with a Series of Aromatic. Hydrocarbons," J. Phys.
     Chem. 79_,  1763  (1975).

26.  R. Atkinson, D. A. Hansen, and J. N. Pitts, Jr., "Rate Constants for the
     Reaction of OH Radicals with CHF2C1, CF2C12,  CFC13, and H2 Over the Tem-
     perature Range 297-434°K," J. Chem. Phys. 63^,  1703  (1975).

27.  R. Atkinson and J. N. Pitts, Jr., "Rate Constants for the Reaction of OH
     Radicals with Propylene and the Butenes Over  the Temperature Range 297-
     425°K," J. Chem. Phys.  63^ 3591  (1975).

28.  R. A. Perry, R. Atkinson, and J. N. Pitts, Jr.,  "Rate Constants for the
     Reaction of OH Radicals with CHFC12 and CH3C1  Over  the Temperature Range
     298-423°K, and with CH2C12 at 298°K," J. Chem. Phys. 64, 1618  (1976).


                                     700

-------
29.  R. Atkinson, R. A. Perry, and J. N. Pitts, Jr., "Rate Constants for the
     Reaction of OH Radicals with 2-Methyl-2-Butene Over the Temperature Range
     297-425°K," Chem. Phys. Lett. 38, 607 (1976).

30.  R. A. Perry, R. Atkinson, and J. N. Pitts, Jr., "Rate Constants for the
     Reactions OH + H2S -> H20 + SH and OH + NH3 -* H20 + NH2 Over the Temperature
     Range 297-427°K," J. Chem. Phys. 64, 3237 (1976).

31.  R. A. Perry, R. Atkinson and J. N. Pitts, Jr., "Rate Constants for the
     Reaction of OH Radicals with n-Butane Over the Temperature Range 297-
     420°K," J. Chem. Phys. 64, 5314 (1976).

32.  R. Atkinson, R. A. Perry, and J. N. Pitts, Jr., "Rate Constants for the
     Reactions of the OH Radical with N02 (M = Ar and N2) and S02 (M = Ar) ," J.
     Chem. Phys. 65., 306 (1976).

33.  R. Atkinson, R. A. Perry, and J. N. Pitts, Jr., "Rate Constants for the
     Reaction of OH Radicals with Ethylene Over the Temperature Range 299-
     425°K," J. Chem. Phys., submitted for publication (1976).

34.  R. Atkinson, K. R. Carnal!, A. C. Lloyd, A. M. Winer, and J. N. Pitts, Jr.,
     "Kinetics and Mechanisms of the Reactions of the Hydroxyl Radical with
     Organic Compounds in the Gas Phase," Accts. Chem. Res., submitted for
     publication (1976).

35.  G. J. Doyle, A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr.,
     "Gas Phase Kinetic Study of Relative Rates of Reaction of Selected Aromatic
     Compounds with Hydroxyl Radicals in an Environmental Chamber," Environ.
     Sci. Technol. £, 237 (1975).

36.  A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr., "Relative
     Rate Constants for Reaction of the Hydroxyl Radical with a Series of
     Alkanes, Alkenes, and Aromatic Hydrocarbons," J. Phys. Chem. 80_, 789  (1976).

37.  A. M. Winer, A. C. Lloyd, K. R. Darnall, and J. N. Pitts, Jr., "Relative
     Rate Constants for the Reaction of the Hydroxyl Radical with Selected
     Ketones, Chloroethenes, and Monoterpene Hydrocarbons," J. Phys. Chem. 80,
     1635 (1976).

38.  A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr., "Relative
     Rate Constants for the Reactions of OH Radicals with Isopropyl Alcohol,
     Diethyl, and Di-n-propyl Ether at 305 ± 2 K," Chem. Phys. Lett., in press
     (1976).

39.  K. R. Darnall, A. M. Winer, A. C. Lloyd, and J. N. Pitts, Jr., "Relative
     Rate Constants for the Reaction of OH Radicals with Selected C6 and C7
     Alkanes and Alkenes at 305 ± 2 K," Chem. Phys. Lett., submitted for pub-
     lication (1976).

40.  A. M. Winer, A. C. Lloyd, K. R. Darnall, and J. N. Pitts, Jr., "Relative
     Rate Constants for the Reaction of OH Radicals with Tetrahydrofuran and n-
     Propyl- and sec-Butyl  Acetate," in preparation.

                                     701

-------
41.   C.  H.  Wu, S.  M.  Japar, and H.  Niki, "Relative Reactivities of HO-Hydro-
     carbon Reactions from Smog Reactor Studies,"  J.  Environ.  Sci.  Health A11,
     191 (1976).

42.   W.  A.  Glasson and C.  S.  Tuesday,  paper presented at the 150th National
     Meeting, ACS, Atlantic City, NO,  September, 1965.

43.   C.  T.  Pate,  R.  Atkinson, and J.  N. Pitts., Jr., "The Gas Phase Reaction of
     03  with a Series of Aromatic Hydrocarbons," J. Environ. Sci.  Health Al1, 1
     (1976).

44.   R.  Atkinson  and J.  N. Pitts, Jr., "Absolute Rate Constants for the Reaction
     of  0(3P) Atoms  with Selected Alkanes, Alkenes, and Aromatics  as Determined
     by  a Modulation Technique," J. Phys.  Chem. 78^ 1780 (1974).

45.   D.  H.  Stedman and H.  Niki, "Ozonolysis Rates  of Some Atmospheric Gases,"
     Environ. Lett.  1, 303 (1973).

46.   "Chemical Kinetic and Photochemical Data for Modeling Atmospheric Che-
     mistry," R.  F.  Hampson,  Jr. and D. Garvin, Eds., NBS Technical Note 866,
     June, 1975.

47.   D.  H.  Stedman,  C. H.  Wu, and H.  Niki, "Kinetics of Gas Phase  Reactions  of
     Ozone with Some Olefins," J. Phys. Chem. 77_,  2511  (1973).

48.   S.  M.  Japar, C. H.  Wu, and H.  Niki, "Rate Constants for the Reaction of
     Ozone with Olefins  in the Gas Phase," J. Phys. Chem. 78^, 2318 (1974).

49.   A.  P.  Altshuller and I.  R. Cohen, "Structural Effects on the  Rate of
     Nitrogen Dioxide Formation in the Photoxoidation of Organic Compound-Nitric
     Oxide Mixtures in Air,"  Int. J.  Air Water Pollut.  7_, 787 (1963).

50.   E.  A. Schuck and G. J. Doyle, "Photooxidation of Hydrocarbons in Mixtures
     Containing Oxides of Nitrogen and Sulfur Dioxide," Report No. 29, Air
     Pollution Foundation, San Marino, CA, 1959.

51.   E.  R. Stephens and W. E. Scott, "Relative Reactivity of Various Hydro-
     carbons  in Polluted Atmospheres," Proc. Amer. Petrol.  Inst. 42  (III), 665
     (1962).

52.   C.  S. Tuesday, "Atmospheric Photooxidation of Olefins.  Effect  of Nitrogen
     Oxides," Arch.  Environ.  Health _7, *I88 (1963).

53.  A.  P. Altshuller,  I.  R. Cohen, S.  F.  Sleva,  and S.  L.  Kopczynski,  "Air
     Pollution Photooxidation of Aromatic  Hydrocarbons," Science J35,  442
     (1962).

54.  S.  L. Kopczynski,  "Photooxidation  of  Alkylbenzene-Nitrogen Dioxide  Mixtures
     in Air,"  Int. J. Air Water  Pollut. 8., 107  (1964).

55.   S.  L. Kopczynski, quote in  Reference  4.
                                      702

-------
56.  "Board Approves Reactivity Classification," California Air Resources Board
     Bulletin, March, 1976, p. 4.

57.  "Adoption of a System for the Classification of Organic Compounds According
     to Photochemical Reactivity," California Air Resources Board Staff Report
     No. 76-3-4, February 19, 1976.

58.  H. Niki, P. Maker, C. Savage, and L. Breitenbach, "Fourier Transform
     Spectroscopic Studies of Organic Species Participating in Photochemical
     Smog Formation," Int. Conf.  Envir.  Sensing and Assessment, Las Vegas,
     Nevada, September 14-19, 1975.

59.  W. P.  L. Carter, K.  R. Darnell, A.  C. Lloyd, A. M. Winer, and J. N. Pitts,
     Jr., "Evidence for Alkoxy Radical Isomerization in Photooxidations of C4-Cfi
     Alkanes Under Simulated Atmospheric Conditions," Chem. Phys. Lett. 42, 22
     (1976).

60.  K. R.  Darnall, W. P.  L. Carter, A.  M. Winer, A. C. Lloyd, and J. N. Pitts,
     Jr., "Importance of R02 + NO in Alkyl Nitrate Formation from C4-C5 Alkane
     Photooxidations Under Simulated Atmospheric Conditions," J.  Phys. Chem. 80,
     1948 (1976).

61.  J. N.  Bradley, W. Hack, K.  Hoyermann, and H. Gg. Wagner, "Kinetics of the
     Reaction of Hydroxyl  Radicals with  Ethylene and with C3 Hydrocarbons," J.
     Chem.  Soc.  Faraday Trans. I  69, 1889 (1973).

62.  I. M.  Campbell, B. J. Handy,  and R. M.  Kirby, "Gas Phase Chain Reaction of
     H202 + N02  + CO," J.  Chem.  Soc. Faraday Trans. I 71_, 867 (1975).

63.  I. M.  Campbell, D. R. McLaughlin, and B. J. Handy, "Rate Constants for
     Reactions of Hydroxyl Radicals with Alcohol Vapours at 292°K," Chem. Phys.
     Lett.  38, 362 (1976).

64.  D. D.  Davis, S. Fischer, and R. Schiff, "Flash Photolysis-Resonance
     Fluorescence Kinetics Study:   Temperature Dependence of the Reactions
     OH + CO -> C02 + H and OH + CH,4 -> H20 + CH3," J. 'Chem. Phys.  6]_, 2213 (1974)

65.  D.  D.  Davis, W.  Bellinger, and  S.  Fischer,  "A  Kinetics  Study  of  the  Re-
     action  of  the  OH  Free  Radical with  Aromatic  Compounds.   I.  Absolute  Rate
     Constants  for  Reaction with  Benzene  and Toluene  at 300°K,"  J.  Phys.  Chem.
     79_, 293  (1975).

66.  D.  D.  Davis, S.  Fischer, R.  Schiff,  R.  T.  Watson, and  W.  Bellinger,  "A
     Kinetics Study  of the  Reaction  of  OH Radicals  with Two  C2 Hydrocarbons:
     C2H4 and C2H2,"  J. Chem. Phys.  63_,  1707  (1975).

67.  D.  D.  Davis, G.  Machado, B.  Conaway, Y. Oh,  and  R. Watson,  "A Temperature
     Dependent  Kinetics Study of  the Reaction  of  OH with  CH3C1,  CH2C12,  CHCL3,
     and CH3Br," J.  Chem.  Phys. 65.,  1268  (1976).
                                     703

-------
68.   S.  Gordon and W.  A. Mulac, "Reaction of the OH(X2n) Radical  Produced by the
     Pulse Radiolysis  of Water Vapor," Intern.  J.  Chem.  Kinetics, Symposium No.
     1,  289 (1975).

69.   R.  A. Gorse and D. H. Volman, "Photochemistry of the Gaseous Hydrogen
     Peroxide-Carbon Monoxide System.   II:  Rate Constants for Hydroxyl Radical
     Reactions with Hydrocarbons and for Hydrogen Atom Reactions with Hydrogen
     Peroxide," J. Photochem. 3., 115 (1974).

70.   N.  R. Greiner, "Hydroxyl Radical  Kinetics by Kinetic Spectroscopy.  VI.
     Reactions with Alkanes in the Range 300-500 K," J.  Chem. Phys. 53_, 1070
     (1970).

71.   C.  J. Howard and K. M. Evenson, "Rate Constants for the Reactions of OH
     with CHit and Fluorine, Chlorine,  and Bromine Substituted Methanes at
     296°K,"  J. Chem.  Phys. M, 197 (1976).
72.  C. J. Howard and K. M. Evenson, "Rate Constants for the^Reaqtions of OH
     with Ethane and
     64, 4303 (1976).
with Ethane and Some Halogen Substituted Ethanes at 296°K," J. Chem. Phys.
73.  J. L. Laity, I. G. Burstain, and B. R. Appel,  "Photochemical Smog and the
     Atmospheric Reactions of Solvents," in Solvents Theory and Practice, R. W.
     less, Ed., Advan.  Chem.  Series 124, 95 T1973).

74.  J. J. Margitan, F. Kaufman, and J.  G.  Anderson, "The Reaction of OH with
     CH4," Geophys.  Res.  Lett.  1, 80 (1974).

75.  E. D. Morris, Jr.  and H. Niki, "Reactivity of Hydroxyl Radicals with
     Olefins," J. Phys. Chem. 75, 3640 (1971).
76.  R. P. Overend, G. Paraskevopoulos, and R. J. Cvetanovic, "Rates of OH
     Radical Reactions.  I. Reactions with H2, CH4, C2H6, and C3H8 at 295°K,"
     Can. J. Chem. 53., 3374 (1975).

77.  R. A. Perry, R. Atkinson, and J. N. Pitts, Jr., unpublished results (1976).

78.  F. Stuhl, "Rate Constant for the Reaction of OH with n-C^Hjo," Z. Natur-
     forsch. 28A, 1383 (1973).

79.  D. H. Volman, "General Discussion III," Intern. J. Chem. Kinetics, Sym-
     posium No. 1, 358 (1975).
                                     704

-------
                                                                           14-3
                  APPLICATION OF REACTIVITY  CRITERIA IN
               OXIDANT-RELATED EMISSION  CONTROL  IN THE USA

                    B. Dimitriades and S.  B.  Joshi*

 ABSTRACT
      The. uAe. o& the. x.e.activ-ity cAiteAton -in oxtdant-Ae£ate.d e.miAAi.on c.ontsiol i
 fie.£x.ami.ne.d hi the. tight ofa fiddmt AeAeafiah ^-inding-i, on pottutant tnanApoxt and
 on oxtdant chemiAtsiy.  Oceavtence. o& pottutant ttianApont, iA i,nteApAete.d to
 dictate, that tie.ac£ivttieA o^ oAganicA  be. me.aAuAe.d undeA tongeA iAAadiation and
 QfmateA ofLgantc.-to-nitAoge.n-oxA.deA fiatio condition* than px.e.v-iouAty thought.
 Re.activJjtieA undeA &ac.k condition* have. be.e.n me.ai>uAe.d by the. EnviAonme.ntat
 ?Aote.ction Agency &OA Ae.veAat ox.gani.cA and fieAuZtA afie. oiecf kojtz to &u.ppo?it a
 new tu)o-c.la{>t> tie.actlvJJ;y c£a££t{itc.ation o& oftgantc. amiAAtonA .   ?nobte.mt>
 ciate.d wtth the. &mog chamber. meaAusie.me.nt o^ -6ac.fi tie.activ4tA.eA  and with the.
 ofa kinetic. indic.eA o^ ?ie.activity aste. diAc.uA&e.d.

                              INTRODUCTION

      The reactivity concept was born out of early laboratory studies that
 showed that organic emissions vary widely  with respect to their ability to cause
 photochemical oxidant formation in the ambient air.  As of today, the concept
 has been in use in control practices in the U. S. in the form of regulations
 calling for selective control of organic solvent emissions.  The best known of
 these regulations is the Rule 66 (1).  It  was designed for use in Los Angeles,
 but it was adopted — intact or slightly modified on -- by several air pollution
 agencies outside California. The Federal Government has issued a drastically
 different guideline for use of the reactivity concept (2), but has not objected
 to the use of Rule 66 by the various State and local governmental agencies.
 Both the Rule 66 and the Federal Government guideline are not entirely con-
 sistent with the latest scientific evidence regarding the role of organic
 emissions in the ambient oxidant problem  (3).

      In their present form, these regulations impose two limits to organic
 solvent emissions:  a high limit (e.g. 3000 Ib/source/day) for the nonreactive
 emissions and a low limit (e.g. 15 or  40 Ib/source/day) for the reactive ones.
 Enforcement of two such limits means that  the thrust of the regulatory measures
 is not so much toward emission reduction as it is toward substitution of non-
 reactive emissions for reactive ones.  In  fact, in theory at least, these regu-
 lations could cause an increase in the amounts of emissions discharged into the
 atmosphere.  This possibility is now thought to have some adverse consequences.
* B. Dimitriades, U.S.  Environmental  Protection Agency, Research Triangle
   Park, North Carolina
  S. B. Joshi, Northrop  Services,  Inc.,  Research Triangle Park, North  Carolina

                                      705

-------
     Because of these problems,  the United States  Environmental  Protection
Agency (USEPA) is in the process of reexamining the  use of the reactivity
concept, in the light of the considerable research findings reported in the last
4 to 5 years.   The following is  a presentation and discussion  of some thoughts
that evolved out of such a reexamination.

     REASSESSMENT OF THE USE AND UTILITY OF THE REACTIVITY CONCEPT

RECENT RESEARCH FINDINGS AND IMPLICATIONS

     There are two aspects of the reactivity issue that need to  be  reexamined.
First, there is the question of  whether or not application of  the reactivity
concept has thus far been demonstrably beneficial.   The second aspect pertains
to the relative roles of the "reactive" and "nonreactive"  organics  when such
organics are subjected to pollutant transport conditions.  These  aspects need to
be reexamined both separately and in conjunction.
      Recent air quality data in the U.S. have indicated that in some localities
the air quality improved, in others it deteriorated, and in others  it  showed  no
effect in  the years since the initiation of emission control.  From detailed
studies of the impact of emission control on oxidant air quality (4,5), it can
be deduced that the impact of the reactivity-related solvent control regulations
cannot be  directly assessed.  This is simply because such impact is masked by
the overwhelming effects of other factors such as a reduction in organic auto
emission  (which is indiscriminate rather than reactivity-related),  an  increase
in nitrogen oxides emissions, growth and/or shifting of human activities, etc.
The judgement submitted here is that the impact of any reactivity-related
control measure upon air quality can be assessed only indirectly based on
laboratory evidence;  direct atmospheric observation is inconclusive.

      The  second aspect of the reactivity issue, the one that deals  with the
relative  roles of the "reactives" and "unreactives," likewise can be examined
only  indirectly based on laboratory evidence.  Relative contributions  of the
"reactive" and "unreactive" organics to ambient oxidant and, in general, the
relative  roles of such pollutants cannot be determined from aerometric data.

      In conclusion, laboratory evidence -- that is, smog chamber evidence --
with  all  its  uncertainties arising from its indirect nature, continues to
provide the only basis for reassessing the use and utility of the reactivity
concept.   For such a reassessment, therefore, it is necessary first that the
recent laboratory evidence be presented and critically examined.

      Recent evidence relevant to the issue consists mainly of smog  chamber data
suggesting that, under the relatively high organic-to-NOx ratio and prolonged
irradiation conditions prevailing in transported air masses, many of the thought-
to-be unreactive organics could, in fact, cause significant oxidant buildup
(6,7).  This  is illustrated by the reported data (6,7) and more pointedly by  the
data  of Table 1, obtained from an on-going EPA study.  These interoretations,
however,  of the recent smog chamber data and associated conclusions have been
questioned by some investigators on grounds related to the smog chamber back-
ground contamination problem.  Specifically, it has been claimed that, in
                                     706

-------
general, nonreactive organics exhibit enhanced reactivity in smog chambers
because of the presence of a substantial background of hydroxyl (OH) radicals in
such chambers (8).  Furthermore, such OH concentrations, wet?e estimated (from
the EPA data on methane (CHj disappearance) to be 1.5 x 10  - 1.7 x 108 mole-
cule/cm3; that is, they were comparable to or higher than the concentrations
observed in the real atmosphere (8).

     Such claims, although qualitatively correct, do not invalidate the original
interpretation of the recent smog chamber data, namely, that under optimum
conditions many of the thought-to-be, unreactive orqanics can in fact cause
significant ozone (O3) buildup.   To explain, consider all possible situations
that cause chamber background OH buildup.   First, background OH could form from
photolysis of nitrogen-compounds desorbing off the chamber walls.   Such a
source, however, is unlikely to be important relative to the source created by
the NO  reactant itself in the chamber.  Second, and more likely,  background OH
could ^orm from photochemical reactions of wall-desorbed organics  (and inor-
ganics) and the N0x reactant.  However, considering that such desorbed organics
are at almost immeasurably low concentrations, it may be deduced that this
source of OH is also unavoidably present in the real atmosphere as natural
background contamination.   Alternatively,  the bias caused by the background OH
radical problem can be reduced -- but not  eliminated -- by subtracting from
measured reactivity values the background  air reactivity.  Finally, irrespective
of the chamber background OH problem, it has been established that conditions of
prolonged irradiation and an optimum organic-to-NO  ratio do enhance the re-
activities of the less reactive organics.

     In conclusion, the laboratory evidence suggests that in consideration of
the pollutant transport phenomenon, the existing reactivity classification of
organics should be revised so as to reclassify most of the "unreactive" organics
into the single class of "reactives."  The question of precisely where the
borderline should be separating the two classes is a crucial but extremely
complex one, and it cannot be answered definitively based on the evidence pre-
sently available unless some simplifying assumptions are made. In  answer to this
question, these authors offer one procedure for defining the boundary between
reactives and unreactives.  This procedure is based on use of reactivity data
obtained in an appropriate chamber under optimum irradiation time  and reactant
organic-to-NOx ratio conditions. This procedure is intended to identify as "non-
reactive" those organics which under no circumstances will yield oxidant/Oa
concentrations exceeding 0.08 parts per million (ppm).  This procedure is illus-
trated in the next section, using the data of Table 1. Note, however, that the
conditions under which these data were obtained may not be the optimum ones.
Additional tests using a variety of organic-to-NO  ratio conditions should be
made — upon the least reactive organics only -- in order to obtain the re-
quisite "optimum " values.

REVISED REACTIVITY CLASSIFICATION

     From the data of Table  1,  it would appear rational  at  first glance that  the
boundary point separating the reactive organics from the unreactive ones  should
be the reactivity value for  propane.  Such a judgment, however, presupposes that
a 4 ppm/0.2 ppm mixture of propane/NOx would yield the same concentration of  03
in the real atmosphere as in EPA's smog chamber.


                                      707

-------
TABLE 1.  REACTIVITIES OF ORGANICS MEASURED IN THE EPA SMOG CHAMBER-/(11)

Reactant
Organic-
Maximum 03
cone. , ppm—
Peak-03 time,
hrs
Average rate of
organic disap-
                                                        pearance, % per hr—
                                                                          d/
Background air
Background air +
0.2 ppm N0x
Methane
Acetonitrile
Benzaldehyde
1 ,1 ,1-trichloro-
ethane
Chloroform
Benzene
Methylenechloride
Ethane
Propane
Acetylene
Isobutyl acetate
Acetone
n-Butane
Methanol
Acetic Acid
t-Butanol
Methyl ethyl ketone
i-Propanol
Acrylonitrile
Perchloroethylene 0.
0.03£/
0 01
\J m \J I
0
0
0
0
0
0.02
0.03
0.03
0.08
0.10
0.18
0.18
0.23
0.25
0.26
0.28
0.30
0.32
0.37
49



31
28
3.1
4.6
2.4
1.5
1.8
3.0
3.6
5.3
4.3
12.1
11.9
12.3
13.7
13.8
8.3
8.6
5.5
1.8



0.05
0.02
2.4
0.1
0.8
3.1
5.7
0.5
2.0
6.3
4.3
0.9
1.4
1.3
2.0
1.1
1.5
3.3
5.3
12..8

— Chamber was a 400-liter Pyrex glass cylinder irradiated externally with
  22 Blacklight and 7 Sunlamp fluorescent lamps.

—Initial reactant concentrations were 4.0 ppm and 0.2 ppm for the organic
  and N0x, respectively.  Each organic was tested at least in duplicate.
  Irradiation was continued until 03 peaked out.
c/
— Reported data represent observed maximum 03 concentrations less the max-
  imum 03 concentration observed for background air (0.03 ppm).

— Corrected for dilution.  Perchloroethylene, methylene chloride, and chloro
  form showed considerably higher rates of disappearance after peaking of 03.
e/
— Value represents highest 03 yield observed in 6 replicated tests
  randomly dispersed within the reactivity measurement program.
                                     708

-------
     While the above presupposition may or may not be correct -- it can  never be
checked directly -- the uncertainties involved are not necessarily prohibitive.
Comparison of the smog chamber and natural  sunlight radiation data indicates
that the light intensity in the EPA chamber is comparable to that of natural
sunlight.   Note that small  differences in the radiation factor between  smog
chamber and real atmosphere are of no serious consequence because, for  low
reactivity systems, such differences affect mainly the peak-03 time and  very
little the maximum 03 concentration.  The effect on peak-03 time, however,
(assuming  "unlimited" irradiation time under pollutant transport conditions)
should not cause the reactivity measurement results to be drastically different
than those in Table 1.  Relevant important factors, other than radiation,  are
the chamber background contamination, the chamber wall surface, and dilution.
The contamination problem has already been discussed; adjusting the reactivity
measurement by subtracting  the background air reactivity is simplistic  but
probably the most acceptable solution to this problem. Thus, for example,  after
such adjustment, the chamber reactivity of methane is zero, essentially  in
agreement  with the photochemical model predictions (8,9).  Furthermore,  the
surface of the EPA chamber  (Pyrex glass) is not expected to alter drastically
the dominant steps of the oxidant formation process.   Finally, dilution  in the
smog chamber tests may not  be as intensive as in ordinary real atmospheres.
However, the difference can be tolerated with the justification that (a)  it is
not believed to be extremely large, and (b) it makes  the chamber data represent
the "most  unfavorable" situation, a situation often used in control deliberations,

     Summarizing the reasoning in the preceding paragraph, the smog chamber
reactivities listed in Table 1 should not be drastically different than  those
expected to be manifested in the real atmosphere.  This means that the  sug-
gestion that the propane reactivity is the boundary point separating the "re-
actives" from the "unreactives" would not be grossly  wrong.  Furthermore,  what-
ever overall error exists is probably in the direction that will produce high
reactivity values, that is, a downward misplacement of the boundary point.
Fortunately, however, raising the boundary point above the propane reactivity
value, e.g., by as much as  100% (i.e., to 0.16 ppm 03), does not seem to produce
a much different classification because the 0.08-0.16 ppm interval is not  ex-
pected to  be densely "populated" by reactivity values of organics.  Thus,  for
example, all halogenated paraffins are expected to have reactivities comparable
to those of chloroform, methylene chloride, and 1,1,1-trichloroethane (i.e.,
below that of propane), whereas all alcohols, ketones, acids and esters  are
expected to be comparable to or more reactive than acetone, methanol, and  acetic
acid, (i.e., much more reactive than propane).  These expectations, of course,
need to be verified by testing commonly emitted organics in addition to  those
listed in  Table 1.  This raises the next question, concerned with the reactivity
measurement method.

     The reactivity measurement method commonly used  thus far is the smog
chamber method entailing irradiation of the test organic-NO  mixture and mea-
surement of resultant oxidant yeild. Use of certain kineticxentities (such as
rate of organic precursor disappearance or rate constant for the OH-organic
reaction)  as substitutes for or indices of reactivity has also been proposed
(10).  In  general, the chamber data on oxidant yield  are inherently more valid
but require extensive and tedious smog chamber tests  covering a wide range of
                                     709

-------
reactant "organic-to-NO  ratio conditions.   The kinetic indices of reactivity are
valid only to degrees depending on their correlation with the oxidant yield
data.  In the case of the simple, two-class classification scheme proposed here,
the most important reactivity data needed are those for the least reactive
organics, namely, for the organics comparable to or less reactive than propane.
For those organics, the oxidant yield reactivity is extremely difficult to
measure reliably because such a measurement is sensitive to interferences from
background contamination and organic reagent impurities.  Measurements of the
rate of organic reactant disappearance and of the rate constant for the OH-
organic reaction apparently can be made more reliably (10), but the results do
not seem to correlate well with the oxidant yield data, at least, judging from
the"data in Table 1.   Thus, of the 18 organics listed in Table 1, benzaldehyde,
benzene, methylene chloride, and acetylene showed rates of organic: reactant
disappearance that were disproportionately high relative to their oxidant
yields (Table 1).  It is not known to what degree this inconsistency reflects
experimental error or peculiarities in the chemistry of these compounds.  Clearly,
additional research is needed (a) to establish the relative oxidartt-yield re-
activities of low-reactivity organics more reliably and for a larger variety of
reactant organic-to-NO  ratio conditions, and (b) to further explore the cor-
relations of oxidant yields with rates of organic precursor reactions and with
OH-organic reaction rate constants for a variety of organic compounds, especially
halocarbons.

                              CONCLUSIONS

1.  Laboratory evidence continues to provide the only basis for assessing the
utility and recommending the use of the reactivity criterion in the oxidant-
related control of organic emissions.

2.  Occurrence of the pollutant transport phenomenon suggests that the existing
reactivity classification of organics be revised to consist of two classes,
reactives and unreactives, defined as follows:
    •  The unreactive organics are those which under optimum irradiation and
       organic-to-NO  ratio conditions in an appropriate smog chamber yield no
       more than 0.0§ ppm of 03.

    •  Reactive organics are all organics except those found by measurement or
       theoretical estimation to be unreactive.

Alternatively, if the recent EPA smog chamber data are to be accepted, unre-
active organics are defined as those with reactivities lower than the reactivity
of propane.

3.  Reactivities of unreactive organics are extremely difficult to measure
reliably.  Uncertainity problems are caused either by experimental error asso-
ciated with the smog chamber measurement or by the questionable validity of the
kinetic indices of reactivity.
                                     710

-------
                                  REFERENCES

1.   Los Angeles Air Pollution Control  District.   Rule 66.

2.   Federal Register, 36., 15502,  Aug.  14,  1971,  Appendix B.

3.   Proceedings of the Solvent Reactivity  Conference.  EPA-650/3-75-010,  Nov.
    1974, U.S.  Environmental  Protection Agency,  Research Triangle Park,  N.C.
    27711.

4.   Trijonis, J., T.K. Peng,  G.T.  McRae, and L.  Lees.  Emissions and Air Quality
    Trends  in the South Coast Air Basin.  EQL Memorandum No.  17, Environmental
    Quality Laboratory, California Institute of  Technology,  Pasadena, Calif,
    Jan. 1976.

5.   Wayne,  L.G., K.W. Wilson  and  C.L.  Boyd.   "Detection and  Interpretation of
    Trends  in Oxidant Air Quality."  Final  Report to EPA from Pacific Environ-
    mental  Services, Inc.  Contract No. 68-02-1840,  July 1976.

6.   Altshuller, A.P., S.L.  Kopczynski, D.  Wilson and W.A.  Lonneman.  Photochem-
    ical Reactivities cf n-Butane and other Paraffinic Hydrocarbons, JAPCA, 19,
    p.  787, 1969.

7.   Heuss,  J. "Smog Chamber Simulation of  the Los Angeles  Atmosphere."   Scienti-
    fic Seminar on Automotive Pollutants.   EPA-600/9-75-003,  Feb. 10-12,  1975.
    U.S. Environmental Protection Agency,  Research Triangle  Park, N.C.  27711.

8.   Weinstock,  W.,  and T.Y. Chang. " Methane and NonUrban  Ozone."  Paper pre-
    sented  at National APCA Mtg in Portland, Oregon, June  29, 1976.

9.   Dimitriades, B., M. Dodge, J.  Bufalini,  K. Demerjian,  and A.P.  Altshuller,
    Envir.  Science and Technology (In press) 1976.

10.  DamaH ,  K.R.,  A.C. Lloyd, A.M. Winer,  and J. Pitts, Jr.  "Reactivity Scale
    for Atmospheric Hydrocarbons  Based on  Reaction with Hydroxyl Radical."
    Envir.  Sci  & Technology,  10.,  No. 7, pp.  692-696, July  1976.

11.  Reactivities of Organics  Under Pollutant Transport Conditions,  S.8.  Joshi
    and B.  Dimitriades, EPA report in preparation, 1976.
                                     711

-------
                                                                                14-4
                    PHOTOCHEMICAL  REACTIVITY CLASSIFICATION  OF
                     HYDROCARBONS  AND OTHER ORGANIC COMPOUNDS

                                    F. F. Farley*

ABSTRACT

      A thorough. Ae.vi.ew o{] oxA.dant data, ^n.om &moQ chambeA. AtudieA ove.A the.
te.n  ye.anj> wot, made, by the. We^teAn OiJL and Gat> M&ocia^tion  in ondeA. to provide,
the.  but available. ba^iA fan. a. photoche.mi.cal n.e.a.ctA.vit.ij cla&Ai.&i.cation
volatile. hydnocanbonA and otkeA  ofiQa.vu.n compound^.   Ozone,  maxima
AeviejM Akow&d ne.activity n.ateA van.yi.ng none. than one.-hundne.d fald i.n a  con-
tA.nu.ouA &pe.ctn.wm  ^fiom mztkana to the. kighl.y tiaactlvn 2
fi&>uit!> Auggutzd that a ?ie.at>onable. numbeA ofa fts.ac.tiv^ty cJtaAt>&> should be
A ejected to dov&i 4acfi a bfioad fiange..   Ai wa-6 done HOAJU.&L by V'Lm e^theA. /tead^tue on. u.ntie.adtive. doei
not  /seem /ad^cjj-ced;  non. doei the. uneven dJj>tAA.buution ofa compounds and the. ove.?i-
           tion o& the. thA.e.Q.-date.Qon.y cJLa&At^t cation adopte.d  by the. CaLi&otinA.a
     ReiouAce^ 'Boan.d.   In &ac.t, the. "ftnai Rupont o^ the. Comm-cttee on Photo-
          Re.activ-ity to the. kin Re4ou/ice4 Boaftd" ^a.\jon^> a "5-tieM.e.d -6cneme."
P-ittA  ha& at(>o p/Lopo-iecf a {^\)e.-cl.cu>& siaactivAty Azale. bcu>e.d  upon ^ue on.d ofa
magnitude, in kydA.ox.yl. tiadtcal. x.e.action ttateA.   A &A.ve.-cate.gon.y sie.activity
cation Mould attow ke.aj>i.bte. AubAtitLition ofi £ei4 sie,acttve.  ^on mon.e. reactive.
voiatite. organic  compounds, e.g.,  in 4 uA^aee- coating 4o£ven£$.  Suck &ubt>titu-
tionf>   .
Caution khouJid be,  exeAcx^sed i.n  at>£igni.ng modeAjote. on high Ae.activ-itieA  to
pana^inic hydnocanbonA un>t
-------
product yields exists.  A more or less continuous spectrum of rates and yields
have been found, ranging all the way from methane, with its extremely low rate
of reaction, to highly reactive species such as 2-methyl-2-butene.  It seems
reasonable, therefore, to classify hydrocarbons and other organic compounds for
regulatory purposes according to their rates of photochemical reaction, or more
appropriately, according to their ozone (03)--forming potential.

     When smog chamber data are extrapolated to the real atmosphere, considera-
tion should be given to results obtained from atmospheric reactivity studies
such as recent airborne measurements.

     The Western Oil  and Gas Association  (WOGA)  review was  directed primarily
toward reactivity in  the traditional  sense.   This  assumes  that the photochemical
reaction rates  and  oxidant yields  in  one  solar day's irradiation are of prime
importance in  judging the impact of a given  organic compound on  air quality.   We
recognize the  importance of considering this question in a  broader sense in
light of the recent findings related  to oxidant  and oxidant precursor transport,
high  rural oxidant, etc.   One theory  put  forward to explain high rural  oxidant
points to the  slower reacting compounds,  unreacted during  the first solar day,
as being the primary precursors  during the  next  solar day.   While we agree that
this  is a possible  explanation,  we know of  no data which directly show that the
slower reacting compounds are the  prime source of  the 03,  and not the many other
compounds present in  the air mass,  such as  organic reaction products of the
faster reacting compounds.   If the slower reacting compounds are significant
precursors,  this would tend to diminish the  utility of a reactivity scale based
on 03 formation in  smog chamber studies for  situations where multi-day irradia-
tions could  lead to elevated oxidant  downwind of an urban  area.

                               SMOG CHAMBER  DATA

     To determine reactivity classes, all  available 03 maximum concentration
data collected in smog chamber studies over  the  past ten years were pooled.  The
data, shown  in Tables 1 and 2, are normalized to toluene =  1.  In the case of
data from Gulf, where toluene data were not  available, the average propylene
value from the other labs reporting propylene was  used in the normalization.

     These chamber studies, for the most part, involved the irradiation of pure
compounds in the presence of nitrogen oxides (NO ) at a molar HC-to-NOx ratio
(HC/NOx) of about 2.   This ratio is a reasonablexapproximation to that found in
urban atmospheres.   For example, studies  carried out by Kopczynski, et al. (8)
in Los Angeles in 1968 showed an average molar HC-to-NO  ratio of 2.0.  Data
presented in a California Air Resources Board (ARB) stafV report (9) dated
November 13, 1974 substantiate this value.   In this report, average 6-9 a.m.
measured values for reactive hydrocarbons and total NO  are given for 1965 to
1972, and projected values are given  up to 1985.  Allowing for the non-methane
hydrocarbons not included in this  reactive  hydrocarbon tabulation gives a ppm C-
to-NOx ratio of about 8-10, which  again,  gives a molar ratio of about 2 since
the average  carbon  number of non-methane hydrocarbons in urban atmospheres is
about 4.5.

     On a day-to-day basis, HC-to-NO  ratio  does vary considerably from the
average, and values from 1  to 4 coulS be expected.  The implications of this HC-
to-NO  ratio effect are discussed  in  more detail later.

                                      714

-------
coLocooocy>oocM   cri





fs
CM
II

X
O
z
o
in
i—
OO
1 1 1
h-H
cc
o
t—
o:
o
CO
j ,_

OO II
O LU
rv ii i
>• _1
O
O
Lu O
t—
LU
Ol , I
LJLJ
M M
O i—"
	 1
s: <:
S c§
h-. O
X ^
s:

~^*_
o
CQ
ce
2
o
cc
Q
1C
,_!
LU
— I
co
«3^
1—




































































































*
>)
o
*->
 r—
LO «d-
C5 O
CO CO 1 —
O O O CO «d-
CD O O O O






co r^ co
LO LO ID
o o o


1
1
1 LO LO
cr> oo i—
0 0 i—

CO
o
•
o






00 O CO
O i— i—

o o o










cu
c
CEXO-l->+->
n5Q-3CUCL)COr-OU
JZ OCQDuQ-rC OOO
-!-> S- 1 1 1 1 >> 1 1
UJQ- C C-r- CO C'r-
O O
O r—




-— -
5
•
0

o o
O i—





r 	 s
o
CO
o
r^. "^— *
i — O
O r—

CO
•
O
O — '
CM O
O •—





-K
•X
^ 	 ^
*»o
CO
o
•* 	 ^
CD ^^
• .
0 r—













CO CU
c c
cu cu
N ~^
C r—
O) O
CQ 1—
OO
0








1 —
Cft
o









0

o





o
<-











CM
CTl
•
0









CO
c
cu
c
CO

, —
>5
^:
4->
LU
r^
o




















o

0




























cu
£=
cu
N
c
cu
^
>^

0
t_
Q.
1
c
r-«OLOLO o LO
O r- r—









co o rv. co i^.
IOOO LO"d"i — COOOLO
OO OOr— r— Or—




LO LO
CM LO LO LO P^ CM
r- O 0 •— i— CO











CTl
[^ LO ^J~ r^*
• • • •
O i— r— r—
CU
0)
N
C"
CU
-d
O) •—
c cu cu cu >>
CU C CU C C -E
N CU C CU CO •*-*
C N CO N N CU
CO C N E C E
r^.aa3.a-Qcucucu'£
>,r— J3 r— i— C E E 1—
O. >,r- >, >, CO CO CO 1
O-M >>+-> -M r— r— r— LO
t- 13 E ^ — 3 ^^ ^> ^^ **
D-CQetCQCQXXXCO
1 1 1 1 1 1 1 1
•r- E •(-> T- 4J O E Q-r-


























-a
CD
^j
c~

J3
E
0
(J



































                 715

-------
           LO
         s_
         o
        •P
         to

         o
        J3
O)
O
O
                     CM i— i— i— i— CM i— (M
                                                  CM
                 oo
                              LO
            (O
OO
cn co
             in m UD    o    10

            •l-^i-^l—    CM    r—
           OO
                     CTi
                                     OO
           C\J
                     CM
                  i— CMr— r— CMCMCM
                                                  CM
                                                    •

                                                  oo
                                      CU
                                             Ol
                                            4->
                                             OJ
                                  (U


                                  OJ
                                      (D
                                cu  cu -P     >>    -i-
                      >-p  x CQ ca  i   x  re
                   >> CL 3  CU  I   I  O>  CU  S-
                                 CQ

                                 oo
                  -P  S-  I   I   I   I   I   I   O>
                  UJQ-T--1--POCMCMI—     i—
                                                              O)
                                                              £
                                                              o
                                                              ISl
                                                              o
                                            re
                                            3
                                            -P
                                            o
                                            ai   •
                                            S- £
                                            re a.
                                               Q.
                                            
                                            OJ C
                                            (/) T-
                                            
                                             CD
                                             O
                                             c
                                             (U

                                             0)
                                            4-
                                             cu
                                            a:


                                             21
                                                              S-

                                                              OJ
                                                              cu
                                                              -a
                                                              o
                                                              o
 S-
 o
-P
 re
 s-
 o
-Q
 re
                                                                     v>
                                                                     cu
                                                                     c
                                                             c/> to
                                                             s- s-
                                                             o o.
                                                             -P -P
                                                             O O'
o -—- oo        cu i—
   CM	.— i— r—   •>
3 —    re re i— r-~
re    i— s- s-  a)—*
cu M- •— a) cu -P
s- i—  cu c c -P 
-------

t/\
*•/ /
UJ
1 — 1
o
0
CQ

_i

00

O i—
1— 1
> UJ
•z.
rv* I i 1
o •=>
It I
I_L_ '
UJ I—
•z.
o o
M I—
0
Q
• UJ
x hvi
^C *"•"*
s: _i
 ^3- CO
^^^ -_ _^ 	


LO
•
r— r—
.101
O LO
•
o


r--. LO
O LO CO 
Of— •> f— > f— \
^— ' 1^^ V.^







LO i
fO ^^ ^~
•P j
-P r— i— .E
CO U O -P
0 >, >> CO
>r—
4-^ >s
ZJ -P
f~^ ^g
O J3
(/> 1
•r- E

>> >s
-E -E
-P -P
CO CO
^ ^
LO •*






1 1






1 1










1 1








LO •>!-
O i—





CO
E CO
O E
-P O
CO -P
•*s CO
XX
r—
^>' 	
Q. >,
0 -P
i. 3
O.J3
1 |
E -P

>> >>
J= -E
_|.i t \
QJ QJ
21 21
o r^ oo LO r^
	





i i i i i






Ii i i i
i i i i








o

•—







i~-> oo LO r-~
O r- r- i—






CO
E
O
-P CO CO
CO E E
.*: o o
-P -P
i— CO CO
>,.^ -K: co
fc -a
IO 1— 1— •!-
o >, >> x co
to Q.-P O E
-i- O 3 O
s~ .a •— s-
>, 1 CO -P -E
JC E T- -i- QL
•P 1 1 CO O
CO •!- -i- CO CO
s: Q Q s: >-•
OO LO LO
CO CO O CM
r~i /"~i




^_
II • 1
^~




LO LO
CO CO O CM
_






^
*
CM
!•!!
•—







CM
|| |
• 1 1
O










r—
o

o o "o
-E O i— -E
O i— O O
<-> fO -E O
•— O i—
to i — O (O
^>^~
i— Q. IO r—
^^ O ^s
**y *~ ^\
-P S- r- -P
3 a. >^ 3
CQ O -E CQ
1 CO -P 1
-P i-< UJ E
LO
CM






1





LO
CM










1





















P— •
O
-E
O
O
r—
(O

1—
>5
3
_Q
o
t/j
i — i
CM 'CS-






1 |





LO
^" ^t-









p^
• |
'—

















r—
O
E
to
i— -P
O E
-E CO
O Q.
0 1
i— CM
(0 1

•— ' >>
4J -P
3 CO
ca 2:

CO ^-
                                            -o
                                            0)
                                            o
                                            o
717

-------





















*
>^
S-
o
4->
O3
O
_Q
*~-^ 03

CU
3
E

4->
E
O
U

.
CM
LlJ
	 1
CO
^^
r—






























CO
Dl
03
&"•
cu
>
c^








^J-






















CM






r—































LO LO
LO r--. co co co ^ r—
O O CD O O r— i—










1 1 1 1 1 1 1 1









LO LO
LO 00 CO CO O
O O O O r—









I"*"*' I"**1* oo *^j" I™™
1 • 1 ... 1
O O t— r— r—




CM
1 1 1 1 • 1 O 1
o













CD
CD CD 4->
4->  tO
CU  tO 4->
I ^ 4-^ 4-3 
0 4-> 0 <0 CD CD
03 CU r— 03 r— O >
O >> r— >> fl3 r—
r— >4-> O
>, O >j4-i 3 r— to
4->r- S- 4-> 3.Q >,O
3 ^> <^ ""S «^ c~ r^
CQjCOOQOOCUr—
1 40 CO 1 CO CD .E CD
40 LiJ i— i Ei — IOOQ-O

LO
CM O> CM r — 1^ r^~ LO
O O O O CD P— CM










III 1 1 1 1 1









LO LO
CM CT> <* LO
!•• Ill"*
O O r— CM










III 1 1 1 1 1





CM r— r^ CTv
Oil .... |
O 0 0 r—











cu
CD E
CU T3 O
-O -i- "O
•r- E T-
CD E tO r— C
4-> 
CD >> ,4-> !—>,>, >,!—
r— .E CD >>-E ^: .E >>
>,4->E ^=4->4->03.E
.ECDT- 4->CL>a>S-4->
4->EQ 0)T-2:4->CD
CU T- 1 T- S- 1 CU T-
SO^ Q 1— E|— Q













































































•—
^
O
o
p—


t—t
fV
oo

CD

h-

-)e
*




f 	 ^
CD

E
CD
S-
CD
M-
a>
S-

co
4-)
03
CD

•P-
_J
*^^^

CD
T3
0
O
>>

O

S-
o

03
— 1


0


•
CO
C
*l —
1 —

q-
o

4->
^
o

CO
•°
0
4-5

(/)
E
cu

03
CO

1
, 	







E
CO
f)

O

C
•r~
-4_3
(O
s-
cu
Q.
O

M —
o

tu
CO
3
03
O
CU
JD

CU
E
O
•r~
4_3
CO
CU
3
CT
























*• 	 x
CO
^^^>

1 —
r—
0)
_c
00

1
CO











d
•r-

C/)
E
O

-M
03
•r—
S-
03
>

•o
E
03

CO
CU
4->
i— .
^5
O
•r—
<4-
Cf«
•r-
•a



















^ 	 ^
LO
i
«*
o
I—
•* 	 rf*

Q
0
a.
<
<

I
<*











it
t/t
c:
O
•r—
-M
10
S-.
4-'
E
O>
O
E
O
CJ

a>

•r™
x:

E
Cl)
DT
0
$-.
-}-J»
•r-1
c:


























































•
OJ
r___

CO
o
r"
CU
>-
(U
4—
CU
fy^

E
•r—

-a
E
^
0
<4-
03
03
"°
£
>
•r-
4^
O
03
CD
S-

4-^
E
CO
o
CO
q_
O

>«
S-
O3
E
E
*^
CO
4J
E
CU
E
CD

E
O
O
718

-------
     The average normalized values found for each organic compound in the
various smog chamber studies are plotted in Figures 1 and 2.  Averaging the data
is not without some drawbacks, since systematic chamber-to-chamber differences
still appear to a certain extent, even after normalization.  As a result, the
values shown in the "average" column in Tables 1 and 2 for compounds which have
been tested in only one or two facilities may reflect a bias peculiar to the
facility in question.   However, apart from a few compounds, for example, n-
pentane, and i-octane, methyl-n-butyl ketone, and isopropyl alcohol, which may
be out of line, the data should provide a reasonable estimate of relative
reactivity.

                           REACTIVITY CLASSIFICATION

     In Figure 1, the hydrocarbons are plotted according to class:  paraffins,
aromatics, and olefins.  Generally speaking, the paraffins show the lowest
potential for 03 formation, aromatics show intermediate 03-forming potential,
and the olefins show the greatest potential.  The oxygenated solvents are also
plotted in Figure 2 according to class:  ketones, alcohols, acetates, and mis-
cellaneous.  Within the classes, there are some definite trends in reactivity
which appear to be structurally related.  Based on the data in Figures 1 and 2,
and following structural guidelines, the reactivity classes shown in Table 3
were defined.

     The data in Tables 1 and 2 and in Figures 1 and 2 show a wide range of
photochemical reactivity, over one-hundred fold.  This wide range of reactivity
covering a more or less continuous spectrum of rates, ranging from methane to
highly reactive olefins, suggests that a reasonable number of reactivity classes
should be selected to cover such a broad range.  The WOGA approach was to
define five classes as was done earlier by Dimitriades (12) and  others  (24).   The
five-category classification represents a technically sound approach and pro-
vides adequate definition of the broad range of reactivity values observed, but
at the same time it does not over-emphasize the accuracy of the data by at-
tempting to subdivide further.  A five-category classification avoids the
uneven distribution of compounds and the oversimplification of the three-cate-
gory classification recently adopted by the ARB (21).  In fact, the "Final
Report of the Committee on Photochemical Reactivity to the Air Resources Board"
(22), issued after ARB adopted a three-category classification, favors a "5-
tiered scheme with an additional sixth or zero class."  Pitts, et al. (23)  also
proposed a five-class reactivity scale based on hydrocarbon disappearance rates
due to reaction with hydroxyl radical (OH).  They "found it convenient and
useful... to employ the order of magnitude divisions which lead to a five-class
scale."

     As noted in Table 3, the assignment for C4 paraffins is somewhat uncertain.
The data in Table 1 for n-butane at HC/NOx = 2 show very low maximum 03 values
for four separate studies, with a significant value found in only one study.  At
higher HC-to-NOx ratio, greater 03 yields have been observed as is discussed
later.

     A statistical T test was applied to the data to determine if the difference
in means between adjacent categories was significant at the 95% confidence
level.  The lack of a significant difference would indicate that there is no

                                      719

-------
CO
CU P
P CO
i

CO X! 4J
•> 4J CO
r-l CU r-l
E ^>
• i _P co
CM 4J P
CO CO CO
• ErH
CO co >i
P (X
4J O 0)
cu J-i C
H PM CU CU
p -1-1
^ 3
CO CO P 1
pa CM
1 1
r-l 0

coco









































cu
p
0)
N
p
cu
X)
r—i
^»
Jp
4-1
QJ
E
•H
5-1
H
I
in
CO
cu
P <-4
cu cu
P 4J *
CO P fO
X PQ CU
CO 1 P
jr~] CM CU
i 1 X
co
4-* p
p ^ cu
•rl T-4 >
o < cu
CL U P
co co P
CO CtJ «rl
4J r-4 P CU
CQ O J-l 'l-l
"O 00
K^^-l S3
60 4.J -H _
P T-lr-4
•rl > CO •
T3 *r< O 'CJ
0) 4-J }-l
O O >^ CO •
CU TO X> O P
S-< QJ PQ O
P_, S-l 4J -r-l
R W 4J
co a) cu cu P
r-l 4J E O X)
CO CO P J-l *rl
VJ O M P J-l -
0) «H -H O -U
6 T) CO CO CO
p p co cu -i-i
S3 «rl CO [^ T3


PC
• • 1
1 — I
co co cu cu cu
* P P P
CO CU CU CU
rH i-l r-l —
X? XX
4-1 1 1 CO
w OS S OP! p 
• O rd r-l r-l CO CO W
H r-l >, >~,Xl P CU CO 1
CO >i4J O.r-1 CO P P P CU _
*JSPO>~.NCUCU P
COUpQ^-iCUpMN .CUCO
bJIP^OCUPP PX
PIS-IXJCUCU CMCOCU











in
1
CM ^

1]
QJ
c
QJ
^3
'o

o 3

CM -a
p-l QJ
EH N
M ^

*~* E
§|

53 "
05 c\j
m
• w "
H z
rS J
s
Ps +J
% r-l i-4 OOP
CO «P4J>^>~> Or-ICQ
CO P E J-» 1 0 -U
CO'tQCOP -HS^O CU
co i i X) 00 cu P
.,-( 4J 1 ' 1 C CO
4J • P CO 4J
CM X P
CO CM ' CU CU —
CM P3 fr( Q)
CM CM 1 1 C
CU . C «H HJ
ej_.l
• » -J
cu • 3 -
N CM PQ CU
p CM 1 P CL
0) P CO P
PQ Cf K
.OX
1 | ^
1 * CM PU tb
• r-4 1
H3
•^

,_!

cu
m ^
a
O -r-
LL.




)
«
i



o
SDIIVWO^V
 720
SNIJJVHVd

-------An error occurred while trying to OCR this image.

-------
















CO
Q

-^
O
O-

o
o

f ^
1 — 1
;^
^f
fr^
CtL
0
Lu
O

^y
O
1 — 1
l__
^^
o

Lu
1 — 1
co
co

_j
CJ

1 —
1 — 1
1 — 1
1 —
o


to
to
ra

f__j









;=•
i — i

to
t/)
ro

O








i— i
t — i
i — i

to
to
ra
r—
f ^












I — 1
1 — 1

to
to
ra
f—
O










1— 1

to
to
ra

o








>^
-E 4-5
C7)*r—
•r- >

4J
>> 0
S- ra
cu cu
> a:








^.
4-5

-E >
CT)T-
•r— 4-)
n: o
ra
cu
CtL







cu >>
4-> 4->
ra *i"~
S- >
CU T-
-a 4->
0 0
S ro

cy







>^
4-*
•r™
>
•r—
-M
O
03
O)


•^
a
\







cu >,
i— 4->

•r- >
01-r-
•r- 4-5
r— O
CD ra
CU CU
2= a;





0
•
i—

ii

S-
o
4-5
O
ra
u_





LO
1^
•
o

II

S-
o
4J
o
ro
u_




o
LT)
•
O

II

s-
o
4-)
a
ra
LI-






LT)
CNJ
•
o

II

S-
o

o
ra
U-






















CL
cu s-
O r— CU
x >, ^:
CU CU -V to 4->
E to i— CU CU
" CU E "O ra E
to i — -r- E ro CU i —
E >-><4- ro S- N >,
•r- _E CU 4J E .E
14- 4-> r- 1 CU CU 4->
cu a> o ••- i— -Q a>
1— T- S_ -r-
O Q 1— Q



to E
CU ra
E S-
CU 3
N <*-
E O
CU S-
cu J3 -a
E r- >>
CU >>JZ
i— -i^ ro
>,r— S_
J^ ra 4-5
4-> -i- CU
LU Q 1—

>> 4-> CU CU
10 S- Q. -O 4-J
CU ro CU -i- ro
E T- 0 •— E 4->
CU 4-> X >, i— ro CU
NS-.— CU -^ >, 4->0
E CU >,-~^-^!— ^4 CU ra
CU 4-> -i^ i— rO •— O
-Q i — to>> t/)ro raCU
•— 4-> ra r- Q. >,i— I/) >
>^ r^* O O t- O "O CU ^~ r~~
^ cu >> c~ ^_ fa *"~ cu E ^> o
i— (JS-OCL-OO-EO-CIO
r0XrOUOC(JU4->4->O
OCUEr—tOOi— ECUCUi—
E^ 	 ••!- rO T- O ra ra .:<: EI —
OS- CU S- T- CU
SO- CO OQ Q O
r_
0 0
•r- to i— J=
i— •— CU >> r— 0
W>,0 C i— ^ >, CUO
E-ii >, O >-, r— ^ Er-

4-rato CUi — to toroto "O
M- CUT3tO-^iraCU:>iCU CU T-I—
ra ^} E CU CU 4-5 S— 4-5 >} 4-5 n™- r— >}
S— S— CU 4-5 E r—~ ^5 ra ro ro S— rO ^^ O Q^
ra ro N ro O >•> S— 4-5 ~O [ ^ to 4-5 c~ S— O
o (r- ^ ^ | < sx ro CU E CU *r— CU 1 ^ S— S—
4-5 CU ^ CU t~~ E C_) O O 4-5 O CU >3 r">
+ S_J34-5^<:eCT- ra O ra S-rOS: Q-O
i^) (U ra IS- CU CU 1 to
Oh- CO ECL. CO h- "Z. i— i


to
E CU
•i- TD
M- T-
<+- E CU
ra i— ro 4-> CU
S- >> E CU CU ra 4->
ro -ii S- E E O ra
Q_ r— O -r- T~ N 4-5
rO to M- E E: E CU
* CU r— >Oi — i— -Q rd
O CU CU CU S- -E >ir— >,
r— E E ra O-E >>x: r— i—
O >»CU O-i- O4-J-E4-5 >>>)
4J 4_> rsl 4-> 4-> r— CU4-5 CUJE C
CUECUS-rOECUT-4->CU
^H O CU O CU -i- -r- S-. CU -E
o«=tcn<:i— QQI— so.
.?
5>
•r-
4-5
tj
rO
cu
S-

tf-"
o

a)
CD
E
(0
s~

ro

i/)
3:
O
_E
(/I

t/l
ai
c:
rci
4—'
«E
•°
1-
O

4-^'
C
CU'
£
cu
o
ra
r—
o.

cu
.E
4-5

s_
o
cu
r— •
r^ •
rO ^ '
i — i —
•i —
ra CU
> I—
rO -Q
ro
S h-
O
C CU
cu
cu co
O — '
c
CU i — i
-O 1-1
> -a
CU E
ro
o
•r— 1 — 1
tl—
•r- to
4-> IO
E ro
CU r—
•r- O
O
tO E
CU
CU CU
JC S
1— 4-5
CU
* -Q
722

-------
justification for making such a separation.   This test, which was carried out on
the complete data set and not just on the average values, included any lab-to-
lab variability in the analysis and showed that the difference in means was
significant at a confidence level greater than 95%.  It should be noted that an
initial attempt to separate the diolefin and substituted internal olefin data
from the rest of the olefins as a separate group did not yield a mean which was
significantly different from the mean of the rest of the olefins.

     As mentioned previously, the data in Tables 1  and 2 and Figures 1 and 2
were essentially all obtained at or near molar HC/NOx = 2.   At lower HC-to-NOx
ratios, 03 formation is inhibited due to the reaction of 03 with nitric oxide
(NO).  At higher HC-to-NO  ratios, relative reactivities of the more reactive
compounds do not seem to Se markedly affected.  However, chamber studies in-
dicate that the less reactive compounds, such as low molecular weight paraffins,
can become more important contributors to 03 formation (5,  7).  The available
data for lower molecular weight paraffins at higher HC-to-NOx were reviewed to
determine to what extent they might contribute to 03 formation on days when HC-
to-NO  was greater than the average.  The relevant data are shown in Table 4.
          TABLE  4.   REACTIVITY OF  PARAFFINS AS A FUNCTION OF HC/NOX
                          Max. Ozone for Various Laboratories*
Compound 2
Molar HC/NOY =246
PPM HC/NOY =8 16 24
Ethane -
Propane -
n-Butane 0.03 0.41
i-Pentane -
n-Hexane -

2
8
0.0
0.0
0.03
0.33
0.47
4a
4
16
0.0
0.13
0.60
1.03
1.00
6
6 246
24 8 16 24
0.3 -
0.40
0.77 0.0 0.0 0.0
1.00
0.93

*  Values normalized as in Table 1.
     The most complete data set is contained in a study by Heuss (5).   This
shows a definite trend to higher 03 potential for n-butane, i-pentane, and n-
hexane at HC-to-NOx of 4 and 6, compared to the values at 2.   The Gulf study
shows a similar trend for n-butane (2).  However, the work by Altshuller, et al.
(7) shows essentially no 03 formation from n-butane until the HC-to-NO^ reaches
about 10.  Overall, the data suggest that under typical ambient conditions, as
exemplified by HC-to-NO  of 2, the C}-^ paraffins are properly categorized in
Class I and the C5+ paraffins in Class II.  While at the higher HC-to-NOx there

                                     723

-------
is evidence of increased reactivity for the C5+ paraffins, this does not appear
to be sufficient cause for removal  of the C5+ paraffins from the Class II
category in the general purpose scale.

                        ATMOSPHERIC REACTIVITY STUDIES

     Three airborne studies have been conducted recently that indicate low
photochemical  reactivity for light  paraffinic hydrocarbons.   These studies are
the refinery plume study of Westberg and Rasmussen (16), the Portland plume
study of Londergan and Polgar (17), and and analysis of LARPP (Los Angeles
Reactive Pollutants Projects) data  by Calvert (18, 19, 20).   Further exper-
imental  work is needed to verify the low reactivities for light paraffinic
hydrocarbons found in these airborne studies, and caution should be exercised in
assigning moderate or high reactivities to paraffinic hydrocarbons on the basis
of smog  chamber experiments alone.
                                  CONCLUSIONS

     A review of all available 03 data from smog chamber studies has been used
to prepare a photochemical reactivity classification of hydrocarbons and organic
compounds.  This review was confined to 03 (oxidant) formation to provide the
broadest scientific base for a reactivity classification.  The very wide range
of photochemical reactivities encountered in this and other reviews argues for
a multi-class scale, such as the five-class scale proposed herein.  A five-
category classification provides adequate definition of the broad range of
reactivity values observed and does not over-emphasize the accuracy of the data.

     The significance of recent airborne measurements of photochemical  reac-
tivity of light paraffinic hydrocarbons should be given careful consideration.

                                  REFERENCES

1.   "Development and Utility of Reactivity Scales from Smog Chamber Data",
     Bureau of Mines Report of Investigations #R18023 (1975).

2.   Unpublished data of Gulf Research Laboratories, Harmarville., Pa.

3.   "Photochemical Smog and Atmospheric Reactions of Solvents", 0. L.  Laity, I.
     G. Burstain, and B. R. Appel,  Paper presented at the National ACS Meeting,
     Washington, D.C. (September 1971).  Advances in Chemistry Series, Number
     124, Solvents Theory & Practice, 95 (1973).

4.   "Hydrocarbon Reactivity and Eye Irritation", J. M. Heuss and W. A. Glasson,
     Envir. Sci. Technol. 2_, 1109 (December 1968).

5.   "Photochemical Reactivity of Mixtures Simulating Present and Expected
     Future Concentrations in the Los Angeles Atmosphere".  J. M. Heuss, pre-
     sented at the 68th Annual Meeting of the Air Pollution Control Association,
     Boston, Massachusetts (June 1975).

6.   "Final Technical Report on the Role of Solvents in Photochemical Smog

                                     724

-------
     Formation", A. Levy and S. E. Miller, National Paint, Varnish, and Lacquer
     Association, Washington, D. C.  (April 1970).

7.   "Photochemical Reactivities of n-Butane and Other Paraffinic Hydrocarbons",
     A. P. Altshuller, et al., J.  Air.  Poll. Control Assoc. 1_9_, 787 (October
     1969).

8.   "Photochemistry of Atmospheric Samples in Los Angeles", S. L. Kopczynski,
     et al., Env. Sci. Techno!. 6., 343 (April  1972).

9.   State of California, Air Resources Board Staff Report #74-21-4A,  (November
     1974).

10.  "Effectiveness of Organic Solvents in Photochemical  Smog Formation", M. F.
     Brunelle, J. E. Dickinson and W.  J.  Jamming, Air Pollution Control  Dis-
     trict, Los Angeles County (July 1966).

11.  "Investigation of Photochemical Reactivities of Organic Solvents",  K. W.
     Wilson and G. J. Doyle, Final Report, SRI Project PSU-8029, Stanford
     Research Institute, Irvine, California (September 1970).

12.  "Proceedings of the Solvent Reactivity Conference",  Dimitriades,  B., et
     al., EPA 650/3-74-010, (November 1974).

13.  "Photochemical Smog Reactivity of Solvents", Levy, A., Advances in  Che-
     mistry Series.

14.  "Photochemical Aspects of Air Pollution:   A Review", Altshuller,  A.  P.,
     Bufalini, J. J., Environ. Sci. Techno!. (January 1971).

15.  "Photochemical Reactivity of Solvents", Hamming, W.  J., A Paper Presented
     to Aeronautics and Space Engineering and Manufacturing Meeting, Los  Angeles,
     California (1967).

16.  "Measurement of Light Hydrocarbons in the Field and  Studies of Transport of
     Oxidant Beyond An Urban Area", Westberg,  H. H. and Rasmussen, R.  A.,
     Washington State University Report to EPA, Contract  No. 68-02-1232,  Sep-
     tember 4, 1974.

17.  "Measurement Program for Ambient Ozone, NO  and NMHC at Portland, Maine,
     Summer 1974", Londergan, R. J. and Polgar,xL. G., TRC Report to EPA,
     Contract 68-02-1382, November 1975.   "Surface and Airborne Ozone  and
     Precursor Concentrations from a Medium-Sized City",  L. G. Polgar  and R. J.
     Londergan, presented at the 69th Annual Meeting of the Air Pollution
     Control Association, Portland, Oregon, June 27 - July 1, 1976.

18.  "Hydrocarbon Involvement in Photochemical Smog Formation in Los Angeles
     Atmosphere", Calvert,  J. G.,  Environ. Sci. Techno!.  JJD, No. 3, 259  (1976).

19.  Letter to Editor, Ludlum, K.  H. and  Bailey, B. S., accepted for publica-
     tion, Environ. Sci. Technol., May 12, 1976.
                                     725

-------
20.  Letter to Editor, Calvert,  J.  G.,  accepted for publication, Environ.  Sci.
     Techno!., May 17, 1976.

21.  "Adoption of a System for the  Classification of Organic Compounds According
     to Photochemical  Reactivity",  State of California Air Resources Board Staff
     Report 76-3-4, February 19, 1976.

22.  "Final Report of the Committee on  Photochemical Reactivity to the Air
     Resources Board of the State of California", S. W.  Benson, J. G.  Calvert,
     J. G.  Edinger, C. Foote,  A. J. Haagen-Smit,  H.  S. Johnston and J. N.  Pitts,
     Jr., April 1976.

23.  "Development and Application of a  Hydrocarbon Reactivity Scale Based  on
     Reaction with the Hydroxyl  Radical", J.  N. Pitts, Jr., A.  C.  Lloyd, A.  M.
     Winer, K. R. Darnall and  G. J. Doyle, presented at the 69th Annual  Meeting
     of the Air Pollution Control Association,  Portland, Oregon, June  27 - July
     1, 1976.

24.  "A Critical Review of Regulations  for the Control of Hydrocarbon  Emissions
     from Stationary Sources", Milton Feldstein,  J.  Air. Poll.  Control Assoc.
     24, 477  (1974).
                                     726

-------
           SESSION 15
ATMOSPHERIC CHEMISTRY AND PHYSICS
                J.G.  Calvert
      Ohio State University
                727

-------

-------
                                                                             15-1
                     DECOMPOSITION  OF CHLORINATED HYDROCARBONS
                      UNDER SIMULATED ATMOSPHERIC CONDITIONS

                               F.  Korte and H. Parlar*
ABSTRACT
iy
          papeA de,aLt> uiitk induA&Lial onganic efie.nu.ca£6 wkick  one. not ge,neAal-
               at> aifi poll.uta.ntf, although. the,y one, att>o pfie^znt in aiti.  The.
        th^  peAAtt>te.nt induAtsviat  cfie.t?u.cat6,  e.g. chloAinate,d  Au.bAtanc.nt> Like,
PCB'A, WT,  tfC8, ?e,ntackto fiphe.no t,  an the. Ao-caU,e,d c.hiofiinate.d kydnocatibon
peAtictdeA unde.fi atmoAphe,fu,c condition^  cute, dit> coined.  ffiom modtt co£co€a-
tionA w-itk WT Lt -U known that the. c.once.ntsiationA actually ^ound tn the.
e.nv.iAonme.nt  one. conAtdeAably towex than  c.aicutate.d.  On the. bcu> o& pro-
duction faigufLeA , theJte. mu&t e.XAJ>t  a non-tde,nti^e,d A-Lnk -in the, e.nvJJionme.nt.
In e,x.peAJjme,ntA unde.fi Aimaiate-d atmo&pheJu.c. condition* , the, mineJiaLization ofa
&uc.h pe.U-u,te,nt &ubt>tanc.eA af> pe.ntac,hio ipke.no L, dieJLdJLin and otheAA can be.
                ThuA, atmo£phe,tu.c. mineAotization might be. the. fainat i>-ink ^on,
                      whtch are. exposed to AunLight eAt.he.fi a^te,fi e.vaponation on
                      -iu/t^ace.  The. kt.ne.ticA o& bfie.akdown to cafibon dioxA.de. and
                                       veJiaJL^ model. che,micatt> .

                                    INTRODUCTION
de.moMtAate.d.
many otke.fi
on the.  Aotid
hydfiochtofiic acid -u, ptieAe.nte.d  faon.
      Problems caused by environmental  chemicals have  a  wide range of chemical
biological  and meteorological  aspects.   This report focuses mainly on the
chemical  aspects.

      Environmental chemicals  are  introduced into the  air in several ways:
through  air transport, from aeroplane  application, by evaporation from plants
and  soil  and via transpiration by plants.  For instance, four weeks after
foliar  application of 1.2 mg  of 14C-labelled endrine  per plant on white cab-
                           mg
bage,  residues  in plants and soil  amounted only to six  per cent
had disappeared into the atmosphere.
                                                                    94 per  cent
      In  the atmosphere, environmental  chemicals can  be  converted through
photochemical  reactions.  The  ratio between the particle form and the free
molecular gaseous form  is not  known.   It is difficult  to determine the exact
form  in  which  the molecules  are  found in the various atmospheric layers,
since  they can exist either  as single molecules, as  clusters, or can be ad-
sorbed to aerosols.  For this  reason  development of  a  model  to simulate
*Institut  fu'r  bkologische Chemie  der  Gesellschaft fur Strahlen- und
 Umweltforschung  mbH Munchen; and  Institut fu'r Chemie der  Technischen Universitat
 Munchen,  Federal  Republic of Germany.
                                       729

-------
atmospheric conditions is advised.

     What chemical changes do environmental  chemicals undergo after they are
released into the atmosphere?  Because of the difficulties involved in direct
atmospheric studies, experiments in the laboratory is the approach that allows
for conclusions to be derived on the behaviour of a chemical  in the environ-
ment.  Photochemical experiments under simulated atmospheric  conditions lead
to an understanding of the relative reactivities of environmental  chemicals,
and of the formation pathway and identities  of photoproducts  -- the latter
usually occurring at low concentrations.   Photochemical  reactions  in  solu-
tion simulate conversion of these compounds  when dissolved in plants waxes.

                             RESULTS AND DISCUSSION

     Since 1968, the authors have been studying the behaviour of cyclodiene
insecticides under simulated atmospheric conditions.   It is known  that these
substances exist in the atmosphere, on the surface of plants  and in soil,
and that under these conditions UV-irradiation may lead to structural con-
formations.  The purpose of the study reported here was to investigate the
mechanisms by which compounds like  aldrin, dieldrin and photodieldrin—chosen
as model substances—chemically change under simulated atmospheric conditions,
and to identify the products resulting from such changes.  Various photore-
action types of cyclodiene insecticides are  known (1).  Two of these reactions
are shown in Figures la and Ib.
                                                            •no-Zo-reaction
 Figure l(a).  The i\a-2a reaction of cyclodiene insecticides.
                                                            (photoreversible
                                                             hydrogen  shift)
 Figure  l(b).   The  photoreversible  hydrogen  shift  of  cyclodiene  insecticides.

     The TTcr-2a-reaction:  The cyclodiene insecticides contain a chlorinated
double bond, which can be excited by long-wave UV-light (300 nm).  By inter-
action with the methylene bridge in the non-chlorinated part of the molecule,
the corresponding photoisomeric product can be formed by a ira-2a-reaction.
The excited double bond abstracts the opposite H-atom, whereby a new 2a-bond
is formed.

     The photoreversible hydrogen shift:  It could be shown that, during
                                     730

-------
sensitized irradiation of chlordane derivatives, intramolecular reversible
hydrogen shifts are possible as well  as the Tia-2a-reactions  typical  for this
class of substances.   Irradiation of  chlordane derivatives  at low temperatures
showed that below -30°C,  the hydrogen transfer isomers  are  the only  photopro-
ducts.  The parent compounds are obtained also by irradiation of hydrogen
transfer isomers.

     Dechlorination reaction:   In protonated solvents as well as in  the solid
phase, cyclodiene insecticides are photochemically dechlorinated at  the chlori-
nated double bond.  It may be  assumed that the dechlorination reactions pro-
ceed from the singlet state of the molecule.

     Reactions with reactive oxidizing agents:  It is known  that cyclodiene
insecticides in the atmosphere are found after their application (gaseous
or adsorbed to aerosols).  These substances change their light absorption  char-
acteristics under adsorbed conditions; for instance, the adsorption  maximum
of photodieldrin (195 nm in hexane) under adsorbed conditions on silicon dioxide
(Si02) moves to longer wavelengths (260 nm).  This adsorption shift  reveals
that even photodieldrin in an  adsorbed form can be excited  by wavelengths
even higher than 260 nm,  so that a possible reaction under  the photochemical
conditions of the lower atmosphere (with wavelength x>290 nm) cannot be
excluded.

     Photoreactions in the atmosphere further depend on altitude (because  of
the increasing radiation intensity of the short-wave UV-irradiation  at higher
altitudes and the presence of other reactive substances).  Due to the vast
number of possible reactions in the atmosphere and the  lack of experimental
information, at the present time it is impossible to estimate the rates of the
individual steps.

     Table 1 shows a compendium of photoreactions of aldrin (2).-

      Non-sensitized irradiation of solvents of  dieldrin  (protonated solvents)
with wavelengths  above 290 nm yielded mainly  dechlorination  products.  On the
other hand, the  isomeric product was found in sensitized reaction.  From the
photochemical standpoint dieldrin  is inert in comparison to  aldrin,  and dech-
lorination and isomerization  reactions proceed  more  slowly.

      Nitrogen dioxide  (N02) is a prominent inorganic air pollutant because of
its generation resulting from combustion of fossil  fuels.  Its'  possible role,
therefore, in atmospheric photochemistry of chlorinated  hydrocarbons should
be considered.

      The  compounds in  Figure  2 were  isolated  from solutions  of  dieldrin and
nitrogen  tetroxide in  fluorocarbon after irradiation with UV-light having
long wavelengths  (\>300  nm) (3).   Irradiation of dieldrin and nitrogen te-
troxide  in carbon  tetrachloride under similar conditions yielded the same
compounds.  The  OH-compounds were  detected after irradiation of a solution of
dieldrin  and  ozone (03)  in fluorocarbon.   In  the presence of N02, dieldrin
distributed in nitrogen  in the gaseous state  was converted mainly into photo-
dieldrin  by UV-light.  Irradiation of dieldrin  gaseously distributed in ozoni-
zed air yielded,  as in ozonic solution, several products of  higher molecular

                                     731

-------
weight, which, however,  were  not  isolated.
                   TABLE 1.  PHOTOREACTIONS OF ALDRIN

Phase


Film on
glass
Time of
irradi-
ation
2 hrs.

Type
of
lamp
HPK
125
Wave- Conversion
length
nm
290 74

Photo-
products
(Yield)
I (41)
II (5)
                                                        III (2)
                                                        V  (1)
                                                        Polymer
                                                        (not ider>-
                                                        tified) (25)
      Film on 1  month
      glass
Sun-     290
light
97
I  (24)
II (4)
III  (9)
Polymer (_60)
Adsorbed 16 hrs.
on Si00
L.

HPK
125 290


I (4)
84 II (55)
III (2)
IV (4)

            Aldrin
             III
                                          IV
                                   V
                                    732

-------
                     0
                                     ON02
                                                                          0
     Figure  2.   Photochemical  reactions  of  Dieldrin  in  the presence of
     During irradiation of dieldrin in solution,  adsorbed or in the  gas
phase, photodieldrin is produced.   Dieldrin was  chosen for irradiation  in
order to determine the conditions  which cause a  complete decomposition  (mineral-
ization) because it is considered  to be one of the most persistent chemicals.

     Tests in the gas phase, however, revealed that isomeric photodieldrin  is
produced in large yield.  This can be explained  by a much longer life time  of
the electronically excited state of the dieldrin  molecule in the gas phase.
These conditions produce a large yield for compounds formed by the more
favorable bridging reaction in comparison to the  bimolecular reactions with
oxygen species.   In presence of 03 a hydroxy product has been formed in  smaller
quantities.  Irradiation of adsorbed dieldrin in  an oxygen stream at wavelengths
below 300 nm revealed that dieldrin is almost quantitatively degradated  to
carbon dioxide (C02) and hydrochloric acid (HC1).  Small amounts of photodiel-
drin were found as well.  Furthermore, we have found that ethanol and n-
hexane also, adsorbed on silica gel, undergo photomineralization with wave-
length above 300 nm to give C02.

     In addition to our investigation with cyclodiene insecticides,  we  have
investigated the behaviour of hexachlorobenzene,  pentachlorobenzene, pentachlo-
rophenol, 1,1,l-trichloro-2, 2-bis (p-chlorophenyljethane (DDT), 1,l-dichloro-2,
2-bis(p. chlorophenyl)ethylene (DDE), 2,4,5,2',4',5'-hexachlorobiphenyl  and
2,5,2',5'-tetrachlorobiphenyl in the presence of  oxygen with wavelengths 230
nm (quartz glass) as well  as with  wavelengths 290 nm (pyrex glass) (4).   A
higher conversion rate was found for substances  adsorbed on particulate  matter
than for substances deposited on a glass surface  as solids or as a thin  film
(see Tables 2 and 3).
                                    733

-------





^>"
C3^
LU
o;
i—
CO

^
UJ
CU
>-
X
o

I?l

1 — 1
1—
o
1 — 1
CO
o
LIJ
X
1 — I
h-

ID

U_
o

CO
I—
_J
	 J
CO
LU
rv*


•
CM

LU
	 ]
CO

h-




















































































CO
+J CM
0 i—
=3 0
T3
O
C,
CL.

E
O r—
•r- C_J
+-> 1C
rO
rvl
• i —
i—~
ro
S-
CD
E CM
•i- O
s: cj

-a
, 	
o
CO r—
rd
ID •!-
CD S-
i- CD
CD •+->
> ro
o 2:
O
CD
r*^


-o
'O r—
CO rd
•r-
r- S-
ro (D
•r- +->
•»-> ro
-.- s:
1 — 1




E
O co
•r— E
+-> O
ro •!-
•r- +->

ro "O
S- E
S- O
i— i c_>










•^3
c~
23
O
CL

O
CJ
•
-a
•
E




cn
E

cn
i —





cn
E

CO
LO





cn
E

to
CO









cn


CO



	 ^
CO
>^
ro
-a

I--.
S 	 '

N

S-
rd

o-

,
-o

E





•
•a

E






.
-o

E





cn
E

CO










cn
E

O




, — *
CO
>>
ro
-a

p^
N — ^

>^
CD
S-
>_
Q-




1
Ln
n
_ |
^3- O
" S-
- O i—
CM i — >,
** r~ E
LO O CD
" rd -E
•* X Q.
« CD T-
CM 3T -Q

•
-a
•
E





•
-a
•
E






•
-o
•
E





CO
E

cn
cn









cn
E

, —
o
•~


*• 	 v
CO
>^
03


p-s^,
•s 	 -*

X
CD
S^
>^
Q.






1
o
1 S-
- 0
LO i — i —
•-c >,
- CJ E
CM ro CD
" S- -E
LO -4-> D-
«> CU -i—
CM 1— -Q
cn
E

CO
1 —




cn
E

CO
1 —





CD
E

ID
^1~





cn
E

CO
CM









cn
E

CO
CO


, 	 „
CO
>^
rd i
-a

CM
* — ^

M
| ^
S-
fO
3
cr

•
T3
•
E





•
•a
•
E






•
-a
•
E





cn
E

1 	
00









cn
E

CM
CO



, 	
CO
>^
rO
-o

r^.
^ — ^

X
CD
S-
>^
Cu







i
0
i-
0
i— CD
-E E
CJ CD
ro N
X E
CD CD
rc .a

•
•a
. •
E





.
"O
•
E






•
-a
•
E





cn
E

CM
<£>









cn
E

"=3-
IO



, — ,
CO
>^
ra
T3

r^s
* — ^

X
cu
S-
>^
Q.






1
O
S-
o
^~
-E CD
O E
ro CD
-(-> N
E E
CU CD
f~\ r~>

•
"O
•
E




cn
E

{&






cn
E

CO
•—





cn
E

cn
to









cn
E

O
co



. — .
CO
>»)
rd
•^l

r —
^ — ^

X
O)
S-
>^
d.






1
o
S-
0
1 —
-E
O r—
ro O
•fJ C
E CD
CD JZ
D- Q-

•
-o
•
E




cn
E

CM






cn
E

CM
r~





cn
E

cn
co









cn
E

«^j-
LO



^ 	 X
CO
>x
ro
T3

r^x.
- — *•

X
CD
S-
>x.
CU















h-
a
Q

•
T3
•
E




cn
E

00






cn
E

0
<—





cn
E

LO
00









cn
E

00
cn



* — x
CO
>}
ro
"O

1^.
* — *

X.
ai
^_
>j
Cu















LU
Q
Q

































































•o
CD
+J
CJ
CU
I 1
T— J
CU
-a

4-^
O
E

II

•
-a
•
c
734

-------
















a
UJ
CO
a:
O
GO
Q
<=c

^t
(_}
1 — 1
h-
o
I— i
03
O

UJ
X

CJ
h-H
h-

s:
o
o;
*z£ 	 i
LU
U- CD
O

^y f ^
O i— i
1—1 	 1
I— 1— 1
<; GO
i — i
a z:
 S-
 >X
•*s rO< — *
•o -ox
0 X
s- r-^ E
Q- 0
0 S- -i-
+J CU 4->
O 4-> fd
JZ 4- -r-
o fo "O



,. — ^
X
X
en x
E E
•r- f-, O
E T-
•r- S- 4->
rd CU rd
E ~f~* '• —
CU 4 — "O
S- ro ro
S-
>>T3 S-
+J CU -r-
•i- -O
4-> 1- -
E O CO
rd CO >,
3 "O rd
cr rd -a



^ 	 V
X
en x
E E
•i- "=f O
E •!-
•i- S- 4->
fd cu rd
t — 1) ,( 	
CU 4- "O
S- fd rd
t-
>>"O S-
4-> CD T-
•i- JD
4-> S- -
E O oo
rd CO >>
3 -O fd
CT rd "O



>•>
4_>
•i —
4_>
c~
fd
3.^ —
crx
r— O)
rd -Q
•r- S-
4-> 0
•r- CO
E TD
i — i rd







-o
E
^
O
a.
E
O
CJ











0)

o
c~















f —
en

CM
t —
V 	

en
E

CM
i —












* — ^
en

in
CM
• — *

en
E

^j-
C\J








. —
en

O
0
^2,

en
E

CO
o
^—



1
o
s-
0
1 —
c~
O r—
rd O
4-> E
E CU
CU -E
a. Q-











ai
E
o
c~















*—^
en

vo
co
V 	

en
E

tn
LO
CM











, 	 .
en

i — ^
r^
N 	

en
E

CO
cr>
CM







, —
en

O
0
^

en
E

UD
co
r—












i—
Q
Q

a>
CU E
E 0
O E
E CU
0) -E
JC Q-
CX 0
O N
N E
E CU

-Q 0
O S-
1~ 0
0 r-
i — en jz en
-E E o E
O -r-
•c— co S- r^~
Q CO |—








,*— «•*
en

cr>
r—
S 	

en
E

0"}

4-3 *r~
^ TD
r- C
0 O
CO CJ

"O CU
i- CO
rd O>
T3 -E
E 4->
rd
4-> S-
fcO CU
-o
-E E
-M a
• 1 —
S -0
cu
C C 4->
O O 3
co -i- i —
•r- 4-> CU
S- CJ
fd E 4->
O-M- O
j~- JJJ p-
r- OX
cu cj cu cu
en i s-
"o i> rd
rd E ^D
CJ rd -E
•i- CU O
r— f — *. c~ -i —
•i- CJ 4J JZ •
CO Q 3. "O
LU 4- a>
Cn 	 O "T3
0) 3
O >> co co i —
O J^ 4-> -i- CJ
r— Q. E S- X
rd CU rd CU
E S- E
O en cu oo cu
OS- -O -Q
T3 4-> 3 C
QJ rd CO 3 4->
JQ E ro O O
s- o cu a. E
0 S- E E
CO .E O E
•a o >, o ro
rd CO -Q CJ
ro S-
oo en co rd ««
rd ro r— "O
54- o CU
O i— O-4-5
O) i—  a>
E E 2 S- 4->
rd ro O O)
4-> O) oo E "O
CO E rO
_Q 4-) 4_)
3 >, E rd O
00 JD O -E E
•r— 4-J
CD oo 4-3 rx>
.E fd ro >> S-
4-> S S- 4-> ro
4-> *i —
,
CO O O) -i-i—
ro -r- O J3 4->
O 4-> E •<- E
rd O 00 CD
-E 4-> O 00 3
O -r- O CT
ro 4-> E CL CU
CU E 2 CO
ro O CU E
E 3 E -E 0
i — i Cf -^ \ — O


f — ,
^ — * X
— X X
XX X
735

-------
     These differences can be attributed to bathochromic shift,  i.e., changes
in the relative extinction or appearance of new absorption bands as a conse-
quence of the adsorption on silica gel, as well as to the greater dispersion
of environmental chemicals/molecules in the adsorbed phase resulting in a higher
environmental chemicals/oxygen contact.  The disappearance of the reactant
substances (see Table 3) cannot be explained by vaporization, nor is it due
to the formation of photoproducts.  Since C02 and HC1 are formed during
"pyrex"-irradiation of pentachlorophenol, DDT, and DDE as solids on glass
(see Table 2) and also the UV-absorption bands of these compounds are located
nearly in the same region as those of the substances adsorbed on silica gel,
we deduce that this is a case of mineralization.


                                    REFERENCES

 1.    PARLAR,  H.   Photochemical reactions  of  Chlordane  Derivatives.   Dissertation.
      University  of Bonn,  1971.

 2.    GAB, S.  Unpublished.

 3.    NAGL, H.  G.   Reactions  of Dieldrin  with  Nitrogentetroxide  in  solutions.
      Dissertation.   University of Bonn,  1971.

 4.    GAB, S.,  H.  Parlar,  S.  Nitz  and  F.  Korte.   Unpublished.
                                     736

-------
                                                                           15-2
                PHOTOOXIDATION OF THE TOLUENE-N02-02-N2 SYSTEM
                            IN A SMALL SMOG CHAMBER

         H.  Akimoto,  M.  Hoshino, G. Inoue, M. Okuda, and N. Washida*

ABSTRACT
     Photooxidation  o&  the. io£uene-N02-02-N2 &y&te.m hat, be.an £tudie.d -in a 67-
tite.fi -imog chamber,  witk find-Lotion p/iovide.d by a faltL OAC o& a high piewuAe
Xenon Lamp through a Pyfie.x window.   lnAadA.aXA.on ofi ain. mixtuAtA  containing
34 ppm tolu.e,ne.  and  10-20 ppm nitwge.n dioxide. yie.l.d&d ai, Abaction  pAodactb
be.nzatde.hyde., o-cAeAot  (the. m- and p-iAomeAt> wteAe £e^4 than a ^  peAce.nt
ofc the. o-JJ>omoJi} , benzyl ntiyiate.,  m-YiWio toluene., and mWiocAnAolA,  When
nl&iouA acid waA addzd  oi a 4ou/ice o^ additional hydAoxyl. gtioupA,  the.  m-
ni&Lotolu.ol& yieJLdA,  fan. the.
&ame. tnitiat c.onc.e.n&iation ofa nitsioge.n dioxide..  Product yieJLdA  have, been
        ous function* oft .te-actant conc^nttation-d and the. fraction mec.ha.n-um


                                 INTRODUCTION

     The  photooxidation of aromatic hydrocarbons has been the subject  of sev-
eral studies  (1-3)  in which the aromatic hydrocarbon/N02/air reaction  system
was used  as a model  of  urban atmospheric reaction systems.  However, in most
of these  studies, the emphasis was on the so-called "photochemical reactivity"
property  of the aromatics; investigations of reaction products characteristic
of the aromatic reactant have rarely been carried out.

     The  purpose of  this study is  to elucidate the reaction mechanism  of the
photooxidation  of the toluene-N02-02-N2 system in the concentration range  of
34 ppm of toluene,  10-310 ppm of nitrogen dioxide (N02) in nitrogen (N2)
and/or oxygen (02) at 1  atm.


                                 EXPERIMENTAL

     The  reaction chamber was a 1660 x 240 mm (inside diameter)  Pyrex  tube of
67 liter  capacity.   It  was evacuable to less than 1 x 10"6 torr.   Each  end of
the reaction chamber was sealed with a Pyrex window 20 mm thick, through which
a photolyzing light  beam was  transmitted into the gas sample.  The light
source was a 500 W xenon short arc lamp, and a parallel light beam of  200  mm
was obtained with an elliptical  mirror, a lens , and an off -axis parabolic
mirror.   Figure 1 shows  the schematic diagram of the small  smog  chamber used
in the present  study.
*National  Institute  for  Environmental  Studies, Ibaraki, Japan.

                                     737

-------
                                                                      GC or
                                                                        GC/MS
                  C : REACTION CHAMBER
                  G : Ti GETTER PUMP
                  R : ROTARY PUMP
                  U : CONCENTRATION SPIRAL TRAP
XE I\RC LAMP HOUSE
SORPTION TUMP
GAS BURETTE
ION PUMP
FORELINE TRAP
SAMPLING FLASK
         Figure  1.   Schematic diagram of the small  smog  chamber system.


     The reaction  mixture was first sampled into  a constant volume glass
sampling bulb (about 700 cc) and then concentrated in  a gas chromatograph  (GC)
sampling tube.   The sampling tube was a Pyrex spiral of 2 mm inside diameter
and about 4 m extended length.   The concentration  of the reaction mixture was
achieved by evacuating the air and N02 while cooling the tube with ethanol-
liquid  N2 at -60°C.   After the concentration, the  sampling tube was heated
with hot water  and the sample was fed directly  into a  gas chromatograph.  This
procedure was found to minimize the thermal reaction of N02 with toluene and
cresols  in the  sampling tube.

     Product  identification was  performed with a gas chromatograph mass spec-
trometer  (JEOL,  JMS-D100),  using a 3 m GC column of SE-30 on  Shimalite W.   The
GC oven  temperature  was  raised  from 80°C to 200°C  at a  rate of 4°C/min.
Quantitative  GC  analysis  was carried out using the same column as  above.


                            RESULTS AND DISCUSSION

     In  the  present  study,  the  wavelength of the irradiation  light is limited
to about  350  nm  and  longer.  In  this spectral region, the phptopxidation of
the toluene-N02-02-N2  system is  initiated by the photodissociation  of N02.
                                      738

-------
         N02 + hv + NO + 0.                                            (1)

The radiation intensity gave a photodissociation rate coefficient for NO? of
1.1 x ID-3 sec-i_

     Typical QC patterns of the products in the toluene-N02-N2 and toluene-
N02-air systems are shown in Figure 2.  In the reaction system of toluene (34
ppm)-N02(ll - 207 ppm)-02/N2 (1 atm), benzaldehyde, o-cresol (other isomers
are less than a few percent of o-cresol), benzylnitrate, m-nitrotoluene,  and
nitrocresol (mainly 6-nitro-o-cresol) are formed as reaction products.   In the
presence of enough 02, such as in the air atmosphere, oxygen atoms produced in
reaction (1) react mainly with 02 to form ozone (03), and in a small fraction
of time the remaining oxygen atoms react with toluene.   Since 03 and nitric
oxide (NO) react fast enough to regenerate N02 dissociated in reaction  (1),
the concentrations of N02 and NO reach their photoequilibrium levels within 20
minutes after the irradiation starts.   The photoequilibrium concentration of
N02 was found to be about half of the initial N02  concentration in the  air
atmosphere, and remained nearly constant during the entire irradiation.   Since
the N02 concentration remains constant,  the yield  of each product theoretic-
ally should increase linearly with irradiation time.   This was  confirmed
experimentally.

     The yield ratio of nitrocresol  to cresol increases  linearly with the
equilibrium concentration of N02, as shown in Figure 3.   But the yield  ratio
of benzaldehyde to the sum of cresol and nitrocresol  was found to be nearly
constant and independent of the equilibrium  concentration of N02, as shown  in
Figure  4.  Because of this independency,  the yield of other products was
normalized to that of the sum  of  cresol and nitrocresol.  Typical data of the
product yield are shown in Table  1.
  TABLE  }.    RELATIVE  PRODUCT  YIELDS  IN THE  PHOTO-OXIDATION OF THE TOLUENE-
          N02-02-N2 SYSTEM AS NORMALIZED TO THE SUM OF CRESOL AND NITROCRESOL

N02
(ppm)
180
180
50
50
10
2
HN02
(ppm)
_
-
-
-
-
15
Atmos-
phere
(1 atm.)
air
N2
air
N2
air
air
Benz-
aldehyde
0.18
-
0.17
-
0.19
0.47
Benzyl -
nitrate
0.057
-
0.044
-
0.019
_
a-Nitro-
toluene
_
0.14
-
0.081
-
_
m-Nitro-
toluene
0.025
0.029
0.056
0.039
0.071
0.28
                                     739

-------
    Figure  2.   Typical  GC  pattern of  the  products:  upper  curve  -  toluene
    34  ppm,  N02
    N?  1  atm.
54 ppm,  air 1 atm;  lower curve -  toluene 34 ppm,  N0? 54 ppm,
     In the reaction system of toluene (34 ppm)-N02 (50 ~ 310 ppm)~N2 (1 atm),
a-nitrotoluene is formed in addition to cresol, m-m'trotoluene, and nitrocresol,
but the production of benzaldehyde is completely suppressed.  The relative
yield of a-nitrotoluene to the sum of cresol  and nitrocresol increased with
the increase of the initial concentration of N02, as shown in Figure 5.   The
production of a-nitrotoluene suggests the presence of the benzyl radical and
implies that the hydrogen abstraction from the methyl  group of toluene by the
oxygen atom does occur.  Although Jones and Cvetanovic (4), Grovenstein and
Mosher (5), and Gaffney, et al. (6) did not detect any product originating
from primary hydrogen abstraction, it is possible that under their experimental
conditions, the benzyl radical may not have been chemically trapped.
                                    740

-------
             0.8
          s
          n
          33
          m 0.4
             0.2
  Figure 3.
  of N02.
                         D
                                        02
                                     O 160 Torr
                                     A 418
                                     D 760
                     o
                        20
                    40        60
                     N02  (ppm)
80
100
The ratio of nitrocresol to cresol vs. equilibrium concentration
     The  reaction  scheme  of the  toluene-N02-02/N2  system under our experimental
conditions may  be  summarized as:
                                                                 (2)
In Table 1 it is interesting to note that the relative yield of m-nitrotoluene
decrease with the increase of N02 concentration both in the 02-free and 07-
present systems.  This is also demonstrated in Figure 6.  The formation of tri-
nitrotoluene would be explained by the reaction initiated by hydroxyl  (OH)
radi cal'.
                                     741

-------
                                     HN 2           2            (J)

                                   CH3
The competition between reaction (3) and the following reactions

          OH + NO + M + HN02 + M                                        (4)

          OH + N02 + M -> HN03 + M                                       (5)

 would explain the decrease of the relative yield of m-nitrotoluene at higher
 N02 concentrations.  This assumption was supported by experiments on the
 toluene-HN02-NO -air system, where OH was formed in the photolysis of nitrous
 acid (HN02).

          NH02 + hv -*• NO + OH                                           (6)

 As shown in the last row of Table 1, in addition to cresol  and berizaldehyde, tri-
 nitrotoluene is a main product.

      According to the above reaction scheme, the formation  ratio of
 (a-nitrotoluene + benzylnitrate + benzaldehyde) to (cresol  + nitrocresol)
 should be constant in the toluene-N02-02/N2 system and is  equal to the branch-
 ing ratio of abstraction to addition in the reaction of oxygen atoms with
 toluene.  Figure 7 shows the ratio in the experiment for various initial
 concentrations.  Thus, the abstraction-to-addition ratio in the reaction of
 oxygen atoms with toluene is determined to be 0.22 +_ 0.02.   This value should
 be considered as the upper limit since the contribution of  the OH radical re-
 action may not be negligible.


                                  REFERENCES

 1.    A.  P.  Altshuller,  P.  W.  Leach,  Int.  J.  Air  Hater Poll.,  8:  37, 1964.

 2.    J.  M.  Heuss,  W.  A.  Glasson,  Environ.  Sci.  Tech.,  2: 1109,  1968.

 3.    S.  L.  Kopczynsky,  Int.  J.  Air Water Poll.,  8:  107,  1964.

 4.    G.  R.  H.  Jones,  R.  J.  Cvetanovic,  Can.  J.  Chem.,  39: 2444,  1961.

 5.    E.  Grovenstein Jr.,  A.  J.  Mosher,  J.  Amer.  Chem.  Soc.,  92:  3810,  1970.

 6.    J.  S.  Gaffney, R.  Atkinson,  J.  N.  Pitts Jr.,  J.  Amer.  Chem.  Soc.,
      98:  1828,  1976.


                                      742

-------
             I0-3

             Nl
             O
             rn
             o
             3J
             m

             8


             z

               0.2
- o
               o.i
                                     -i	1	1	1	1	r
                              02


                          o  160 Torr (Air)

                          A  418


                          D  760
                                 I     I     I	I	1	L.
                           20
                             60
                                                         80
                                     N02  (ppm)
100
Figure 4.   The ratio of benzaldehyde to the sum  of cresol and nitrocresol

vs. equilibrium concentration  of NO?.
             o 0.20

             z
             | 0.15


             m
             2
             m
             ^
             o

             2 0.10
               0.05
             IS
                       -J-	1
                                           TOLUENE  3 4 ppm


                                           N2       1 aim
                              100           200


                                  NO 2
                                    300
Figure  5.   The ratio  of a-nitrotoluene  to  the sum of cresol  and nitro-

cresol  vs.  initial  concentration of
                                       743

-------
\J.\JO
3
? 0.06
H
33
O
3
(—
m
m aCK
\
n
*"l
m
+
z
i? 0.02
I*
c
— 	 	 1 	 , 	 1 	 , — 	 	 , 	 , 	 , 	 , 	 , 	
_ 0
\ 02
h o\° D A7
\ 0 160
\ • 418
\ A 760
• V
\
\^
- D \.
^~^~^^
^^--~^^ o
o~~^^- — ___









) 20 40 60 80 100
N02 (ppm)
Figure 6.
cresol vs.
0.3
IT
z
H
+
00
z
roO.2
> ,
\
O
-v
+
?0.1
The ratio of m-nitrotoluene to the sum of cresol arid nitro
equilibrium concentration of N0? .
i i i i i i ii


A •
^ 	 Q 	 	 	 — 	 J 	
/ • N02/02
r A 2.2 x10~5
• 2.2
o 2.4
• 3.3
A 30.6
D 00
i i i i i i i i
200 400 600 800







0? (Torr)
Figure 7.   The ratio of (a-nitrotoluene + benzylnitrate  +  benzaldehyde)
to (cresol  + nitrocresol)  vs.  partial  pressure of 02-
                                   744

-------
                                                                           15-3
                THE CHEMISTRY OF NATURALLY EMITTED HYDROCARBONS

                        B.  W. Gay, Jr. and R. R. Arnts*
ABSTRACT
     To  bztteA. undeAAtand u.no4ua££t/ high fuuAaJL oxA.da.nt c.onc.e.ntA£utionA, the,
              a nambeA o{> natuA&tty mittad hydfiocaAbonA, mcunly
                ?hotoc.he.micat Atudi&> weAe. conducted on. &e.te.cte.d natuJwJL  hydsio-
         to  det.nn.mine. nateJ* ojj disappearance, abi£i,ty to piodace. owne., and
ne.a.ction. pioductA.   Re4u££s i>howe.d that ozone, formation is a. Sanction  o^ the.
sin^tictt  hydsioc.asibon-to-nsLtsiogen-oxA.deJ, natJio, that te.fipe.yiu n.&a.ct fiap^idty  with
ozone., and  that Qe.o(ju>  peJiox.ga.cet.yLvut.fwite. U> a common te.ipe.ne.
product.
                                 INTRODUCTION

     The emissions  and  reactivities of natural compounds have particular  sig-
nificance in  understanding unusually high rural oxidant concentrations  and  a
more obvious  phenomenon—the aerosol haze characteristic of rural areas like
the Smokey Mountains  in North Carolina or the Blue Ridge Mountains of Virginia.
Natural emissions emitted  into the atmosphere by various forms of vegetation
include oxygen(02), hydrogen(H2),  low molecular weight hydrocarbons  (methane,
acetylene, ethylene,  etc.), high  molecular weight hydrocarbons (olefins and
aromatics), alcohols,  aldehydes,  and organic acids.  Rasmussen et al.,  (1)  has
estimated that the  ratio by weight of natural to anthropogenic emissions  on a
worldwide basis  is  about 4 for unreactive hydrocarbons (biological processes
in swamps being  the largest source) and about 6 for reactive hydrocarbons
(biological processes  in forests)  (1).   Even though the combined worldwide
emissions of  natural  hydrocarbons  (HC) is immense, emissions from individual
trees are minute.

     The Environmental  Protection  Agency (EPA) is conducting and supporting
laboratory and field  studies to provide a data base on natural hydrocarbon
emissions, ambient  oxidant profiles, hydrocarbon reactivities, and other  vital
parameters necessary  to assess the influence of naturally emitted hydrocarbons
on the ambient air  quality of rural and urban areas.   The photochemistry  of
C}-C4 hydrocarbons  and  oxygenated  hydrocarbons, such as aldehydes, ketones,
and acids, has been previously examined with respect to atmospheric  chemistry
(2,3).

     The photochemical  research reported in this paper concerns a number  of
high molecular weight hydrocarbons consisting essentially of terpenes.  The
*Environmental  Protection  Agency,  Research Triangle Park, North Carolina.

                                      745

-------
main source of emitted terpenes is conifers.  Since forests of conifers cover
large areas of the world, terpenes possibly could play an important role in
the formation of rural oxidants.

     The natural hydrocarbons selected for study were isoprene,  rnyrcene, a-and
3-pinene, p-cymene, A-carene, d-limonene, and terpinolene.   All  of these
hydrocarbons, except isoprene, are terpenes.  Furthermore,  they  have been
identified in the atmospheres of loblolly forests during field studies.  To
assess rates of disappearance, ability to produce ozone, and reaction products,
each hydrocarbon was exposed to simulated ultraviolet (UV)  light in the presence
of nitrogen oxides (NO ).  The formation of ozone (03) and  other gaseous
products was monitoredxby a variety of analytical methods.


                                EXPERIMENTAL

SMOG CHAMBER STUDIES

     Smog-chamber-type irradiations of the natural hydrocarbons  in question
were conducted in 250s, 50.8 ym (2 mil) FEP Teflon bags.   These bags were
filled with hydrocarbon-free (<0.1 ppmC hydrocarbon) and NO^-free (<1 ppb)
air.  The calculated amounts of terpenes and nitric oxide (NO) necessary to
produce a given concentration (ppm) of each were injected into the air stream
with a syringe while the bag was being filled.   The rectangular  chamber in
which the Teflon bag was suspended between light banks was  thermostated to 25°
+_ 1°C during the irradiations.  For these experiments, 22 blacklamps (energy
maximum 3660 A) and 4 sunlamps (energy maximum 3160A) were  used.  The light
intensity, as measured by the photolysis of nitrogen dioxide (N02) in nitrogen
(4), was 0.46 min-1.   The natural hydrocarbons were monitored by gas chroma-
tography (GC) with the use of flame ionization detection.  The NOX and 03
were measured by chemiluminescence instruments.   Peroxyacetylnitrate (PAN) was
monitored by gas chromatography with the use of an electron-capture detector.
The oxygenated hydrocarbon products, such as acetaldehyde (CH3CHO) and acetone
((CH3)2CO) were monitored by gas chromatography with the use of  flame ioniza-
tion detection.  Wet chemical analysis spot checks were made for N02 and
formaldehyde (H2CO) with the use of the Saltzman (5) and chromatropic acid (6)
methods, respectively. The purity of the terpenes was determined by gas chroma-
tography, and when needed the terpenes were vacuum distilled to  obtain the
purest fraction.  All other compounds were research grade and were used with-
out further prepurification.

LONG PATH INFRARED STUDIES

     For the long-path infrared studies, the photochemical  chamber also served
as an infrared absorption cell for i_n situ measurements.  The chamber length
was 9.1 m and the diameter was 0.31 m.  The chamber was constructed of six
borosilicate pipe sections, with sample inlets between each section.  Plexi-
glas end plates and Teflon gaskets were used throughout.  The volume of the
chamber was 690x,, and it could be evacuated to less than one torr and operated
at 760 torr.  Surrounding each of the six sections were cylindrical  light
banks containing a total of 96 lamps.  In these experiments, 48 blacklamps and
24 sun lamps were spaced evenly around the length of the chamber.  The light


                                     746

-------
intensity, as measured by the photolysis of N02 in nitrogen, was 0.51 min"1.
Samples were mixed and introduced into the chamber via a glass manifold.
     The chamber contained an eight-mirror optical system to achieve long path
lengths.  An optical path of 357 m was used.   A rapid-scan Fourier Transform
Spectrometer that uses a Michel son interferometer, germanism-coated potassium
bromide (KBr) beam splitter, and liquid-nitrogen-cooled detectors covered the
spectral region of 700-3200 cm'1.  A more detailed discussion of the system
and techniques used is given elsewhere (7).
                                   RESULTS
SMOG CHAMBER EXPERIMENTS
     The photooxidation of the natural hydrocarbons in Teflon bags was con-
ducted at HC/NOv ratios of approximately 3 and approximately 20 (ppm compound/
ppm NOX).  The concentration of NO was kept constant at 0.33 ppm,  and the
hydrocarbon concentration was varied.  Results for the HC/NOX ratio of approxi-
mately 3 are listed in Table 1, in decreasing order of ozone formation.   Also
shown is the percent of hydrocarbon reacted at 60 minutes, and the time to
reach one-half of the initial hydrocarbon concentration.   In all cases, the
oxidant maximum was reached before the end of the irradiation (200 min).   The
photooxidation of propylene, a commonly run compound used for smog-chamber
validation, is included. The amounts of 03 produced from cyclic terpenes
ranged from 240 to 308 ppb, while the open-chain terpene myrcene produced 322
ppb; 556 and 642 ppb were produced from propylene andjisoprene.
            TABLE  1.   HYDROCARBON  REACTIVITY  AND OZONE FORMATION


Compound



Ratio HC/NOX % HC reacted
in 60 min.
Time (min)
to reach
(HC)i

2


Max 03
(ppb)
 Isoprene           3.2
 Propylene          3.0
 P-cymene           3.2
 Myrcene            3,2
 D-limonene         2.7
 A-carene           5.3
 a-pinene           4.0
 3-pinene           3.1
 Terpinolene        2.8
72
50
27
89
96
71
78
12
98
 46
 60
128
 21
 40
 33
 41
114
 34
642
556
450
322
308
274
271
255
240
                                    747

-------
     Irradiations carried out at a HC/NOX ratio of approximately  20 are shown
in Table 2.  The cyclic and open-chain terpenes at this  ratio are again lower
in 63 production when compared to p-cymene and isoprene.
             TABLE 2.  HYDROCARBON REACTIVITY AND OZONE FORMATION
 Compound
Ratio HC/NOX
% HC Reacted in
    60 min.
Max 03 (ppb)
P-cymene
Isoprene
3-pinene
;<-pinene
D-l imonene
Terpinolene
Myrcene
20.8
19.8
20.3
19.5
20.6
21 .3
20.3
7
34
21
25
33
34
27
495
257
125
43
30
15
1.5

     The results of varying the HC/NO^ ratio while keeping (NO) = 0.33 ppm for
the compound d-1imonene are shown in table 3.
LONG PATH INFRARED EXPERIMENTS

     Irradiations in the long-path infrared cell  were conducted at a HC/N02
ratio of 6.2-6.7 to determine the gas phase products resulting in the photo-
oxidation.   There is a depletion of the reactants during the photooxidation. A
good material balance was not obtained.  This fact is due to the inability of
the infrared instruments to monitor aerosols suspended in the gas phase.


     When isoprene (8 ppm) and N0? (1.3 ppm) were irradiated in the chamber,
at 60 minutes, 68 percent of the isoprene had reacted.   The major products
observed were H?CO, CH3CHO, carbon monoxide (CO), carbon dioxide (C02), PAN,
and formic acid (HCODH).  Table 4 lists the percent of hydrocarbon reacted and
products observed at 60 minutes.  In the irradiation of isoprene, no strong
bands or moderately absorbing bands were unidentified.
                                     748

-------
     TABLE 3.  EFFECT OF VARYING LIMONENE/NOy RATIO ON MAX 03 FORMATION

HC/NOX Ratio
0.14
0.45
0.63
1.13
2.25
2.70
4.53
4.93
20.6
% HC reacted in Tlm£
60 min. reac
71
63
50
83
95
96
88
90
33
; (min) to
:h (HC)i
2
38
48
60
48
43
40
35
31
NA
Max 0 (ppb)
295
251
364
430
410
320
219
204
30


TABLE 4. REACTIVITY AND
LONG PATH INFRARED AT
PRODUCTS BY
60 MIN.


Compound
3-pinene
y\-carene
Isoprene
ex-pi nene
Myrcene
D-l imonene
Terpinolene
HC/NOX % HC H2CO HCOOH
ratio reacted
5.3 58 0.94 0.39
6.7 60 0.46 0.39
6.2 68 2.39 0.44
6.7 72 0.33 0.32
6.2 73 0.84 0.16
6.7 91 0.9 0.35
6.2 95 1.2 0.31
CO C02
0.23 0.4
0.50 0.9
0.77 0.2
0.39 0.6
0.26 0.24
0.46 0.49
0.49 0.69
CH PAN (CH3)2
CHO CO
0.05 0.018 0
0.33 0.26 0
2.0 0.32 0
1.57 0.20 0
2.35 0.70 0.76
0.23 0.24 0
0.35 0.28 0

Initial hydrocarbon cone. =
Products as ppm compound
ppm
                                    749

-------
      Each  of the other irradiated
 Unidentified absorption bands for
compounds had unidentified absorption bands.
the 3-pinene photooxidation were a strong
 absorber at  1285 cm"1 and two moderately weak bands at 1650 and 1740 cm l.
 Other unidentified bands were:  a-pinene, with moderate absorbing bands at
 1230, 1280,  1375, 1650, and 2890 cm'1; A-carene, with moderate bands at 855,
 1375, and  1650 cm'1; myrcene, with a strong band at 1842 cm'1 (acetone was
 identified as a product in this reaction);  d-limonene, with weak bands at
 855, 1630, and 1650 cm"1; and terpinolene, with weak bands at 855, 1290, 1375,
 and 1630 cm-1.  The contribution to the material balance of these systems, if
 the chemical composition of the unidentified absorption bands were known, is
 minimal.
                                 DISCUSSION

     The carbon and nitrogen balance in all of the irradiations was poor.  In
the irradiation of d-limonene with the use of GC and wet-chemical methods, 3-
20 percent of the initial carbon could be accounted for as observed products
after the 03 maximum had been reached.  Infrared studies could account for
only 4 percent of the carbon.  The nitrogen balance in the Teflon bags ac-
counted for 36-50 percent of the initial nitrogen while infrared studies could
account for only 20 percent of the nitrogen.   Results were similar for the
other terpenes. JThe formation of aerosols when terpenes and NOX are irradiated
with UV light is  well  known (8).   However,  the mechanisms of formation and
chemical  composition of the aerosols  are largely unknown.   It is  apparent that
much of the carbon and nitrogen not accounted for is present in the aerosols.

     A comparison of the data in Table 1 and  2 shows cyclic and open-chain
terpenes  lower in 03 formation than p-cymene  and isoprene.   If a  comparision
is made of ozonolysis  rates (9) to the decreasing order cf 03 maximum in Table
2, the compounds  are arranged in increasing reaction rates from top to bottom.
The only  exception is  terpinolene, whose ozonolysis rate is faster than
myrcene.  The apparent  low 03 formation of the terpene at higher HC/NOX ratios
seems  to  be due to the rapid reaction of newly formed 03 by reaction with the
terpenes.

     Table 3 shows the results of maximum 03  formation when the HC/NOX ratio
is varied for d-limonene.   A dramatic maximum when plotted in formation is
seen at a ratio of about 1.7.   At this ratio  or greater, the 03 and d-limonene
reaction  is faster than 03 formation  processes.


                                 CONCLUSION

     When terpenes and NOX are irradiated with UV, aerosols are formed that
contain most of the initial carbon and nitrogen of the system.  Little is
known  about the formation mechanisms  and chemical composition of the aerosols.
Gaseous products  observed are aldehydes, CO,  C02, HCOOH, (CH3)aCO, some com-
pounds whose infrared  spectra are unidentified,  and, in all the terpenes
                                     750

-------
studied, peroxyacetylnitrate.  There is an optimum HC/NOX ratio that results
in a maximum formation of 03.  This optimum ratio is much lower than observed
in the ambient rural air.  Terpenes do react rapidly with 03.

     On the basis of present-day knowledge and field studies, which show very
low concentrations of naturally emitted hydrocarbons, it is unlikely that nat-
urally emitted hydrocarbons could be responsible for any significant fraction
of oxidant formed in the urban areas.  In rural areas, these hydrocarbons par-
ticipate in aerosol or haze formation, while the significance of oxidant for-
mation is probably also small.  More experimental field and laboratory work is
necessary before final conclusions can be reached, due to the extremely high
HC/NO^ ratio observed in rural areas.  This ratio is far too large to produce
any significant 03 levels.  All of the 03 produced would necessarily have to
react with the excess terpenes available.

                                ACKNOWLEDGEMENT

      The authors wish to express  their appreciation to Mr.  Theodore Winfield
 for his assistance in the smog chamber experiments.

                                 REFERENCES

1.   Rasmussen,  K.,  M.  Taheri,  and  R.  Kabel.   Sources and Natural  Removal
    Processes  for Atmospheric Pollutants.   Center for Air Environmental
    Studies,  Univ.  Park,  Pennsylvania,  1974.

2.   Leighton,  P.  A.   Photochemical  Aspects  of Air Pollution.  Chapter III,
    X.   Academic  Press,  New York,  1961.

3.   Altshuller, A.  P.,  and J.  J.  Bufalini.   Photochemical  Aspects of Air
    Pollution:  A Review.   Photochem.   Photobiol.,  4: 97, 1965.

4.   Tuesday,  C.  S.   The Atmospheric Photooxidation of Nitric Oxide and Trans-
    Butene-2.   In:   Chemical  Reactions  in  the Lower and Upper  Atmosphere,
    Interscience  Press,  New York,  1961.   pp.  15-49.

5.   Saltzman,  B.  E.   Colorimetric  Microdetermination of Nitrogen  Dioxide  in
    the Atmosphere.   Anal.  Chem.,  25:   1949,  1954.

6.   Altshuller, A.  P.,  D.  L.  Miller,  and S.  F.  Sleva.  Determination of
    Formaldehyde  in  Gas  Mixtures  by the Chromotropic Acid Method.  Anal.
    Chem.,  33:  621,  1961.

7.   Hanst,  P.  L., A.  S.  Lefohn, and B.  W.  Gay.   Detection of Atmospheric
    Pollutants  at Parts-per-Billion Levels  by Infrared Spectroscopy.  Appl.
    Spectres.,  22:   188,  1973.

8.   Went,  F.  W.   Blue Hazes in the Atmosphere.   Nature, 187:  671, 1960.

9.   Grimsrud,  E.  P.,  H.  H.  Westberg,  and R.  A.  Rasmussen.  Atmospheric Reactiv-
    ity of Monoterpene Hydrocarbons,  NOX Photooxidation and Ozonolysis.   Inter.
    J.  Chem.  Kinetics,  Sym.  1, 1975.   pp.  183.

                                     751

-------
                                                                             15-4
                      MEASUREMENT OF PHOTONS INVOLVED  IN
                       PHOTOCHEMICAL OXIDANT FORMATION

               D.  H.  Stedman, R. B.  Harvey, and  R.  R.  Dickerson*
ABSTRACT

     We have. con6ideA.e.d a pkototy&iA ofa the.  mol.e.cu£eA ni&toge.n dioxide., ozone.,
VU&LOUA acid,  faoftmai.de.hyde., hydrogen pe.tioxide.,  and nitric acid, which OAe.
important to ain pottution modeJLtA.ng.  E-t>timatej>  ofa thoJJi phototy^iA ftateJ>
u)e.tte. made, fatiom known paA.ame.te.u.  Expe.ftime.ntal.  te.chniqu.eA to dateAmine, th&Ae.
.i.atej> have, been deviled and tested faon. nitn.oge.n dioxide.,  nitA.ouA oxide., and
ozone..  Exte.n&ive. fizAulttA faofi ni^&toge.n dioxide,  and -borne. p?ie.£iminatty data on
ozone, afie. ptie.Ae.nte.d.

                                 INTRODUCTION

     Photochemical  smog is intrinsically dependent  on  the solar flux.   The
rate of photolysis  (J)  of any pollutant present in  small  concentrations is
given by  the formula:

              ,00
          J =     I(X) $(X,T)  e(x,T)dX                                    (1)
where I(x) is the  time and wavelength-dependent solar  flux,  $(X,T) is the
temperature and wavelength-dependent dissociation quantum yield, and e(x,T)
is the photon absorption cross-section.  While in principal  all  the above
parameters can be  measured a priori and the result  can  be calculated for any
given model, there are for some molecules significant  uncertainties in the
appropriate structured integral over x.  Thus, it is best to measure any im-
portant rates directly to compare with theory.


                             MOLECULES CONSIDERED

     Several molecules that have been identified as being of importance in
photochemical smog are listed in Table 1.  The theoretical  rate  of photolysis
averaged over 24 hours and one year at 30°N is given for  each molecule.

     We have determined an experimental technique whereby each of these rates
may be determined  in  situ using flow actinometers.  The methods  are summarized
in Table 2.
*The University  of Michigan, Ann Arbor, Michigan

                                      753

-------
   TABLE 1.  THEORETICAL RATES OF PHOTOLYSIS (AVERAGED OVER 24 HOURS AND
             ONE YEAR AT 30° N) FOR IMPORTANT PHOTOCHEMICAL SMOG MOLECULES
       Reaction
Javg sec"
Ref.
N02 H
03 +
CH20
HONO
H202
HN03
h hv -> NO
hv -> 0 D
+ hv> H +
+ hv-> OH
+ hv-> 20H
+ hv-> OH
+ 0
+ 02
CHO
+ NO

+ N02
3.
5.
9.
3.
9.
5.
1 x 10"3
1 x 10"6
4 x 10"6
1 x 10"4
0 x 10"7
5 x 10~7
Hampson
De More
Calvert
Levy
Schumb
Levy

   ,ABLE  2.    EXPERIMENTAL  TECHNIQUES  FOR MEASURING THE  RATES  OF  PHOTOLYSIS
              OF  SELECTED MOLECULES  IJ SITU WITH  FLOWING ACTINOMETERS

Molecule
N02
03
CH20
HONO
H202
HN03
Flowing
Gas
Air/N02
N20/0
N?/CH?0
HC1/HONO
HCL/H202
Air/HN03
Method
Measure NO by chemi lumi nescence.
Measure NO by chemi luminescence.
Measure H by resonance fluorescence.
Measure NO by chemiluminescence.
Measure Cl by resonance fluorescence.
Measure NO by chemiluminescence with converter.

                   DETAILS OF THE NITROGEN DIOXIDE STUDIES

     Nitrogen dioxide (N02) is the most important photochemically active mole-
cule in photochemical smog (1, 2, 3).   Therefore, we have worked on measurement
of J(N02) for some time.   In the earlier studies, Jackson et al.  (4) prepared
the first flow actinometer.  In brief,  Jackson measured Jj by exposing a 1 ppm
N02 flow for four seconds to direct sunlight, through an ultraviolet (UV)

                                     754

-------
transparent quartz tube.  The resulting production of [NO] in the air stream
gave an indication of Jx through the kinetics of the following reaction:
         d[NO]/dt = Ji[N02]                                            (2)

               i  -   ANQ    1
                  -         At

where ANO is the increase in the nitric oxide (NO) concentration of the air
stream after exposure to the sun, [N02]  is the initial  N02 concentration,  and
At is the exposure time.                °

     Jackson plotted his measured Ji against an Eppley UV pyranometer output
(4) that had a spectral sensitivity range that overlapped the N02 spectrum.   A
straight line correlation was obtained with considerable scatter where Jj =
0.019 +_ 0.02 E, E being the Eppley reading in watts meter"2.   Further engineer-
ing improvements made recently on the apparatus produced data that shows a
curved correlation between measured Jj and the Eppley output.   Problems not
considered by Jackson, concerning the Eppley and N02 absorption overlap,
background surface albedoes, the Eppley's cosine response and inherent errors
in relation 9, can account for the curvature seen in the new data.


AN IMPROVED ACTINOMETER

     A schematic of the new actinometer is shown in Figure 1.   Two 60 Hz dia-
phragm pumps, arranged in parallel to enhance the flow,  pushed ambient room
air into the system.  The stream of air is first passed  through a drying agent
and into a constant temperature enclosure containing the N02 source,- a permea-
tion tube containing liquid N02.  This permeation device was usually operated
at 34°C.   It was necessary to dry the air passing over the permeation tube  in
order to keep the background NO level in our airflow low.  The following re-
actions were probably giving rise to the unsteady and unacceptably high N0/N02
ratio of 0.005 that was originally observed:

         2N02+ H20 -> HN03+ HN02

         2HN02+ NO + N02+ H20

Based on tabulated thermodynamic data and assuming that pure N02 permeates, we
find that

         [NO]/[N02] = k[H2OjV3

where k ^0.1 atrrf /3.  Thus, at normal [H20] of 0.01 - 0.02 atm, [NO]/[N02]
^10 5 atm., then with the one-third power dependence one would expect a ten-
fold reduction in [NO]/[N02].  The observed values indicate qualitative agree-
ment with this mechanism and show that water (H20) must be avoided if pure  N02
in air is required.
                                     755

-------An error occurred while trying to OCR this image.

-------
     Using dried air and a potassium dichromate (K2Cr207) impregnated filter
for NO -> N02 conversion (6), we were able to bring our N0/N02 ratio down to a
steady, acceptable 0.001.   A by-pass around the drying agent - constant temp-
erature box -K2Cr207 filter can be included, as shown, in order to prolong the
life of the drying agent and K2Cr207 filter.

     The airflow was then passed in teflon through a flow meter and outside
the building to the photolysis chamber, which was a 0.27 cm radius, 1-meter
long quartz tube bent in a u-shape.  A three-way teflon solenoid valve was
placed between the two ends of the quartz tube.  One position of the valve by-
passed the quartz tube and directed the stream of air directly to our nitrogen
oxides (NOX) detector, a Thermo-Electron model 12A chemiluminescent analyzer
altered for direct capillary inlet to the vacuum system.  The other setting of
the valve sent the flow through the quartz tube to the NOX detector.  Thus,
the same capillary was used to monitor both the background [N02] and ANO.
The quartz tube was suspended approximately one foot above a 4'x6' horizontal
white plywood board to achieve a constant albedo background.

     The entire plumbing was made from teflon or glass.  Metal was avoided be-
cause of the corrosive properties of N02.  Care was also taken to make sure
that our air stream was in complete darkness except when passing through the
quartz tube.  The exposure time, At, to the sun can be controlled by either
adjusting the pump speed or adjusting the length of the quartz tube exposed to
the sun.   Adjusting the N02 concentration in the air stream can be accomp-
lished by either changing the pump speed or the permeation tube temperature.
Thus, there are more than enough variables to obtain any N02 concentration and
exposure time At desired.   Typically we used a At of 0.734 sec., and an N02
concentration around 24.6 ppm; this gave the most error-free approximation of
Ji from Equation 3.  Calibrations were taken on [N02] , background NO, and
flow speed every 90 minutes.                         °

     Keeping a steady flow speed and obtaining a steady, low background NO
level were the most difficult variables to control, the probable error in the
total system being +_ 6 percent.


MINIMIZATION OF THE ERRORS IN EQUATION 3

     If all other factors entering into the flow actinometer photolysis are
taken into account, then a system can be built in which systematic errors
almost totally cancel.  The resultant [N02]  used is 25 ppm with a flow time
of 0.75 sec.  Note that since              °

          ANO
         [N02]

is used to define Ji from Equation 3, the absolute calibration of the NOX
meter is not an important parameter to the first order.  Instead, the 100%
conversion in the NOX converter is more important, and can be tested as des-
cribed by Stedman (7).
                                     757

-------
                     ACTINOMETER VERSUS EPPLEY RESPONSE

     We have established a definite curved correlation between the actinometer
measurement of Jj and the UV Eppley photometer measurements as seen in  Figure
2.  The points represent data taken between noon and sunset on a clear  day at
Ann Arbor, Michigan, June 4, 1976.   The line in Figure 2 represents a theoreti-
cal prediction which agrees well with the data points, considering that the
line is completely independent of the points and depends only on the calibra-
tion factors supplied by Eppley and traceable to NBS.

     The theoretical determination of J: as a function of zenith angle e was
made using the formulation of Leighton as described by Stedman et al.  (3).
Similarly, the Eppley photometer reading, E, in mwatts cnf2, was determined
from
             A/- max
         E -        (cos(e) Id(A,e) + 1/21 (x,e))d ,
             A-1 min
where Amax = 385
spectrum, Id and
                 nm and Anll-n = 290
                 Is are the direct
respectively, and (cose) and (1/2)
the Eppley photometer.  (Note that
nm define the limits of the Eppley action
and scattered radiation reaching the ground,
take into account the cosine response of
we have assumed the scattered radiation to
be homogeneous over the hemisphere.)
formulation of Leighton (1).
                                      Id and Is were calculated using the
     The dominant factor determining the shapes of our theoretical  line in
 Figure 2 is the Eppley photometer's cosine response for the hemisphere directly
 overhead, whereas the actinometer is equally sensitive to light from all
 directions.  This correlation will hold for any location provided that the
 weather is clear.  The effect of cloudiness is shown in Figure 3.   One set of
 points represents data taken with clear skies from sunrise to about 8:30 a.m.
 EOT.  Clouds began moving in aftrtr 8:30, with overcast conditions prevailing
 after 11:30 a.m.  A definite curved relationship can be seen between the
 Eppley observed light coming directly from the sun.  Later, overcast conditions
 diffuses the sunlight more evenly over the hemisphere of the sky, arid the
 resulting decrease in total sunlight reaching the ground can be seen in a
 resulting linear decrease in measured J: between 11:30 a.m. and 1:30 p.m.,
 usually a time of roughly constant Ji.

     The mismatch of the Eppley and actinometer spectra could result in a
 slight curvature of the Eppley-actinometer relation.  The Eppley  responds  to  a
 wavelength peak at 360 nm whereas the N02 action spectrum peaks at  400 nm.
 The  result is that the Eppley is again more sensitive to the solar  zenith  than
 the  actinometer (1).


 Jj  FROM EPPLEY MEASUREMENTS

     Based upon data presented earlier in this report, Figure 2 will give
 accurate Jl values from a known UV photometer output for a given  solar zenith
                                      758

-------
  319NV    H1IN3Z
O 4Q
O
fO
                      m
O
CO
ID
o
o>
                                                               CU
                                                               cu
                                                             en 3

                                                             •i- CU
                                                             ~O -Q
                                                             fO
                                                             CU C
                                                             S- CU

                                                             > fO
                                                                              cri
                                                                              SZ
                                                                            fl -i—
                                                                            cn-o
                                                                            c: rr3
                                                                            T- O)
                                                                            T3 S-
                                                                            fO
                                                                            OJ -M
                                                                            S- n3
                                                                            ID
                                                                            i — S_
                                                                            CL O
                                                                        Q_
                                                                        CL
                                                             OJ
                                                             CU


                                                             O) ^
                                                             _Q to

                                                             C !-
                                                             O fO
                                                             •i- O)
                                                             •4-> i—
                                                             to o

                                                             CU ro
                                                             5-
                                                             i- S-
                                                             O O
                                                             O C4-
                                                                            r~- co
                                                                            CTi to
                                                                            i— O
                                                                            <=f O)
                                                                              i —
                                                                            QJ CD
                                                                            S-
                                                                            O
                                                                              Q)
                                                                              N
                                                                        rO
                                                                        O
                                                                       •r- -O
                                                                       •*-> CU
                                                                        cu s_
                                                                            c ^
                                                                            C E
                                                                        cu
                                                               oo
                                                                              
                                                                          01
                                                                       •a
                                                                          o
                                                                       -a cu
                                                                        cu s-

                                                                        5— ~O
                                                                        cu
                                                                        tO "O

                                                                       o to
                                                                        ID
                                                                        S-
                                                                 "o 3
                                                                 c:

                                                                    cu
                                                                 c >
                                                                 O -r-
                                                                 O en

                                                                    CU
                                    759

-------
                                                               QC >,
                                                               CM re
                                                               o o

                                                               S- 03
                                                               O)
                                                               -M S-
                                                               CD C
                                                               E <•!-
                                                               O
                                                               O -M
                                                               re o
                                                               C OJ
                                                               re i-
                                                                 Q.
                                                               >,
                                                               OJ r—
                                                               i— re
                                                               n. o
                                                               CLT-
                                                               LU -i->
                                                                 CD
                                                               C S-
                                                               O) O
                                                               CU OJ
                                                               2 -c:
                                                               4-> -(->
                                                               OJ
                                                               JD O)
                                                                 ^Z
                                                               c: +->
                                                               o
                                                               fO
                                                               r— O)
                                                               CD C  •
                                                               S- -r- CXJ
                                                               S-. i—
                                                               O   O)
                                                               U d) S-
                                                                 -C 3
                                                               TD 1— O)
                                                               (1)   -i-
                                                               >   U_
                                                               i-  .
                                                               O) IO E
                                                               c/i r~- o
                                                               JD CTi S_
                                                               O •— M-
[2.uuo   UDMUU]  An    A9|dd3
                           760

-------
angle on a clear day.  On a cloudy day we can make an approximation that
overcast conditions behave like an opaque isotropic wedge filter, i.e., they
give a linear relation between Jj and Eppley as illustrated by the triangles
in Figure 3.  Clearly, this wiTl not be absolutely correct, but the data of
Figure 3 show that it ic a reasonable approach to the difficult problem of
variable cloudiness.  Thus to find Jl5 find the value of Eppley reading (Emax)
and associated J± (Jmax) which would hold for a clear day (either from pre-
vious data injthe. s^arne location or from the known solar zenith angle shown in
Figure 2).   Then J\  =  (Eppley  reading)  x  Jmax/Emax-   If working  in  an  area  of
high surface albedo  (a), all "data-derived J: should be multiplied by (1 + a),
where a is obtainable by direct measurement or by estimation from the data in
Table 1.

     We recently received a report from L. Zafonte, P. 0.  Rieyer, and J. R.
Holmes of the California Air Resources Board (8).  They have been using a
similar apparatus, and report similar values of Jj.  They do not optimize the
intrinsic chemical errors in their system by our technique, but rather correct
for their errors by using the known reaction rates.


                      MEASUREMENTS OF OZONE PHOTOLYSIS

     There are two processes in the photolysis of ozone (03).   The most im-
portant is the production of 0(1D) in the near ultraviolet, since 0(1D) is an
important source of atmospheric hydroxyl  (OH) radicals:

         03 + hv (300 - 310 nm) •* 0(1D) + 02

         0(1D) + H20 + 20H

     The actinometer that we have devised depends on the reaction of 0(TD)
with nitrous oxide (N20).  We photolyse a stream of N20 with a small amount of
02/03 mixture.  The reaction scheme is

         03 + hv -> 0(TD) + 02

         0(1D) + N20 + N2 + 02

                     -> 2NO

With equal rates on both paths, the yield of NO formation per 0(1D) is taken
to be unity.  The NO continues to oxidize thus:

         NO + 03 -* N02 + 02

         N02 + 02 -»• N03 + 02

         N02 + N03 £ N205

         N205 + H20 £ 2NH03

     The problem is then to measure the NOX yield in the presence of excess

                                      761

-------
03.   We have tried a number of routes  and a reasonably  successful  method  is  to
decompose the 03 and higher oxides  at  350-400°C on  a platinum/aluminum oxide
(Pt/Al203) catalyst, followed by conversion to NO in a  ferrous  sulfate (FeSO^)-
packed trap.  Measured data are still  preliminary but are  in  the  same  range  as
our theoretical  predictions based on our most recent $(\,T).


                               ACKNOWLEDGEMENT

     We acknowledge the support of NSF under grant  #ATM76-03793.


                                 REFERENCES

1.   P. A. Leighton, Photochemistry of Air Pollution, Academic, New York, 1969.

2.   J.G. Calvert,  "Test of the Theory of Ozone Generation in Los Angeles
     Atmosphere,"  Environ. Sci . Techno!., 10, 248 (1973).

3.   D.H. Stedman,  W. Chameides, and J.O. Jackson,   "Comparison of Experimental
     and Computed  Values for JCNOg)."  Geophys. Res. Lett., 2, 22  (1975).
4.   J.O. Jackson, D.H. Stedman, R.G. Smith, L.H. Meeker, and P.O. Warner,
     "Direct N02 Photolysis Rate Monitor," Rev. Sci. Instrum., 46, 276 (1973).

5.   D.D. Wagman, W.H. Evans, U.B. Parker, I. Halow, S.M. Bailey, and
     R.H. Schumn, NBS Technical Note 270-3, Nat. Bur. Stand., (1968).

6.   D.L. Ripley, J.M. Clingenpeel, and  R.W. Hum, "Continuous Determi-
     nation of Nitrogen Oxides in Air and Exhaust Gases," Int.  J. Air
     Water Poll., 8., 455 (1964).

7.   D.H. Stedman, "A Flow Independent Procedure for the Gas Phase Titration
     of an Ozone Source."  J. Air Poll. Assoc., 26_, 62 (1976)

8.   Zafonte, P.L. Reiger, and J.R. Holmes, "Notrogen Dioxide Photolysis
     in the Los Angeles Atmosphere," Submitted to Environ. Sci. Technol . ,
     (1976).
                                     762

-------
                                                                           15-5
       ACTINIC  SOLAR FLUX AND PHOTOLYTIC RATE CONSTANTS IN THE TROPOSPHERE

                 J.  T.  Peterson,  K.  L.  Demerjian, K. L. Schere*

ABSTRACT

     The results  o^ detailed theoretical calculations ofi the. volumetric &lu.x. of,
solaA  Radiation  -in  the atmosphere aAe  discussed, and the. de.veJLopme.nt o^ a
generalized  scheme,  to  appty those data to the. calculation o& photolytic note.
constants  over a diurnal period  -is  described.  Ike, vertical variation of, the.
late. constants f,ofi  nitrogen dioxide and formaldehyde through the. lowest be.veA.al
ktlometers of,  the atmosphere, -is  emphasized.   Be.co.o6e of, changes -in AolaA ^ax,
thue.  'tate. constant?, one. typicaHij 20  to 70  peJtce.nt Qie.at.eA at 1 km aLtita.de.
than at the.  e.aAth'& iu/L^ace,  faon example.  The. the.ox.etic.alty deAived note.
conAtantA meAe comparted to  AimiloA ambient mea&ntementA and theAe wai good
agreement  betive.en them.   The neAultA pfie&ented h&te have relevance to photo-
chemical aiA quality simulation  models and smog chambeA Studies.

                                  INTRODUCTION

     Important parameters of photochemical air quality simulation models and
smog chamber studies are the rate constants  (k) for the photodissociation of
certain species.  They describe  the rate (time -1) at which a molecule dis-
sociates when  irradiated by ultraviolet (UV) solar radiation.  Since k is
directly dependent  on  incident solar intensity, it provides a link between solar
radiation  and  air quality levels in both mathematical models and chamber studies.
Thus,  accurate knowledge of the  spatial, diurnal, and annual variability of k
throughout a metropolitan atmosphere is necessary in various studies of photo-
chemical air pollution.   This paper describes the results of recent detailed
calculations of  solar  radiant energy in an urban atmosphere and the development
of  a generalized scheme to  use the radiation values to calculate photolytic rate
constants  over a diurnal period.  The  rate constants for nitrogen dioxide (N02)
and formaldehyde (H2CO)  are emphasized.

                           RADIATIVE FLUX COMPUTATIONS


     For photochemical  applications, the  direction  at  which  solar  radiation  is
incident upon  an N02 molecule  is  of  no  concern.   The  primary consideration  is
the flux through a  spherical  volume, rather  than  that  through  a  plane  horizontal
surface, which  is used for most  geophysical  applications.  Thus, this  actinic,
^Environmental Protection  Agency,  Research  Triangle Park,  North Carolina.
 All three authors  are  on  assignment  from the  National  Oceanic and Atmospheric
 Administration, U.  S.  Department  of  Commerce.

                                      763

-------
or volumetric, flux is the sum of the radiative intensity over all angles of
incidence and represents the solar radiation that would be incident on an
ensemble of molecules with random orientations.

     The actinic solar flux for a model atmosphere has been calculated for solar
zenith angles for 0° to 86° and for wavelengths from 290 to 700 nm (Peterson,
1976a).  A detailed radiative transfer model, based on the work of Braslau and
Dave (1973), was used for the flux computations at 40 model levels from the
earth's surface near sea level to the top of the atmosphere.   Aerosol scat-
tering and absorption, ozone absorption and Rayleigh scattering were account-
ed for.  Total atmospheric ozone was assumed to be 0.295 cm-atrn.  The normal
optical thickness of the aerosols was 0.254 to 500 nm wavelengths.  This
corresponds to annual average nonurban conditions over the eastern U. S. and
is slightly less than that for annual average U.S. urban conditions (Flowers
et al., 1969).  It also corresponds to conditions in Los Angeles on "clean"
days during the smog season (July through October).  Thus, the calculations
were based on typical aerosol  amounts so that computed actinic fluxes would
have general applicability.

                      PHOTOLYTIC RATE CONSTANT COMPUTATIONS

      A generalized scheme and computer program for calculating photolytic rate
 constants  has been developed  by Schere and Demerjian (1976).  It uses as input
 the actinic flux values,  J,  described above,  in addition to  the absorption cross
 section, a, and primary quantum yield, ,  for each species.   For a particular
 species, i, the rate constant for photodissociation in the lower atmosphere, is
 expressed  as follows:

           k.(e,h)  - zj(A,e,h) •  a,(A) •  <|>.(x) ,                (1)
            I         A            II

 where  0  is solar zenith angle,  h is the  height above ground  and A is wavelength.
 Values of  a and <}>  have been  compiled for seven species which dissociate when ex-
 posed  to radiation from 290  to  800  nm wavelength:   N02, nitrous acid, (MONO),
 nitric acid (HON02),  03,  H2CO,  hydrogen  peroxide (H202), and acetaldehyde
 (CH3CHO).  Specific reactions  considered  were:

           N02 + hv -> NO + 0(3P)                                 (2)

           HONO + hv -> HO + NO                                  (3)

           HON02 +  hv -> HO +  N02                                (4)

           03  + hv  -> 0(3P)  +  02                                  (5)

           03  + hv  -> 0(1D)  +  02                                  (6)

           03  + hv  -> 02(1A) +  0                                  (7)

           H2CO + hv -> H + HCO                                  (8)
                                     764

-------
           H^CO + hv > H? + CO  .                                (9)

           H?0? + hv > 2HO                                      (10)

           CH3CHO + hv -> CH3 + HCO                              (11)

           CH3CHO + hv - CH4 + CO                                (12)

     The user-oriented computer program for calculating photolytic rate con-
stants was developed to  provide output for a diurnal solar cycle.  Rate con-
stants are first generated  for the specific zenith angles for which actinic flux
values are available, i.e., 0, 10, 20, 30, 40,  50, 60,  70, 78,  and 86°.   Actual
solar zenith angles are then calculated for the diurnal  period  for a  specified
location and date.   Finally, the corresponding  rate constants are determined
throughout the diurnal period by using a cubic  spline fit to interpolate for any
zenith angle.

     The rate constant program is  formulated to calculate values for  the
lowest tens of meters near the earth's surface.  However, actinic flux
data are available for ten additional  levels of the radiative transfer
model:  0.15, 0.36, 0.64, 0.98, 1.38,  1.84, 2.35, 2.91,  3.53, and 4.21 km
above the surface.   The actinic flux values at  each zenith angle for  these
levels can be supplied to interested readers upon request.

                                    RESULTS

     Since the wavelength-dependent absorption  coefficients and quantum yields
for any species are fixed, the variation of that species' rate  constant in space
and time depends solely on the variability of the actinic flux.  Besides its
diurnal dependence, the actinic flux is also sensitive  to changes of  surface
albedo, atmospheric total 03 amount, cloudiness, surface elevation,  atmospheric
aerosol concentrations, and altitude above the  surface  (Peterson 1976a).  This
last factor—the altitudinal dependence of actinic flux--is itself strongly
dependent on aerosol concentrations and their change with height.

     In the ultraviolet spectral  region, the total actinic flux generally
increases with height through the  lowest five to ten kilometers of the atmos-
phere.  The upward- and downward-directed actinic components show similar
height dependence.   The downward component typically exhibits its greatest rate
of change near the surface, where  it is depleted by absorption  and back-scat-
tering from high aerosol  and pollutant concentrations.   The upward component
also increases rapidly with height near the earth's surface because  of the low
reflectivity of the earth (typically about 5% in the ultraviolet) and the large
amount of aerosol and Rayleigh scattering.  But, because of its small  absolute
value, the upward component has only a minor effect on  the percentage change
with height of the total  actinic flux.

     Of the several species listed previously which dissociate  when  exposed to
sunlight, N02 and H2CO are two of  the  most important for photochemical air
pollution (Dodge and Hecht, 1975).  Equation 1  was evaluated for these compounds
                                      765

-------
for each 10 nm wavelength interval using the actinic flux at the earth's surface
and 40° zenith angle.  The results were normalized for each species to a maximum
value of 1.0.  The resulting wavelength sensitivity of the two rate constants  is
shown in Figure 1.  The sensitivity of the rate constants at the shortest wave-
lengths goes to zero parallel with the amount of incident UV radiation.  In each
case, the long wavelength cutoff is determined by the respective quantum yield
and absorption cross-section.  The H2CO rate constant is sensitive to a rather
narrow spectral interval centered at about 330 nm.  In contrast, the N02 rate
constant responds to a much broader spectral and region longer wavelength.  The
importance of this difference in wavelength sensitivity will be discussed below.
    1.0
    0.8
  t 0.6
  _

  £
  cc
0.4
    0.2
                      I
                           I
I
I
     0 —

     280
             300
                      320
                              340      360       380


                                   WAVELENGTH, nm
                                                       400
                                                                420
                                                                        440
  Figure 1.  Normalized sensitivity of  photodissociation  rate  constants for
    N02  (solid) and H2CO  (dashed)  versus wavelength.   For each species, the
   wavelength dependent values were obtained  from  the  product  of absorption
     cross section, quantum yield, and  actinic  flux  at the earth's surface
                              for 40°  zenith angle.
     Using Equation  1, the photodissociation  rate  constants  for N02  (Equation 2)
and H2CO (Equation 8) were calculated  as  a  function  of  zenith  angle  for the
surface and first ten model  levels above  the  surface.   Selected results are
presented in Tables  1 and 2  for N02 and H2CO,  respectively,  and in  Figure  2 for
N02.  Because of the change  of actinic flux with height,  the rate  constants for
both species show significant increases with  height.  The rate constants general-
ly increase with height and  zenith angle, except for the  86° case.   For N02,  for
                                      766

-------
example, k is 21 to 70% greater at 0.98 km than at the surface, depending on
zenith angle.  Corresponding percentage increases for H2CO are even greater for
zenith angles of 79° and less, but are smaller for the largest zenith angle.
  UJ
  O
        e0 = 86°
                            78°
60°
50°  40° 30° 0°
             0.1      0.2      0.3      0.4     0.5      0.6      0.7

                      NO2 PHOTODISSOCIATION RATE CONSTANT, min 1
             0.8
                     0.9
  Figure  2.   Calculated N02  photodissociation rate constants as a function of
                     altitude and solar zenith angle (0,J.
     The differences between the NO? and H2CO rate  constant  increases  from  the
surface to higher atmospheric levels result primarily  from the  different wave-
length sensitivity of the two species.  As shown  in  Figure 1, H2CO  responds
more to shorter wavelength radiation than does N02.   Since Rayleigh (molecular)
scattering depends on the inverse fourth power of wavelength, the shorter
wavelength energy undergoes more depletion while  passing  through the  atmosphere.

     The increased values of the rate constants above  the surface can  be quite
significant for ground level pollutant concentrations.  For  example,  precursor
pollutants emitted aloft, such as from a tall chimney,  into  a stable  atmospheric
layer will be exposed to more solar radiation than  near the  surface.   If subse-
quent increases occur in the mixing height, reaction  products of pollutants
trapped in stable layers can be mixed downward to the  surface,
example, precursor pollutants emitted near the surface  can be
            In  another
          transported aloft
and be exposed to more intense radiation when vertical mixing  is  occurring  in
the lower atmosphere.  Subsequently, the reaction  products  can  be transported
back to the surface.
                                     767

-------
oo
        \
O LU CM
       00
       O)

-------
1 	
1
 

o
h-
LU
c •>





CTi
CO
r*^.



CM
.
o
r-.



LO
.
^j-
to



co
•
CD
to



r--
•
CO
LO

oo
co
LO
E
r—

-------
     An alternate way of  presenting  the  height dependence of the photodissocia-
tion rate constant is shown  in  Figure  3.   The diurnal  variation of k.,n  was cal-
                                                                     NV?
culated for Los Angeles,  California, on  September 1,  using the actinic flux for
the surface and at 0.98 km  (fourth model  level).   Although the largest absolute
differences are during midday,  the greatest percentage changes occur during
early morning and late afternoon.  Thus,  these data show that photochemical
reactions generally take  place  at a  significantly faster rate 1 kilometer above
the earth than at the surface.
           80
               70   60
                        50
                             40
                                    ZENITH ANGLE, degrees

                                   30     25 7    30
                                                              60
                                                                   70
                                                                        80
        0600
              0700
                         0900
                              1000
                                    1100   1200    1300

                                     TIME (PST). hour of day
                                                    1400
                                                          1500
                                                               1600
                                                                     1700
                                                                          180C
   Figure  3.   Calculated N02 photodissociation rate constants versus  time  for
       Los  Angeles,  California on September I  for the earth's surface and
                 0.98 km above the surface (fourth model level).


     The results presented  in  the previous  discussion were derived from cal-
culated actinic fluxes  based  on  average  atmospheric aerosol characteristics
typical of eastern U.S.  nonurban  conditions.   In reality,  aerosol concentrations
over metropolitan areas  are often two  or three  times  greater than our model
values.  Consequently,  the  vertical  gradient  of actinic flux, and hence the
altitudinal change of rate  constants,  is often  substantially greater than that
shown above (Peterson,  1976b).   In fact, for  aerosol  concentrations double those
used here, the increase  of  k  from the  surface to 0.98 km was calculated to range
from 29% for a zenith sun to  103% for  a  solar zenith  angle of 78°.

     The rate constant  data previously described have all  been  theoretically
generated.  To test  the  overall  reliability of  our  computational  scheme, in-
                                      770

-------
 dependent measures  of  the  N02  rate  constant were  obtained  by  Research Triangle
 Institute personnel  exposing a chemical  actinometer  to  sunlight  using the
 technique of  Sickels and Jeffries  (1975).  An  example of the  comparison  between
 the rate constants  measured with the actinometer  and calculated  by  the method
 described herein  is  shown  in Figure 4.   A more complete comparison  is discussed
 by Schere and Demerjian (1976).  The circles represent  the measured values.  The
 dashed line represents rate constants calculated  for cloudless conditions.
 Since clouds were present  on this day (April 23,  1975), the calculated values
 were adjusted downward by  the  ratio of the measured UV  radiation on this day to
 the measured UV radiation  during an adjacent cloudless  day.   Although the  data
 have considerable scatter, the  two  sets  of data evince  general agreement.
     10.0
  4
  o
   tvl
   O
   Figure 4.  N02 photodissociation rate constants versus  time  for  Research
      Triangle Park, North Carolina on April 23,  1975.   Circles  represent
       five-minute average values measured with a chemical  actinometer.
   Dashed line represents calculated values for cloudless  conditions.   Solid
    line represents calculated values for prevailing  cloudiness  conditions.


     Another set of empirical N02 rate constant data has been recently published
by Zafonte (1976) for El Monte, California, 17 miles east of Los Angeles.
During June and July 1975, he obtained a maximum value of about 0.55, based on
a measurement of UV radiation.  During midsummer the noontime solar zenith angle
at Los Angeles is about 10°.   The corresponding calculated rate constant for the
                                     771

-------
 surface  is  0.574  (see Table  1).   However, even  though  the  uncertainty  in ~th~e~"
 calculated  values  exceeds  several  percent when  they  are  applied  to  a specific
 situation,  a  small  correction  (-3.2%)  should  be  applied  to the calculated  rate
 constants to  account for  the greater  than average earth-sun  distance during  that
 time  of  year.   The resulting measured-calculated rate  constant agreement is
 remarkably  good.

                                  DISCUSSION

     Two recent studies using the photochemical  diffusion model developed by
Systems Applications, Inc. (SAI) have shown that computed  03 concentrations  in
the Los Angeles Basin are quite sensitive to  incident solar radiation as it
affects k.|Q   (Peterson and Demerjian, 1976; Liu  et al.,  1976).  However, in  the

SAI model,  kwn  is invariant with height.  In view of the  large veiriation of
            INU2
kMn  with height discussed above, we recommend that photochemical models be
 NU2
tested for  their sensitivity to vertical changes of radiation-dependent rate
constants.  In  the two reports mentioned, solar  radiation was linked to computed
pollutant concentrations only through k.,Q .   Photochemical models with improved

kinetic mechanisms including other species subject to photodissociation, such as
H2CO, are are now  being developed.  Since species like H2CO, which  respond to
short wavelength UV radiation, are more dependent on altitude than  k^g , the

improved models may show even more sensitivity to the spatial variations of
solar radiation.

     The results presented in this paper also have  relevance to smog chamber
studies.   Zafonte  (1976) has suggested that typical  k.,^  values in chambers and

the length of time for chamber runs usually do not  agree with k.,Q  values mea-

sured near the earth's  surface.   Our calculations indicate that the discrepancy
between ambient and chamber kKin  levels is even greater for the top of the mixed
                             IN Up
layer compared to that at the earth's surface.  Our data also point out the
necessity to maintain continual  measurements and calibrations of radiation
levels as a  function of wavelength within chambers.   Photodissociation rates for
N02 and H2CO are important in kinetic mechanisms, yet the two species are
sensitive to different UV wavelengths.  Thus, although an N02 chemical  acti-
nometer may  be useful  for general  checks of chamber light levels, it alone would
not be sufficient for an overall calibration check.

     To date,  only a few ambient direct measurements of k.,n  have been made  in
                                                         INU2
the atmosphere.   More such measurements need to be  taken in a comprehensive and
systematic way to determine typical rate constant values in a variety of con-
ditions and  to provide data for comparison to theoretical estimates.  In this
context, experimenters  should record and publish the state of all parameters
that can affect the actinic flux during their measurement periods.  These in-
clude date and location, cloudiness, aerosol  concentration or optical thickness
(visibility  is a useful  surrogate), surface elevation, significant obstructions
                                     772

-------
to the horizon, and the underlying surface.  The actinometer should  be  located
over a uniform underlying surface and should be distant  from highly  reflecting
structures such as white walls.  With such care in  acquiring and  reporting  am-
bient measurements of photodissociation rate constants,  the results  of  different
experimenters can be compared along with  theoretical estimates.   Finally, we
recommend that measurements be made of photodissociation  rate constants  verti-
cally through the lowest several kilometers of the  atmosphere,  such  as  by
mounting a chemical actinometer on an aircraft.  More  confidence  could  be placed
on the theoretical computations of the vertical dependence of the  rate  constants
described in this paper if they were experimentally verified.

                                  REFERENCES

1.   Braslau, N. and J. V. Dave, 1973.  Effect of Aerosol on the Transfer of
     Solar Energy Through Realistic Model Atmospheres.   Part I:  Non-Absorbing
     Aerosols.  J. Appl. Meteor., ]_2, 601-615.

2.   Dodge, M. C. and T. A. Hecht, 1975.  Rate Constant Measurements Needed to
     Improve a General Kinetic Mechanism  for Photochemical Smog.   Int.  J. Chem.
     Kinetics, 7_, 155-163.

3.   Flowers, E. C., et al., 1969.  Atmospheric Turbidity Over  the United
     States, 1961-1966.  J. Appl. Meteor., 8, 955-962.

4.   Liu, M. K. et al., 1976.  Effects of Atmospheric  Parameters on  the  Con-
     centration of Photochemical Air Pollutants.  J. Appl. Meteor.,  15_,  829-835.

5.   Peterson, J.  T., 1976a.   Calculated Actinic Fluxes (290-700nm) for Air
     Pollution Photochemistry Applications.   Rept.  EPA-600/4-76-025.   U. S.
     Environmental Protection Agency, Research Triangle Park,  N. C.

6.   Peterson, J.  T., 1976b.   Dependence of the N0? Photodissociation Rate
     Constant on Altitude.   Submitted to Atmos. Environ.

7.   Peterson, J.  T.  and K.  L.  Demerjian, 1976.  The Sensitivity of Computed
     Ozone Concentrations to Ultraviolet Radiation in  the Los  Angeles Area.
     Atmos.  Environ.  1_0_, 459-468.

8.   Schere, K. L. and K. L.  Demerjian, 1976.   An Algorithm for the Generation
     of Selected Photolytic Rate Constants over a Diurnal Range.  To be pub-
     lished by U.  S.  Environmental  Protection  Agency,  Research Triangle Park, N.
     C.

9.   Sickles, J.  E.  and H.  E. Jeffries, 1975.   Development and Operation of a
     Device for the Continuous  Measurement of <|>k for Nitrogen  Dioxide.  Publ. No
     396, Dept.  of Environ.  Sci.  and Engin.,  Univ.  No.  Car.,  Chapel Hill, N.C.

10.   Zafonte, L.,  1976.   Nitrogen Dioxide Photolysis in the Los Angeles Atmos-
     phere.   Submitted to Environ.  Sci.  and Technol.
                                     773

-------
                     SESSION 16
MATHEMATICAL MODELS OF OZONE/OXIDANT AIR QUALITY - I
                       '-  K. Demerjian
           Environmental Protection Agency
                         775

-------
                                                                            16-1
                 PHOTOCHEMICAL AIR QUALITY  SIMULATION MODELING:
                      CURRENT STATUS  AND  FUTURE  PROSPECTS

                               K. L.  Demerjian*
ABSTRACT

     Mathematical, model*  that ftelate  pollutant. miAt>iont> to ambient (Wt quality
through thu the.otie.tic.al ttie.atme.nt  ofa  the.  chmical. and physical ptiocik&e.t> o^
the atiw^plie.fie axe, ne.v-ie.wed.  AH evaluation o{] c.utv>ie.ntly available, models 
p-ie&ented and vasu,ou6 t>hofitcomiiiQ&  o& ^tate-ol-the-aJtt models ate. di&ui£&e.d.
The St. Louc4 Re.gi.onat kin.  Pollution  Study mode.l veHi^i-C.a.'tion pfiogsiam it, out-
lined and a (it i CUM-ton o{} the approach -it,  pie.&ente.d.


                                 INTRODUCTION


     Pollutants emitted into the atmosphere as a result of man's activities are
transported, dispersed, transformed,  and  deposited via complex physical and
chemical processes.  Mathematical  models  provide a unique technique for evalu-
ating the impact of anthropogenic  emissions on air quality and will prove to
be formidable tools in implementing the mandates of the Clean Air Act.  The
purpose of this paper is  to review and evaluate currently available photo-
chemical air quality simulation models and to  discuss their potential applica-
tions and future prospects.

     The purposes for air quality  simulation models and their ultimate poten-
tial applications are listed below:

     •     Research:  Gaining an understanding of pollution processes on which
           to base air pollution abatement strategies.

     •     Environmental  Legislation:   Deciding what emission standards should
           be set for emissions from  stationary, mobile, or indirect sources,
           in order to meet already agreed-upon ambient standards.

     •     Implementation Planning:   Evaluating the effectiveness of emission
           control regulations  in  achieving ambient air quality standards.
 *Environmental  Protection Agency, Research Triangle  Park,  North  Carolina.   He
  is  on  assignment to the EPA from the National Oceanic  and Atmospheric Admin-
  istration,  U.S.  Department of Commerce.


                                      777

-------
     •     Impact Assessment:   Predicting the consequences  of policy decisions
           in terms of resulting air quality.

     •     Source Identification:   Determining which major  sources  most sig-
           nificantly deteriorate  air quality in  the modeled  region.

     t     Monitor Siting:   Evaluating optimum site  selection for stations
           in an air monitoring network.

     t     Transportation and  Land Use Planning:   Determining the relative
           air quality effects of  various alternatives  for  physical  develop-
           ment.

     0     Episode Control  System:  Provide real-time simulation  and control
           strategy capabilities for an alert warning system  in order to avert
           episodes of extremely high pollutant concentrations.

     Though some of these applications have been  realized for inert species,
primarily through the use of Gaussian plume models,  complex numerical-
atmospheric diffusion equation models have found  limited application as a re-
sult of their complexity, extensive resource requirements,  and most impor-
tantly, their unverified status.  The issue of model verification and its role
in the Regional  Air Pollution  Study (RAPS) in St.  Louis is  discussed later  in
this paper.

                    MODEL CLASSES, SCALES AND CHEMICAL  MODES


     Air quality simulation models (AQSM) fall into  three broad meteorological
classes.  These are:

     ^     Diagnostic - models which require input of predetermined meteorolo-
           gical data in time and  space over the  period of  the simulation
           study.

     •     Prognostic - models which generate meteorological  fields in time and
           space given initial and boundary conditions  for  the modeling domain.

     •     Climatic - models which utilize climatological data as input in
           developing meteorological fields for the  AQSM.

     AQSM scales, which range from micro to synoptic, have  been broken into
the four functional modeling domains below:

     Local (micro)                     (400 m X 400  m X 100 m)
     Urban (meso)                      (40 km X 40 km X H)
     Regional (meso)                   (800 km X  800 km X H)
     Continental (meso-synoptic)       (2800 km X 2400  km X 2 km)
     where H = mixing height

     The modeling domains outlined above could easily be expanded or con-
tracted by a factor of two with little or no impact  on  the  overall  formulation

                                     778

-------
of the model.

     The chemical mode of an AQSM is either non-reactive or reactive.  Non-
reactive refers to those models that only consider inert pollutants or pollu-
tants whose chemical transformation can be expressed by a first order reaction
rate.  Pollutants typically considered as non-reactive, as defined above, are
total suspended particulates (TSP),  sulfur dioxide (S02),  carbon monoxide (CO),
and sulfate (SO^").   Reactive models, the subject of this  review, have been
termed here as photochemical air quality simulation models (PAQSM).   This is
the result of the major role played  by photolytic reactions  in the atmospheric
transformation processes of pollutant species.   Pollutants typically consid-
ered in PAQSM include reactive hydrocarbons (HC), CO, nitric oxide (NO),  nitro-
gen dioxide (N02), and ozone (03).   Historically, the development of PAQSM has
focused on the urban scale, where the ozone problem was thought to originate
and reside.  Recent field programs  indicate that ozone and its precursors are
not isolated to the confines of the  urban complex, but are transported over
scales of hundreds of kilometers and time periods of several  days in many in-
stances.   Based on these findings,  development work in regional  scale photo-
chemical  air quality simulation models has begun.

                               MODEL FORMULATION
     The development of a mathematical  relationship for simulating the trans-
port, dispersion, transformation, and disposition of pollutant emissions into
the atmosphere can assume varied ranges of complexity depending on the appli-
cation to be considered.   Figure 1  gives some indication, conceptually, of the
complex nature of the PAQSM and shows the interactions of the various physical
and chemical processes operative in polluted atmospheres.  Techniques available
for treating transport and dispersion include gradient-transfer, Gaussian,
similarity, and statistical approaches.  Though each has been considered in
atmospheric modeling, the Gaussian and gradient-transfer approaches have found
the most prevalent use.  In particular, PAQSMs have considered only gradient-
transfer approaches, mainly due to the time-dependent nature of the chemical
transformation processes.

     The need for practical (operations oriented) PAQSM for describing the
physical and chemical dynamics of the atmosphere has necessitated the develop-
ment of simplified approaches in treating reactive species in the turbulent
planetary boundary layer.  This need for simplification has resulted in the use
of the so-called atmospheric diffusion equation in PAQSM approaches (Equation
1).

  9S-        3C.       3C.       3C          (K3C)        (K3C.)
                  + V  -  + W  -  -  -     -  +
                       -            -
  at         3x        3y        3z      3x     3x      ay
         (KV ac".)  + R. (c. ..... Fn,T)  +  ^.(x.y.z.t)                   (1)
                                     779

-------
                          HETEOROGENEOUS
                            PROCESSES
                  PHOTOCHEMICAL
                    REACTIONS
ANTHROPOGENIC
  EMISSIONS
  NATURAL
  EMISSIONS
  ADVECTED
 POLLUTANTS
SOURCES
               TOPOGRAPHY
               ROUGHNESS
                       HOMOGENEOUS
                        PROCESSES
         DEPOSITION
             &
        RESUSPENSION
             THERMOCHEMICAL
               REACTIONS
 AEROSOL
PROCESSES
                     TRANSFORMATION
                    MATHEMATICAL MODEL
                                               TRANSPORT
                                              TURBULENCE
                                               RADIATION
                       TEMPERATURE
                     CLOUD
                     COVER
                                                               WIND
                    SINK
SCAVENGING
~1

SES

TED
KA1ION


PRECIPITATION
  Figure 1.   A  schematic  of the major  components  contributing  to  the  photo-
               chemical  air quality  simulation problem.
                                             780

-------
where

  c. = ensemble mean concentration for species i

x,y,z= Cartesian coordinates

u,v,w= ensemble mean velocities

K,j,K = horizontal and vertical eddy diffusivities

  S. = rate of injection (or removal) of species i by a source (or sink)

  R. = rate of production (or consumption) of species i through chemical

  T  = Temperature

  t  = time

The atmospheric diffusion equation is the result of the application of simpli-
fying assumptions to the equation of conservation of mass.  Since its deriva-
tion is given in many texts (for example, References 1,2), it will not be
covered here, but some discussion of its entailing assumptions will be con-
sidered in the following section.

                      PAQSM CURRENT STATUS AND EVALUATION


     A synopsis of currently available photochemical air quality simulation
models is given in Table 1.  The models covered are a representative sample of
the scales and techniques currently available.

     The evaluation of any air quality simulation model begins with assessing
the validity of the basic assumptions used in deriving the models' working
mathematical equations.   In addition, identifying potential sources of error
in various components used in the formulation of the transformation, emission,
and transport modules of the PAQSM's is critical to evaluating the overall per-
formance of models and in assessing potential sources for model inaccuracies.
Seinfeld(2)  and Demerjian(3) have recently reviewed this subject area and a
summary of their respective findings is given in Table 2 and 3.

     The sensitivity of a PAQSM to various errors associated with input para-
meters to the model has been reported by Liu et al.(ll).  A summary of the
results, reported in Table 4, gives some insight into the potential inaccura-
cies that may be expected in PAQSM predictions.  One cautionary note:  the
sensitivity study reflects only the response of the model to perturbations in
its input parameters.  No inference can be drawn between model sensitivity and
that of the real atmosphere until verification studies confirm the soundness
of the fundamental physical and chemical constructs of the model.

     Initial and boundary conditions represent critical inputs to the PAQSM.
In the case of the grid point approaches, boundary pollutant concentrations
must be specified for each hour for all grid cells that lie along an inflow

                                     781

-------An error occurred while trying to OCR this image.

-------
  TABLE 2.  ERRORS ASSOCIATED WITH THE USE OF THE ATMOSPHERIC DIFFUSION
                                EQUATION
True form of the turbulent fluxes u' c •,  v 'c.',  and w' c? is unknown

Turbulent fluctuating chemical  reaction  terms  are neglected

Mean velocities u, v and w used in solutions are not true ensemble means

Chemical reaction mechanism R.  does not  accurately reflect those chemical
processes occurring in the atmosphere

Solutions on grids must be spatial consistent  with the wind and emission
resolution
    TABLE 3.  SOURCES OF ERROR IN THE DEVELOPMENT OF TRANSFORMATION,
                EMISSION AND TRANSPORT MODULES FOR PAQSM
Transformation - Chemical Kinetic Mechanism

  Smog Chamber Related Errors

     Inadequate or no control and measurement of levels of H20 in the chamber.

     Impurities in background chamber air.

     Inadequate or no measurements of the spectral  distribution and intensity
     of the chamber irradiation system.

     Inaccurate or ambiguous analytical methods.

     Non-homogeneity due to inadequate stirring or poor chamber design.

     Adsorption and desorption of reactants and products on chamber walls.

     Chemical reactions occurring on chamber surfaces.

     Inadequate control and measurement of chamber temperature.


  Chemical Mechanistic Errors

     Uncertainties in experimental determinations of specific reaction rate
     constants.

                                (continued)

                                   783

-------
                            TABLE 3.   (continued)

    Chemlcal Mechanistic Errors (continued)

       Variations of rate constants with temperature either uncertain or
       unknown.

       Inadequacies in lumping due to the non-representativeness of lumped
       class reactions relative to specific species within the class, e.g.,
       reaction rates, products, and stoichiometric coefficients.

       Inaccuracies in the mechanism due to insufficient verification studies


    Source Emissions

       Inaccurate or no specification of source location.

       Uncertainties in emission factors.

       Inaccurate or no temporal resolution of emission.

       Inadequate or no verification of emission methodologies.


    Transport

       Uncertainties in the measurement of wind speed and direction.

       Inadequate or non-representative spacial  measurements  of wind  speed
       and direction.

       Uncertainties associated with  wind field analysis techniques.

       Inadequate or no spacial and temporal  measurements of  vertical
       temperature.

       Inaccuracies in the specification of eddy diffusivities in time and
       space.

       Inadequate or no spacial and temporal  measurements of  solar radiation.
boundary from the ground to the inversion base.   The horizontal  boundary con-
ditions are established, when possible,  using nearby monitoring  data  as  a
guide.  When such information is unavailable, estimates used are made solely
on a judgmental  basis.   Concentrations are assumed invariant in  the   I  direc-
tion for any given x, y and t.   Initial  concentrations  are specified  in  each
ground level grid cell  by interpolation  using the data  collected at monitoring
stations during  the hour at which the model  run  is initiated.   For  trajectory
approaches, only initial concentrations  need to  be specified.   The  usual


                                     784

-------
GO

o
i—i

<:
I—I



o
O
O
O
Q_
o;
CD
o

>-
GO
z:
LU GO

   LU
LU
D~
I—
   
-------
approach is to initiate the trajectory in the vicinity of an air monitoring
station and to use the observed measurements as initial  concentrations in the
trajectory cell.

     To illustrate the importance of boundary concentrations on grid point
model calculations, comparative simulations for Los Angeles on September 29,
1969 were made by first using the standard prescribed boundary concentrations
and by then using a reduced set of nominal boundary concentrations.   All other
conditions remained the same.  Ozone average concentration maps for  the Los
Angeles basin between the hours of 1:00 and 2:00 P.M. are shown in Figures 2
and 3 for the respective simulations.   A comparison of the results shows that
only minor differences are observed at the eastern and northern edges of the
basin where the maxima occur, but significant differences are observed at the
western and central portions of the basin.  The overall  effect of the reduced
boundary conditions was to bring an additional  904 square miles of the Basin
below the oxidant standard, which is equivalent to 237% improvement  in oxidant
air quality in the modeling region on  an area basis.

     The impact of initial conditions  on the grid point approach is  illustrated
in Figure 4.   In  this case, in addition to the  reduced boundary condition (as
in Figure 3), initial conditions prescribed in  the model  base case run were
reduced by 50 percent.  A comparison of results in Figures 3 and 4 shows re-
duction in the predicted ozone levels  at the northern and eastern edges of
the basin (the meteorological conditions for this day are such that  these are
the only two  areas where some portion  of the early morning initial pollutant
concentration has been retained) to be of the order of 20 - 30 percent.

     The importance of initial conditions on ozone predictions using a trajec-
tory model approach are illustrated in Figure 5.  In  this case, comparative
trajectory simulations were run for Los Angeles on September 29, 1969 for the
original initial  concentrations prescribed for  the cell  and plus and minus 20%
of those values.   The results indicate an approximate equivalent percentage
change in the ozone maximum by the model.

               REGIONAL AIR POLLUTION  STUDY:  MODEL VERIFICATION

     The St.  Louis Regional Air Pollution Study (RAPS)(12) is a five-year field
program sponsored by the EPA and initiated in 1972.  Three major objectives
sought by the program are:

     •     Develop, evaluate and verify air quality simulation models on an
           urban  scale covering urban/rural stationary and mobile sources.

     •     Develop, evaluate and verify models  of local-scale phenomena that
           complement urban-scale models.

     t     Archive all data collected  under the program in the form  of a
           readily retrievable data base to use in evaluating future air
           quality simulation models and effects models.

     Model verification studies have been very  limited in nature due to the
lack of adequate  air monitoring data bases against which  to test models.  This


                                     786

-------
IT     ec
                                                                                8
•« r- f^ tfi o •*> IM «no>


* " %

1 1
1 1
1/1 x° " ' * i "
* — 1
|» IT «n IM 0 K IS



	 	 	 -*-'/



4
«M in fs, — — — — _r>4 ivj fsifsi rsi «i IN r* _ ., _

tsi






^

----„---/ N? - - --

s. — j \; . " ;
<1 fSJ «J — — S *~ \ — — —
£ / \
of \
o i/t < fs. — .fflr^r^F^f^ ^ -o •* f"» . 1
1,N~-J--- V »___
* tr \ X
1 
1 "x ^ -*^ a
o» ^^^ u
r*l|M — — — — — — ' — -^V l^™^ —
1 " \ 1 ^
' J \ 1 °
1 \ I J
- V ----- - .----»,»- ---a
- * *
N \ •-
_ ^ — cr r* — fvj — OP-1/1— --— ^
" '," ¥
1 Ul
1 z
rwlrMCj-CN — r^h-— CiO — — r-
.* LJ
" -a 0
/ u C.

r-im — — — -^^— — — — U
>_
cu to

QJ -(->
-M to
"• Q- >,
O) CO
U~l 	

C
t/) oo


-Q O
CO -(->
OJ "3
r— S_
cu -t->
en c
sr ai
to O
0 U
	 i
0) i- •
JC nD •• '
-!-> T3 r—
c cu

•r- O O

to

O O) C
-MOO
ro E Q_
s-
4_) ~o "O
c s- -i-
CD T3 S—
(J "O C3
C C
O rC
O -M O
to C
CD ' — '
c cn
O C "
M -r- tO
O l/> C
^ O
cu cn +->
-M LO n3
o cn u

-a r—
CD •> Q.
s- cn o_
a. CM =i


CM


CD
S-
                              O-    -^   (M   —  C?
                                                                 (M   —   O
                                                         787

-------
a • -

' 77
5 	
M M rt
N MM

f
H «H ^ LO
^J (M Ul i-4
S 8 >3
»- «
^ X CO M
^ M r-l
J
M
01 £C ^ w4
t~1 (O *4 r4
U

« z « o
H * t-l r*
H

^ ^ ^
u> en eo en
*-* M .H <
Z J
0 /
irt ui m « «
••* N M (/I |
2/
/
» 0 -4 JiO
•H M rt /
/
/

-< IM M "/ "*
1
I
1
•-t CM i-l |
I
1
r-t 111 C W » CO *
 WOfO N Mff»J Ifl . H \ C» IK ! C- | * | « N IM IM N t-l M
I i \ i ' ' I
i I * i ' ' /
n M M M nm. «.MO|»I»' on / M CM H M (M r J CM
i-l M M M * O 03> M 13 A0 / in 1 t~- M r-l r-l < r^ CM F*J IM
M M M M M »x * XU, jn K, , H 2 j_
v*-*. " / ' 1 ^
	 rj r H" HMM -°

' ^-- ^.-^ ^ / w 4->
"""•» '* ^/ rl r-l Cl CM Cl
r4^^^^^li-INrl--<< CD
+ t U~l
d***"' "* =1
I *"
flT>>"MK»^MCMM'NCMIMlHMIM O
1 _J

0)
rf-MMK»fslf»| MCg
fO

S -(->
>- d
5 OJ
f~ ri & ft n r*i ri tv r* *-* r-i t<> tv TM tv (_)
° d
o
u
MMIMIMNrsl r^rHi-*i-4 IMM IMfM
a>
d
o
M fj r, IM n « ^. rt r< rH IM CM CM IM N
O
X
f-l ^
z CD
Ul ^_)
-1 o

1 ~°
J 0)
•- s_
MMHMNMNIM Q_
Dt
Z .
h- M r\J r-J ft fM Z rv-i
E CO
S ^
M M M TJ rj ^~
Z Z!
* .?
< M fO IM 1 1
—
z
II
^1





CO
E
O)
-4->
CO
>,
OO


•
C/}
d
o

4->
•i—
-a
d
o
o
>>
i.
ra •
~o • — ^
d 1 —
13 QJ
0 T3
JD O
y

r—
03 -M
d d
•1 — •! —
E 0
o a.
d
"O
-a -r-
a> s-
u trs
Z3
TD


1 —
" Q-
O-i Q-
CNJ =t








788

-------
                                                            S-     -t->
                                                            QJ     c:
                                                            -O     T-
                                                            E r-  O
                                                            O>  re  Q-
                                                                   S-
                                                                  CD
                                                            CO
                                                            •r- -O   •
                                                            oo  c:  o
                                                            re  03  c
                                                            -Q     1—1
                                                                oo
                                                            oo  C   "
                                                            01  o  E
                                                            1— T-  O
                                                            O> 4~>  -i-
                                                            OVi-  4-J
                                                            c ~o  re
                                                            eC  C  U
                                                                O  T-
                                                            00  O  I—
                                                            O     Q-
                                                            _J  >, Q-
                                                                S_    C  E
                                                                Z!  O)
                                                            C  O  -4->
                                                            •i- JD  00
                                                                   >,
                                                            oo i—  oo
                                                            c  re  -—
                                                            o  c
                                                             re  o c?s
                                                             s-  c o
                                                            4->     UO
                                                             E T3
                                                             OJ  O)  >,
                                                             U  O -Q
                                                             E  3
                                                             O TD -O
                                                             U  CU    o
                                                            +J UD -r-
                                                             o cn -M '—

                                                            -5     -o  CD
                                                             O)    "  E "O
                                                             s- en  o  o
                                                            D_ CXI  O 21
                                                             Ol
                                                             s_
                                                             CD
789

-------
    \        \         \
                                                                                     OD<
o
0
                                                                                                       o
                                                                                                       en
                      CM _
                            CM
                         2
                      {J OC o
                      _: O _:

                      OD<
                                                                                                                         CT>
                                                                                                                         C\J

                                                                                                                          S- CD
                                                                                                                          CU -C   •

                                                                                                                         "i "^^
                                                                                                                          CD C  CD
                                                                                                                         +-> •!- "O
                                                                                                                          Q.     O
                                                                                                                          cu c s:
                                                                                                                         to o
                                                                                                                             •I-  >)
                                                                                                                          C -M  5-
                                                                                                                         •r— ro  O
                                                                                                                          to S	P
                                                                                                                          05 +->  U
                                                                                                                         CO C  CD
                                                                                                                             CD •<->
                                                                                                                          to O  (13
                                                                                                                          CD C  5-
                                                                                                                         r— O f—
                                                                                                                          O) O
                                                                                                                          en
                                                                                                                          c: i—  Q-
                                                                                                                           I/I
                    C      CD
                    O ^S Qi

                   4-> CXI i—
                    ro  +\  fO

                   4-> "O  CD
                    c  c  c:
                    CD  03  CD
                    O     CD

                    O  ro
                    o  c
                      •I—   •
                    CD  CD.—

                    O  S-  CD
                    MOO
                    O
                       cn >,
                   TD  C  S-
                    CD
                                                                                                                                 O
                                                                                                                          •r-     CD
                                                                                                                          ~O CTi T-3
                                                                                                                          CD UD 03
                                                                                                                          S- CTl i-
                                                                                                                          Q- i— 1—
                   LD

                    CD



                    01
O         O
CD         CO
                                            uidd'[CQ]
                                                                 790

-------
has proved to be a considerable deterrent in the application of the more com-
plex models, due to their resource requirements and the unverified status of
their results.  The RAPS will provide the first adequate data set to allow
model performance based on statistical criteria.

     The monitoring network that came on line late in 1974 consists of twenty-
five monitoring stations spatially distributed in a spiral configuration
covering approximately a 6400 km2 area.  Routine chemical measurements include:
03, NO, nitrogen oxides (NO ), CO, methane (CH^j, total  hydrocarbon, total
sulfur, S02, hydrogen sulfiae (H2S), total suspended particulate and dichotomous
particulate samples.  Physical parameters measured on a routine basis include:
wind speed and direction, temperature, dew point, vertical temperature gradi-
ent, pressure, solar radiation, and aerosol back scattering coefficient.  Data
tapes are routinely prepared based on one-minute and hourly averages of the
measured parameters.

     A schematic view of the elements of the RAPS air quality simulation model
verification program is given in Figure 6.  Since the verification of each and
every urban-scale air quality simulation model  currently in existence is an
unrealistic task, the decision was made to select judiciously from the model-
ing stockpile a representative sampling of modeling techniques.  The selected
models will be then adapted to the St. Louis region and three test days will
be considered for model debugging and tuning.  Models will then be turned over
to the Agency for the verification study.

     The selection of days from the RAPS data base to form a subset of verifi-
cation days to perform the studies will be done via a series of filtering
programs.   The filtering criteria may include such tests as:

     o     days in which criteria pollutant X had Y% or greater valid data,

     o     days in which criteria pollutant X exceeded an hourly average
           value of Y ppm;

     o     days in which criteria pollutant X exceeded a 24 hour average
           value of Y ppm.

The specific numbers to be used by the filters  are currently being developed.
The subset of verification days will  then be stratified by season.  Non-
reactive models will be tested Against randomly selected days within each
season.   An attempt will  be made to distribute  the verification test days pro-
portionately among the four seasons.   The reactive models will  use randomly
selected days, but in a disproportionate manner with respect to season.  The
larger sample selections will come from the summer months, and lesser sample
selections will come from the spring, fall, and winter seasons.

     After the selection of the model verification test days, a five-day subset
will be chosen and input data to the models for those days prepared.  The five-
day subset will be used for preliminary testing and should allow spotting any
major model inconsistencies.   Any obvious minor model refinements at this stage
will be considered.  Since all the models selected will  have undergone some
form of evaluation and testing, major unresolvable inconsistencies are not ex-
                                      791

-------
    SELECTION OF
    VERIFICATION
       DAYS
                                                    MODEL
                                                  SELECTION
     1
OBSERVATIONAL
    DATA
     INPUT DATA
    PREPARATION
                                                                1
                            PRELIMINARY
                            MODEL TESTING
                                                    MODEL
                                                  ADAPTATION
                            PRELIMINARY
                            PERFORMANCE
                            EVALUATION
                                                           REFINEMENTS
                                I
                             STATISTICAL
                            VERIFICATION
                               MODEL
                            EVALUATION
                                                           SENSITIVITY
                                                            ANALYSIS
                                                          MINIMUM DATA
                                                          REQUIREMENTS
Figure 6.
Elements
Program.
                   of the RAPS Air Quality Simulation Model Verification
                                  792

-------
pected.  Should such a case arise, the verification studies for the particular
model in question will stop and the preliminary test results will be sent to
the model developer for study.

     Once the preliminary model testing phase is completed, the statistical
verification studies will begin.  Approximately 50 computer simulation days
will be considered for each model.  Comparisons between computer predicted and
observed hourly averaged concentrations of pollutants will be made on a day-
by-day basis.  Specific statistical tests for model verification studies have
been discussed by Brier(13), Nappo(14), and Liu et al.(ll).  In light of these
studies, several  approaches are under consideration for the comparative
studies.  Model verification will provide the first adequate set of statisti-
cal criteria on which to judge model performance and will, in the final analy-
sis, provide the necessary information for assessing and selecting models for
various application purposes.
 1.
 2.
 3.
 4.
 5.
 6.
 7.
Seinfeld, J.H. and S.D.
Adv. Chem. Series 113.
1972.  pp. 58.
      REFERENCES

Reynolds.  Simulation of Urban Air Pollution
American Chemical  Society, Washington, D.C.
Seinfeld, J.H.  Accuracy of Prediction of Urban Air Pollutant Concen-
trations By Diffusion Models, Assessing Transportation-Related Impacts.
Special Report 167, TRB.  National Academy of Sciences.  Washington,
D.C., 1976.  pp. 34.

Demerjian, K.L.  Photochemical Diffusion Models For Air Quality Simulation:
Current Status, Assessing Transportation-Related Air Quality Impacts.
Special Report 167, TRB.  National Academy of Sciences.  Washington, D.C.
1976.  pp. 21.

Reynolds, S.D., P.M. Roth, and J.H. Seinfeld.  Mathematical Model of
Photochemical  Air Pollution:  I.  Formulation of the Model.  Atmos.
Environ. 7.  1973.  pp. 1033.

Roth, P.M., P.J.W. Roberts, M.K. Liu, S.D. Reynolds, and J.H. Seinfeld.
Mathematical  Modeling of Photochemical Air Pollution:  II.  A Model and
Inventory of Pollutant Emissions.  Atmos. Environ., 8.  1974.  pp. 97.

Reynolds, S.D., M.K. Liu, T.A. Hecht, P.M. Roth, and J.H. Seinfeld.
Mathematical  Modeling of Photochemical Air Pollution:  III.  Evaluation
of the Model.   Atmos. Environ., 8.  1974.  pp. 563.

Eschenroeder,  A.Q., and J.R. Martinez.  Concepts and Applications of
Photochemical  Smog Models.  Adv. Chem. Series, Vol.  113.  American
Chemical Society, Washington, D.C.  1972.  pp. 101.

MacCracken, M.C., and G.D. Sauter, eds.  Development of an Air Pollution
Model for the  San Francisco Bay Area.  Final  Report to the National
                                     793

-------
      Science Foundation.   University of California,  Livermore,  California.
      1975.

 9.    Sklarew, R.C., A.J.  Fabrick,  and J.E.  Prager.   A Particle-in-Cel1  Method
      for Numerical  Solution of the Atmospheric Diffusion Equations,  and Appli-
      cations to Air Pollution Problems.   Systems,  Science,  and  Software, Inc.
      La Jolla,  California.   1971.

10.    Hanna,  S.R.   A Simple  Dispersion Model  for the  Analysis of Chemically
      Reactive Pollutants.   Atmos.  Environ.,  7.  1973.  pp.  803.

11.    Liu,  M.K., D.C.  Whitney, J.H. Seinfeld,  and P.M. Roth.   Continued  Research
      in Mesoscale Air Pollution Simulation  Modeling:   Vol.  I -  Assessment of
      Prior Model  Evaluation Studies and Analysis of  Model  Validity and  Sensi-
      tivity.  EPA Report  600/4-76-016a.   Systems Applications,  Inc.,,  San Rafael,
      California.

12.   Pooler, F. Jr.  Network Requirements for the St. Louis Regional Air
      Pollution Study.  J. Air Poll. Control Assoc.,  24.  1974.  pp.  228.
      see also  Burton, C.S., and G.M. Hidy.   Regional Air Pollution Study
      Program Objectives and Plans.  EPA-650/3-75-009.  Rockwell International
      Air Monitoring Center, Newberry Park,  California.  1974.

13.   Brier, G.W.   Statistical Questions Relating to the Validation of Air
      Quality Simulation Models.   EPA Report 650/4-75-010.  1975.

14.   Nappo, C.J. Jr.  A Method For Evaluating the Accuracy of Air Pollution
      Prediction Models.  Proceedings of Symposium on Atmospheric Diffusion
      and Air Pollution, American Meteorology  Society, Santa Barbara, Calif.
      1974.
                                     794

-------
                                                                          16-2
         THE SYSTEMS APPLICATIONS, INCORPORATED URBAN AIRSHED MODEL:
                  AN OVERVIEW OF RECENT DEVELOPMENTAL WORK

                               S. D.  Reynolds*


ABSTRACT

     Jk paper summarize* work on pkotoc.hemic.al model development carried out
at System* Application*, Incorporated over the pa*t three year*.  A* a revolt
o& thi* work, the original System* Application*, Incorporated Urban Airshed
Model ha* undergone *-ign^^i,cant modification and re^-inement.  The*e refine-
ments are di*cu**ed, and plan* &or a**e**
-------
of the validity of grid.and trajectory modeling concepts  (Liu  et al.,  1976).
The following sections are devoted to discussions  of the  overall model  formu-
lation and the treatment of its components.


                      GOVERNING EQUATIONS OF THE MODEL

     The fundamental equation governing the  concentration of each pollutant
species is the so-called mass continuity equation.   After carrying out  several
mathematical operations and invoking a number of assumptions,  this equation
can be written in the following form:


         3   _  3          3   _           „

         -if+ ui  -iif-' 4 Ki  Txf-+ Y   +-      <]'


where the symbols <   > and ~ represent ensemble and  spatial averaging opera-
tions, respectively, and

     C  = the approximate value of the instantaneous concentration of pollutant
          a at a point*

     U- = mean wind  velocity components

     K. = eddy diffusivity components

     S  = rate of emission of pollutant £ from elevated sources

     R  = contribution of the mean  concentrations  to the  net reaction rate  of
          pollutant  £

     R' = contribution of the turbulent concentration fluctuations to the net
          reaction rate

     R" = contribution of subgrid-scale concentration variations to the net
          reaction rate.

Ground-level emissions and pollutant uptake  processes are treated in the
boundary conditions.

     Because the governing equations are nonlinear,  they  must  be solved numer-
ically.  Furthermore, the use of numerical techniques generally  requires that
the modeling region  be subdivided into an array of three-dimensional grid
cells, where each cell may have horizontal and vertical dimensions on the
order of a few kilometers and several tens of meters, respectively.  Before
the general mass continuity equation can be  solved,  it must be "filtered" to
remove all small scale variations that the grid cannot resolve,  both in the
*Because a number of assumptions  have been  made  in  the  derivation  of Equation
 1, it is only a model  of the "true"  equation.   We  wish to emphasize its
 approximate nature.

                                     796

-------
concentration field and in the independent parameters, such as K.  and U..
The necessary filtering can be accomplished by averaging the continuity1 equa-
tion at each point over a volume equivalent to that of a grid cell.   This
spatial averaging is denoted by the symbol ~ in Equation 1.   In addition,
Equation 1 has been time-averaged over an interval equivalent to that used in
each step of the numerical solution procedure.  Thus, the concentration pre-
dictions obtained from a grid model represent spatially- and time- averaged
quantities.

     To clarify the origins of the three reaction terms, we note that they
might typically be of the form:


         
-------
                           COMPONENTS OF THE MODEL

CHEMISTRY

     A chemical kinetic mechanism is one of the most important components  of a
model for predicting concentrations of photochemical pollutants.,   A 31-step
mechanism, called the carbon-bond mechanism, has been developed at SAI by
Whitten and Hogo (1976) and adapted for use in the Urban Airshed Model.

      Because of the association of reactions and reactivities with carbon
bonds, the range of reactions and the range of rate constants in a kinetic
mechanism can be narrowed considerably when each atom is treated according to
its bond type.  This concept is the basis for the carbon-bond mechanism.  In
this mechanism, hydrocarbons are divided into four groups:  single-bonded
carbon atoms, fast double bonds (i.e., relatively reactive double bonds),  slow
double bonds, and carbonyl bonds.   Single-bonded carbon includes not only
paraffin molecules, but also the single-bonded carbon atoms of olefins, aro-
matics, and aldehydes.  Double bonds are treated as a pair of carbon atoms.
An activated aromatic ring is considered as three double bonds in the present
formulation of the mechanism, and because of a similarity in reactivities,
aromatics are lumped with the slow (ethylene) double bonds rather than with
the fast double bonds.  The carbon-bond mechanism has been applied to available
smog chamber data for propylene/NOx, l-butene/NOx, butane/NOx, propylene/
butane/MOx, and toluene/NOx systems by shitten and Hogo (1976).  Test results
indicate that this mechanism performs significantly better than those previously
employed in the Urban Airshed Model.


     Also included in the mechanism are five reaction steps describing S02
oxidation  and two reaction steps  accounting for the formation of organic  and
nitrate aerosol products.  Sulfate aerosol products are assumed to form as a
result of the gas phase reactions  leading to the production of sulfur trioxide
(S03), which rapidly combines with water to give sulfuric acid (H2S04).  We
have assumed that the ozone-olefin reaction is the limiting step in the pro-
duction of organic aerosol, and that only olefins with at least six carbon
atoms participate significantly in this process.  Nitrate aerosol production
is assumed to be proportional to the N02 concentration.   Further details of
the treatment of chemistry in the Urban Airshed Model may be found in the
report by Reynolds et al. (1976b).

ADVECTION

     In the original version of the Urban Airshed Model, the horizontal  wind
components were assumed to be invariant with height.  Sensitivity studies
reported by Reynolds et al. (1976a) indicate that inclusion of wind shear
effects in the model can alter N02 and 03 predictions significantly.  Conse-
quently, we expanded the capabilities of the model to treat wind shear
phenomena.

     Theoretical wind shear relationships were derived by Lamb (1976a) using
the predictions of a planetary boundary layer model developed by Deardorff
(1970).   These relationships will  be useful in instances when wind measure-


                                     798

-------
ments aloft are not available.  For situations in which wind measurements are
available both at the surface and aloft, we developed an objective technique
for preparing appropriate three-dimensional wind inputs to the model.  [See
Reynolds et al. (1976b)].  This technique is unique in that it incorporated
provisions for treating perturbations to the flow field caused by thermal
effects.
TURBULENT DIFFUSION

     Eddy diffusivity coefficients have been derived through use of a novel
methodology developed by Lamb (1976a).  In this procedure, a point source is
assumed to emit an inert pollutant into a turbulent boundary layer.  Employing
flow fields predicted by the model of Deardorff (1970), particles are released
from a point and followed as they are transported downwind.  From the particle
trajectories, it is possible to calculate the pollutant concentration field
downwind of the release point.  Given the concentration and mean flow fields,
we "invert" the mass continuity equation and solve for the diffusivity profile
through use of optimal control theory techniques.   Preliminary diffusivity
results are reported by Lamb (1976a) and the final relationships incorporated
in the Urban Airshed Model by Lamb (1976b).


OTHER MICROSCALE PHENOMENA

     The diffusivity work cited above is but one example of a group of studies
dealing with the treatment of phenomena that have characteristic spatial or
temporal scales too small to be resolvable in an explicit, deterministic
manner in urban-scale models.  The two primary objectives of our microscale
work have been (a) to identify and parameterize microscale effects that influ-
ence the spatially-averaged pollutant concentration field (such as that pre-
dicted by a grid model) and (b) to develop means for predicting the distribu-
tion of pollutants in the immediate vicinity of point and line sources.   In
order to make significant headway in this area, it was first necessary to
develop an appropriate mathematical framework for treating microscale phenom-
ena.  This framework is described by Lamb (1976a).

     To achieve the objectives cited above, we developed three microscale
modules for use in the Urban Airshed Model.  A point-source module treats the
effects of plume rise, the influence of turbulence on chemical reactions in
plumes (i.e., )? and subgrid-scale concentration variations and their
influence on the overall net reaction rate in each grid cell (i.e.,).
Moreover, this module can be used to estimate the "point" value of the pollu-
tant concentration field downwind of a point source.  The second module treats
the subgrid-scale chemistry effects resulting from the emissions of a network
of line sources in the groundlevel grid cells.  In this layer of cells, taken
to be about 20 meters deep, NO emissions from a roadway are allowed to react
only with 03 in the immediate vicinity of the roadway.  The third microscale
module estimates the concentrations of photochemical pollutants along highways
and in street canyons.  In this module, consideration is given to upwind
pollutant concentrations, vehicular emissions, turbulent wake effects, chem-
istry, transport, and the design of the roadway (elevated, at-grade, or


                                      799

-------
depressed).  A more detailed description of these modules is given by Lamb
(1976b).


SURFACE REMOVAL PROCESSES

     A methodology for calculating the removal of gaseous pollutants by sur-
face sinks was incorporated in the Urban Airshed Model.  Surface removal was
assumed to take place in two steps — transport to the surface followed by
adsorption on the surface.   Parameterization of this two step procesis was
accomplished by defining a resistance to mass transport (the reciprocal of the
tranport velocity) and a resistance to surface removal (the reciprocal of the
deposition velocity measured under conditions when transport is not a limiting
factor).  The transport resistance is estimated from the theory of turbulent
transfer in the atmospheric boundary layer, and the surface resistance is
obtained from experimental  data on the uptake of pollutants by various types
of surfaces.

     Although deposition rates of S02 have been measured under a relatively
wide range of conditions, only a few types of surfaces have been examined.
For other pollutants, few measurements are available.  For the present, we
made crude estimates of the uptake resistances of typical urban surfaces for
all pollutants. Model sensitivity studies based on these estimates indicate
that surface removal processes may have significant effects on the predicted
concentrations of S02, N02, and 03.   This work is reported by Reynolds et al.
(1976b).


NUMERICAL INTEGRATION METHOD

     Refinements to the numerical integration method used to solve the govern-
ing equations have focused on the implementation of a more accurate technique
for computing horizontal transport.   Tests of a number of techniques by Reynolds
et al.  (1976a) indicated that the SHASTA, Egari and Mahoney, and finite element
methods all offered the promise of significantly suppressing the propagation
of numerical  errors that typically plague grid models.

     In making the final selection of a method to be included in the Urban
Airshed Model, the finite element method was abandoned because it could not be
adapted to the solution of 13 coupled, nonlinear partial differential equa-
tions in three spatial dimensions without a rather large expenditure of effort
to develop an appropriate computer code.   Further testing and evaluation of
the remaining methods by Reynolds et al.  (1976b) resulted in the selection of
the SHASTA method.  Although the Egan and Mahoney method was somewhat more
accurate, the SHASTA method was found to require significantly less computing
time.
                                   SUMMARY

     Developmental work carried out during the past three years has resulted
in significant refinements of many components of the SAI Urban Airshed Model.

                                     800

-------
At the present time there is a clear need to evaluate the accuracy of the
model's predictive capabilities.   To make this assessment,  we are currently
adapting the model for usage in St.  Louis.   An important element of this work
will be the design of appropriate means to interface the model with the
Regional Air Pollution Study (RAPS)  data base.  When this work is completed,
it will be possible to perform extensive evaluative studies.   In addition,  the
computer codes are being revised  somewhat to make them more efficient and
easier to implement on a variety  of host computers.  Documentation for the
codes is also being prepared as an aid to the potential  user of the new
version of the Urban Airshed Model.
                               ACKNOWLEDGMENT

     The following members of the SAI Environmental  Studies Group have made
significant contributions to the developmental work  summarized in this paper:
J. Ames, G. E. Anderson, D.  R. Durran, T.  A.  Hecht,  H.  Hogo, T.  N.  Jerskey,
J. P.  Killus, R.  G. Lamb, M.  K.  Liu, J.  P.  Meyers, P.  M.  Roth, p.  C.  Whitney,
G. Z.  Whitten, and M.  A. Yocke.   Funding for this work was provided by the
Environmental Protection Agency (Contracts  68-02-1237 and 68-02-2216).


                                  REFERENCES
 Deardorff, J. (1970), "A Three-Dimensional  Numerical  investig
      Idealized Planetary Boundary Layer," Geophys.  Fluid Dyn.
      pp. 377-410.
Investigation of the
 ' '  ~     Vol.  1,
 Lamb, R.G. (1976a), "Continued Research in Mesoscale Air Pollution
      Modeling: Volume Ill—Modeling of Microscale Phenomena," EPA-600/
      4-76-016c, Systems Applications, Incorporated, San Rafael, Calif.

 _______  (1976b), "Continued Development and Validation of a Second
      Generation Photochemical Air Quality Simulation Model: Volume II—
      Microscale Modeling Studies," Final Report for EPA Contract
      68-02-2216, Systems Applications, Incorporated, San Rafael, Calif.
      (in preparation).

 Liu, M.K., D.C. Whitney, J.H. Seinfeld, and P.M. Roth (1976), "Continued
      Research in Mesoscale Air Pollution Simulation Modeling: Volume I—
      Analysis of Model Validity and Sensitivity and Assessment of Prior
      Evaluation Studies," EPA-600/4-76-016a, Systems Applications,
      Incorporated, San Rafael, Calif.

 Reynolds, S.D., M.K. Liu, T.A. Hecht, P.M. Roth, and J.H. Seinfeld (1973),
      "Urban Airshed Photochemical Simulation Study: Volume I—Development
      and Evaluation" (with appendices), EPA-R4-73-020a-e, Systems
      Applications, Incorporated, San Rafael, Calif.

 Reynolds, S.D., J. Ames, T.A. Hecht, J.P. Meyer, D.C. Whitney, and
                                     801

-------
     M.A. Yocke (1976a), "Continued Research in  Mesoscale Air Pollution
     Simulation Modeling:  Volume II--Model  Development and Refinement,"
     EPA 600/4-76-016b,  Systems Applications,  Incorporated,  San  Rafael,
     Calif.

Reynolds, S.D., et al.  (1976b), "Continued  Development and Validation  of
     a Second Generation Photochemical  Air  Quality  Simulation Model:
     Volume  I — Refinements in the Treatment, of Meteorology,  Chemistry,
     Pollutant Removal  Processes, and Numerical  Analysis," Final  Report
     for EPA Contract  68-02-2216, Systems Applications, Incorporated,
     San Rafael, Calif,  (in preparation).

Roth, P.M.,  S.D. Reynolds, P.J.W. Roberts,  and J.H.  Seinfeld (1971),
     "Development of  a  Simulation Model  for Estimating Ground Level
     Concentrations of  Photochemical  Pollutants," Report 71-SAI-21,
     Systems Applications, Incorporated,  San Rafael,  Calif.

Whitten, G.Z., and H.  Hogo (1976), "Mathematical Modeling of Simulated
     Photochemical Smog" Final  Report for EPA Contract 68-02-0580,
     Systems Applications, Incorporated,  San Rafael,  Calif.
                                   802

-------
                     SESSION 17
MATHEMATICAL MODELS OF OZONE/OXIDANT AIR QUALITY - II

                Ck&Uu/nan:  K. Demerjian
            Environmental Protection Agency
                         803

-------
                                                                           17-1
              A SURVEY OF APPLICATIONS  OF  PHOTOCHEMICAL  MODELS

                              J.  E. Summerhays*
ABSTRACT
     Photoc.hnmic.at mode.ts o& several types are identi^ed  and  their applica-
tions (Via fie.V4.iW&d.  The.  faocus u>  on the.  various  application*  oft  the. mono.
sophisticated models, -i.e.., Vi&kin and the. Systems  Applications,  Incorporated
model,.  In particular, an emission induction study  by the.  author  using  the.
Systems Applications, Incorporated model  is reviewed.   This  study su.gge.sts
that hydrocarbon control.  (1) can have, a dramatic  e^ect in re.du.cing ozone, con-
centrations, (2) tends to delay the. time,  ofa peak  ozone. concentration, and (3)
moves the. area o^ highest concentrations  farther  downuiind.


                                INTRODUCTION

     There is much yet to be learned about the chemical  and  meteorological
processes involved in the production of tropospheric  ozone (03).   Nevertheless
numerous mathematical models that  estimate 03 concentrations have already been
developed.  Several of these models have  been used  in validation  studies,
i.e., studies comparing model estimates to observed concentrations.   On the
other hand, few of these models have been applied in  studies assessing  the
impact of changes in hydrocarbon (HC) and nitrogen  oxides  (NOX) emissions.
This paper is intended as an overview of  these applications, including  an
extensive review of an unpublished application by the author (1).

     Two models in particular, the "SAI Model" (2)  (developed  by  Systems
Applications, Inc.) and Difkin (3) (developed by  General Research Corp.)  have
been applied in a wide range of studies.  These models fit into a sophisti-
cated class of models called photochemical dispersion models,  giving explicit
consideration to the chemistry, advection, and diffusion of  the pollutants in
a photochemical system.  Other photochemical dispersion  models have been
developed by Lawrence Livermore Labs (4), by Pacific  Environmental  Services
(5), and by Systems, Science and Software (6).

     Many simpler photochemical models have also  been developed.   These in-
clude three statistical models:  (a) a rollback model, developed  by John
Trijonis  (7), (b) a "repro-model"  which is a statistical condensation of
results using the SAI model, developed by Technology  Services  Corp.  (TSC) (8),
and (c) a statistically-derived short-term predictive model, developed  by
*U.S.  Environmental Protection Agency, Durham, North Carolina.

                                     805

-------
Tiao, Phadke and Box at the University of Wisconsin (9).   Also, three photo-
chemical box models have been developed:   (a) several  versions of Hanna and
Gifford's box model that are adapted to consider photochemical pollutants (10,
11, 16), (b) a model developed by S. Hameed and others at NASA's Goddard Space
Flight Center (12), and (c) a model being developed by Ken Demerjian (this
model has not been described in the literature).   A model that is difficult to
categorize is the one developed by Chameides and Walker (13).   This model is
designed to estimate rural 03 concentrations based solely on measured rural
methane and NOX concentrations, assuming no further dispersion.  Two other
models that have been used to estimate 03 concentrations  in a large number of
studies are linear rollback (14) and Appendix J (15).   However, these two
models are simplified models and are not considered in this study.

     Validation studies are also not considered in this paper.  Nevertheless,
the reader may wish to refer to validation studies in  Los Angeles for the SAI
model, Difkin, the Pacific Environmental  Systems (PES) model,  the Systems
Science Software (SSS) model, and the Hanna-Gifford model, reported respec-
tively in References 2, 3, 5, 6, and 10.   Validation studies have also been
conducted for Difkin and Hanna-Gifford in Denver (16), for Difkin in San
Francisco (17), and for the Lawrence Livermore Lab model  in San Francisco.
                             MODEL APPLICATIONS

     Table 1 summarizes the applications that have been made of photochemical
models.  As may be seen from this table, the applications of Difkin are more
numerous and variegated than the applications of any other model.   One of the
first studies conducted using Difkin (18) is a type of study that should be
conducted with all photochemical models —estimating the impact of a general
reduction of HC and NOX emissions.   The study conducting this analysis using
Difkin also assessed the importance of the initial concentrations, i.e., the
concentrations assumed at the beginning of the model run.  This study showed a
strong dependence of estimated 03 concentrations on whether or not the initial
concentrations were reduced, i.e.,  whether or not emissions reductions began
prior to the beginning of the modeled time period.  The trajectories; were not
long enough for 03 concentrations to reach their potential peak values, but
in any case, a comparison of highest values attained showed that a 30% general
emissions reduction without any change in initial concentrations resulted in
only a 16% reduction in estimated 03 concentrations.

     This study also found some interesting results from a study of various
patterns of reducing HC and NOX initial concentrations.  Not unexpected was
the result that reducing the HC/NOx ratios reduced the estimated 03 concentra-
tions.  However, in one case in which both HC and NOX initial concentrations
were reduced without changing their ratio, the estimated 03 concentrations
actually increased.
     Hameed1s box model was used to study the same strategies as modeled by
General Research Corp  (GRC) (12).  His results were more like what would be
expected:   emissions reductions caused a reduction in 03 concentrations simi-
lar to that found by GRC, but a reduction of initial concentrations caused a
sizable decrease in 03 concentrations.
                                     806

-------
TABLE 1.  APPLICATIONS OF PHOTOCHEMICAL MODELS

Model Name or
Developer
Difkin

S. Hameed
Difkin
Difkin
Difkin
Difkin
Difkin
Difkin
Difkin
Difkin
Lawrence Livermore

SAI
SAI
SAI
SAI
Applier
of Model
GRC

Hameed
SRI
GRC
GRC
SRI
ERT
Cal trans
GRC
GRC
LLL

EPA
SAI
TSC
EPA
Reference
Number
18

12
17
19
20
21
22
24(& 23)
(23 only)
25(& 23)
4

1
26
8
27
Subject of
Application
4 different emission reduction stra-
tegies in Los Angeles and a few
initial concentration reductions
The same 4 strategies as above
1980 Concentrations in San Francisco
"Cold Start" emissions
Substituting electric cars for
gasoline cars
A Los Angeles emergency episode plan
Non -conventional fuel recovery plants
A proposed freeway
A refinery
Various growth rates in Santa Barbara
Various patterns of changes to
emissions
Various emission reduction strategies
for 6 days in Los Angeles
A proposed transportation control
plan for Los Angeles
Developing a repro-model for 9/30/69
in Los Angeles
Various radiation levels
                      807

-------
     As mentioned earlier, Table 1 shows the wide variety of other topics that
have been studied using Difkin (17, 19-25).  These studies estimate the effect
on 03 concentrations of a variety of causes of change in HC and NOX emission
levels, including addition of new sources, alteration of old sources, and
growth in the number of sources.   These studies usually involve two steps:
(a) estimating the effected change in emission levels, and then (b) estimating
the effect of this change in emission levels on 03 concentrations.   For example,
"cold start emissions" were estimated to cause about 8 to 15% increase in
emissions and no change in initial concentrations.  This change in  emissions
was then estimated to cause a 2 to 7% increase in peak 03 concentrations.
Similarly, for many of the other studies, the change in emission levels is
first estimated, and then the effect on estimated 03 concentrations due to
that change is typically found to be less than but in general  proportion to
that change.   Note that this result is in general accord with the results of
the general emission reduction study.

     Note that the final three studies using Difkin are all  cited at least in
part as Reference 23.  The summary by A. Q. Eschenroeder of photochemical
modeling studies appears to be the only available source of information on
these studies.

     The Lawrence Livermore Lab (LLL) model (4),  Liraq-2, was developed more
recently than most of the other photochemical models.  This  model  is compara-
ble in many ways to the SAI model, but has been adapted to San Francisco, not
Los Angeles,  and treats the vertical concentration distribution and the chem-
istry quite differently.  Application studies using Liraq-2 (also discussed in
Reference 4)  indicated that emission reductions upwind, in San Francisco, had
a substantial effect of reducing Livermore 03 levels.  Another study suggested
that the addition of a refinery in the northern Bay area would have little
relative effect on emissions and correspondingly, little impact on  air quality.
In a third study, a 40% increase in nitric oxide  (NO) emissions decreased the
03 concentrations in the source regions, but the  03 concentrations  achieved
later in the  downwind portions of the Bay area were not substantially changed
by this change in emissions.

     The SAI  model has not been used for as great a variety of studies as
Difkin, but three different groups have studied the effect of a general emis-
sion reduction according to the SAI model.  Two of these studies have been
published (8, 26), and one, done by this author,  is unpublished (1).   In
addition, an  interesting study was published recently analyzing the importance
of changes in solar radiation (27).  Significant  changes to the amount of
solar radiation due to changes in season or addition of cloud cover generally
caused an equally significant change in 03 concentrations.   However,, temporary
morning cloudiness seemed to have relatively little effect on 03 concentrations
downwind of the area of greatest emissions.


                         AN EMISSION REDUCTION STUDY

     As noted on Table 1, the three studies of general emission reductions
were conducted by Summerhays of the Environmental Protection Agency (1), SAI
(26), and TSC (8).  The results of the study by Summerhays have not been pre-

                                     808

-------
viously published, and so this study is discussed at some length.
of the other two studies fit in well with the author's results.
                                                 The results
     In the first part of the study by Summerhays, the model was operated
assuming no emission reductions for each of the six days for which sufficient
data is available to run the model.  These six days are all in the 1969 "smog
season" in Los Angeles.  Then the peak estimated 03 concentration at each
fifteen sites was compared with the peak actually measured at each site.
                                                        of
     Table 2 provides statistics showing this comparison separately for each
day.  Ideally, of course, one would find Y = 1.00 x + 0.0 with a correlation
coefficient of 1.0.  It will be noted that in some cases the model estimates
quite accurately and in other cases the model does not estimate as accurately.
It is interesting then to look at the wind pattern for each of these days.  A
typical pattern in the fall in Los Angeles is for a southwesterly Seabreeze to
develop during the day.  The model seems to do better on days when this pattern
occurs than when it does not.  The findings discussed below are also typically
best illustrated on those days when this typical pattern developed, although
results on the other days do also support the findings.   In any case, it is
obviously wise to avoid basing studies solely on one day of meteorological
conditions.
     TABLE 2.   CORRELATION STATISTICS BETWEEN PREDICTED AND OBSERVED OZONE
                                   PEAKS (in pphm)
Day
The Best Fit Line*
Correlation
Coefficient
9/11

9/29

9/30

10/29

10/30

11/4
Y =  .29 x +  9.96

Y = 1.01 x +   .35

Y =  .68 x +  1.15

Y = 1.03 x +  2.27

Y = -.38 x + 13.29

Y = 1.22 x -  1.81
*Y = Model Prediction, x = Observation
**Statistically significant at the .05 level
    .27

    .88**

    .71**

    .63**

   -.26

    .48
     In the second part of the study by Summerhays, nine different strategies
involving various combinations of HC and NOX emission reductions were tested.
                                     809

-------
These results were then compared to the case vn'th no emission reductions,
i.e., the actual 1969 emissions.  One of the interesting results of this study
was that the first increment of HC control caused a more than proportional
reduction in 03 concentrations at "downwind stations" and very little effect
at "upwind stations."  Here downwind and upwind are relative to downtown Los
Angeles, the center of most emissions in the Los Angeles area.  Figure 1 shows
the peak estimated 03 for four strategies at twelve selected sites on Septem-
ber 29.  It can be seen from this graph that, at the downwind stations, the
initial 50% HC control causes generally in the range of 65 to 85% reductions
in peak estimated 03.  Subsequent emission control then seems to reduce peak
estimated 03 concentrations by a proportion similar to the emission reduction.
However, estimated 03 concentrations at the upwind stations seem to be unaf-
fected by even the most stringent emission reduction strategies, implying that
this 03 was primarily advected from the region upwind of the airshed boundary.

     In contrast to the situation for HC, control of NOX had very little im-
pact on 03 levels.  Figure 2, showing the impact of nine control strategies at
five selected sites, demonstrates that NOX control may cause either an in-
crease or a decrease in estimated 03 concentrations.  In any case, the effect
is relatively minor.

     One of the main advantages of using a photochemical model to estimate the
effect of changes in emissions is the ability to consider spatial and temporal
variations both in the inputs and in the estimated concentrations.  No changes
were made to the spatial or temporal distribution of emissions in this study,
but the emissions control caused some interesting changes in the temporal
and spatial distribution of estimated 03 concentrations.  In particular,
reductions in HC not only caused reductions in 03 peaks but also delayed these
peaks and caused a shift of the region of highest concentrations farther down-
wind.  For September 29, when this effect was most marked, the peak at any one
site typically occurred two hours later, and the peak concentrations were
found over ten miles farther downwind after controls compared to before
controls.


                                   SUMMARY

     This report has reviewed several interesting applications of photochemi-
cal models.  The emission reduction studies using the SAI model, Difkin, and
other models, and the variety of special studies using Difkin, all serve to
illustrate the potential for the use of photochemical models for estimating
the impact of changes in emission levels.

     This report would not be complete, however, without emphasizing the
developmental nature of photochemical modeling at this time.  The SAI model,
especially, has been undergoing extensive changes, including revamping the
consideration of chemistry, and Difkin will be undergoing some of the same
changes.  In this context, it is important to remember that the studies sum-
marized in this report demonstrate what the respective models find to be the
effect of certain changes, not necessarily what would be found in the real
world.  Presumably as photochemical modeling is further researched, these
models will have increasing reliability in simulating reality.  As we seek to


                                     810

-------An error occurred while trying to OCR this image.

-------
~O *~~~~ CI
C -0 -r-
•r- C 3
3 -r- Q.
0- 3 ^- 3
— 3 C -0 — -
-o 3 C
^— » c S- O -r- oo
"O -i- 
c 3 -C Q- •—
•r- C +-> S- 3 CU
3 3 -i- O 	 CD
C O O) C C
cn 3 -o c -c: =t
c o --' — o
•i— "O as in
-O — -* E O) O
O C ••- CQ —J
O (O OJ (L>
(/) _Q _£Z Cr> -l->
c: 3 S- 
CM



tJ *i
0

o LO o













r
O

00
O

4->
O)
o
o
0 •
CO
n a>
O -1-
cn
^i O)
03 -M
Ol rd
CX i-
1 >
TD CO
CD
ro O
E ^~
•r~ -4-!
00 O
 -r-
-C 03
t— >

C\J
OJ
S-

en
1 1



(WHdd) SNOIl\/yiN33N03 ^0 W3d Q31VWI1S3
              812

-------
understand and solve our tropospheric 03 problem, photochemical  modeling
should be an increasingly important tool to be used in achieving that goal
                                REFERENCES
     Summerhays,  J.,  "A  Modeling Study of the Effect of Emission Controls on
     Ozone  Concentrations," Offi-ce of Air Quality Planning and Standards,
     Research  Triangle Park, N.C.  27711 (1975; unpublished manuscript).

     Reynolds,  S.  D.; M.  K. Liu; T. A. Hecht; P. M. Roth; J. H. Seinfeld,
     Urban  Airshed Photochemical Simulation Model Study, EPA-R4-73-030a,
     Systems Applications  Inc. , July 19"73.

     Eschenroeder, A. Q.;  J. R. Martinez; R. A. Nordsieck, Evaluation of a
     Diffusion  Model  for Photochemical Smog Simulation, EPA-R4-73-012a,
     General Research Corp. /October 1972.

     McCracken, M.  C.; D.  J. Wuebbles; J. J. Walton; W. H. Duewer, and
     K.  E.  Grant,  The Li vermore Regional Air Quality Model:  I.  Concept and
     Development  and  II.   Verification and Samle Application in the San
             ~~
pe
bs,
     Francis co~B~ay Area,  Lawrence Livermore Labs, Nov. and Dec. 1975.

 5.   Wayne, L.;  R. Danchick; M. Weisburd; A. Kokin; A. Stein, "Modeling
     Photochemical Smog on  a Computer  for Decision Making." J. Air Pollution
     Control  Association, Vol. 21 # 6, pp. 334-339, Pacific Environmental
     Services,  June  1971 .

 6.   Sklarew, R.C.;  A. J. Fabrick; J.  E. Prager,  "Mathematical Modeling of
     Photochemical Smog Using the PICK Method," J. Air Pollution Control
     Association.  Vol. 22 #11, 865-869, November  1972.

 7.   Trijonis,  J.,  "Economic Air Pollution Control Model for Los Angeles
     County in  1975,"  Environmental Science and Technology, Vol. 8 #9, pp.
     811-826, September 1974.

 8.   Horowitz,  A.; W.  Meisel ; D. Collins, The Application of Rearo-Modeling
     to the Analysis of a Photochemical Air Pollution Model. EPA-650/4-74-6oi
     Technology Service Corp. , December 1973.

 9.   Tiao, G. C. ;  M. S. Phadke; and G. E. P. Box, "Some Empirical Models for
     the Los  Angeles Photochemical Smog Data," J. Air Pollution Control
     Association,  Vol. 26 #5, pp. 485-490, May T976~;

10.   Hanna, S.  R.,  "A  Simple Dispersion Model for the Analysis of Chemically
     Reactive Pollutants,"  Atmospheric Environment, Vol. 7, pp. 803-817,
     August 1973.

11.   Hanna, S.  R.,  "Modeling Smog Along the Los Angeles-Palm Springs
     Trajectory,"  to be published in Advances in  Environment Science and
     Technology.


                                     813

-------
12.  Hameed, S.; S. A. Lebedeff; R. W.  Stewart,  A Model  Study  of the  Impact
     of Emission Control Strategies on  Los Angeles Air Quality,  Goddard
     Space Flight Center, NASA.

13.  Chameides, W.; and J. C. G. Walker, "A Time-Dependent Photochemical
     Model for Ozone Near the Ground,"  J. Geophysical  Research,  Vol.  81,
     pp. 413-420, January 1976.

14.  de Nevers, N.; J. R. Morris, "Rollback Modeling - Basic and Modified,"
     J. Air Pollution Control Association, Vol.  25: p. 943, September 1975.

15.  Appendix J. Federal Register, August 14, 1971.

16.  Photochemical Oxidant Modeling:  Detailed Technical  Report, GCA/
     Technology Division, Vol. II ,EPA-450/3»75-069b,  April 1975.

17.  Ludwig, F. L.; J. H. S. Kealoha, Present and Prospective  San Francisco
     Bay Area Air Quality, Stanford Research Institute,  SRI Project #3274.

18.  Martinez,  J.  R.;  R.  A.  Nordsieck: A.  Q.  Eschenroeder,  Impacts of
     Transportation Control  Strategies on  Los Angeles  Air Quality,
     EPA-R4-73-012, Vol.  d.,  General  Research Corp., May 1973.

19.  Martinez,  J.  R.;  R.  A.  Nordsieck; A.  Q.  Eschenroeder,  "Morning Vehicle
     Start Effects  on  Photochemical  Smog,"  Environmental Science  and
     Technology, Vol.  9  #5,  pp.  433-477,  May T97B~:
20.  Martinez, J.  R.;  R.  A.  Nordsieck,  Air Quality Impacts of Electric Cars
     in Los Angeles, General  Research Corp., GRC #RM-1905, August 1974.

21.  Dabberdt, W.  F.,  et  al . , Evaluation of Air Pollution Emergency Plans.
     Stanford Research Institute  (prepared for California Business Properties
     Association), May 1975.

22.  Nordsieck, R.; E. A. Berman;  J. Markins; and G. M. Hidy, Impact of
     Energy Resource Development  on Reactive Air Pollutants in tne Western
     United States, Environmental  Research and Technology, Feb. 1976.

23.  Eschenroeder, A.  Q., "Forecasting  Regional Photochemical Air Pollution,"
     Proceedings.  Long Term  Maintenance of Clean Air Standards, pp. 147-187,
     February 1975.

24.  Ranzieri, A.  J.;  P.  D.  Allen; W. B. Crews, A PI FKIN Sensitivity
     Analysis for  Transportation  Air Quality Studies, California Pi vi s i on
     of Transportation.  This has  not yet been released.

25.  Nordsieck, R. A.: J. R.  Martinez,  Population Growth Impacts on Air
     Quality in Santa  Barbara, General  Research Corp., GRC #RM-192'1T

26.  Reynolds, S.; J,  Seinfeld, "Interim Evaluation of Strategies for
     Meeting Ambient Air  Quality  Standard for Photochemical Oxidant,"


                                    814

-------
     Environmental  Science anu  Technology, Vol. 9 #5, pp. 433-447, May 1975,

27.  Peterson, J.  T.;  and K.  L.  Demerjian, "The Sensitivity of Computed
     Ozone Concentrations to  U.  V.  Radiation in the Los Angeles Area."
     Atmospheric Environment, Vol.  10,  pp. 459-468, May 1976.
                                     815

-------
                                                                            17-2
                TESTS OF THE DIFKIN  PHOTOCHEMICAL/DIFFUSION MODEL
                USING LOS ANGELES REACTIVE  POLLUTANT PROGRAM DATA

                       G. H. Taylor  and  A.  Q.  Eschenroeder*
ABSTRACT
      UA-ing  airborne and Aur^ace data obtained -in the LoA AngeleA Reactive.
Poltu-tant Program conducted in the, autumn  ofi  1973,  a AerieA o^ trajectory
AJ.matatA.onA MOA carried out uA-ing the PIFKIN pn.ogx.am (a A e.mi-Lagrang-i.au mode.t
utJ.tizi.nQ che.mic.at kineticA and K-theory di^uA-ion  calculationA] aA a pie-
Liminany Atep in modification o{, DT.fK.lH to embody recent update.*) in kA.ne.ticA
and didfauA-ion theory.   The. tAa.jncton.ij ofa an air. parcel WOA determined
LOA An.geJ.eA Re.ac-ti.ve. PoU.ut.ant. Pnognam te.tn.oon  dat.a,  and pottutant
u)e.tie,  den^ive.d fin.om e.miAA4.onA -inve.nto?u.e,A fion. Auc.ctAA4.ve. gn^id Aquane.A
the ain  pan.ce.1.  Ob^eAued vatucA ofa Aoian.  z&n^ith angle., vertical e.ddy
and -initsiaJL po-tt.uta.nt pn.o{,.iieJ> compn^iAe.d the. addLttonai -input paA.ame.teJiA
the. A-imuiation.  Integration U)OA conducted eveny minute, and pn^inted output
WOA obtained ^on. eveny 20-m-i.nute -intenvat.  Predicted concentrations at the
variouA  veAtA.ca£. levels wene compared w-ith keticopten. and Aunfiacc obAerva-
tionA  obtained in the LoA AngeleA Reactive Pottutant Program Atudy.  S^m-t-
          and di^erenceA were noted and dtAcuAAed,  and AuggcA£ionA ^or i.mprove
       to be utilized -in the forthcoming model Mere  outlined.

                  DESCRIPTION  OF THE  MODEL AND THE DATA  BASE

     The  Los  Angeles  Reactive  Pollutant Program  (LARPP),  sponsored  by  the
Coordinating  Research  Council,  Inc.,  with cooperation from the  Environmental
Protection  Agency  (EPA),  National  Oceanics  and Atmospheric Administration
(NOAA),  and  California Air Resources  Board  (CARB), was  aimed at supporting
"the development of sound, unbiased  scientific information relating to air
pollution"  (1).   A total  of 35  operations was  conducted,  with  primary  interest
focused  upon  tracking  and measurement of a moving air parcel  within the marine
layer  for the  purpose  of providing a  data base for air  quality  models.   To
accomplish  this,  constant-pressure balloons (tetroons)  were tracked by radar;
their  positions  were  relayed  to instrument-equipped helicopters and an air
quality  van, which attempted  to follow the  tetroons as  closely  as  possible
and, therefore,  monitor the air parcel  surrounding the  tetroons.   Careful
planning  and  auditing  were conducted  in obtaining the experimental  data;  as  a
result,  the  completed  data archives  represent a well-documented source of
information  for pollutant analysis and model validation.
   Environmental Research  and  Technology, Inc., Santa Barbara, California.

                                       817

-------
      In  recent years,  regulatory  agencies  have  required the development of
 photochemical diffusion modals  dealing with both atmospheric chemistry and
 diffusion  processes.   The  fact  that all but a few of the regulated pollutants
 undergo  atmospheric  transformations requires that these more complex models
 supplant the  diffusion models used previously.  Only where chemical reactions
 are  not  significant  in the  determination of atmospheric concentrations can the
 simpler  models  (generally  Gaussian in nature) be used with confidence.

      Evolving during the middle and late 1960's, the DIFKIN model  (a photo-
 chemical DIFfusion/KINetics  program) was based  on a Lagrangian approach in-
 volving  air trajectories over the pollutant emissions sources.  Preliminary
 analyses indicated excellent correlation between predicted and observed pollu-
 tant concentrations  (2); nevertheless, certain  aspects of the model have
 become outdated  due  to advances in chemical and diffusion theory.  For that
 reason,  scientists at  Environmental Research and Technology, Inc.  (ERT) have
 begun the  process of writing a  new program (under the acronym ARTSIM**) to
 entail recent technical improvements.  In  order to reassess the performance of
 DIFKIN using  a  complete, well-documented data base, LARPP data were utilized
 in setting up a  series of  simulation runs  to recalibrate the model and deline-
 ate  possible  areas of  improvement for use  in ARTSIM development.

      DIFKIN carries  out three major functions:  calculation of an air trajec-
 tory using wind  data;  determination of source emissions along the trajectory
 path;  and  the computation  of concentrations of pollutants at times and for
 several  vertical levels within  the air parcel (3).  Calculation of the trajec-
 tories was bypassed, and a  trajectory determined by tetroon positions was
 input directly.  Inputs necessary for the  current study included surface
 pollutant  fluxes, photochemical rate constants, vertical eddy diffusivities,
 and  initial pollutant  profiles.

      Since DIFKIN calculations  are based on a finite-difference solution to a
 K-theory diffusion equation  with  time- and height-dependent parameters, verti-
 cal  eddy diffusivities must  be  input for several vertical levels and updated
 to account for changes in  atmospheric structure.  Diffusivities used in the
 current  study were obtained  using a procedure based on vertical variations in
 temperature and  wind speed;  the diffusivity values were updated hourly.

      Surface  emission  fluxes were required for the primary pollutants nitric
 oxide  (NO), carbon monoxide  (CO)  and reactive hydrocarbons (RHC), and were
 obtained from Reference 4;  data consisted  of stationary and mobile emissions.
 The  mobile emissions were  broken  down into freeway and surface street emis-
 sions  for  each hour in 2x2 mile squares.

      It  is necessary to provide initial concentrations of six pollutants (NO,
 nitrogen dioxide (N02), RHC, CO,  ozone (03), and nitrous acid (HNO?) at five
 vertical stations (the levels used in this simulation were at the surface and
 250,  500,  750, and 1000 meters  above the surface)* as DIFKIN input.  Surface
 concentrations were obtained from LARPP van data and 250 meter data were ob-
 *These elevations  were established by the flight patterns  adapted during the
  selected portions of Operation 33.
** Atmospheric Reaction and Transport Simulation.

                                     818

-------
 tained from LARPP helicopters;  initial  concentrations  at the  remaining  three
 levels were estimated from other available data (and generally  were  quite
 low).   Adjustment of data became necessary in  order to preserve the  photo-
 stationary equilibrium set up by the following three reactions:

          N02 + hv •> NO + 0                                             (1)


          0 + 02 + M -» 03 + M                                           (2)


          NO + 03 -> N02 + 02                                            (3)

 Since Equation 2 is known to be a very fast reaction,  Equation  1  and Equation
 2 are often combined into the single reaction

          NO-,, + hv  > NO + OT                                            (4)

Letting k, be  the rate constant for Equation 4  (a function of solar zenith
angle  and, therefore, time-dependent) and k3 the rate constant for Equation 3
(assumed to  be constant with time), a photostationary balance has been postu-
lated  (5) between the three pollutants involved, such that

         [03 pphm]  •  [NO pphm]     ,
              [N02 pphm]         = k7

The above result is based on laboratory experiments of fully-mixed parcels.
Since an air parcel in the atmosphere may be heterogeneous if mixing is inhib-
ited (such as during more stable atmospheric conditions), and because local NO
sources affect measurements, the observed ratio in Equation (5) may not corre-
spond to the ratio of rate constants.  During the present study, for example,
the observed ratio derived from LARPP data was  occasionally  an order of magni-
tude larger than the k1/k3 values.


              CASE STUDY:  LARPP OPERATION 33, NOVEMBER 5, 1973

     Of the 35 LARPP sampling days, Operation 33 was a particularly useful
one, due to long duration (over 7 hours of helicopter flight time) and lack of
logistics problems.  A group of three tetroons was launched at 0715 PST from
Downey, 17 kilometers southeast of central Los Angeles.   Upon instructions
from the radar operators, the LARPP helicopters were dispatched to locations
of the tetroons, and flew four-leg square patterns at incremental altitudes.
The patterns surrounded the centroid of the three tetroons.

     Vertical temperature structure and wind speed data, necessary for calcu-
lation of vertical eddy diffusivities, were obtained from 0530 PST rawinsonde
flights at El Monte (EMT) and Los Angeles Tnternational  Airport (LAX), 1330
PST El Monte soundings, and from LARPP measurements.  Table 1 is a list of
diffusivities  input for each  necessary level  at every  hours.   Since diffusivi-


                                     819

-------
ties are specified at only six vertical stations, it is usually necessary to
interpolate the actual values to avoid aliasing problems.

     In order to preserve photostat!onary state requirements outlined earlier,
adjustments of initial pollutant concentration was necessary.   The observed
[03][NO]/[N02] ratio was much higher than the kj/k3 value.   Experimentation
indicated that best results were obtained by decreasing the 03 value and
increasing N02, leaving NO as observed.  Table 2 is a list  of  observed and
corrected concentrations for the five vertical levels.   kj  values utilized as
input parameters were obtained from Reference 6.

                                    RESULTS

     DIFKIN output consists of concentration profiles at five  vertical levels
(the same altitudes specified by the input concentrations)  undated at a user-
specified time interval (in this case 20 minutes).  Surface concentrations of
03, NO, N02, RHC and CO, as a function of time, appear in Figures la through
13; both predicted (DIFKIN) and observed (LARPP) concentrations are included.
Figures 2a through 2e give concentrations for the same pollutants at the 250-
meter level.

     Since the observed concentration ratio appearing in Equation 5 was approx-
imately an order of magnitude greater than that predicted by k]/k3, initial 03
input was much smaller than measured values (0.3 pphm instead  of 4 pphm), and
N02 input was much larger (16 pphm instead of 7 pphm).   As  a result, 03 pre-
dictions remain below observed values for approximately 3 hours, after which
excellent correlation is found; N0? predictions, however,  remain much higher
than observations throughout the simulation (see Figures la and Ib).  Nitric
oxide values, which were not manipulated, are seen to correlate very well with
measurements (Figure Ic).   Quite clearly a problem in accuracy is inherent,
due to the kl/k^ anomaly:   successful simulation of 03, NO  and N0? will depend
on  the development of mixedness corrections for the rate constants,  In the
present case,  two of the three were well accounted for  (at  least after 1000
PST); by changing the initial balance  it may be possible to adjust the simu-
lation to accurately assess any two of the three.

     Regarding the airborne  (250 meter) measurements, though an imbalance
occurred similar to that outlined above, the difference between observed and
adjusted initial concentrations was considerably less than  that at the surface.
Such behavior  is to be expected since  the degree of mixing  at 250 meters above
the surface is greater than that at the surface in the  proximity of primary
pollutant sources.  After 1000 PST, for example, both 03 and N02 predictions
remained relatively close to observed  values  (Figures 2a and 2b), as in the
ground level case; measured NO levels  were closely approximated by the program
(Figure 2c).

     Carbon monoxide is a relatively nonreactive pollutant  which can be util-
ized to assess the accuracy of diffusivity values  and/or surface pollutant
fluxes.  As is evident in Figures  le and 2e,  DIFKIN predictions provide excel-
lent approximations of measured concentrations throughout the morning  hours,
indicating  that diffusivity and flux inputs are reasonable.  Between 1200  and
1400 PST, the  model overpredicts somewhat, indicating that  assumed surface


                                     820

-------
TABLE 1.  VERTICAL EDDY DIFFUSIVITIES USED IN DIFKIN SIMULATION (m-7min)

IEIGHT ABOVE
SURFACE (m) 0700
0
125
375
625
875
1000
126
72
69
42
30
27
.0
.0
.0
.6
.6
.0
0800
126.0
72.0
69.0
42.6
30.6
27.0
0900
330.
76.
69.
42.
30.
27.
0
8
0
6
6
0
TIME
1000
348.0
590.0
69.0
42.6
30.6
27.0
1100
348.0
666.0
186.0
42.6
30.6
27.0
1200
360.0
720.0
306.0
42.6
30.6
27.0
1300
360
720
306
42
30
27
.0
.0
.0
.6
.6
.0
1400
360.0
720.0
306.0
42.6
30.6
27.0
.
TABLE 2.  OBSERVED AND CORRECTED (FOR DIFKIN INPUT) VALUES OF 03, NO, NO,
~
MFTPHT ARfWF OBSERVED CONCEN-
HFTGHT ABOVE TRAT I ONSj^phm)^
SURFACE (m) °3 N0 N0?
0 4.0 20.0 10.0
250 6.0 7.0 3.0
500
750
1000
RATIO ACTUAL
[0,] [NO] kL
[N02J k3
8.0 .203
14.0 .203
.203
.203
.203
CORRECTED CONCEN-
TRATIONS (pphm)
03
0.3
1.0
4.0
4.0
4.0
NO
20.0
7.0
0.5
0.5
0.5
N0^:
29.6
34.5
9.85
9.85
9.85
                                 821

-------
         10-
 (pphm)
                                                           •   •
fpphm)
 c.)
   NO
 fpphm)   6_
 d.)
   RHC
 (pphmC)
 e.)
   CO
  (ppm)
          0
0700     0800     0900
1200     1300     1400
                                     1000     1100
                                        LOCAL TIMi-
Figure 1.   Surface DIFKIN predictions (solid lines) and LARPP observed
           concentrations (circles), November 5,  1973.
                                    822

-------
h.)
c.)
d.)
c.)
10 -


 5


 0


30
10


 0-

20


15-


10

 5 -


 0
        50
   CO
0 -

4 _

2 _


2 -


1 -

0
0700
    Figure 2.
          0800      0900
                                    1000     1100     1200

                                       LOCAL T1MF:
                                                     1300     1400
       250 m.  DIFKIN predictions (solid lines)  and  LARPP  observed
       concentrations (circles), November 5,  1973.
                                     823

-------
fluxes and/or upper-level (375 m) diffusivity values are excessive.

     Although reactive hydrocarbon predictions agree quite well with 250 m
measured data, surface observations are somewhat higher than the model output
(see Figures Id and 2d).   Two possible explanations are offered:  traffic
levels may be slightly higher than assumed; and the measurement van, which
operated on or near the surface streets, may be heavily influenced by micro-
scale "puffs" of emissions which exceed ambient levels.  Both occurrences are
difficult to assess or rectify; it may be necessary to apply a correction
factor to predictions to account for this anomaly.


                                 CONCLUSIONS

     The chief factor impeding accurate air pollutant predictions using the
DIFKIN model appears to be the NO-OrNO? balance represented in Equation 5.
Since experimental results assume (and utilize) a well-mixed atmosphere,
actual measurements during more stable conditions (generally night and morning
hours) can yield balances which differ greatly from experiment.   It would be
highly desirable to establish a prediction scheme for the pollutant balance as
a function of stability and height,  perhaps for incorporation into the ARTSIM
development.  The ratio in Equation 5 is highly dependent on the degree of
homogeneity of the pollutant mixture.   A physically-based empirical correction
depending on source emission configurations and atmospheric turbulence will  be
necessary to rectify this problem.

     Although the well-known problems of emission source inventory accuracy,
diffusivity accuracy, and realism in the kinetic model still provide a chal-
lenge, this test of a photochemical  diffusion model  against three-dimensional
atmospheric data has provided encouraging insights  into the directions of
model improvement.   The expanded data base provided by the LARPP study is
certain  to provide further guidance in the development of air quality simula-
tion models.
                                 REFERENCES

1.   J. F. Black, "Background on LARPP," A Paper at the Los Angeles Reactive
     Pollutant Program Symposium, Santa Barbara, California, November 12-13,
     1974.

2.   A. Q. Eschenroeder, J. R. Martinez, and R. A.  Nordsieck, Evaluation of a
     Diffusion Model for Photochemical Smog Simulation (Final Report) /General
     Research Corporation, CR-1-273, October 1972.

3.   J. R. Martinez, R. A. Nordsieck, and M.  A. Hirschberg, User's Guide to
     Diffusion/Kinetics (DIFKIN) Code (Final  Report), General Research Corpora-
     tion, CR-2-273/1, December 1973.

4.   R. A. Nordsieck, Air Quality Impact of Electric Cars
     Apj>ejKJij<_A, General Research Corporation, RM-1905-A, August 1974.


                                     824

-------
5.    P.  A.  Leighton, Photochemistry of Air Pollution, Academic Press, New York,
     1961.

6.    K.  L.  Schere, Personal Communication, October 1975.
                                    825

-------
                                                                            17-3
                    DEVELOPMENT  OF  A MARKOV CHAIN MODEL FOR
                       PHOTOCHEMICAL OXIDANT PREDICTION

                                J.  R.  Martinez*
ABSTRACT

     A method U> d&>csu.be.d wkeAcby the. time. &&u.e.t> ofa houAZy photochemical.
oxidant Lt> mode.ie.d by a MaAkov  chain.   The. Masikovtan i,tn.mctuA.e. oft the. time
&QAA.2.& 4A tested UAJJIQ data  ^ofi Houston,  Texcus.   The output o& the. Markov
model. it, thown to be. tie.late.d to the. t>tatu>ti.cat csute.tu.on -imposed by the.
National Ambte.nt bJJi QuaJUty Standard  faoti photochemical 0)tidanti>.  A technique.
to e.Atimate. the. pafiamete.^, o& the. Masikov  chain -ib bsu.e.filLy dc4ca44ed.


                                 INTRODUCTION

     The National Ambient Air Quality  Standard (NAAQS) for photochemical oxi-
dant is based on a probabilistic criterion that requires that the hourly
average concentration of 0.08 ppm not  be  exceeded more than once a year.
Expressed in terms of probabilities, the  NAAQS imposes that

         Prob(x > 0.08) £ 1/8767

where x is the hourly average oxidant,  Prob (x > 0.08) denotes the probability
of x exceeding 0.08 ppm, and 8760 is the  number of hours in a (non-leap) year.
In view of this requirement, it would  be  helpful to those involved in air
quality management to be able to predict  the probability of exceeding the 0.08
ppm threshold in a specified area over a  one-year period.  This would enable
one to estimate whether the  NAAQS would be met in the future.

     This report describes progress toward the development of a model of
photochemical oxidant using  a Markov chain.   The model's output is the prob-
ability of exceeding 0.08 ppm during a year in a given geographical region.
The forecast of such statistical data  complements results of other models
which predict absolute concentration levels.

     In developing the Markovian model  we seek answers to several questions,
namely:

     •   How is the Markov chain defined?

     •   Does the time series of oxidant  values satisfy the Markovian
         assumption?
^Environmental Research  and  Technology,  Inc.,  Santa Barbara, California.

-------
     •   How can one predict the values of the parameters of the Markov chain?

     In the sequel we attempt to provide answers to these questions.  Specific
examples using data for Houston, Texas  are provided.


                              MODEL FORMULATION
DEFINITION OF STATES

     The motivation behind the Markov chain approach stems from the fact that
the NAAQS for oxidant basically distinguishes between two mutually exclusive
conditions or states as regards air quality.  One condition occurs when the
hourly oxidant level exceeds 0.08 ppm; the complementary condition occurs when
oxidant concentrations are at or below 0.08 ppm.   Consequently, it is natural
to think in terms of a two-state system where one state denotes exceedence of
the .08  ppm threshold and the other state denotes nonexceedence.   More
formally, let s(n) denote the state of the atmosphere at the nth hours.  Then
we define two states, denoted by 0 and 1, as follows:

         s(n) = 0 if x  < 0.08 ppm,
                      n —

         s(n) = 1 if xn > 0.08 ppm,                                           (

where x  is the average oxidant concentration at the nth hours.

     Note that state 1 represents the occurrence of oxidant concentrations
which exceed the 0.08 ppm threshold specified by the NAAQS.  Thus, the NAAQS
is not satisfied if in a one-year period the probability of state 1 occurring
is greater than 1/870.  The output of the Markov chain model consists of the
probability of occurrence of state 1, thereby providing a direct indication of
the degree of compliance with the NAAQS.


THE MARKOVIAN ASSUMPTION

     The assumption which forms the basis of the Markov chain is that the
occurrence of a state at time (n + 1) depends solely on the state of the
system at time n.  The states of the system at times earlier than n are
assumed not to influence the state of time  (n + 1).  This is a very strong
assumption which is not always satisfied since a system's memory may extend
over more than one time interval.  In our application to photochemical
oxidant, we must test the autocorrelation function of the oxidant time series
to determine the degree of dependence between states at successive times.
Such analysis has been carried out for Houston and the results are shown in
Table 1.  It is apparent from Table 1 that the 1-hour dependence is strong.
As might be expected, the dependence decreases as the time lag increases.  The
magnitude of the autocorrelation for a 1-hour time lag provides evidence for
considering the oxidant time series as a sequence of dependent events which
can be approximated by a Markov chain.  Other tests have been performed which
bolster the use of the Markovian approximation (1).

                                      828

-------
     TABLE 1.   AUTOCORRELATION FUNCTION OF OXIDANT TIME SERIES FOR HOUSTON
                 AT MAE DRIVE, MAY-SEPT.  1973-1974, 0800-1900

TIME LAG (HOURS)
1
2
3
1973
.7
.5
.3
1974
.7
.4
.3

     For completeness, at this point we should mention that in Houston most
of the exceedences of the 0.08 ppm level occur during the months  of May through
September, always between the hours of 0800 and 1900.


OUTPUT OF MARKOVIAN MODEL

     In accordance with the mandates of the NAAQS, we are interested in finding
the probability of state 1  occurring over a large number of transitions.   This
is termed the limiting state probability and is found by using standard methods
described in References 2 and 3.   The limiting probability of state 1  will  be
denoted by i^.   Once obtained, one checks whether n} <_ 1/8760, thereby assess-
ing whether or not the NAAQS is satisfied and, if not, how far the region  is
from being in compliance with the standard.


MODEL PARAMETERS

     The behavior of the Markov chain is governed by a set of state transition
probabilities defined by:

         Pid (n, n + 1) = Prob [s(n + 1) = j|s(n) = i], j = 0,1             (2)

Thus p. . (n, n + 1) is the  conditional probability that the system will be in
state jjat time (n + 1) given that it is in state i at time n.  Our two-state
system has four transition  probabilities which can be put in matrix form as
shown below.

                         p00(n, n + 1}  p0i(n, n + 1)

         p(n, n + 1) =                                                     (3)

                         p10(n, n + 1)  pn(n, n + 1 )

It should be apparent that  each row of P(n, n + 1) adds to unity.

     A moment's reflection  about the physical nature of the system suggests

                                      829

-------
that the transition probabilities must be functions  of time.   This  is  because
sunlight intensity is a main contributing factor to  oxidant formation  and
because oxidant buildup times are typically several  hours  in  duration.   Thus,
for example, the value of Pn(n, n + 1), the probability that the system is  in
state 1 and remains so during the next hour, has to  be greater at noon than  at
sunset.  The Houston data clearly exhibit this time  dependence,  as  is  shown  in
Figures 1 and 2.  These diagrams contain plots of p01  and  p10, respectively,
for 1974.  The 1973 data yield similar plots.   It can  be seen that  the time
variation coincides with expectation:   the transition  from 0  to  1 is  most
probably around noon and that from 1 to 0 is most probable at sunset.   We note
that Figure 2 shows no data points for times earlier than  0900.   This  is
because state 1 never occurred prior to 0900 in either 1973 or 1974.   However,
as shown in Figure 1, at 0800 there is a nonzero probability  for the  transition
0 -* 1.   Another feature of interest shown in Figure  2  is that p10 shows a
generally stable value around 0.37 between 1300-1700.   Thus,  in  this  time
    o
   D-
                                                                  1900
                                  Time  (CST)
         Figure 1.   Hourly variation  of PQI>  Houston,  May-September 1974.

interval, the probability of remaining in state 1  is  high.   Finally, we note
that after 1900 we are certain of being in state 0.   Indeed,  between the hours
2000-0800, including the endpoints, state 1  never occurred  in the two years  of
available data examined.   The information displayed  in Figures  1  and 2 indicates
                                     830

-------
         0.8 -I
             I
 o
 i-H
CL
         0. 1 _

                 T~
                                ~T	T "
                           1100      i.inn
            0700    ()'.)()()

                                   Time (CST)

        Figure 2.  Hourly variation of P^,  Houston, May-September  1974.
that we have
11 hours.
          a  diurnal  cycle  in  the transition probabilities  with a period of
     It is apparent from the discussion that the only parameters  required  by
the model are the state transition probabilities P-jj(n, n +  1).   One will
naturally wonder how these parameters are obtained in order  to be able  to  pre-
dict the behavior of the system.  This brings us to the important question  of
causality and the relationship of the Pij's  to such factors  as emissions  and
meteorological conditions.  In a later section we discuss a  method  for  finding
functions which relate the p.. 's to causal factors.
STRUCTURE OF THE MARKOV CHAIN

     The periodicity of the time variations of the transition probabilities  is
the key factor that leads to a model structure where  the  transition  probabili-
ties appear to be stationary on a diurnal basis.  The formulation  involves
redefining the states of the system and  increasing their  number.   The  proce-
dure is as follows.  Suppose the cycle has  a  period of  K  hours, where  K  -  11
for Houston.  At each one of those hours we have either state 0 or 1 occur-
ring, but the transition probability changes  from hour  to hour.   Let us  define
                                      831

-------
       o
       o
       o
       (Nl
       o
       o
       CTt
       o
       o
       oo
                                                      O)
                                                     -o
 o
-l->
 CO
 Z3
 O
rn

 s.
 o
C/3
o:
3
O
I
                                                     
-------
a new set of 2K distinct states and number them so that the odd-numbered
states correspond to state 1 and the even-numbered states to state 0.  Thus,
at every one of the K hours we have two new states, an even-numbered and an
odd-numbered state.  For Houston, the expanded chain will have 22 states and
is illustrated in Figure 3.

     Since the Houston data show thet the system is always in state 0 from
2000 to 0800, in Figure 3 the hours' from 2000 to 0800 are represented by a
single state, also denoted by 0.

     To obtain from the expanded chain the probability of occurrence of the
original  state 1, we simply find the probabilities for the odd-numbered states
and add them.
                             PRELIMINARY TESTING
COMPARISON OF OBSERVED AND COMPUTED STATE PROBABILITY

     Preliminary testing was directed at resolving whether or not the model
would reproduce the observed limiting state probabilities given a set of tran-
sition probabilities calculated from the data.   This, of course, does not con-
stitute a model validation test.  Rather, it is a test of the algorithm.   The
results are shown in Table 2 for 1973 and 1974 using data for Houston.  Table 2
shows that the limiting state probabilities obtained from the Markov chain are
in very close agreement with the observed probabilities.   The tentative conclu-
sion that might be drawn from these results is that given a set of physically
plausible transition probabilities, we may use the previously described Markov
chain to estimate the long-run probability of occurrence of exceedence of the
0.08 ppm threshold oxidant concentration.


    TABLE 2.   COMPARISON OF OBSERVED AND COMPUTED PROBABILITY OF OCCURRENCE
    	OF STATE 1 FOR HOUSTON, MAY-SEPT.  1973-1974

                   YEAR          OBSERVED          COMPUTED
                   1973            .162              .164

                   1974            .109              .110
                   ESTIMATION OF TRANSITION PROBABILITIES

     For purposes of prediction, it is necessary to obtain a function which
relates the transition probabilities of the Markov chain to factors such as
emissions and meteorological variables.  In this section our discussion will
be limited to methodology, since our work on this topic is in progress.

                                      833

-------
      Basically,  the problem is  one  of  finding  a  regression  relation  for the
 transition  probability as  a function of several  variables.   The  logistic func-
 tion  appears  to  be  a convenient model  for the  desired  function.   With  this
 model,  the  transition probability assumes the  following  form:

                                        N
                              exp[b   +   I  b.xj
                                       k=]  K
         Pij(t;xi,X2---xN) =	N	                 (4)
                             1 + exp[b   +  I  b,X|<]
                                      o   k=1  K K

where i, j= 0,1, the quantity XR is  the kth independent variable, k = 1, 2,
	N, and N is the number of independent variables.   The coefficients b, are
the parameters of the function and are  estimated from the data.  The metnod of
maximum likelihood can be used to find  normal equations which are solved for
the parameters b^ (6,7).  Sufficient statistics are available for this computa-
tion.

     We note that in Equation 4 the  time dependence is  implicit;  one would
obtain a function such as Equation 4 for each hour of the day instead of ob-
taining a single such function with  time as an independent variable.   Also, in
Equation 4 the variables Xi, would be emissions, temperature, solar radiation,
concentrations of other species, wind speed and direction, etc.

     Once such a function is obtained,  one can manipulate the transition pro-
babilities to determine the effects  of changes in emissions.  The modified
transition probabilities are then spplied to the Markov chain model to obtain
the limiting probability of the oxidant concentration exceeding 0.08 ppm
during a specified time interval.  This limiting probability is then compared
with the frequency measure mandated  by the NAAQS.


                                 REFERENCES

1.   T.  W. Anderson and L. A. Goodman,  "Statistical Inference About Markov
     Chains," Ann. Math. Stet. , Vol. 28, pp. 89-110 (1957).

2.   D.  L. Isaacson and R. W. Madsen, Markov Chains:  Theory and Applications,
     John Wiley and Sons, New York,  1976.

3.   R.  A. Howard, Dynamic  Probabilistic  Systems,  Vol .  I:   Markov Models,
     John Wiley and Sons, New York,  1957.

4.   W.  Feller, An  Introduction to Probability Theory and its Applications,
     V°L_ !> 2n^ed., John Wiley and Sons , New York, 1957.

5.   A.  T. Bharucha-Reid, Elements of the Theory of Markov Processes and Their
     Applications, McGraw HTT1 Book  Company, New York, 1960.

6.   D.  R. Cox, The Analysis of Binary Data, Methuen and Co., Ltd., London,
     1970.

                                      834

-------
7.    S.  H.  Walker and D.  B.  Duncan, "Estimation of the Probability of an Event
     as  a Function of Several Independent Variables," Biometrika, Vol.  54,
     pp.  167-179 (1967).
                                     835

-------
                                                                            17-4
              A  PRELIMINARY INVESTIGATION OF THE EFFECTIVENESS
                       OF AIR POLLUTION EMERGENCY PLANS

                        W.  F.  Dabberdt and H. B. Singh*
ABSTRACT
     A Atudy wot, made,  ofi  the. e^ecttvene-6.6 o& cuAA.e,ntty conceived
abutment plant, ^OA  oxidant control.   €pLt>ode.A -in the. South.  Coa&t  MA.
over appfioxA.mate.ly fiive, ye.au (1970 to Se.pte.mbeA. 1974} aAe. c.haAacte.A^ze.d by
thoMi {jfte.qu.ency ofa oc.c.uAAe.nc.e.,  theJJi donation, and theAA. e.xte.nt.   U-6-uig  the.
June. 1974 South Coast  MA Ea^in e.piA_ode.& at, a tej>t coAe., and a photochemical
ajji quality bimuJjation mode.t,  an objective evaluation o\ tke. e^ec^cuene-44  oft
&hoAt-t&im AtAate.gteJ>  w?o4 £ondu.cte.d.   It Lb concluded that,  because  ofa the.
chemical and me.te.oAotog^icat comp£ex^ce-6 o^ the. e.podeJ>, th&y aAe. e.xtAe.me-1-y
di^-icutt both to pAe.dict and to contAol. oveA. a &hoAt time, pe.Atod.


                                 INTRODUCTION

     The air pollution emergency plan provides the basis for taking  action  to
prevent or abate episodes where air pollution concentrations reach levels that
could endanger, or cause  significant harm to, the public health.

     In a recent study, Stanford Research Institute (SRI) undertook  a review
and assessment of air  pollution emergency episode planning in general and the
proposed 1975 California  Air Pollution Emergency Plan (CAPEP) in particular.
The scope of the study was restricted to emergency plans relating  to photo-
chemical oxidant concentrations and to the associated proposed short-term
control strategies.

     The CAPEP relies  heavily on prediction, in that abatement action is to be
taken if oxidant concentrations are "predicted or reached".  The requirement
to take action on predictions alone makes CAPEP significantly different  from
Environmental Protection  Agency (EPA) recommendations.  Figure 1 presents an
example of the action  by  stages that can be taken to prevent or abate air
pollution episodes.

     In the technical  discussions that follow, historical episodes in the
South Coast Air Basin  (SCAB) are used in evaluating the control plan.  Two
particularly relevant, and interrelated, technical aspects are presented:
^Stanford Research  Institute,  Menlo Park, California.


                                      837

-------An error occurred while trying to OCR this image.

-------
     These data were then used to identify the days and stations that had
Stage 1, 2 and/or 3 oxidant values and to tabulate the duration in days of
each episode by stages.  It must be recognized that a ten-day Stage 1 episode,
for example, might include two three-day Stage 2 episodes and one one-day
Stage 3 episode.  The results of the analysis are shown in Tables 1 and 2,
which present the frequency and temporal and geographical extent of these
episodes by their stages.

     Several major observations can be made from the analysis of the five-year
data shown in Tables 1 and 2:

     •   Only seven Stage 3 episodes occurred in the five-year period.

     •   Stage 3 oxidant levels did not persist for more than one day.

     •   No Stage 3 condition existed at more than one station during any
         emergency episode.

     0   Of the 50 Stage 2 episodes, 78% lasted two days or less.
         None lasted more than four days.

     t   More than 90% of the Stage 2 episodes were limited to two
         stations or one station.
               EFFECTIVENESS OF SHORT-TERM CONTROL STRATEGIES

SCOPE

     Notwithstanding the infrequent occurrence of episodes, their limited
extent, and the inexact forecast techniques used (Dabberdt et al.,  1975), an
important question still remains to be answered:  Given that there  is a reason-
able level of confidence that Stage 2 or Stage 3 oxidant levels will  be ex-
ceeded tomorrow at known locations, what is the probable effectiveness of
short-term control strategies (Figure !)?  To seek a resolution to  this ques-
tion, a numerical photochemical simulation model was used, in conjunction with
historical aerometric episode data from the SCAB, to predict the oxidant
impact of several alternative short-term emission reduction scenarios.

     A case-study approach was used to evaluate the effectiveness  of  various
short-term control strategies.  The week of 23-28 June 1974 was marked with
poor air quality in the SCAB, primarily due to adverse meteorological condi-
tions, weekday emissions, and abundant sunlight.  Six stations — Azusa, Bur-
bank, Pasadena, Pomona, Upland, and Whittier — reported the poorest air
quality conditions.  A maximum oxidant value of 44 pphm was measured  on 27
June at the Upland station.

     The analysis was designed primarily to answer two questions:

     1.  How effective are the currently conceived short-term emergency
         strategies for oxidant control likely to be?
                                    839

-------
TABLE 1.  FREQUENCY OF OCCURRENCE  OF  THREE AIR  POLLUTION EMERGENCY STAGES
FOR OXIDANT, BY EPISODE DURATION,  FOR THE CALIFORNIA SOUTH COAST AIR BASIN



Year
1970







1971










1972






1973










19/4









Episode
duration
(days)
1
2
3
4
5
6
10
67
1
?
j
4
5
6
7
8
9
10
27
1
2
3
4
5
9
11
1
2
1
4
!)
6
7
11
14
14
18
1
2
3
4
7
8
9
10
22

Stage 1 : alert
(0.20 to 0.34
ppm)
7
3
1
3
4
3
1
1
5
b
4
0
1
1
1
2
1
1
1
14
5
5
3
2
1
2
9
6
3
1
1
/
1
1
2
2
1
7
1
3
9
1
3
p
1
1

Stage 2: warning
(0.35 to 0.49
ppm)
9
7
5
0
0
0
0
0
3
0
1
1
0
0
0
0
0
0
0
5
0
1
0
0
0
0
7
2
0
1
0
0
0
0
0
0
0
2
4
1
1
0
0
0
0
0
"
Stage 3: emergency
(;0.50 ppm)

3
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
Q
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
                                 840

-------
       TABLE  2.   EXTENT  OF  STAGE  2  (WARNING) AND STAGE 3  (EMERGENCY)
          OXIDANT LEVELS, BY  EPISODE  DURATION,  FOR THE CALIFORNIA
         SOUTH  COAST  AIR BASIN  (MAXIMUM  DAILY NUMBER  OF MONITORING
                           STATIONS PER  EPISODE)

Freauency of
Stage 2
duration
Number of
Year stations
1970 1
2
3
1971 1
2
3
1972 1
2
3
1973 1
2
3
1974 1
2
3
occurrence
Stage 3
duration
(days)
1
7
2
0
2
1
0
3
2
0
7
0
0
2
0
0
2
3
4
0
0
0
0
0
0
0
1
1
0
3
0
1
3
1
2
2
1
0
0
1
0
0
0
0
0
0
1
0
4
0
0
0
0
0
1
0
0
0
0
1
0
1
0
0
1
3
0
0
1
0
0
0
0
0
2
0
0
1
0
0
(days)
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

     2.   What meteorological and chemical factors influence the
         effectiveness of such short-term strategies?

Simply stated, the approach entailed:  (a) simulating historical atmospheric
conditions that existed on selected days by using the DIFKIN (DIFfusion/KINetics)
Lagrangian air quality simulation model and (b) projecting the impact of
short-term emission reductions on peak oxidant concentrations.  The prime
motivation in selecting a trajectory model was the ease with which it permits
the establishment of a cause-and-effect relationship between emissions and air
quality. Operational details of the DIFKIN model have been provided by Martinez
et al.  (1973) and Ludwig and Kealoha (1974).  Two representative trajections
were selected for the analysis and evaluation:  the first trajectory passed
through the Upland station on 27 June at the time (1430 PST) when the peak
oxidant level was measured; the second trajectory passed through the Whittier
station similarly at 1330 PST on 28 June.  The Upland trajectory (see Figure
2) was selected because it was dominated by emissions from mobile sources.
The Whittier trajectory, on the other hand, represented a combination of
                                     841

-------
 •This number represents the average time (standard time) between 0400 and 050O hours
 and is treated as 0430 hours in the model.                                         T A-653522-111

    FIGURE 2   UPLAND (JUNE 27,  1974)  AND WHITTIER (JUNE 28, 1974) TRAJECTORIES


emissions from  power plants, oil refineries, distributed stationary sources,
and mobile sources.

     Interpolation  between soundings available  from  the Los  Angeles Inter-
national  Airport  and El  Monte stations was  performed by using the model
proposed  by Edinger (1969) to obtain mixing-height data along the trajectory
paths.   Eddy diffusivities within the mixing layer,  as  well  as  the inversion
layer,  were calculated according to Hosier  (1969).

     Considerable effort was made to generate an  up-to-date  1974 emissions
inventory for the SCAB.   After considering  the  nature of the case trajec-
tories, the 66-x-38-mile grid shown on Figure 2 was  selected.  The emissions
inventory data  were divided into five source categories:

     1.   Freeways [(nitrogen oxides (NC»x), Reactive hydrocarbons (RHC), and
          carbon monoxide (CO3

     2.   Surface  streets (NO. RHC, CO)
                             /\

     3.   Distributed sources (NOX, RHC)

     4.   Power  plants (NO )
                          s\

     5.   Oil  refineries (NO. RHC).
                            A
                                      842

-------
 Distributed sources represent emissions from industrial,  commercial,  and  resi-
 dential  sources.   Point sources  were uniformly distributed over the  2-x-2-mile
 grid area.   A complete description of the emissions  inventory  was  provided by
 Dabberdt et al.  (1975).

      Table 3 shows the estimate of pollutant emission reductions under the
 application of selected control  scenarios.  Scenarios I and II were designed
 to correspond to complete and effective implementation of Stage 3 and Stage 2
 controls, respectively.  Scenarios III and IV, on the other hand,  represent
 exclusive mobile source and stationary source controls, respectively.  These
 emission-reduction estimates were based on analyses and data reported by  TRW
 (1973), the Census of Transportation (1972), the Census of Population (1970)
 and the Los Angeles Regional Transportation Study (1967); a detailed summary
 was given by Dabberdt et al. (1975).

          TABLE 3.  PROJECTED POLLUTANT EMISSION-REDUCTION SCENARIOS

Scenario
Pollutant category
Pollutant
Reduction
(percent)
Comments
        I      Mobile sources             70.5
              Distributed stationary
               sources                   25
              Power plants               50
              Refineries                 35

        II     Mobile sources             47.6
              Distributed stationary
               sources                   0
              Power plants               50
              Pvfiseries                 35

        III    Mobile sources             70.5
              Distributed stationary
               sources                   0
              Power plants               0
              Refineries                 0

        IV     Mobile sources             0
              Distributed stationary
               sources                   25
              Power plants               50
              Refineries                 35
Appropriate to Stage
episode controls
Appropriate to severe
Stage 2 controls
Typical of severe
vehicle controls alone
Typical of comprehen-
sive stationary source
controls alone
     Air quality data were acquired from the Air Resources Board (ARE) for the
air monitoring stations operated in the SCAB.  Table 4 shows the initial
conditions for the two trajectories (0530 PDT).  Vertical pollutant profiles
were not available.  Accordingly, two types of "cases" of vertical profiles
were considered.   The vertical profiles used in Case 1 are similar to those
                                      843

-------
used by Eschenroeder et al.  (1975) for validating the model  in the SCAB; they
assume maximum initial oxidant concentrations at the surface.

     In a recent study, Johnson and Singh (1975) demonstrated the formation
and persistence of layers of ozone (03) aloft and the implications for the
multiple-day, ground-level impact of such layers.  Because of the significance
of 03 aloft, Case 2 simulations were conducted for locations where 03 was
assumed to be left over from the previous day.  These estimates of elevated 03
maxima were based on previous-day oxidant levels.  Table 4 shows the estimated
vertical 03 profiles for the two cases.

          TABLE 4.  INITIAL CONCENTRATIONS FOR THE CASES UNDER STUDY
   rwhittier Station  (W), 28 June 1974; Upland Station (U), 27 June 1974]

        ~  ~~              Vertical levels at 0530" hours, standard~~time
                                              (pphm)

Case
1




2




Initial
conditions
NO
RHC
N02
0
HN02
NO
RHC
NO,
°3*
HN02
Level 1
W
7.0
42.0
5.0
1.0
4.0
7.0
22.0
5.0
1.0
4.0
U
20.0
33.0
8.0
1.0
4.0
20.0
25.0
8.0
1.0
4.0
Level 2
W
4.4
23.4
4.1
1.0
4.0
4.4
12.1
4.1
12.0
4.0
U
12.7
18.3
6.5
1.0
4.0
12.7
13.8
6.5
12.0
4.0
Level 3
W
1.0
10.4
2.7
1.0
4.0
1 .9
5.0
2.7
24.0
4.0
U
5.6
8.3
4.2
1.0
4.0
5.6
6.3
4.2
24.0
4.0
Level 4
W
1.8
8.6
2.2
1.0
4.0
1.8
4.6
2.2
22.0
4.0
IT
5.0
6.8
3.4
1.0
4.0
5.0
5.2
3.4
22.0
4.0
Level 5
W
1.8
7.0
1.5
1.0
4.0
1 .8
3.3
1.5
20.0
4.0
U
4.0
4.8
2.3
1.0
4.0
4.0
3.7
2.3
20.0
4.0

^Estimated from previous-day oxidant values.

Note:  Level 1:  0-125 m; Level 2:  125-375 m; Level 3:
       Level 4:  625-875 m; Level 5:  875-1125 m.
375-625 m;
DISCUSSION OF ASSUMPTIONS

     The most pressing problem was to establish unequivocal  initial  conditions
for trajectories originating on land.   A lack of adequate network density or a
clean sea breeze near the coastal  areas can result in a fair amount  of uncer-
tainty.   Vertical pollutant profiles have traditionally been unavailable, and
only recently have attempts been made to obtain comprehensive data bases.
Furthermore, there is considerable inaccuracy in assigning a RHC fraction to a
relatively clean air mass.   Current information from Los Angeles indicates
that 58% of total hydrocarbon (THC) (as carbon) is reactive, with an average
carbon number of 6.22.  Although this is consistent with our choice  of initial
conditions, no guarantee exists that this fraction does not vary from location
to location.  An additional problem is associated with emissions from point
                                     844

-------
sources. The accuracy with which trajectories can be drawn from surface wind
measurements may not be adequate, and a slight shift in trajectory path can
produce some error.  Furthermore, elevated sources (power plants) are not
adequately treated in any of the present photochemical air quality models, and
all emissions are assumed to be at ground level.  Despite the many assumptions
entailed, it is our experience that the general applicability of such models
remains valid and that they are sound tools for decision-making purposes.


RESULTS

     To summarize the analytical procedure, the simulation model was first
tuned to replicate observed oxidant maxima on two trajectories.  Tables 5 and 6
 show the agreement between observed and simulated values.   Next, the oxidant
impact of four emission control scenarios was evaluated for each of two cases
of initial pollutant profile conditions.  Tables 5 and 6 also summarize the
projected effectiveness of the various control scenarios for different trajec-
tories and different initial conditions.  From these model  simulations,
results described below were derived:

     o   Analysis of the Upland trajectory (dominated by transportation
         emission sources) demonstrates that, even if the proposed Stage
         3 emergency plan could be successfully implemented, a one-day oxidant
         reduction of no more than 17% would be achieved.   The maximum effec-
         tiveness of proposed Stage 2 controls would be limited to a 7% oxi-
         dant reduction (Table 5).

     o   Analysis of the Whittier trajectory (characterized by emissions
         from transportation, distributed, and point sources) indicates that
         implementation of any of the emergency strategies  (Stages 3 or 2)
         might actually have an adverse or negative effect  (i. e., up to a 40%
         peak oxidant increase) characteristic of South Coast episodes over
         the short time period (Table 6).  A reduction in only mobile sources
         (Scenario III) would result in a 3% to 5% reduction in oxidant values.

     o   It is also demonstrated that episodes on a given day will be
         influenced by the previous day's photochemical events and will simi-
         larly affect the events of the following day.  Elevated 03 layers
         from the previous day will essentially remain intact in upper stag-
         nant layers and will not only disperse downward during the following
         day, but will  also, by reacting with ground-level  nitric oxide (NO),
         be able to significantly affect peak surface 03 production (Case 3,
         Table 5).


                                   SUMMARY

     From the consideration of episode characteristics, it is clear that Stage
3 (emergency) oxidant levels historically have not persisted for more than a
single day at a time; accordingly, the prediction of the onset and termination
of a Stage 3 emergency must be timely, accurate, and reliable.  Therefore, to
be effective, a Stage 3 control strategy must work almost immediately.  Stage


                                     845

-------
        TABLE 5.   EFFECTIVENESS  OF  CONTROL  STRATEGIES:
                          TRAJECTORY  FOR 27 JUNE  1974
                                       UPLAND STATION

Historical
Measured ozone
hourly maxium*
Case (pphm)
1 44

2 44

conditions
Simulated ozone
hourly maximum*
(pphm)
46

45


Scenario
I
II
I
II
Simulated
ozone
maximum*
(pphm)
38
43
38
42

Percent
change
-17%
- 7%
-17%
- 6%

*14-1500 PST.

       TABLE 6.
EFFECTIVENESS OF CONTROL STRATEGIES :   WHITTFR STATION
            TRAJECTORY FOR  28 JUNE 1974

Historical conditions
Measured ozone
hourly maximum*
Case (pphm)
1 33



2 33



Simulated ozone
hourly maximum*
(pphm)
34



31





Scenario
1
11
m
IV
1
11
111
IV
Simulated
ozone
maximum*
(pphm)
43
47
33
45
37
41
30
39


Percent
change
+28%
+40
-3
+33
+17
+29
-5
+24
 3 episodes are not isolated but often occur immediately after Stage 2 episodes.
 The difference between CAPEP-and EPA-recommended episode plans becomes signi-
 ficant here.  According to EPA recommendations, a Stage 3 episode will prob-
 ably never be called, because one must first measure the occurrence of Stage 3
 episode conditions and then predict that these conditions will recur within 24
 hours. But Stage 3 episodes have never lasted two or more days.  According to
 CAPEP, however, if conditions show signs of deteriorations at the Stage 2
 level, it is possible to institute Stage 3 controls on the prediction that the
 next day will achieve Stage 3 levels.

     It is clear from the above discussion that there is a need for accurate
 air pollution forecasts of 18- to 24-hour duration.  At present, however, no
 objective techniques with an adequate verification record are available

                                     846

-------
 (Dabberdt et al.,  1975).  One difficulty is that, although photochemical air
 pollution is a three-dimensional problem,  it is often treated as two dimen-
 sional.  For example, if a reservoir of 03 exists entrapped between the upper
 stagnant layers and is  likely to disperse downward, surface measurements of 03
 will not help to predict such a reservoir because of the surface destruction
 of 03 by precursors and surface features.  Similar limitations apply to sur-
 face meteorological information.

     A further problem  associated with the effectiveness of short-term strat-
 egies for oxidant  control is the complex kinetics of oxidant formation and the
 inapplicability of traditional rollback concepts.  Case 1, Table 5 clearly
 shows that even drastic emission controls (Stage 3) can result in only a 17%
 maximum oxidant relief.  Even in the absence of an elevated reservoir of 03,
 the impact of reduction in emissions is inherently limited by the high
 atmospheric pollutant loading and the inherent nonlinearity of the photochemi-
 cal system.   The potential effect of "new" pollutants is therefore limited by
 the residence time of "old" ones.

     The results for the Whittier trajectory are intriguing in that some of
 the control  strategies may actually cause short-term increases in oxidant
 levels.  As is clear from Table 5,  a reduction in mobile sources only (Scenario
 III) would result in a  3% to 5% reduction in oxidant values.   Under the appli-
 cation of all the other strategies, the oxidant values for the Whittier station
were actually projected to increase by as much as 40% for Stage 2 controls
 (Case 1).  This projected increase is primarily attributable to the fact that
 the Whittier trajectory passes over at least one power plant and the reduction
 is net 03 formation resulting from the institution of emission reduction
 strategies would be more than compensated by the reduction in 03 destruction
 that would probably result from reducing the power-plant NO emissions.   It is
 important, however, to emphasize that the total  air pollution potential in a
 given air mass is always greater in cases where no abatement strategies are
 applied than in cases where some strategies are applied.   The problem is that
 this potential is not necessarily reduced over a short time period.   A look at
 the RHC-NO  ratio for the Whittier station showed that, if the trajectory were
 followed over subsequent days, eventually a net oxidant reduction by the
 application of these strateiges would probably have been realized.  This
 reduction, however, might occur several days later and many miles downwind.
The effectiveness of the short-term strategies must be deemphasized, in that
 they may provide little or no short-term relief.

     So far we have assumed that the strategies could be enforced on a short-
 term basis under ideal  conditions.   In actuality, the enforcement problems
encountered would be considerable.   For example, during a Stage 3 emergency,
particularly since a strict enforcement mechanism has not been implemented, it
 is unrealistic to expect that the population, faced with an unexpected vacation
 day, would not use automobiles.  Socioeconomic factors would likely play an
 important role, since the cost of a Stage 3 abatement action in the SCAB has
 been estimated by Dabberdt et al.  (1975) to be about $100 million in lost
wages alone.   Similar Stage 2 actions could result in lost wages up to $45
million.  Indeed, such socioeconomic factors may ultimately control the nature
 and scope of the available short-term control options.
                                     347

-------
                                 REFERENCES

 1.   Census  of Population,  1970:  Characteristics of the Population, California.

 2.   Census  of Transportation,  1972:  Truck Inventory and Use Survey, California.

 3.   Dabberdt, W.  F.,  et  al., 1975:   "Evaluation of Air Pollution Emergency Plans,1
     Final  Report,  SRI  Project  4056,  Stanford Research Institute, Menlo Park,
     California.

 4.   Edinger,  J. G., 1969:   "Changes  in the Depth of the Marine Layer over the
     Los  Angeles Basin,"  J.  Meteor.,  Vol. 16, No. 3, pp. 219-225.

 5.   Hosier, C. R., 1969:   "Vertical  Diffusivity from Random Profiles," J.
     Geophys.  Res., Vol.  74, No. 28,  p. 7018.

 6.   Johnson,  W.,  and  H.  Singh, 1975:  "The Effect of Ozone Layers Aloft on
     Ground-Level  Concentrations" (manuscript to be submitted to Geophysical
     Research  Letters  for publication).

 7.   Los  Angeles Regional Population  Study, 1967:  "Base Year Report Origin
     Destination Survey."

 8.   Ludwig, F. L., and J.  H. S. Kealoha, 1974:  "Present and Prospective San
     Francisco Bay Area Air Quality," prepared for Wallace, McHarg, Roberts
     and  Todd, Inc., and  for Metropolitan Transportation Commission, SRI Project
     3274,  Stanford Research Institute, Menlo Park, California.

 9.   Martinez, J.  R.,  R.  A.  Nordsieck, and M. A. Hirschberg, 1973:  "User's
     Guide  to  Diffusion/Kinetics (DIFKIN) Code," Final Report, EPA Contract
     68-02-0336, General  Research Corporation, Santa Barbara, California.

10.   TRW, 1973:  The Development of an Air Pollution Episode Contingency Plan
     for  the Metropolitan Los Angeles Air Quality Control Region, prepared for
     EPA  Region  IX, Contract No. 68-02-0048, TRW (Transportation and Environ-
     mental  Operations).
                                    848

-------
                       SESSION 18
OXIDANT-PRECURSOR RELATIONSHIPS AND THEIR INTERPRETATION
  IN TERMS OF OPTIMUM STRATEGY FOR OXIDANT CONTROL - I

                  ChcuAman:  J.N.  Pitts
                University of California
                           849

-------
                          A  "J" RELATIONSHIP  FOR  TEXAS

                                  H. M. Walker*

ABSTRACT

     An uppcifi timit cu/tue neiat/ionAhtp  k(tt  be.e.n contacted ^on motto, than 900
ozone-non-methane lujdnocanbon data paLu  obtained by the. Texa/s A,(A Control.
Boa iff monitolAng netmo-ik Aince  1973.  UnLika  the.  AP-84  cu/ive,  the  Te,xat> MA.
Contio? Boa^id cuAve. LuteAeAtA the. .0 6-9  a.m.  non-me,thane. hydA.ocaAbon ax-ci
at .125 ppm ozone.  It aJt&o  xeacheA a maximum at  1.1  ppm 6-9 a.m.  non-me.thane.
hi/d'ioca'ibon, and then deAcendb  to valued  be. How the. National. A.mbte,nt Ax>
Standaid fool ozone at l.eve.1^ above 5.0  ppm  non-me.thane,  hydrocarbon.  Fi^ty-
nine data pa<>4 {^nom the Houston A-ci Pollution ContAol  Vu>&u.ct ne,tu)oxk move,
(lie maximum out to 2.4 ppm,  but do not  change the t>hape ofa the catve.   The
pabbibCe implications o{} the Texas "J"  ctiive.  ale  discussed.

                                  INTRODUCTION

     The concept of the upper limit curve as  a device for developing a mean-
ingful empirical relationship between atmospheric oxidants and atmospheric
hydrocarbons (HC) was first  annunciated by  Schuck and Altshuller et al. in
an article  in the Journal of the Air Pollution Control  Association published
in May, 1970 (1).  While these  authors  gave several  graphs, the one (Figure 1)
relating maximum daily one-hour average oxidant  in ppm  to 6-9 a.m. average
non-methane hydrocarbons (NMHC) concentrations in ppm carbon was probably
the most significant.  Shortly  thereafter,  this  graph,  with some change in
the positioning of the line, was published  as Figure 4-2 in AP-84, the
Criteria Document for Nitrogen  Oxides  (2) (Figure 2).  On August 14, 1971,
the Environmental Protection Agency  (EPA) finalized its requirements for
State Implementation Plans.  This document  contained Appendix J -- a curve
relating the degree of HC reduction  required  to  attain  the 0.08 ppm oxidant
standard to the current ambient oxidant level.  Appendix J was derived from
Figure 4-2  of AP-84 (Figure  2).  Thus,  the  upper  limit  curve correlation became
the technical basis of the  legally-required strategy for HC control in the
United States.

     A brief review of the  conceptual basis of the upper limit curve analysis
seems appropriate.  Technically, a dependent  variable (oxidant level) might
be a function of ten independent variables  (perhaps HC  concentrations, nitro-
gen oxides  (NOs) concentration, temperature,  incident sunlight,  HC composition,
   lidity, mixing height, initial ozone  (Oo),  free radical initiators, etc.).
*Monsanto Company, Alvin,  Texas.

                                      851

-------
    0.30
    0.25
  E
  Q.
  0.
  Z
  •<
  o
  X
  o
  oc.
  UJ
0.20
    O.15
  O
  r
    0.10
    O.OS
                                                              LOS ANGELES'
                                                LOS ANGELES^.«**• DENVER
                                        WASHINGTON*   x» *            —

                                                   . ^  • LOS ANGELES
                                         ^*  . PHILADELPHIA
                                         LOS ANGELES
       PHILADELPHIA •        S
                         s
                        0
         PHILADELPHIA^/
         WASHINGTON S         »
    _WASHINGTO^yp'H|LADELpH|A

    WASHINGTON/  '  *

           V  '    ..
        WASHINGTON
             .     •
        	  /
      •  •  *  •     •  •  »
      •  •  »     •
•*  • *»       •
«  A«  *       »  •       *
*  •    » *•        »
i..: •:  *•  -    •
                 .    *
•  •.»»••*.••*»
                    0.5
                                 1.0
                                               1.5
                                                            2.0
                                                                         2.S
                       6-9 a.m. AVERAGE NONMETHANE HYDROCARBON
                                 CONCENTRATION, ppm C

     Figure 1.   Maximum daily oxidant as a function  of  early  morning non-
        methane hydrocarbons, 1966-1968 for CAMP Stations;  May through
                         October 1967 for Los Angeles.


The upper  limit concept  assumes  that,  if  enough data is available covering  a
wide range of  all  variables,  some  points  will  be found where  the combined
effect of  all  variables  except  the one  under study will be most favorable
for a maximum  level of the  dependent  variable.   Thus, if one  plots  enough
data relating  maximum  daily oxidant to  6-9  a.m. NMHC, the plotted upper  limit
                                      852

-------
   0.30
   0.25
   0.20
E
a.
a.

CO
a
X
o
   0.15
   0.10
   0.05
                                                             LOS ANGELES
             APPROXIMATE UPPER-LIMIT
             OBSERVED OXIDANT
                                                            A
                                                           DENVER
         LOS ANGELES
'WASHINGTONA    A
                 * LOS ANGELES

    A A PHILADELPHIA
    LOS ANGELES
  PHILADELPHIA,

      PHILADELPHIA
     WASHINGTON   A
          f\ n   •*
                PHILADELPHIA
WASHINGTON   AAA    A^AA    A
                       A  A A   A
                     ^*A       A
WASHINGTON  *^A MA       A A
                    A    A ^      A
                                              A A
                                            AAA

                                         AAA   A*
                                                                   A A
                   0.5
                         1.0           1.5
                       NONMETHANE HC, PPm C
                2.0
2.5
   Figure 2.  Maximum daily  1-hour-average oxidants  as  a function of 6-to-9-
          a.m. averages  of nonmethane hydrocarbons  at CAMP stations,
          June through September, 1966 through 1968, Los Angeles, May
                 through October 1967.  (Figure 4-2  of  AP-84.)


 of the observed values are deemed to represent the  maximum oxidant  level
 possible at the corresponding HC concentrations.   Such  a relationship can
 become the basis of a strategy for controlling the  dependent variable on  the
                                     853

-------
basis of the chosen single independent variable  --  as  indeed it has.   Probably
no piece of technical  analysis has  ever become more controversial.

     While the basic concept is logical and acceptable,  the practical  pro-
blems with the approach are enormous.   Among the more  significant weaknesses
are:

     •    The approach is totally dependent upon the most extreme values of
          either variable -- thus the  final result  tends to depend primarily
          upon those values most suspect by conventional methods of data
          correlation  and data validation.

     •    Because of the reaction time required  for oxidant formation  and
          wind effects, the two variables are not measured for the same air
          sample.  Typically, the observations are  removed in time by  3-8
          hours and in distance by 10  to 100 miles.

     •    There is no way to judge whether or not sufficient data is  at hand
          to yield an adequate upper limit curve.  Ultimately this becomes a
          matter of statistical analysis.  Perhaps  a useful control relation-
          ship might be derived from a set of maxima that occurs once  in 25
          observations, whereas a relationship based upon levels that  occur
          but once in 500 observations might represent a meaningless  study
          of rare phenomena.

     0    The dependent variable chosen may not  actually be functionally
          related to the independent variable being analyzed.

     Laboratory data has, in general,  qualitatively supported the HC/oxidant
relationship of the Schuck and AP-84 upper limit curves.  Aerometric  data
have, in general, failed to confirm the relationship and have yielded  many
conflicting results with uncertain conclusions.   As a result, Appendix J
has fallen into disrepute and is no longer supported by any major group.
The replacement rollback relationship  does not have an experimental basis of
support and seems to be used merely as a legal tool, attractive because of
its great simplicity.

                           THE TACB UPPER LIMIT  CURVE

     Early in 1975 a staff member of the Texas Air  Control Board (TACB),
G. K. Tannehill, constructed an upper  limit curve based on limited 1973-74
TACB "Connie" monitoring network data  (3).   This curve had a greatly  different
shape than the original upper limit curves and it had different implications
(Figure 3).

     This work was never published, but it was cited to an EPA technical
panel in a public discussion held in St. Louis last January, and it provoked
strong disagreement.  One of the objections was  that it was not based  upon a
sufficiently large data sample.  Since much more data is now available from
the Texas network, and because of its  wide geographical  coverage and  the ex-
tremely high standards employed by the Texas monitoring network, a repeat of
                                      854

-------
I
T
a
fc

ca
-P
•.10
               Figure   -2  curve
     •  •
                           Texas upper-limit curve
                            • *•
i i.  :•;:{•. -. i'   .  •
n  >•:•»•.'.-•   ••"
Iji'-'i!.''1'/'   •   •.
i«!    '-'s1':.   •   •'   ••
                                      •

                                      t
                 c>
                                       t.e
                 Nonine thane  HC,  ppm C
      Figure 3. The TACB upper-limit curve compared to the Figure 4-2
                       curve of AP-84.
 the study based upon the latest information was deemed justified.  This paper
 reports the results of that study.


    For simplicity and clarity, the plotting was done year-by-year and the

 results were finally combined by tracing.  Figure 4 presents data for the two-
 year period 1973-1974.  There simply was not enough valid data from 1973 to
 attempt an independent analysis.


                           855

-------An error occurred while trying to OCR this image.

-------
     The procedure followed was straightforward.  Copies of computer pages
giving hour-by-hour values for both ozone and NMHC were furnished by the TACB.
Computer-derived 6-9 a.m. NMHC averages were also provided.  The oxidant values
were eye-scanned for daily maxima.  These data were first tabulated and then
plotted.  All days were plotted that had valid values for 6-9 a.m. NMHC and
maximum oxidant and reached 0.08 ppm ozone or more, or equalled or exceeded
1.5 ppm 6-9 a.m. NMHC.  In these respects all of these plots are completely
comprehensive.  All of the points meeting these criteria appear on the appro-
priate graphs.

     Initially it had been planned to also include all points when the daily
mean NMHC equalled or exceeded 1.0 ppm.  However, this proved excessively
laborious; therefore, this criteria was abandoned rather early in the study.
Some points from this group, however, were plotted, which in many cases led
to points where ozone peaked below 0.08 ppm and the 6-9 a.m. NMHC was also
below 1.5 ppm.  This accounts for the scattered points in the lower left-
hand corner of the 1975 graph.  These were not deleted even though they in
no way affect the upper limit curve, because they serve as a reminder that
the lower left-hand block really contains thousands of points, should one
have the perseverance to plot them.

     Commenting specifically on the 1973-74 plot, it essentially used the
same data as Tannehill's earlier analysis with essentially the same upper
limit curve resulting.

     Plotting the 1975 data yielded a similar result (Figure 5).  At this
point let me discuss a detail that may provoke later questions.  TACB reports
NMHC values only to one-tenth percent on the basis that the accuracy of the
analytical method does not justify reporting hundredths of a percent.  Dis-
cussion with TACB staffers revealed that all of their computer data processing
involves the use of truncation rather than rounding.  Thus, the instrumental
data is first averaged for the hour and then truncated to the significant
tenth ppm.  In creating the 6-9 a.m. NMHC average, these truncated values
are averaged and again truncated to the significant tenth.  This system, on
the average, will understate the latter value by 0.1 ppm but under some cir-
cumstances can understate by as much as .167 ppm.  Some values are not under-
stated at all by the system.  TACB is now modifying these programs.

     While this procedure provides strong legal protection against any risk
of overstating pollutant data values, it is not desirable for a scientific
analysis.  When this feature was discovered, the first three years of data
had already been plotted.  A total repeat with hand-averaging was simply not
possible.  However, high accuracy in the hydrocarbons (HC) values is only crit-
ical at the lowest levels.  Therefore, new averages were calculated by hand
for all 6-9 a.m. NMHC values originally reported as .0, .1 and  .2 ppm.  No
attempt was made to round back to the nearest tenth ppm, but the data were
simply replotted at the averaged value to the nearest one-hundredth ppm.  Thus,
we have points at .03, .07, .13,  .17, .23 and .27 on the left-hand section
of the graphs, but mostly only at integral tenths of ppm in the remainder.
In general, about 1/3 of the points so processed increased by .03; 1/3 in-
creased by .07; and 1/3 did not increase.  Therefore, on the average, the bias
in the data removed by this effort was .03 ppm.  There was no way to compensate


                                     857

-------

                                                      -M
                                                      OJ
                                                       Ol
                                                      •I—



                                                       o


                                                      CD


                                                      h-

                                                      un

                                                      CT>
                                                      ns
                                                      01
                                                      O
                                                      JD
                                                      s_
                                                      fO
                                                      O
                                                      o
                                                      S-
                                                      T3
                                                      CTl
                                                       I
                                                      l-D
                                                      o
                                                      M
                                                      O
                                                      QJ
                                                      01
3NOZO AHIVO WOWIXVW


    858

-------
for the original truncation error, which should amount to 0.05 ppm on the
average.   Therefore, it is believed that all values of less than 0.3 ppm are
on the average understated by 0.05 ppm and most of the remainder of the values
are understated by 0.08 ppm.  This has nothing to do with any possible instru-
mental or sampling errors.

     Commenting on the upper limit curve for 1975, there was uncertainty
about one point that appears wild.  For the moment two choices, A and B, are
shown for the curve.

     The curve for the first six months of 1976 (Figure 6) differs little
from those of the previous three years.  It is a striking testimony to the
increased scope and efficiency of the TACB monitoring system that the first
half of 1976 yielded 310 valid, plottable data pairs, whereas the entire
year, 1973, yielded only 30 such pairs.  Of course, there are now many more
stations than in 1973; but of equal importance is the increased on-stream
factor for the HC analyzers.

                    THE COMPREHENSIVE TACB UPPER LIMIT CURVE

     Figures 4 through 6 have traced the development of the data year by year.
Figure 7 gives the final result -- the comprehensive Texas upper limit curve
based on all valid data available from all fourteen Connie monitoring stations
of the TACB network.  This is almost the final result of this study.

     In drawing the final curve, the single Houston data point at .170 ppm
0-j, 3.9 ppm HC was ignored as an outlier.  This was largely done on the
basis that to use this point would have resulted in a curve inconsistent with
the curves for the other years.

     What can we say about this curve?  First it differs from the original
curves in two major respects.  The left end of the curve intersects the Y
axis at .125 ppm 0^, suggesting that there is no way to reduce ozone maxima
below that level by HC control alone.  This high intercept is established
because this analysis did not exclude HC points below 0.3 ppm as did the
original.   There are 123 validated pairs of values in this range.  This cloud
of points effectively blocks any approach to the origin after the manner of
the AP-84 curve.

     There are ten days when all three hours contributing to the 6-9 a.m.
average were reported as .0; 16 days when two of the three hours were .0
and the other hour was .1, and 13 days with a .0, .1, .1 or similar pattern.
As was pointed out earlier, the computer system used does not guarantee that
any of the values were really .0.  Statistically, several would have been
expected to be .0 and the average for these 39 lowest points would be expected
to be .05 ppm.  It would not, however, be reasonable to attempt to drop the
curve to the origin down a slot only 0.05 ppm wide.  As a practical matter,
it would have been more justifiable to flatten the curve and intersect the
ax is at 0.16 ppm.

     The technical implication of this is, of course, that background ozone
or ozone not generated because of local HC concentrations can amount to as

                                     859

-------
                                                          o
                                                          ID
                                                                       CTi
                                                                       OJ
                                                                       >>
                                                                        I
                                                             I
                                                             CO
 to

 o ^
 fi &_
 t- O
 13 2

 O OJ
 s_ c:
T3
 >> O)
                                                          o
                                                         ' OJ
                                                                       E O
                                                                         CD
                                                                      O~> c_>
                                                                      UD
                                                                       O)

                                                                       o
                                                                       N
                                                                      o
                                                                      UD


                                                                       O)
                                                                       Z!
                                                                       en
NOIlVdlN30N003NOZOAniVa WOWIXVW
                       860

-------
                                                        J



                                                        10
                                                        T 
• ' * H II
^^ £
• ^co 3
.... ^ u.
: 	 i
L 	 §
> "0 i
' '. ':".' ', ' ^"5
' 	 <
•• . -•> •: 	 ff)
1
. . . .._-* •- .. J- , . - if)
u>
>,
CVJ

oo ^
s_
. » o
i>o 3
C 4J
O O)
/—i r~
^..J i_
s_
to CD
O •!-
O c:
S- E
T3 O
>,0
c—
CD
• CJ
E 
1
UD
,
C/l
                                                                    O)
                                                                    c
                                                                    O
                                                                    Nl
                                                                    O
                                                                    OJ
                                                                    S-
                                                                    Z5
                                                                    CJ)
Wdd NOIlVdXN30NOO 3NOZO  ATIVQ WOWIXVW
                       861

-------
much as .125 or even .16 ppm.  This is certainly consistent with many of the
recent reports of rural ozone, mountain-top ozone, aerial ozone, etc. where
high values are found with no concomitant HC found in the ranges detectable
by a Beckman 6800-type NMHC analyzer.

     While EPA has recently cast doubts upon the value of all data from the
Beckman 6800-type instrument, it is still the best instrument available for
long-term monitoring of NMHC, and it has provided the national data base to
date, including that used in the original Schuck and Altshuller study as well
as in this study.  Many improvements have been made in instrument operation in
recent years and both of the Texas monitoring groups report that, although
it is a subtraction-type instrument, negative NMHC values are seldom obtained.
This fact and the basic scheme of the instrument suggests that if there is any
instrumental error, there is a greater likelihood that high values will be
produced rather than low values.  This fact was also confirmed by the Scott
study of NMHC analyzers (4).

     All factors considered, it is probable that the intersection of the
upper limit curve with the Y axis at a level well above the national ambient
air standard (NAAS) of 0.08 ppm is a valid result and will have to be reckoned
with.

     The other and more surprising feature of the Texas upper limit curve is
that it reaches a maxima at 1.1 ppm and then descends to values below the
NAAS at HC levels above about 5 ppm.

     The chemical explanation of this may be that above 1.1 pprn the rate of
ozone destruction reactions increases more rapidly with HC concentration than
does the rate of ozone formation.

     This reverse slope also suggests that in areas of excessive HC concen-
tration, reduction will increase ozone maxima until levels below the 1.1
ppm peak are reached.  Abatement beyond this point would, of course, result in
a reduction.  This may have been the reason for the observed increases in
ozone maxima observed in several Texas cities after the institution of HC
control measures, as reported by Tannehill (5).

     Some may argue that the Texas curve drops at HC levels above 1.1 ppm
simply because sufficient points do not exist at the higher levels to assure
that the maximum combination of all other variables has been attained.  This
is a valid question.  In fact, Altshuller noted the tendency for the plot to
deteriorate into a scatter diagram as the number of data points is decreased.

     However, one should note that this study involves about seven times as
much data as the Altshuller study.  There are 263 points at levels of 6-9
a.m. NMHC above the maxima of the curve.  Since not one of this large number
of points indicates a further rising relationship, the analysis seems rather
conclusive in pointing to the validity of the up and over shape for the
curve.
days
It is regrettable that 6-9 a.i
with ozone levels above 0.20.
 NMHC values  were not available for more
In Lho on Lire 3 1/2 years, only five valid
                                     862

-------
pairs were recorded when ozone was above that level.   At the same time, 11
additional instances of such ozone levels were recorded without simultaneous
HC measurement.  Of these, four were above 0.25 and two were above 0.3.  The
highest of all, 0.38, was recorded in Nederland in March, 1973.

                 THE HOUSTON AIR POLLUTION CONTROL DISTRICT DATA

     In an effort to find other data, the Houston Air Pollution Control
District (HAPCO) was contacted.  A data set containing 59 pairs, all from
three city sites during 1975, was furnished.  This information had been reported
by McKenzie et a!., (9) at the Dallas oxidant symposium.  When these were
plotted, they were found to be quite different from the previously reported
data.  Two pair above 0.25 ppm ozone (one 0.29 ppm) were in the set which,
of course, raised the curve maxima somewhat.  The HAPCD HC concentrations were,
in general, several ppm higher than the TACB points.   Figure 8 shows the effect
of adding these points to the previously-developed curve.  Essentially the
shape remains unchanged, but the maxima is higher and has moved out to 2.4
ppm 6-9 a.m. NMHC.

     What is the reason for the apparent inconsistency of the two data sets?
To answer this a comparison was made of data that was on hand from the two
networks.  Specifically, TACB's Mae Drive Station was compared with Houston's
Clinton Drive site.  The two are both industrially oriented and are only about
one mile apart; the Houston city site is slightly over one mile from an
industrial area; TACB's Mae Drive site is about twice this distance.  The
observed ratio of monitored NMHC's was 1.45, Clinton over Mae.  The TACB
Aldine and HAPCD Parkhurst sites are both on the north side of Houston about
six miles apart.  Both are suburban; Parkhurst is at least five miles from  the
ship channel industrial area; Aldine is eleven miles in the same direction.
The monitored NMHC ratio was Parkhurst over Aldine 3.29 to 1.  This was for
20 days during the late summer of 1975 when both sites monitored HC.  The Mae
Drive - Clinton Drive comparison was for 16 days.

     It is thus clear that the HAPCD sites on the average ran more than twice
the NMHC level of the TACB sites.  Since such mesoscale discontinuities in
HC values would not be expected to be matched by a mesoscale change in ozone
maxima, it becomes clear why the HAPCD points gave a curve with a maxima more
than twice that of the TACB curve.

     It has been suggested that the upper limit concept may be city-specific.
This analysis suggests that it is probably site-specific.  The more important
result is that the HAPCD upper limit curve has the same overall shape as the
TACB curve.

     One further analysis was attempted.  All of the ozone data was plotted
versus daily mean NMHC and an upper limit curve was constructed (Figure 9).
This curve is essentially the same as the prior curve, except for a consider-
ably shortened HC axis.  The outlier point moved back into the pack, offering
some justification for the previous decision to ignore this point.  The utility
of the maximum ozone versus daily mean HC plot appears equal to that of maximum
ozone versus 6-9 a.m. HC curve.
                                     863

-------
r
                                                                                                         i in
                                                                                                         I 10
                                                                                                                          un
                                                                                                                          r^
                                                                                                                          cr>
                                                                                                                              -*->
                                                                                                                              c:
                                                                                                                              CD
                                                                                                                           CO Q.
                                                                                                                           C O)
                                                                                                                           O Q
                                                                                                                           ro  O
                                                                                                                           O  i-
                                                                                                                           O 4->
                                                                                                                           s-  sz
                                                                                                                          -a  o
                                                                                                                               cz
                                                                                                                               o
                                                                                                                          CTi i—
                                                                                                                            I   O
                                                                                                                           cu  c
                                                                                                                           c:  o
                                                                                                                           o +->
                                                                                                                           N  00
                                                                                                                          o  3
                                                                                                                               o
                                                                                                                           co
                                                                                                                           a>
                                                          864

-------
                                               O
                                               ro   (/)

                                                   1
                                                             S-
                                                             -C
                                                             4->


                                                             CO



                                                             CTl
                                                             i/i
                                                             s_
                                                             03
                                                             CD
Q
'£ I
I
UJ
z
o t
• IAJ ^
Z
o
z
z
w
^ 2
5

• ri
(/)
o
JD
rt3
O
O
5-
T3
^1

C.
03
CD
E

r—
ro
0 "°
•
. o -...
CD
s_
o
CD

OJ
c
0
o

(/I
fO
X
CD
h-






                                                             O
                                                             INI
                                                             C3
                                                             CD

                                                             13
                                                             cn
NOI1VH1N30NOO 3NOZOAllVa
           365

-------
     All of the results argue strongly that there is a major maxima in the
ozone HC functional relationship.  Is this consistent with laboratory work
and theoretical analyses?  To examine this point it is noted that several
analyses have been made of the ozone-HC-NOx relationship.  An early one was
that of Dimitriades (6) whose 1972 graph in Environmental Science and Technology
indicated a slope to the isopleth line for 0.1 ppm oxidant that implied in-
creased oxidant with reduce HC in the region of low NOX (Figure 10).  This
was more fully developed by Hecht and Seinfeld (7), who were able to plot a
family of ozone isopleths versus NOX and HC (Figure 11).

     On Figure 12, a similar plot by Quon and Wadden (8)  is reproduced.  The
shaded area is the area where 9Q% of all TACB monitoring  points fall.  All
of the Texas sites must be characterized as having very high HC-to-NOx ratios.
These vary widely, but typically range 17.5/1 to 70/1 for the cities in Texas
having the higher HC levels.  Referring to Figure 12, in  the region of the
Texas data where HC are above 1.0 ppm, a simple reduction in these levels will
increase, not decrease, ozone.  Only after the axis of the isopleths are
passed at about 1.0 ppm will HC reduction produce oxidant reduction.  At the
St. Louis discussions, it was stated that a maxima would  be found only above
a 12.5/1 ratio of HC-to-NOx.  The Texas ratios appear to  be generally well
above that level.  Thus, the shape of the Texas upper limit curve is con-
sistent with these theoretical considerations.

                                     SUMMARY

1.   The Texas upper limit curve intersects the .0 hydrocarbon axis at an
     ozone level well above the ambient air standard, reflecting background
     ozone of various types.  This is in agreement with recent field studies.
     The Texas curve has a pronounced maxima.
     theoretical considerations.
                                               This  would be expected from
3.

4.
     The upper limit analysis appears to be site-specific.

     Plotting maximum ozone versus daily mean HC yields results  similar to
     plotting versus 6-9 a.m. HC,  except for the shortened  axis.

                                ACKNOWLEDGEMENTS

     The author wishes to acknowledge the cooperation and help of the staff
of the TACB, particularly Mr. Roger Wallis and Mr.  L.  Butz, in conducting
this study.   The assistance of Mr. L. Wenzel of the HAPCD is  also acknowledged,
                                 REFERENCES

1.   Schuck, Altshuller, Barth and Morgan  JAPCA  20, 297, May 1970,

2.   AP-84 Criteria Document for Nitrogen Oxides, EPA  January 1971,

3.   Federal Register, April 7, 1971,  p. 6700.
                                     866

-------An error occurred while trying to OCR this image.

-------
OJ  —   —   —  —   —
                                                a.
                                                CL
                                                         o c:
03 O
S-
4-> 00
sr CD
CD S-
O Z3
d | v
O X
U T-

CD
C oo
O =5
Nl O
O T-
  S-
E 03
3 >

•i- M-
X O
OJ
E E
  O
T3 -i—
O) -M
4-1 fD
O -i-
•i- TD

CD S-
S- S_
CL-r-

4- S-
O -C
  I
oo CO

+J SZ
CD fD
i—
Q- cn
O sz
CO -I—
i—i S-
  Z5
  -a

i— TD
r— CD
  >
CD CD

rs ^i
CD U
•r- ra
                                                            4-
                                                            o
  O

  O
 • +->
O
2: r—
  ro
•a rs
C cr
(T3 ,
                                                             Q-
                                                             O
                                                             S-
                                                             Q.
(lAIdd) NOadVDOdOAH  1V101
                         868

-------
       0.0
1.0        2.0       3.0       4.0       5.0       6.0
Non-methane Hydrocarbon  Concentration, ppmC
Figure 12.   Ozone concentrations  for  various  initial  hydrocarbons-NOx mixtures.
                       Legend:  ——Calculated  values.
                               	  Measured  values  using  automobile exhaust.
 4.    Reckner,  L.R.,  Survey  of  the  EPA-Reference Method for Measurement of
      Non-Methane  Hydrocarbons  in Ambient  Air   PB-247   515.

 5.    Tannahill, G.K.,   The  Hydrocarbons/Ozone  Relationship in Texas.  Presented
      at  APCA Symposium  March 11, 1976, Dallas, Texas.

 6.    Dimitriades,  B., ES&T, 6_, 253  (1972).

 7.    Hecht, Seinfeld and  Dodge, ES&T, 8_,  327 (1974).

 8.    Quon,  J.E.,  & Wadden,  R.G., "Oxidants in  the Urban Atmosphere,"
      Illinois  Institute for Environmental Quality Report, January 1975.

 9.    Wenzel, MacKenzie  and  Siddiqi,  APCA  Symposium  on  Oxidants, Dallas,
      Texas, September  1975.
                                     869

-------
                                                                            18-2
                      AN  ALTERNATIVE TO THE APPENDIX-J METHOD
                     FOR  CALCULATING OXIDANT- AND N02-RELATED
                               CONTROL REQUIREMENTS

                                  B. Dimitriades*

ABSTRACT

     A method is described and o^eAe.d tu an alte,sinative. to the. Appendix-J
method ffO.fi calculating num^-ical mission control SLe.quAAeme.nti>.   The. ne.W method
•iA baAe.d on  the. uAe.  ofi oxides/kyd'iocaA.bon/nitJioQe.n oxides ie£ationAkipA that
we.ne. derived ^wm &mog chambeA data and w&ut e.x.pande.d uAing photoc-he.m4.cai
modeling te.chntqu.eJ).  Main  advantages ofi the. method OVCA. the. kppe.ndi.K-J method
nou) in uAe. aA.e. the. cauAe.-e.^e.ct nature. o{, the. oxides/hydA.ocaAbon/nitAoge.n
*ie.la.tionshipA uAed and the. quantitative tie.at.me.nt o& the. nitrogen oxide.s-pn.c-
cutiAofi note. ViAadvantageA ane. the  A o me what questionable. appticabiLLty ofi the
&mvQ chamber data and the  method'&  demand ^01 non-available input information.
TheAe advantages and disadvantages  and ot.heA contA.ol-Ke.la.te,d impticat^ionA o&  the.
method'* ox.ides/hijdAocasibon/n.itJiogen oxides diagtiamA afie. discuAAe.d.

                                   INTRODUCTION

     In the  first oxidant  control  strategy recommended by the United States
Environmental Protection Agency (EPA) in 1970, emission control requirements
were to be calculated by the well-known Appendix-J method (Figure 1) (1).
Briefly, the method  is based on use of an "upper limit" curve (Figure 2) re-
lating observed daily maximum one-hour oxidant (0 ) and 6-9 a.m. non-methane
hydrocarbon  (NMHC) concentrations  (2). The Appendfx-J method has been known to
suffer from  several  severe limitations -- discussed extensively elsewhere  (3)  --
and for this reason  EPA  has been conducting an extensive research program aimed
at development of alternative, superior methods.

     One product from the  research  effort is the "smog chamber method" to be
described and discussed  here.   As  with the Appendix-J method, the primary
purpose of this new  method is to estimate the degree of emission control needed
for achievement of the national ambient air quality standard (NAAQS) for oxidant/
Ozone (0}).  The new method also lends itself to predicting changes in oxidant-
related air  quality  as a result of given emission control.  It should be clari-
fied at the  outset that  the entire  method consists of two components used
sequentially:  One component deals  with the dependencies of ambient oxidant/O^
and nitrogen dioxide (NO?) on ambient NMHC and nitrogen oxides (NO  ) concen-
trations; the follow-up  component deals with the dependencies of tfie ambient
NMHC and NO, concentrations on respective emission rates.  The procedures

*U.S. Environmental  Protection Agency, Research Triangle  Park, North Carolina


                                       371

-------
described and discussed here are addressed to  the  first  component  only,  that  is,
to the ambient oxidant/03-to-precursor dependencies.   Emission  reduction is
calculated by simply assuming proportional  relationships between ambient con-
centrations and emission rates.
            0.30
            0.25
            0.20
         OL
         O.
        O
        X
        O
            0.15
             0.10
            0.05
                  APPROXIMATE UPPER LIMIT
                     OBSERVED OXIDANT
              .v.  .  .   . f
           ••.  t  .  :•.••: ••••
                                                 •      • •
                                                   •  •  •
                                                   • •   ••
                                     I
                                   I
                          0.5        1.0        1.5
                                NONMETHANE HC, ppm C
                                            2.0
2.5
 Figure  1.
"Upper limit"  curve  relating  ambient maximum  one-hour oxidant  to
 6-9 a.m.  concentration  of  non-methane  hydrocarbon.
                            DESCRIPTION OF THE METHOD

     The smog chamber method is based on use of oxides/hydrocarbon/nitrogen
oxides (oyHC/NO )  relationships first derived from smog chamber data and
subsequently modified to more closely reflect real  atmosphere  conditions.
Therefore, for complete description of the method,  it is necessary  to describe
(a) the underlying  smog chamber data and the experimental  procedures  by which
they were obtained, (b) the photochemical  modeling  techniques  used  in conjunc-
                                     872

-------
CO
 e

 DO
<
o:
O

X
CD
O
O
o
ac
Q-
O
LU
CC

afe
                                                                            
           c: 01
           o o
           .a c
           S- O
           O3 (J
           U
           O +J
           s- c
                                                                                          x
                                                                                       en O
4-> O)
(13 S-
i— 3
CD (/>

   O)
oj e

s- E
3 r;
<-> E

•-D X

X E
                                                                                       c:
                                                                                       O)
                                                                                       CL
                                                                                       cx
                                                                                      C\J
                                                                                       cu
                                                                                       cn
                % 'SNOISSIW3 OH Nl NOI10nQ3d Q3Minf)3a
                                           873

-------
tion with the smog chamber data to derive the requisite 0 /HC/NO, relationships,
and (c) the use of such relationsips to estimate control requirements.  The smog
chamber data and the modeling techniques are described and discussed elsewhere
(4,5,6).  The application of these diagrams in calculating control requirements
will be described and discussed here.

     The 0 /HC/NOx relationship diagrams used in this method are depicted in
Figure 3 ifi the form of oxidant/03 isopleths.   For application of these labo-
ratory-derived diagrams to the real  atmosphere,  it is necessary that the chamber
concentrations of oxidant, NMHC, and N0x be identified or related with the
concentrations in the real atmosphere.  xSuch relationships have been assumed
here to be as described in the following paragraph.

     The highest afternoon concentration of oxidant/03 observed in a given
region's air is identified with the oxidant/03 concentration observed in the
smog chamber. Also, the early-in-the-morning (6-9 a.m.) average (during the smog
season) NMHC-to-NO^ ratio associated with the observed (real atmosphere) oxidant/
0-.J  is identified with the NMHC-to-NOx ratio in the initial chamber charge.
Thus, ambient oxidant/03 and NMHC-to-NOx ratio data specify in Figure 3 a single
oxidant/03 isopleth, and define a constant slope straight line emanating from
the origin.  The intercept point of these two lines defines the "chamber coun-
terparts" of the ambient NMHC and NO^ concentrations associated with the ob-
served ambient oxidant/03 concentration.  These "chamber counterpart" concen-
trations are not expected to be equal to those in the real atmosphere; however
the degree of their reduction needed to achieve given degrees of oxidant/03 and
N0? reduction are taken to be equal to the degrees of NMHC and NO  reduction
needed in the real atmosphere.

     The detailed procedure for calculating control requirements is illustrated
for a hypothetical region, the relevant air quality data for which are given as
follows:

     i    Maximum one-hour oxidant/03 in "base" year:  0.40 ppm

     •    Average 6-9 a.m. NMHC-to-NO  ratio and 90 % confidence interval
          during the smog season of tfie "base" year:  8.0 ± 4.0

     >    Annual N02 mean for the "base" year: 0.06 ppm


     Because control requirements depend on the NMHC-to-NO  ratio, such re-
quirements must be calculated for all ratio values rangingxfrom 4 to 12.  Such
calculations are illustrated in the following paragraph, using the average ratio
value of 8.0.

     First, the constant slope line ab is drawn (Figure 3), corresponding to a
NMHC-to-NO  ratio equal to 8.0.  The intercept of line ab with the 0.40-ppm-O
          ,x
                                                                             3
                                                                           x ,
isopleth defines point e representing the "chamber counterpart" NMHC and NO
concentrations of 1.57 ppmC and 0.20 ppm, respectively.  Next, the NO  reduction
needed for achieving the NAAQS for N0? (annual mean of 0.055 ppm) is calculated
to be 10 % -- assuming proportionality between 6-9 a.m.  N0x and annual N02 mean
concentrations.  Reducing N0x by 10 % defines point e1 representing 1.57 ppmC of


                                     874

-------
                            OXIDANT/03, ppm:

                    .08  .20 .30 .40 .50 .55 .60  .65
                    1.0
2.0           3.0

   NMHC, ppmC
                                                            4.0
Figure 3.  Oxidant/03 isopleths derived from combined  use of  smog  chamber  and
                     Photochemical modeling techniques.


NMHC and 0.18 ppm of N0x-   hor  achievement  ot  the  NAAQS  for oxidant/03,  the NMHC
must then be reduced along  the  line  e'f to  the point f being  the intercept of
line e'f with the 0.08-ppm-03 isopleth.   Such  a NMHC reduction  is  calculated to
be 8V;:
     Similar calculations for NMHC-to-NO ratio equal  to 4 and  12  give  NMHC
control requirements equal to 62% and 87%,  respectively, the latter  figure  being
the maximum within the given ratio range.   It is concluded, therefore,  that the
control requirement estimates for
and 87% for the NMHC emissions.
 the given region are  10%  for  the  NOY  emissions
                                    DISCUSSION

     Points  that  merit  discussion are  (a)  the strengths and limitations of the
new method  relative  to  those of the Appendix-J method, and (b) control impli-
cations of  the  relative roles of the HC and NO  precursors as such roles are
depicted  in  Figure 3.   Since the smog  chamber method is offered as an alter-
native to the Appendix-J method, any judgement to be made upon the new method
needs only  be a comparative  one.
                                     875

-------
     Tn regard to the conceptual bases, the smog chamber method is advantageous
over the Appendix-J method in that it is based on a cause-effect relationship
between oxidant/03 and the precursors.   In contrast, the Apperidix-J method is
based on a relationship of stochastic nature and therefore its predictions have
limited validity.  Another major advantage of the chamber method is that it
considers and, in fact, permits quantification of the role of the NO  precursor
factor in the oxidant/03-forming process.   A conceptual limitation o£ the smog
chamber method arises from the fact that the ambient atmosphere cannot be
duplicated in the smog chamber. This  problem, however, should not be attributed
wholly to the chamber method.  A part of the problem is caused by the fact that
the ambient atmosphere cannot be defined,  at least not in terms of commonly
available monitoring data.  For this  reason, any method for predicting air
quality will suffer from uncertainties, as direct validation of such predictions
is not possible.   Another conceptual  limitation of the smog chamber method is
that although it predicts levels of oxidant/03 occurrence, it offers no infor-
mation whatever on the frequency at which  such levels occur.  This is a limi-
tation only because the NAAQS for oxidant/0; has been defined in terms of a
concentration level and a frequency of occurrence.

     Aside from its conceptional strengths and limitations, the smog chamber
method has strengths and limitations  related to  its derivation arid to its
application.

     Use of photochemical modeling techniques in the derivation of the method
has extended the method's applicability into ranges of precursor concentration
and radiation conditions that are known to occur in the real atmosphere, but for
which smog chamber data either do not exist or are not reliable.  This and other
advantages as well as errors/limitations associated with the modeling techniques
are discussed in detail elsewhere(6).

     Application of the smog chamber method is not as simple as with the Appendix-
J method, mainly because it requires  more  input information.  Thus, while both
methods require knowledge of the highest (or second highest) one-hour oxidant/03
and annual N02 mean concentrations for the base year, the smog chamber method
needs, additionally, data on the ambient 6-9 a.m. NMHC-to-NOx ratio.  These
latter data are not required to be obtained by law; therefore, at present their
availability is limited. The requisite data can be obtained through ambient
measurements that must be sufficiently abundant to provide a reliable measure of
the range of the NMHC-to-NOx ratio.  Thus, it is estimated that a three-month
research type effort could provide the requisite minimum amount of aerometric
data.  In the lack of local ambient data,  and provided adequate emission in-
ventory data are available, the average NMHC-to-NO  ratio can be approximated
from such inventory data; the variation around this mean ratio value could be
assumed to be equal to the variation  measured in other areas of similar pol-
lution characteristics.  In the lack  of either ambient or emission data, the
requisite ratio range could be given  rational, assumed values.

     Final points of discussion are some implications arising from the diagrams
of Figure 3 and pertaining to the relative roles of the HC and NO  precursors in
oxidant formation in atmospheres with unusually high NMHC-to-HC ratios and in
rural atmospheres.


                                      876

-------
     The diagrams of Figure 3 were derived from smog chamber data reflecting
ambient conditions within a source area (urban area).  Based on these diagrams,
it appears that the relative roles of the HC and NO  precursors can be dras-
tically different depending on the air quality goalxthat must be achieved.
Thus, if this goal is the NAAQS for oxidant, 0.08 ppm 03, then the diagrams
indicate that unilateral control of HC is the optimum strategy for achieving the
goal, since the degree of NO  control needed to achieve the same goal is pro-
hibitive.   The diagrams further indicate that control of N0x may have detri-
mental or beneficial temporary effects but ultimately it isxdetrimental in that
it raises the HC control requirements and, hence, makes achievement of the final
goal more difficult.

     If the air quality goal to be achieved is relaxed, e.g., to 0.15 - 0.20 ppm
0( -- these values are used for illustration purposes only -- then unilateral HC
control is again an effective control strategy, especially for atmospheres with
relatively low NMHC-to-NOx ratios, namely, lower than 5.0. For such atmospheres,
control of NO  is clearlyxdetrimental.   However, for atmospheres with increa-
singly high N^MH-to-NOy ratios, N0x control becomes beneficial.  The critical
NMHC-to-NOx ratio beyond which thexNOy role is reversed has not been estab-
lished, but the overall experimental evidence points to a ratio of 8 to 10,
somewhat higher than the value of 5.6 estimated from the diagrams of Figure 3.

     With regard to the rural oxidant problem, the evidence now available is not
sufficiently complete to permit definition of optimum control strategy. There
are several sources of rural oxidant, the relative strengths of which have not
been established quantitatively.  Transported urban pollutants are almost
certain to constitute the major source in rural areas with an oxidant problem
(7-11).  However, the dependence of such "transported" oxidant on the precursors
has not been established.  Such dependence, or alternatively the effect of
precursor control on rural oxidant, is an extremely complex function of several
factors, as explained in the following paragraph.

     Based mainly on smog chamber evidence, HC control is expected to have two
offsetting effects upon rural oxidant:

          Hydrocarbon control will reduce urban oxidant -- a part of the decre-
          ment will be carried into the rural areas during subsequent days.

          Hydrocarbon control will result in greater amounts of NO  surviving
          the urban process and being carried into the rural areas.  There is
          evidence to show that the potential for oxidant formation in such
          "aged" rural air increases with increasing NO  concentrations (11).
                                                       X
Control of NO  alone has similar offsetting effects when applied upon urban
atmospheres with low NMHC-to-NO  ratio (<5):  (a) it will increase urban oxi-
dant, a part of the increment to be carried.into the rural areas, and (b) it
will result in rural air with lower N0x concentration and, hence, with lower
potential for oxidant formation.  Note, however, that in urban atmospheres with
high NMHC-to-NOx ratio (>10), control of N0x will reduce oxidant both in the
urban and in the downwind rural areas.  In conclusion, except for the case of
the high NMHC-to-NOx ratio atmospheres, the net effects HC and NO  controls upon
rural oxidant are neither known nor predictable at this time.

                                      877

-------
     Considering this complexity of the rural oxidant problem and judging from
the diagrams of Figure 3, it appears that the most rational emission control
strategies at this time are as follows:

     1.   For achievement of the 0.08 ppm-03 air quality standard, the emphasis
          should be on HC control.  The N0x should be controlled only to the
          extent needed to meet the NAAQSxfor NO? and/or to prevent ambient NO
          concentrations from increasing as a result of oxidant-reduction.

     2.   For achievement of a less stringent air quality goal, the optimum
          strategy depends on the urban NMHC-to-NOx ratio, namely:  (a) for low
          ratio atmospheres, the strategy should be as in case 1.. and (b) for
          high NMHC-to-NOx ratio atmospheres, credit should be given to both HC
          control  and NO Control.
                        X

                                   REFERENCES

1.    Federal Register, 36, 115486-15506, August 14, 1971.

2.    U.S.  Environmental  Protection  Agency.   "Air Quality Criteria for Nitrogen
     Oxides."  Air Pollution Control Office Publication No. AP-84, Washington,
     D.C., January 1971.

3.    National Academy of Sciences,  Transportation Research Board. "Assessing
     Transportation-Related Air Quality Impacts."  pp 8-20. Special Report 167.
     National Academy of Sciences,  Washington, D.C. 1976.

4.    Dimitriades,  B.   "On the Function of Hydrocarbon and Nitrogen Oxides in
     Photochemical Smog Formation." U.S. Bureau of Mines, Report of Investiga-
     tions, RI-7433,  September 1970.

5.    Dimitriades,  B.  "Oxidant Control Strategies. Part I.   An Urban Oxidant
     Control Strategy Derived from  Existing Smog Chamber Data."  Accepted for
     publication in Envir. Science  & Techno!. 1976.

6.    Dodge, M. "The Combined Use of Modeling Techniques and Smog Chamber Data to
     Derive Ozone-Precursor Relationships."  Paper presented at the International
     Conference on Photochemical Oxidant Pollution and Its Control, Raleigh,
     N.C., September 13-17, 1976.  Proceedings to be published as USEPA report.

7.    Research Triangle Institute, "Investigation of Rural  Oxidant Levels as
     Related to Urban Hydrocarbon Control Strategies,"  EPA-450/3-74-034, March,
     1975.  U. S.  Environmental Protection Agency, Research Triangle Park, N.C.
     27711.

8.    Blumenthal, D.L., W.H. White,  R.L. Peace, and T.B. Smith.  "Determination
     of the Feasibility of the Long-Range Transport of Ozone or Ozone Precursors.1
     EPA-450/3-74-061, November 1974.  U.S. Environmental Protection Agency,
     Research Triangle Park, N.C. 27711.
                                     878

-------
9.    Cox, R.A., A.E.J.  Eggleton, R.G.  Derwent, J.E.  Lovelock, and D.H. Pack.
     "Long-Range Transport of Photochemical  Ozone in North-Western Europe,"
     Nature, Vol.  255,  No. 5504, pp.  181-121, May 8, 1975.

9.    Research Triangle  Institute.   "Study of the Formation  and Transport of
     Ambient Oxidants in the Western  Gulf Coast and North-Central and North-East
     Regions of the United States."  RTI Report pursuant to USEPA Contract 68-
     02-2048, September 1976. U.S.  Environmental Protection Agency, Research
     Triangle Park, N.C. 27711.

10.  Research Triangle  Institute.   "Ambient Monitoring Aloft of Ozone and
     Precursors in the  Vicinity and Downwind of a Major City Using a Balloon-
     Borne Platform and Aircraft."   RTI report on part of Project DaVinci under
     USEPA contract 68-02-2341,  1976.  U.S. Environmental Protection Agency,
     Research Triangle  Park, N.C.  27711.

11.  Ripperton, L.A., W.C. Eaton,  J.E.  Sickles, II, "A Study of the Oxidant-
     Precursor Relations Under Pollutant Transport Conditions," Final Report  to
     Environmental Protection Agency,  EPA Contract No. 68-02-1296, January,
     1976. U.S. Environmental Protection Agency, Research Triangle Park, N.C.
     27711.
                                     879

-------
                                                                           18-3
                  COMBINED USE  OF MODELING  TECHNIQUES AND SMOG
              CHAMBER DATA TO DERIVE  OZONE-PRECURSOR RELATIONSHIPS

                                  M.  C.  Dodge*

ABSTRACT

     A method is o^eAe.d  &OA deA^iving  ozone,-pAe,cuASoA Aerations hips that can be.
used to design a conttol  stAate,gy &OA uAban ozone. Ae.du.cti.on.   The. method i&
deAi.\ie,d &Aom the. combined Aesults ofi  smog chambeA data, and photochemical model-
Ing te.chniqu.es.  A model,  wot, ^iASt  developed to  fait the. max.imu.rn one.-houA ozone.
le.vels obtained in smog chambeA studies  o&  iAAad2ate.d auto e.xhaust and oxi,des o
nitAoge.n mixtures.  The. model was the,n adjusted  to conditions that mane, closely
approximate, those, ofi the,  pollute.d atmospheAe.   Included -in these, adjustments
the. use. ofa di.uA.nal values o& the photolytic note, constants ^on the. atmoApheA
pollutant^.  Base.d on the. modeling  ftesults,  ozone, isopteths weAe. con&tAucte.d
a wide. Aange. ofi initial pollutant le.vels.   These isopleths can be. use.d to
estimate, the, variious combinations o^  non-methane, hydJiocaAbon and oxhides o&
nitAogen that meet the, fe.deA.al  aiA  quality  standard &OA ozone..

                                  INTRODUCTION

     The use of smog chamber data in  conjunction with modeling offers a means
for estimating the degree of hydrocarbon (HC)  and oxides of nitrogen (NO )
control needed to achieve the air quality standard for ozone (03).  Chamber data
by themselves are not ideally suited  for this  purpose because of their many
limitations.  No one chamber study  has yielded a sufficient data base to char-
acterize  fully all HC and NO   concentration ranges of interest.  Often the
irradiation period of the studies is  insufficient to determine full 03 potential
of the chamber mixture.   Also,  smog chambers are normally operated at constant
light intensity.  The spectral  distribution of the light source is often not
comparable to that of natural sunlight.   In addition, chamber walls can serve as
both a source and sink for pollutants; as a consequence, the results obtained
from chamber studies may  not be applicable  to the real world.  Modeling com-
plemented with chamber data offers  a  means  of circumventing many of these
problems.

     The modeling studies presented in this paper were carried out in two
stages.  In the first stage, a  model  was developed to fit chamber data of
irradiated auto exhaust and N0x mixtures.   In  the second stage, adjustments were
made to yield a model that is more  descriptive of the polluted atmosphere.  The
adjusted model was used to generate maximum one-hour 03 concentrations over a
wide range of HC and N0x  concentrations. Only the results of the modeling
studies will be presented here.  A  suggested use of these results to design an
*Environmental Protection Agency,  Research  Triangle Park,  North Carolina

                                      881

-------
alternative control strategy for urban ozone reduction will  be discussed in a
separate paper (1).

                                     RESULTS

DEVELOPMENT OF SMOG CHAMBER MODEL

      The data base was obtained in the extensive study of irradiated auto
exhaust and NOX mixtures conducted by the Bureau of Mines.  In this study,
a 100-ft3 aluminum chamber was charged with dilute auto exhaust mixed with
varying amounts of added nitric oxide (NO) and nitrogen dioxide (NCh).
Nonmethane  hydrocarbon (NMHC) and NOX reactants ranged from 0.2 to 5.0 ppmC
and 0.08 to 1.4 ppm, respectively.   The initial  N02 concentration was ad-
justed to approximately 10 percent of the initial N0x.  NMHC-to-NO ratios
of the resulting mixtures ranged from 2.3 to 11.6.  Chamber contents were
irradiated for 6 hours with a mixture of black-lights, blue lamps, and sun-
lamps.  The rate of photolysis of N02, KI, in the chamber was determined
to be approximately 0.29 min-1.  A complete description of the experimental
procedures and results obtained in this study are presented elsewhere (2, 3).

     A detailed chemical kinetics mechanism was  employed to model  the auto
exhaust data.   The 75-step mechanism contained two HC species:  propylene and n-
butane.   An aromatic hydrocarbon was not included in the model since a reliable
mechanism for this class of organics has not been developed to date.  A detailed
description of the reaction steps included in the propylene and n-butane me-
chanism can be found elsewhere (4).

     The following conditions were used to adapt the model to the Bureau of
Mines chamber characteristics:

     •    ki was taken to be 0.32 min-1.  This is 10 percent higher than the
          reported value of 0.29 min"1.   Intensity measurements in the chamber
          were conducted in the first 1  or 2 minutes after switching on the
          lights.  Since the lamps were not fully warmed-up and were, therefore,
          operating at less than full  intensity, the measured kL is apt to be
          low.

     •    The remaining photolytic rate constants were assigned values corre-
          sponding to a solar zenith angle of 60° for a 34°N latitude (Los
          Angeles).  At this zenith angle, k;t =  0.36 min"1,  which is close to
          the ki value reported for the smog chamber.

     •    The reported (5) chamber dilution rate of 2.3 percent per hour and the
          half-life of 03 in the dark of 14 hours were used in the model.

     •    The initial nitrous acid (HONO) concentration was  taken to be 6 X
          10-1* ppm.  This value reflects the formation of HONO, from the thermal
          reaction of NO, N02, and water (H20),  occurring in the period between
          injection of the reactants and onset of irradiation.

     •    To simulate "dirty" chamber effects (6), the chamber walls were
          assumed to be a source of propylene.  A reaction was included in the


                                     882

-------
          model to allow propylene to desorb continually from the chamber
          surfaces at a rate of 1 X 10"4 ppm min'1.  This rate reproduced the
          high 03 yield obtained in a chamber run containing only background
          levels of HC (<0.1 ppmC) and 0.081 ppm N0x.


     •    The  reported initial concentration of formaldehyde (HCHO) was included
          in the model.  Acetaldehyde was taken to represent the remaining
          aldehyde content of the auto exhaust sample.   Its molar concentration
          was  assumed to be two-thirds of the initial HCHO concentration.

     Seventeen experiments were modeled in this study.   The one adjustable
parameter used to achieve a fit to the data was the  relative concentration of
initial n-butane and propylene.  The closest fit to  the experimental data was
obtained when  the auto exhaust mixture was assumed to contain 75 percent n-
butane and 25  percent propylene on a ppmC basis.

     The "goodness of fit" was determined by comparing the simulated profiles
for N02 and 03 to the experimentally generated profiles.  The agreement between
the experimental maximum one-hour 03 concentration and the simulated value for
each of the runs is given in Table 1.  The chamber data  were reported in terms
of total non-N02 oxidant (7) measured with a Mast meter  (and corrected for N02
interference).  Such oxidant was estimated to consist of about 90 percent 03 and
10 percent non-03.  The experimental 03 values listed in Table 1 are, therefore,
90 percent of  the reported maximum one-hour average  oxidant concentrations.  The
first entry in Table 1 corresponds to the chamber run containing only background
levels of hydrocarbons discussed earlier.

     In general, the experimentally determined maximum one-hour 03 levels and
the predicted values are in good agreement.   On the average, there is a 14
percent difference between  the experimental  and simulated results for these 17
runs.   The lack of agreement is most pronounced for Runs No. 1  and 2.  The poor
fit for these runs,  however, is deceptive.   The high NO  concentration in Run
No. 1  inhibited 03 formation to the extent that, after ^ hours  when the experi-
ment was terminated, 03 was just beginning to appear.  Under such conditions,
the results are extremely sensitive to any small difference between the experi-
mental  and predicted induction periods.   The model  predicts a slightly greater
induction period than was observed.  It is expected that agreement would have
been much closer if the chamber contents had been irradiated until the 03
maximum was reached.   Run No.  2 contained an exceedingly low HC concentration
and, consequently, this run was highly affected by  chamber contamination.   In
this experiment, the "dirty" chamber effect on 03 production was as important as
the initial  reactant parameters.   The poor agreement between the experimental
and simulated results probably reflects the inability of the model to treat
chamber artifacts accurately.   If Runs No.  1 and 2  are excluded, the simulated
results for the remaining 15 runs are in error by an average of only 10 percent.
The experimental error associated with the smog chamber itself is at least as
great  as this.   Seven replicate runs were carried out during this chamber study.
The average difference in the maximum one-hour 03 concentrations for each of
these  pairs of runs  was 15  percent.
                                     883

-------
          TABLE 1.  COMPARISON OF CHAMBER DATA AND MODEL PREDICTIONS
No.   NMHC/NO
NMHC
NO
Max 1-hour Ozone

Expt.         Calc.
Percent Error
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
*
t
<1.
2.
2.
2.
3.
5.
5.
5.
5.
5.
7.
7.
7.
8.
11.
11.
11.
11.
Units
Units
2
3
5
5
3
1
1
1
3
3
3
4
6
3
4
4
5
6
Of
of
DEVELOPMENT
<0
1.
0.
0.
0.
0.
4.
1.
0.
2.
0.
4.
3.
2.
3.
0.
4.
2.
ppmC
ppm
u
45
187
656
92
415
80
403
657
12
97
75
14
13
21
95
79
13


OF ADJUSTED
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.


081
618
075
265
276
083
046
272
125
398
133
641
414
256
281
083
416
183


0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.


025
073
118
103
141
188
425
240
207
323
199
558
458
355
373
252
623
332


0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.


025
041
075
088
166
151
458
247
176
340
222
539
432
343
393
194
493
308



44
36
15
18
20
8
3
15
5
12
3
6
3
6
23
21
7


MODEL
     In order to relate 03 formation to HC  and NO  emissions,  it is  first
necessary to adjust the smog chamber model  to conditions  that  are more repre-
sentative of the polluted urban atmosphere.   The smog chamber  model  described
above is not suitable for predicting ambient. 03 concentration  for several
reasons:

     •    The present model  treats  chamber  artifacts, such  as  wall  degassing of
          propylene, that have no counterpart in the  atmosphere.

     •    The chamber contents were irradiated for only 6 hours.  The light
          intensity for this period of time in the chamber  is  substantially less
          than the integrated light intensity for a 10-12 hour period on a
          typical summer day.  Of the 17 chamber runs modeled  in this study,
          only 6 (Runs No. 11, 12,  14-17)  reached an  03 maximum during the 6-
          hour irradiation.   03 was still  on the rise in  the remaining runs when
          the experiments were terminated.   Thus, in  most of these runs, the
          potential for 03 production is greater than what  was observed.

                                     884

-------
     •    The chamber runs were conducted at constant light intensity in con-
          trast to the diurnal variation of natural sunlight.

     •    The initial ratio of N02 to N0x varied considerably from run to run
          and was most pronounced for the low-NOx experiments.   Such variation
          can have a marked effect on the maximum one-hour 03 level and makes
          comparison of the data difficult.

     The following adjustments were made to the smog chamber model in an effort
to generate a model more representative of the urban environment:

     •    No initial concentration of MONO was assumed.

     •    Wall degassing of propylene, used to simulate  the effects of chamber
          contamination, was deleted.

     •    Wall degradation of 03, used to describe the dark decay of 03 in the
          chamber, was deleted.

     •    The initial N02 concentration was taken to be  25 percent of the initial
          NO .
            X

     •    Diurnal one-hour average values of the photolytic rate constants were
          used.   These rate constants, as a function of  solar zenith angle for
          the Los Angeles summer solstice, were obtained from a computer routine
          developed for this purpose (8).

     •    Simulations were carried out for a nine-hour period corresponding to
          the hours between 7 a.m. and 4 p.m.   Implicit  in this treatment is the
          assumption that full atmospheric loading has occurred by 7 a.m. and
          that no influx of fresh pollutants occurs in the subsequent nine-
          hours.

     •    A "worst case" approach was adopted to describe the dispersion of
          pollutants.  A zero wind speed was assumed and the only dispersion of
          pollutants considered was that caused by the daily lifting of the
          inversion layer.  The worst mixing occurs along the West Coast where
          there is only a 100-meter difference between the mean summer morning
          and afternoon mixing heights (9).   Over a nine-hour period this
          corresponds to a three percent per hour dilution rate.

     Using the adjusted model, simulations were carried  out for initial NMHC
concentrations of 0.10 to 5.0 ppmC and NO  of 0.005 to 0.6 ppm.  The maximum
one-hour 03 levels obtained in these simulations as a function of NMHC at
constant NO  are shown in Figure 1.

     The plots in Figure 1 were used to determine the various combinations of
NMHC and N0x concentrations that correspond to the air quality standard for 03
of 0.08 ppm?  These values were used to generate the 03  isopleth shown in Figure
2.  The shaded area depicts the various combinations of  NMHC and N0y that are
predicted to meet the 03 standard.
                                      885

-------
S
                                                                         o

                                                                         a
                                                                         a

                                                                         O


                                                                      q  Z
                                 o    CO  (Nj  N  Oj  <-   .-  0  0
                              O  O  O  o O  O

                                  -EO anoH-i
                                                      O   O  O  O
Q
CC

Q
z
<
I-
(/)
                                                                                                                            q
                                                                                                                            fo
                                 U

                                 a
                                 a

                                 CJ
 (A
i—
 0)

 O)
                                                                                                                            q
                                                                                                                            c*i
                                                                                                                        \
co   r>  

                                                                                                                                         13
                                                                                                                                         en
                                                                              886

-------
          0.1      0.2     0.3     0.4     0.5     0.6
                                   NMHC, ppmC
0.7
0.8
0.9
1.0
 Figure  2.   Ozone  isopleth  corresponding  to  the  Federal  standard  of 0.08 ppm.
     Following this procedure, a series of 03 isopleths was constructed that
correspond to various maximum one-hour 03 values between 0.08 and 0.65 ppm.
These isopleths are shown in Figure 3.  A line corresponding to a NMHC-to-NOy
ratio of 5.6 is drawn through the apex point of these isopleths.  This line
divides the figure into a region that is predominantly NO -rich and one that is
predominantly HC-rich.  In the low HC/NOx region to the left of the line, a
reduction in HC levels will  result in the greatest decrease in 03 yields.  In
contrast, throughout most of the high HC/NOx region to the right of the line, a
reduction in NO  levels will lead to the greatest decrease in 03 formation.
                                   CONCLUSIONS

     The 03 isopleths constructed in this study can serve as a basis for developing
a control strategy for the Los Angeles area.  Using appropriate diurnal photo-
lytic rate constants and pollutant dispersion rates, similar isopleths can be
generated for other localities.  Thus, with this method it is possible to
develop specific control strategies for individual areas of the country.

     The accuracy of the results presented in this paper depend on both the
validity of the chamber data used to develop the model and the uncertainties
contained in the model itself.  The error associated with precision for the
chamber data appears to be only about 15 percent.  The errors resulting from
"dirty" chamber effects and other chamber artifacts, however, may be signifi-
                                     887

-------
 cant.   The  model  itself also contains  a number of uncertainties and assumptions
 that may or may not be valid.   The rate constants and products for a number of
 the  reactions  in  the mechanism are in  doubt,  and therefore the predictions are
 subject to  some degree of uncertainty  (10).   The greatest source of uncertainty,
 however, probably arises from the simplifying assumptions invoked in developing
 the  model.   It was assumed that the hydrocarbon content of the auto exhaust
 mixture consisted of only propylene and n-butane.  Such an assumption is pro-
 bably valid since good fits to the experimental data were obtained with this
 simplistic  representation.   The model,  however, was only tested against chamber
 runs with an initial  HC-to-NOx ratio of between 2 and 12.  Within this limited
 range the model does give acceptable fits  to  the experimentally generated ozone.
 However, the simulations carried out for HC/NOx of much less than 2 or much
 greater than 12 are subject to a higher degreexof uncertainty.  Thus, the
 extreme upper  left and lower right portions of Figure 3, which correspond to
 very low and very high HC-to-NOx ratios, should be viewed with caution.
                  '0.08 0.20 0.300.40 0.50 0.55 0.60  0.65
                   1.0
2.0            3.0
    NMHC, ppmC
                                                               4.0
Figure 3.   Ozone isopleths corresponding to maximum one-hour 03 concentrations.



       Questions may also be raised about how representative the adjusted
 model is in relation to the real-world situation.  For the simulations
 carried out in this study, the atmosphere was treated as if it were a
 static chamber, with full loading of pollutants shortly after sunrise and
 no further influx or fresh pollutants during the day.  Such a situation
                                      888

-------
clearly does not conform to the real  world.   The adjusted model, however,
was designed with the intent of describing a "worst case" condition for
the atmosphere.   By assuming stagnant conditions, early loading of pol-
lutants, clear skies, and no influx  of fresh sources of NO that might
tend to reduce 03 levels, the most favorable conditions for Oa formation
it is hoped, were established.   Thus, the maximum one-hour Os concentra-
tions predicted by the model are apt to be higher than those that would
be found routinely in the atmosphere given comparable 6-9 a.m. levels
of NMHC and NOx .   Such an approach is consistent with the existing air
quality standard for OB,  which is expressed in terms of an hourly
average not to be exceeded more than once a year.
                                   REFERENCES

1.   B. Dimitriades, "An Alternative to the Appendix-J Method for Calculating
     Oxidant-and-N02 Related Control Requirements," these proceedings.

2.   B. Dimitriades, Environ. Sci.  Technol., 6., 253 (1972).

3.   B. Dimitriades, "On the Function of Hydrocarbon and Nitrogen Oxides in
     Photochemical Smog Formation."  U.S. Bureau of Mines Report of Investi-
     gations RI 7433, Sept. 1970.

4.   P. A. Durbin, T. A. Hecht, and G.  Z. Whitten, "Mathematical Modeling of
     Simulated Photochemical Smog," Environmental  Protection Agency Report
     No. EPA-650/4-75-026, June 1975.

5.   B. Dimitriades. J. Air Pollut. Contr. Assoc., ]7_, 460 (1967).

6.   J. J. Bufalini, S. L. Kopczynski,  and M. C. Dodge, Environ. Letters, 3^,
     101 (1972).

7.   B. Dimitriades, "Oxidant Control Strategies.   Part I.  An Urban Oxidant
     Control Strategy Derived from Existing Smog Chamber Data," submitted for
     publication (1976).

8.   K. L. Demerjian and K. L. Schere,  "A Computer Program for Generating the
     Diurnal Variation of Photolytic Rate Constants for Atmospheric Pollutants,"
     in Proceedings of the International Conference on Environmental Sensing
     and Assessment, Vol. II. Las Vegas, Nevada, Sept. 14-19, 1975.

9.   G. C. Holzworth, "Mixing Heights,  Wind Speeds, and Potential for Urban
     Air Pollution Throughout the Contiguous United States," Environmental
     Protection Agency Report No. AP-101, Jan.  1972.

10.  M. C. Dodge and T. A. Hecht, "Rate Constant Measurements Needed to Improve
     a Generalized Kinetics Mechanism for Photochemical Smog," Int. J. Chem.
     Kinet. , Symposium No. 1, 155 (1975).


                                     889

-------
                                                                            18-4
      OUTDOOR SMOG  CHAMBER STUDIES:   EFFECT OF DIURNAL LIGHT,  DILUTIION,
         AND CONTINUOUS  EMISSION ON  OXIDANT PRECURSOR RELATIONSHIPS

          H. E. Jeffries,  R.  Kamens, D.  L.  Fox, and B. Dimitriades*
ABSTRACT

     The. in&lu.e.nce.  oft  diurnal tight -inte.nbity,  dilution, continuous -injection,
and combinations  o& these, on the. Vimitraides' Smog ChambeA Method  ^or Cnu.da.nt
Control calculations is  e.xamine.d using e.xpeAime.ntal data ^rom the.  UniveASity
o<) North Carolina outdoor chambeA.   Factor* influencing the. re.lative.  spe.e.d ofa
the. photoc.ke.nu.calL system [heteAoge.ne.ouif> pn.oc.ui>, pkotoac.ce.ptou  other than
nitrogen dioxide., and  nitrogen dioxide, to oxides ofi nitrogen xatio) can change.
ozone, by 20 peJtcent andeA dhiAnai Light condition*.  Ex.peAime.ntat  data AuggeAt
that ditution o{,  a  sie.acting  -i,y^te.m (aWL have. le.AA than proportional. e.^e.ct on
ozone, faorwation,  EzpeAimentA comparing Static condition* uiith continuous
-inje-ct-ion o£ the.  borne.  ma64 into the. borne, volume. oveA 12-houA peAiodi,  Ahowe.d
that the. ozone, pro fale.*  uieAe. bimJtaA.   Maximum ozone, -in dynamic  e.x.peAijne.ntb
wo* 90 peAcent ofa ozone.  i.n static e.x.peAime.ntA when the, total mai>A  o&  oxhides o&
nttroge.n and hydrocarbon uxu, the. borne..  Stmtlar re.battb occurred -in e.xpeAur\entb
with AJjnuJLtane.oub inje.cti.on  and di&ution.


                                 INTRODUCTION

     A method for calculating control  requirements based on smog chamber data
has been proposed by Dimitriades (1).   The original data base consisted of
constant light-intensity, static experiments in the Bureau of Mines smog cham-
ber.  Through the use  of modeling,  these results have been extended by Dodge
(2) to values not originally included and for a single diurnal-light-intensity
pattern with certain chamber characteristics removed.  The purpose of this
paper is to use data from the University of North Carolina (UNC) outdoor smog
chamber to address  some  factors related to the use of static smog  chamber
results in estimating  control requirements.  Three general factors will  be
discussed:  (a) diurnal  and  seasonal light intensity variations, (b)  dilution
of reactant system, and  (c)  continuous injection of reactants.

CHAMBER AND INSTRUMENTS

     The UNC dual outdoor chamber (3)  is constructed of 5 mil (127 urn) FEP
Teflon film supported  by an  exterior wooden A-frame.  It is divided into two
compartments, each  148 m3.   The interior surface is 98% Teflon and 2% aluminum

*H. E. Jeffries,  R. Kamens,  and D.  L.  Fox, School of Public Health, University
 of North Carolina, Chapel Hill, North Carolina; and B. Dimitriades,  Environ-
 mental Protection  Agency, Research Triangle Park, North Carolina.


                                      891

-------
and glass. The chamber operates with natural sunlight, temperature, and humid-
ity, and the air source is natural rural background air of very low reactivity.
In normal chamber operation, the rural background air was not pretreated; in
some experiments, however, it was filtered by activated charcoal.   Ozone (03)
is measured by chemiluminescent reaction with ethylene using 03-nitric oxide
(NO) gas-phase-titration for calibration.   NO and oxides of nitrogen (NOX) are
determined by chemiluminescent reaction with 03.   Nitrogen dioxide (N02) is
determined by difference.


LIMITATIONS

     The NOX meter is subject to interference by other nitrogen compounds.
Comparison tests of the NOx meters used suggested that the primary interfering
compound in the sampling arrangement was peroxyacetylnitrate (PAN), and that
there was essentially no response to nitric acid (HN03) even though measure-
ments by the Battelle subtractive method showed that HN03 was present at the
expected concentration (3).  Interference in urban hydrocarbon (HC) mix experi-
ments by PAN is 5-10% of the initial NOX at the end of the run since only low
concentrations of PAN are formed.  In addition, one of the three NOX meters
exhibited a 25 ppb positive interference for NO after N02 maximum,  It has
subsequently been shown that this was a problem in the detector housing.  Data
obtained with this meter are uncorrected and thus do not obey photostationary
state (PPS) for NO, N02, 03 after N02 maximum; data from other meters appear
to obey PPS to within the limits of the meter.

     Like any smog chamber, there are potential wall effects which may influ-
ence the reaction system.   Modeling studies, however, suggest that the UNC
chamber has rather low rate constants for the heterogeneous reactions that
form nitrous acid (NO + N02 + 2H20 -> 2HN02) and nitric acid (N205  + H20 +
2HN03).   These reactions are the two most important surface-mediated processes
in most mechanisms.  The nitrous acid (HN02) can cause a rapid irritation of
the reaction system and the formation of HN03 increases the rate of removal of
N02 after 03 formation occurs.  On the other hand, some results could be
explained by assuming that the walls may contribute small amounts  of NO
during hot midday portions of runs.  The mechanism by which nitrate on the
walls may be recycled to N02 is not known.


                                   RESULTS

DIURNAL LIGHT INTENSITY

     The primary effect of diurnal light intensity conditions is to set a
limit on the duration of a photochemical system and therefore, in  those
systems that do not completely consume N02, a limit on the amount  of 03 which
can be produced. At the UNC chamber location, the length of a solar day varies
from 9 hours 34 minutes in December to 14 hours 25 minutes in June.  The
maximum total solar radiation intentisy for totally clear days varies from 0.6
cal-cm~2-min"1 (approximately 0.2 to 0.55"1 for $ka for N02 photolysis).  Many
indoor smog chamber runs are of much shorter duration at less than optimum
light-intensity conditions and thus the 03 isopleths produced are  often a
function of the time of irradiation.

                                     892

-------
     In diurnal light intensity experiments, each set of initial  conditions of
HC and NO  has its own unique photochemical ly weighted light- hi story.  Factors
that affect the relative speed of a HC-NOX system, such as a large source of
HN02 initially present, the presence of other photoacceptors (aldehydes), or
high N02-to-NO  ratios, can cause an increased yield in 03 because the system
can reach 03- forming potential during maximum light intensity.   In the UNC
outdoor chamber, side-by-side comparison of static runs, with all  initial
conditions the same except for the addition of acetaldehyde, (at 10% of the
non-methane hydrocarbons (NMHC) can yield a 20% increase in 03.   Increasing
the N02-to-NO  ratio from 0.2 to 0.4 at 0.36 NO  and 3.4 ppmC can  yield a 20%
increase in 03.  The systems producing higher 03 under the above conditions
are faster:  the N0-N02 crossover and N02 maximum occur up to 1.5  h earlier in
the solar day, and 03 formation starts earlier.

     In extrapolating chamber results to the open atmosphere, (for instance by
a modeling technique), it is important to choose not only the length of a
solar day and its maximum intensity appropriate for the time period of desired
control, but also to choose a correct relative speed.   That is,  the concentra-
tion of photoacceptors other than N02 and an N02-to-NOx ratio that will lead
to a proper time scale of events must be specified.   Little information is
available regarding the relative speed of the open atmosphere.   The hetero-
geneous contribution to HN02 formation may be small.  Aldehydes  are usually
present which decrease the role of HN02.   It is difficult to estimate how
much NO is oxidized thermally at high concentrations before it can be diluted
to ambient concentrations.

     Based on UNC experiments, it is estimated that for a given  diurnal light
intensity profile, these factors can lead to 20% variation in the  location of
03 isopleths.  Changing light intensity profiles from one season  to another
can lead to large changes (> 50%) in locations of 03 isopleths.


DILUTION

     In a photochemical system, over a short time period, NO, N02, and 03 are
nearly in photoequilibrium, since other competing reactions are  usually much
slower than the classic photostationary state reactions.   Over a long time
period, however, free radicals compete quite successfully for NO,  thereby
increasing the 03.  That is,

                $k3  [N02]
where [N02]/[NO] is a function of radical flux (organic reactions).   As a
first approximation, as long as the absolute photolysis rate of N02  is large
enough, dilution should not affect the 03, because dilution would not change
the N02-to-NO ratio.  The formation of 03 is essentially a first-order pro-
cess, while destruction of 03 is a second order process; therefore,  dilution
could potentially have a greater affect on the 03-destructive processes lead-
ing to higher 03 yields per precursor molecule.   The oxidation of NO to N02 by
free radicals can be approximated by a first order process to the extent that


                                     893

-------
the free radicals are in steady-state.   Therefore,  the effect  of dilution
depends upon how dilution effects the free-radical-NO-oxidizing  species which
are at very low concentration in near steady-state.
     Figure 1 shows the results of dilution on  a relatively  slow  NOv-synthetic
urban mix dual run in the UNC chamber.   The dilution  rate  was  9.5%-^,  a  rate
slightly less than the NO-to-N02 oxidation rate,  as  evidenced  by  the  slow  rise
in N02 on the diluted side.  Two processes, both  acting to lower  the  concen-
tration, were operating on NO:   dilution and conversion to N02.   Dilution  also
acted to lower N02, but conversion of NO to N02 maintained or  actually  increased
the N02 concentration; therefore, a more optimum  N02-to-NO ratio  was  achieved
earlier in the solar day.  Ozone was therefore formed at a faster rate  in  the
diluted side than in the undiluted side which did not achieve  optimum 03-
producing conditions until later in the solar day when the light  intensity was
decreasing. Apparently, this dilution rate did not greatly influence  the free-
radical concentrations responsible for NO-to-N02  conversion.
    Q.
    EL
     CM
    O
       0.5
       0.4
        0.3
        0.2
       0.1
        0.0
                                   October  7,  1974  HC Mix-NO
                    NO
                                  DILUTION STflRTS
                                                           N02
                                                                ..
                                                                °8
                           8,    9   10    11
                                 hours, EDT
12   13   14   15   16   17
 Figure 1.  Dual run comparing dilution and no dilution.   Initial  conditions:
 NO* (—) 0.504, (--) 0.512 ppm; N02 (—)  0.106, (--)  0.111  ppm;  NMHC (—)
 2.39, (--) 2.43 ppmC urban hydrocarbon mix in the UNC outdoor smog chamber.
 Dilution rate 9.5% per hour.


     This experiment has  been duplicated and  experiments  have  been  performed
at other conditions.   In  those experiments, dilution rates close to  the NO-to-
N02 conversion rate made  more 03 by  dilution,,  and  higher  rates made  less  03.
(Unfortunately, the NOX meter which  exhibited NO interference  was  used in
                                     894

-------
these experiments so the fact that the system may have obeyed PSS at the  end
is not evident.)

     Another experiment in the UNC Aerosol Chamber (Figure 2), illustrates  how
difficult it may be to dilute 03.   The normal run was  over by 1200 EOT, and
the chamber was exhausted.  Although "NO " and NO decreased at a very rapid
rate (SO^h"1) 03 did not decrease until  more than one  hour of dilution had
occurred, and then the decrease was not proportional  to the dilution rate.
                                              i—i—i—|—i—|—i—|—i

                                                  flUGUST  9, 1975
                               + +  DILUTION STflRTS
7    8    9   10   11   12   13
                                                15   16   17   18   19
                                 HOURS.  EOT
  Figure 2.  Effect of dilution in a high concentration run.   Initial  condi-
  tions:  NOx, 1.95 ppm; N02, 0.05 ppm; propylene,  23.3 ppmC  in  UNC outdoor
  aerosol chamber.

     These and other experiments would suggest that dilution  may have  substan-
tially less impact on 03 formed in photochemical  systems than one might  pre-
dict. Dilution patterns of the type discussed here  are not very  typical  of
urban atmospheres.   The effects of dilution and continuous emission will  be
discussed subsequently.


CONTINUOUS EMISSION

     In a typical  urban environment,  reactant material  is  usually emitted
throughout the day.  The volume into which  the reactants are  mixed determines
the concentration  of precursors available for 03  formation.   Static smog
chamber experiments, on the other hand, start with  all  the mass  present ini-
tially.
                                     895

-------
     The UNC dual chamber has been used to compare  static experiments with
those in which the same mass was  added to the  same  volume, but  ramp-injected
slowly over a 12 hour time period (0600-1800  EOT).   In  these experiments,
propylene, a 60:40 mixture of n-butane:propylene, and the UNC mix  (acetylene,
5 paraffins, 6 olefins) were each used as a HC.  All static experiments had a
N02-to-NOx initial ratio of 0.2.   On  the ramp  side,  NO  and HC were  injected
each minute. Between 5 and 10% of the NO was  oxidized to N02 at injection.

     Figures 3-5 illustrate some  typical results.   (Solid lines are static,
dashed lines are ramp).  Table 1  summarizes the  conditions and  results.  In
general, the 03 profiles in the static-and ramp-injected runs are  very similar.
The ramped side usually forms 03  earlier and  is  higher  in 03 until  afternoon,
when the static side usually overtakes and slightly  passes the  ramped 03.  The
03 at 1900 EOT in the ramped run  is usually about 90% of the static 03 unless
cloud conditions in the late afternoon influence the final 03 (these have a
slightly lower percentage).   In two runs, the  ramped side formed more 03.
This was apparently because of a  low  NO injection rate, resulting  in less
total NO and a shift in the HC-to-NO   ratio on the  ramped side, compared to
the matching static run.
                                                                       0.50

                                                                       0.45
flUGUST 18,  1975 _
                              10   11  12  13   H  15  16   17  18
                                  HOURS, EOT
                   19
Figure 3.  Static (—)  and Ramp (--)  injection  dual  run  in  UNC  outdoor  smog
chamber.   Initial conditions at 0545  in static  side  (—):   NO,  0.29  ppm;  NO,.
0.07 ppm; n-C4H10, 0.60 ppmC; C3H5,  0.40 ppmC.   Ramp side  (--)  would achieve:
NO, 0.36 ppm; n-C^o,  0.60 ppmC; C3H5, 0.40 ppmC at 1800  hours if no reac-
tion occurred.
                                    896

-------
                                                RUGUST  18.  1975
                                           propyfene
                                  I....L.I..L.L,l.i.1,1,1.1.1.1 J.i
    0.00
                             10  11  12   13   H   15  16  17  18   19
                                 HOURS.  EOT
   0.00
Figure 4.  Static (—) and Ramp injection (--) dual  run in  UNC outdoor smog
chamber.   Initial conditions at 0545  in static side  (—):   NO, 0.29 ppm; N02,
0.07 ppm;  n-C^HiQ, 0.60 ppmC; C3H63 0.40 ppmC.  Ramp side  (--) would achieve:
NO, 0.36  ppm; n-Ci+Hio, 0.60 ppmC;  C3H5, 0.40 ppmC at 1800  hours if reaction
occurred.
     .600
    .500  -
  E
  Q.
  CL
    .300  -
   - .200
     100
    .000
                                                 1  I  '  I  '  I  '   I
                                                flUGUST 13.  1976  -
   .600
— .500
- .300
— .200
- .100
                      8   9  10   11   12  13  11-  15  16   17   18   19
   .000
                                    HOURS
 Figure 5.   Static  (—) and Ramp (--)  injection dual run in  UNC  outdoor smog
 chamber.   Initial  conditions at 0540  in static side:  NO, 0.151  ppm; N02,
 0.072 ppm;  1.93 ppmC urban hydrocarbon mix.  Ramp side (--) would achieve-
 NO, 0.22  ppm;  1.93 ppmC urban hydrocarbon mix at 1800 hours if  no reaction
 occurred.
                                    897

-------
         TABLE  1.   OZONE  FORMATION IN STATIC AND RAMP  INJECTED OUTDOOR
                                 SMOG CHAMBER EXPERIMENTS


                  Precursor
                  Conditions3             HC/NOtf       0, max, ppm      ramp/static
                     ~~~  "" ~"       L           X        O
       Date        NO     H£     crop       ppmC/ppm      static   ramp
      8/16/75      0.37  0.98     P         2.65        0.68   0.48          71

      8/17/75      0.36  1.00    B/P         2.78        0.28   0.25          89

      8/18/75      0.36  1.00    B/P         2.78        0.18   0.16          89

      8/4/76       0.35  0.97     M         2.70        0.05   0.05         TOO

      8/5/76       0.31  1.93     M         6.23        0.24   0.32C        133

      8/6/76       0.24  0.97     M         4.04        0.17   0.29C        171

      8/12/76      0.24  1.93     M         8.04        0.72   0.65          90

      8/13/76      0.24  1.93     M         8.04        0.49   0.41          84

      8/24/75   '   0.38  4.00     M        10.53        0.57   0.46          81

      8/30/76      0.36  4.56     M        12.00        0.74   0.65          88
      3 initial values for static experiments established at 0540 EOT and values that would
        have existed at 1800 EOT in ramp injected experiments if no reaction occurred.

        P, propylene; B/P, 60% n-butane-40% propylene by carbon; M, synthetic urban
        mixture of acetylene, paraffins, and olefins.

      c NO injection rate for ramped side lower than required to give final NOX value
     The NO  and HC  profiles in these runs are quite different  from their
static  counterparts,  not converging  until near the  end of the run.   This
convergence implies  approximately  the same total  mass reacted over the same
time period.

     In another experiment, a highly simplified  urban condition  was simulated.
It was  assumed that  an  air parcel  moved along a  trajectory over  a uniform
emissions surface from  0600 to 1200  EOT.   At'1200 EOT the parcel  left the
emissions surface.   During its movement over the  surface, from  0900 to 1200
EOT the mixing height doubled in a linear mariner.   Therefore, in  the absence
of reaction,  the concentration of  precursors in  the air parcel would increase
linearly from 0600 to 0900, and then the  concentration (mass/volume) would
remain  constant while an additional  three hours  of  emissions was  added to a
volume  that was doubling in three  hours.

     This run was compared to a static run which  started with the same mass as
would be present, in  the absence of  reaction, at  the end of the  dynamic run
(and also at 0900 EOT).   The run was accomplished by ramp injecting NO and HC
into a  static chamber from 0600 to 0900 EOT, and  then diluting  the chamber by
50% while continuing  to inject NO  and HC  at a rate  equal to the  mass lost by
dilution assuming no  reaction.  The  results are  shown in Figure  6-9.  Except
for 03  at the very end  of the run  and the ramp MO and HC injection, there is


                                        898

-------
                                                                       1.00

                                                                       0.90
flUGUST 30.  1975 _
                      8
9  10   11   12  13
       HOURS, EOT
  15  16  17  18   19
 Figure 6.  Static (—) and Ramp injection (0600-0900)  with SQ%  dilution
 (0900-1200) and continued ramp injection (0900-1200)  at rate equal  to mass
 removal by dilution if no reaction (--).  Initial  conditions in static side:
 NO, 0.31 ppm; N02, 0.05 ppm; urban hydrocarbon mix, 4.56 ppmC.   An  initial
 condition of 0.05 ppm NO was established in ramp side  (--).   NO and six
 injected in (--) so as to achieve same condition at 0900.
almost no difference in the profiles.

     Conditions were well  matched at 0900,  and  n-pentane,  the  least  reactive
paraffin in the system, is almost constant  from 0900  to  1200 EOT.  These
dynamic runs suggest that  it is  the total mass  amitted into the system divided
by the total mixing volume of the system that determined the 03 produced.
These experiments all  had  a constant HC-to-NO,  ratio  for the injected material.
Experiments are currently  underway to test  the  effects of  varying  the HC-to-
NO  ratio around an average value to determine  the  degree  to which the re-
sponse of the system is dependent upon how  the  mass is injected.

     Assuming that, within limits, the maximum  03 produced is  dependent  upon
the total mass of the precursor emitted divided by the total mixing  volume  at
the time the maximum occurs, and that this  03 is 90%  of  that obtained  from  the
equivalent concentration  (mass/volume) irradiated in  a static  chamber with  the
same light-intensity and relative speed, then static  03  isopleths  can  be  used
to estimate control requirements.

     The intersection of an 03 isopleth, 1.11 times the  highest  observed  03,
and a HC-to-NO  ratio line, obtained by dividing the  sum of the  total  HC  emis-
              X

                                     899

-------
   1.10
   1.00
   0.90
 0 0.80
 e
 Q.
 CX

0.70
0.60

(MO
0.30
0.20
0.10
0.00
Figure 7
(—) and
                                      flUGUST 30.  1975
                                      f ene/acety f ene
                                          ethylene
         i  i  i
i  I  i I  i  I  i I  i  I  i  I i  I  i  I  i I  i
1.10
1.00
0.90
0.80
0.70
0.60
0.50
O.tO
0.30
0.20
0.10
0.00
     5   6   7  8   9  10  11  12   13  H  15  16  17   18  19
                         HOURS,  EOT
      Selected hydrocarbon compounds  in urban hydrocarbon mix for Static
     ramp injection (--) dual run  in UNC outdoor smog chamber.
 i
 o
 e
 o
 a
 cr
 o
 CO
2.00
1.80
1.60
l.*0
1.20
1.00
0.80
0.60
O.*0
0.20
0.00
                                          r  I  ' I  T  I  '  I
                                          flUGUST 30, 1975
                  1,1,1,1,1,1,1,1,1,
     5  6   7   8   9  10  11  12  13  1*  15  16  17  18
                         HOURS, EOT
         Figure 8.  Solar radiation for August 30, 1975.
                                                         19
2.00
1.80
1.60
1.^0
1.20
1.00
0.80
0.60
0.^0
0.20
0.00
                              900

-------
O.H

0.12

e 0.10
Q.
O.
w 0.08
o
QD
0= 0.06
CJ
0
tr
° 0.04


0.02

n nn

1 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 '
"—-^ flUGUST 30, 1975 ~
- \ -
y

— /^"V, —
/ ' v^f \
'*'' V^?
// \v
\nV\.
/ .' \ \\\ butene-1
/ \ \ \ \» —
, \ v \ v
\v\ \^
\ \ \ ^"v
	 (' \ v \ \ __

1 V \ - «"-
/ cfs-butene-2\ ^ \ \
- / \ v VN
\ \ X^f 	
. 1' i 1 i 1 , 1 i 1 i 1 , 1 i 1 , 1 , 1 i 1 i 1 , 1 ,
O.H

0.12

0.10


0.08

0.06



0.04


0.02

n nn
         5    6   7   8    9  10  11   12   13  H   15   16  17   18   19
                                  HOURS,  EOT

Figure 9.   Selected hydrocarbon  compounds  in  urban  hydrocarbon  mix  for static
(—)  and ramp injection  (--) dual  run in UNC  outdoor smog  chamber.

sions by the sum of the total NO  emissions over the trajectory of  the air
parcel for the day, would define the static initial HC  concentration and
initial NO  concentration which  would give rise to  the  observed 03.   Note that
these woula also be the concentrations that would be observed  in the air
parced at the time of 03 maximum if no reaction had occurred.   These precursor
concentration and emissions are  related by a mixing volume,  which would be a
critical mixing volume for this  trajectory because'it gave rise to  the highest
03.  Emissions can be reduced such that,  if this  critical  mixing volume
occurred again, ambient concentrations of  precursors would not become large
enough to cause the 03 formed to exceed a  given value.   This is essentially
the technique proposed by Dimitriades, with the exception  that the  HC-to-NO
ratio line is obtained by considering the  total daily emissions.            x
CONCLUSIONS

     Based on experiments performed in the UNC outdoor smog chamber,  static
03-HC-NO  relationships have a good potential  for estimating 03 yields  in the
"real atmosphere" if the factors discussed are taken into account.   Specifi-
cally, a method must consider:

      t   duration and maximum intensity values for diurnal light intensity and
         for the season in which control is desired,

      •   the appropriate relative speed for the photochemical system, which

                                     901

-------
         depends upon the initial  N02-to-NO  ratio,  the rates  of heterogeneous
         processes, and the presence of photoacceptors  other than N02.

     We have hypothesized that,  within the limits of typical atmospheric
conditions, the total mass emitted into the system divided by  the total  mixing
volume at the time of 03 maximum determines the 03 produced.  If experimental
work verifies this hypothesis, then the initial conditions of  the statically-
derived isopleths can be directly  related to the concentration of precursors
that would have been in the ambient air parcel  at the time of  03 maximum if no
reaction had occurred.
                                  REFERENCES

 1.   B.  Dimitrades,  "An  Alternative  to  the Appendix-J  Method  for  Calculating
     Oxidant-and-N02 Related Control  Requirements",  these  proceedings.

 2.   M.  Dodge,  "Combined Use of Modeling  Techniques  and  Smog  Chamber Data  to
     Derive Ozone-Precursor Relationships", these  proceedings.

 3.   H.  E.  Jeffries, D.  L.  Fox, R. Kamens,  "Outdoor  Smog Chamber  Studies:
     Effect of  Hydrocarbon  Reduction  on Nitrogen Dioxide",  EPA-650/3-75-011,
     June,  1975.
                                     902

-------
                                                                           18-5
           THE  USE  OF  TRAJECTORY ANALYSIS FOR DETERMINING  EMPIRICAL
                   RELATIONSHIPS AMONG AMBIENT OZONE LEVELS
                  AND METEOROLOGICAL AND EMISSIONS VARIABLES

             E.  L,  Meyer,  Jr.,  W.  P.  Freas, III, J. E. Summerhays
                             and P.  L.  Youngblood*


ABSTRACT

     Low-level  uppeA aiA  tnaje-ctonieA weAe. computed ^on pancelA o&  ain asmiving
at thne.e. uAban  and  thne.e.  nunal AiteA at 1800 GMT duAing July and  August 1974.
Tke. moAt ne.cent 72  hounA  oft e.ach tnaje-ctony weAe. divide.d into time.  -i,e.gme.nt{,
vanying in le.ngth fanom thne.e. to twelve, haunt,.  Each Augment WOA chanacteAize.d
by the. e.xiAting metcono logical vaAiable^ and the. total hydnocanbon  and nitno-
ge.n oxide e.m£ionA  encountered fanom monmo.de AounceA.  Gnoiwd le.ve.1 ozone.
concentnationA  at the.  end pointf,  oft the. tnaj'e.ctonA.e.6 weAe.  n.e.late.d &tatu>ticaUt.y
to e.miA&ion  and mete.onological van^Lablu e.ncoa.nte.ne.d in e.ach tnaje.ctony Ae.g-
me.nt. ContnaAt  o& unban and nuAal &ite.-f> te.nd& to AubAtantiate. the. impontance.
ofi local pne.cum>on  e.miAAionA.   Ne.veAtheJLeAA,  obAe.nve.d leveJtb o& ozone.  Deemed
     highly  conneJtate.d with m&te.o no logical vaAiableA, panticulaAly  tempuna-
       The. pnime. utility  o& the. method ne.ponte.d heAe^in iA  4een 04 qualitative.-
ly &ubt>tantA.ating ne.lation£hip& which have, been hypotheAize.d, on  a  physical
ba&iA, between  ozone, and  pne.cuAf>on emibAionA, and between  ozone, and mete-ana-
logical vaAiableA.


                                 INTRODUCTION

     Until a few years  ago,  the photochemical oxidant problem was believed to
be one which centered  primarily in the vicinity of a few urban areas.   Control
strategies to reduce the  level  and frequency of harmful concentrations of
photochemical oxidant  were designed primarily on this basis.  During the past
several years,  however, a number of observations in non-urban areas  have
indicated that  violations of the  primary National Ambient  Air Quality  Standard
(NAAQS) for  photochemical  oxidant are far more ubiquitous  (1-5).  This new
information  is  likely  to  have  far-reaching implications on the evolution of
oxidant control  strategies,  although as yet,  specific implications  are diffi-
cult to identify.   At  least two questions need to be addressed before  the full
implications of high observed  levels of photochemical oxidant (measured as
ozone) in rural  areas  can be assessed.   First, how important a determinant of
ambient ozone are locally generated precursors, as opposed to precursors and
*0ffice of Air Quality Planning  and Standards, U.S.  Environmental Protection
 Agency, Durham, North Carolina.   P.  L.  Youngblood is also on assignment  from
 National Oceanic and Atmosphere Administration, U.S. Department of Commerce.


                                      903

-------
ozone which are advected into an urban area?  Second, what is the relative
importance of different meteorological variables at different specified time
segments along an air parcel's trajectory?


                                  PROCEDURE

     Six monitoring sites were chosen to serve as trajectory end points.   It
is believed that the selected sites represent situations in which transport of
precursors and ozone, as opposed to local precursor emissions, is likely to be
important to differing degrees.   In the ensuing analysis, local  emissions are
considered to be precursors emitted within 3-6 hours of an hourly ozone con-
centration observed at the monitoring site.  Thus, locally generated emissions
were frequently confined to the  county containing the monitoring site and
perhaps one or two upwind counties as well, depending on the wind speed.   A
single monitoring site was chosen at each of the following locations.

     o   Indianapolis (thought to represent a city with moderate non-methane
         hydrocarbon (NMHC) and  nitrogen oxides (NOX) emissions, dominated by
         mobile sources and having relatively low upwind emissions outside of
         the county).

     o   Houston (thought to represent a city with very large NMHC and NOX
         emissions, dominated by stationary sources and having relatively low
         upwind emissions outside the county).

     o   Boston (thought to represent a city with high NMHC and  NOX emissionss
         dominated by mobile sources and frequently having significant upwind
         emissions outside of the county).

     o   Poynette, Wisconsin (thought to represent a location having locally
         low emissions with only moderate upwind emissions outside of the
         county).
     o   McConnelsville, Ohio (thought to represent a location having locally
         low emissions with potentially high emissions occurring upwind out-
         side of the county).
     o   Dubois, Pennsylvania (thought to represent a location having locally
         low emissions with potentially high emissions occurring upwind out-
         side of the county).

The selected sites in the three  urban locations (Indinapolis, Houston, Boston)
are all in the predominantly downwind direction from each of the cities (i.e.,
the 71st Street & Takoma, Aldine, and Medford sites, respectively).

     Once trajectory end points  were selected,  upper air trajectories were
estimated using a model developed by Heffter and Taylor (6).  Each derived
trajectory traces the movement of a hypothetical parcel of air which arrives
at a specified monitoring site on a specified day at 1800Z (1400 EOT, 1300
CDT).  The movement depicted by the trajectories represents the average wind
velocity in a layer of air 300-2000 meters above terrain.   One trajectory was
plotted for each day between July 1, 1974 and August 31, 1974 at each of the
sites  for a total  of 372 trajectories.
                                     904

-------
     Each trajectory was next superimposed on national maps depicting total
hydrocarbon (THC) and NOX emission density (t/y-mi2) arising from anthropo-
genic activities in each county in the United States.  County emission densi-
ties were estimated using the National Emission Data System (NEDS) maintained
in the Office of Air Quality Planning and Standards.  Trajectories were then
segmented as indicated in Table 1.  Mean THC and NOX emission densities were
estimated for each segment by averaging the emission densities of the various
counties traversed by that segment of the trajectory, as picutured in Figure
1.  In segment 1, additional consideration is made of emissions carried past
the monitoring site by the surface wind.   This latter estimate was made by
plotting the location of the monitoring site and point sources of THC and NOX
on a county map and suballocating county area source emissions according to
population distribution within the county.  By assuming that the observed
surface wind velocity at 1800 GMT prevailed during the previous hour, it was
possible to estimate the mean THC and NOX emissions encountered by the surface
wind.   The mean emissions encountered by the surface winds during the last
hour were averaged along with the mean emission density encountered along one
segment of the upper air trajectory (calculated as described previously for
other segments) to obtain an estimate of mean emissions during the last 3 hours
prior to arrival of the air parcel at the monitoring site.  Thus, emissions
encountered within 3 hours of the ozone concentration observed at each trajec-
tory's end point reflect the impact of both surface and transport layer winds.

                          TABLE 1.  TRAJECTORY SEGMENTS
  Segment
  Number
Time Range
(EOT) for
Emissions
Time Intervals
    (hours)
Time of Surface
Meteorological
Observations
(EOT)
Hours Prior
to (03)
Observation
1
2
3
4
5
6
7
8
9
10
1100-1400
0500-1100
2300-0500
1700-2300
1100-1700
0500-1100
2000-0500
0800-2000
2000-0800
0800-2000
3
6
6
6
6
6
9
12
12
12
1400
0800
0200
2000
1400
0800
0200
1400
0200
1400
00
06
12
18
24
30
36
48
60
72

     As a result of the procedure described in the previous paragraph, the THC
and NOX emissions encountered in each segment were computed by determining the
                                     905

-------
I	
                                                                  CNI
                                                                   'i
                                                                    II
                                                                   Cl
                                                                           c
                                                                           o
                                                                           o
                                                                           c
                                                                           u
                                                                           CD

                                                                           O)
                                                                          LL.
                          906

-------
mean of the annual emission values of the counties transversed.   However, one
would expect seasonal and diurnal variations as well  as variations by day of
week to occur about this mean.  Subsequent review indicated that seasonal
changes may be relatively small (8).   Therefore, only diurnal  and day-of-week
emission patterns were superimposed on the estimated annual average emission
rate calculated for each segment.  The diurnal and weekly emission patterns
assumed are based on data available for St.  Louis (9, 10).

     Surface meteorological observations recorded at 68 National Weather
Service, Air Force, and Navy weather stations across the eastern two-thirds of
the United States were used to derive appropriate values for the meteorologi-
cal variables considered in each segment of each trajectory.  Observations at
the weather station nearest to each segment's mid-point were used as an esti-
mate of conditions along the segment (see Figure 1).

     In addition to the meteorological data directly available from weather
observations, solar radiation, atmospheric stability conditions, and absolute
humidity estimates were derived for each segment.  Solar intensity was derived
as a function of the time of the day, day of the year, latitude and longitude,
and cloud cover.  The Pasquill-Gifford stability class was  determined from the
derived estimate of solar radiation (or cloud cover observations at night) and
surface wind speed, using the method described in Turner's  Workbook (9, 11).
Table 2 summarizes the meteorological parameters available or derived for each
segment of every trajectory.

               TABLE 2.  METEOROLOGICAL PARAMETERS CHARACTERIZING
                        EACH SEGMENT OF EVERY TRAJECTORY
   Trajectory        Segment         Meteorological  Characteristics


       i                 j              Cloud Cover (CC)
                                       Ceiling (CL)
                                       Visibility (V)
                                       Surface Temperature (T)
                                       Relative Humidity (RH)
                                       Surface Wind  Speed (SP)
                                       Surface Wind  Direction  (D)
                                       Precipitation (PR)
                                       Transport-layer Wind Speed  (W)*
                                       Solar Radiation (SR)
                                       Atmospheric Stability (ST)
                                       Absolute Humidity (A)
  *  Average horizontal  wind speed in  a layer 300-2000 meters  above terrain.


     Correlation coefficients were calculated between the 1800Z ozone concen-
tration and each of the meteorological and emissions variables from each tra-
jectory segment.  Those correlations significant at the 5% level for selected


                                     907

-------
meteorological and emissions variables are shown in Table 3 for individual
sites, for urban vs. rural sites, and for all  sites combined.

     Using the combined data set, correlations between all  possible pairs of
meteorological and emission variables were determined.   Statistically signifi-
cant correlations are depicted in Table 4.  The combined data  were also used
for a stepwise regression analysis.   The stepwise regression procedure was
used to identify the set of meteorological and HC and NOX emissions variables
which appeared to be the most important in predicting 1800Z ozone concentra-
tions.  The results of the analysis have  been summarized in Table 5.


                            DISCUSSION OF RESULTS

     The correlation analyses indicated a substantial degree of variability
among the six sites.  This variability could be attributed to the location of
the individual sites with respect to the  distribution of sources of precursors,
the relatively small range of certain variables (particularly emission vari-
ables) at any one site, and the somewhat  smaller data base available for some
sites.  Therefore, it is probably most informative to group the data by sites
to seek out, on a qualitative basis, relationships among ambient levels of
ozone, precursor emissions, and meteorological variables.  In this vein, it is
of interest to compare urban and rural sites in Table 3.  The correlation
between local precursor emissions and ozone is significant and positive at the
urban locations.  (It should also be noted from Table 4, however, that there
is a high degree of intercorrelation between HC and NOX emissions during any
given time period).   In contrast, the rural sites exhibit no apparent rela-
tionships between local manmade precursor emissions and ozone.   Instead, early
morning upwind NOx emissions, and hydrocarbon emissions on the previous day,
appear to be positively correlated with ozone observed at the rural locations.
Relationships between ozone and meteorological variables are remarkably simi-
lar at the two types of sites.  The major difference in meteorology-ozone
relationships between urban and rural sites indicated by the correlation
analysis is with respect to local transport-layer wind speed (Woo).  The
relationship between local transport-layer wind speed and ozone appears to be
a negative one at the urban sites, whereas no apparent relationship exists at
the rural sites.  This anomaly with respect to wind speed is not surprising.
In urban areas, one would expect a higher wind speed to dilute the nearby
relatively dense emissions more rapidly and carry them beyond the monitoring
site before optimum conditions for photochemical ozone formation are attained.
As with urban sites, the effect of wind speed on ozone levels  at rural loca-
tions would be expected to vary according to the position of the monitor with
respect to emissions.  However, at rural sites the distances between the ozone
monitor and major sources of precursors is expected to be much more variable
than at urban sites.  All of the foregoing observations tend to support the
contention that ozone levels near urban areas are positively related to pre-
cursor emissions in the city and environs.

     The results of the stepwise regression analyses of the combined data set
are also consistent with what one might expect.  Among the variables consi-
dered, temperature (Too) appears to be the single most important indicator of
                                    908

-------













— 1
LJJ


_i
10

1 1 1

;c
1—


1 —
<=c


1 —

~^
o
h— 1
u_
1— 1

LTjT
*— '
I-H
00


OO
o
1— 1
1 —
^f.
Or
1 —
J
^^
LU
o

o
o


LU
0
M
o


o
o
00
1 —
C£
o
u_

oo
o
K— 1
1
1—
^c
	 1

LjJ
S?
0




•
OO


LU
	 I
QQ
e^
| —












O
ro
X
o


CM
X
o

CO

X
0
CM

X
0
0
X
o


o
o
X
o


o

1 1
3;

Si
o
3;
CO

o
n:

CM

(_j

<£>
O

:c
o
o

~r-


*$•
CM
2

O
O
a:
CM
O_
CO

0
Q

ex.
O
0
\~~
00
O
O
O
O

C£


CM
C£
00
O
O
ecu
oo


CM
1—

o
o
H-




OO
UJ
I—
1 — (
oo

LU

O

O
U3 *t UD
CM OO CM








00

•
-








oo oo r--.
r- ^J- CM




^O
^~






ro







•3-
oo
•

00 f^. r—
CM «^- 00


CT> co r*^
CM ro ro

i i i
ko r^. ^j~
CM ro <*

i i i
CM ro
i t


r — r^. o oo
oo CM «d- *d-
i t i i

CO LO LO
CM OO CVJ
1 1 1
00 CO
CM ro

ro in in

i i i


ro CM
ro un

r-* co UD 01
CM CM "^J" *£j~


CM
no


CM r^. CO "^D
LO ro <3~ UD




yi

QJ r—
-I-J O
•r- Q.
OO C  fO
03 -P CO -r-

s- o o c
^D CQ OI ^-H
CO
CM



CO
CM
'









ro i~
CM LD










ro r^ CM
CM CM ro


1JD O i —
r- 00 0->
*

















r**. cj> vo t^«
CM CM OO OO

i i i i
LO
CM

1
o
ro
i


CM CM OO
1 1 1

un
i
r- O
CVJ -3-
co ^o ^o
CM ro ^t-

i i t


O I-- •—
OO LO OO

ro CM ^d- co



r- r— O
oo ro un
> . •

<* un o i —
LD LT) 
•M ui
•r- p— 
r— -r- C QJ
(0 O O C
S- £l O >)
3 D U 0

Ch
r^-


























O
CM


















T—



CO
CVJ

1
^o


1
CO
1


CTt
CM
(

CM
1
0
CM
O
ro

l


r—
ro

en
CVJ



CM


CM
LD







I/I
QJ
•M
.f_
00


i —
*£


OJ
u
(O
s_
3
w
QJ
_C
4-1

•12
=
"o
o
H-
(1)
'o
E
03
X
CU


o
U-



4J
C
CU
E
en

CU

.
"O ~-^
C i—
a
CD

cu ja

(0
•r- CU
S- 0)

>• —
c
o s-
• •- CU
to E
•r- 13
E c
OJ
1 1
Tl C
c; QJ
on
i— QJ
O
cn s-
O 0
i — +->
o u
S- Q)
0 ••-)
OJ fO

QJ -M
fcr

>, 0
<<- 1-
C 0)
0) -4-1

•r- E

l/l •(->
Ol t/1
c cu

"O GU
ro S-
QJ =1
(^
C S-
E CU
Z3 Q-
i— E
O CU
O 4->



CU
4-)
O
^
909

-------
GO
GO
«=C
  I
UJ
>
UJ
00

CO
UJ
_i
co
GO
Z
o
GO
GO
                                                             CM  CO ro  i—  i—
                                                                     i—  CM  i—
                                                                    i—  LD  O
                                                                                            .
                                                                                   n  o  LD  •—
                                                                                D  U3  CM  CO  «±  O
                                                                                O  O  r—  r—  CM  CO
                                                                                X  X  X  X  X  X
                                                                                o  o  o  o  o  o
                                                                                           '   -
                ro  c\j  CM «d-
LU
                                     O  CO c\j
CD

O
               r—      CO  .—  O  O  Lf>
00

O
LD  LT)  i—
           §
a:
o
o
                       O  •—  co  m  co  in
                       O  «*  LT)  CM  OO  CO
                   »—  to  on  cr:
                                             910

-------
               TABLE 5.  STEPWISE REGRESSION RESULTS - ALL SITES

Number In
Model
1
2
3
4
5
6
7
8

R-Square
.2658
.3369
.3619
.3794
.3851
.3914
.3966
.4016

Vari
Too
Too
Too
Too
Too
Too
Too
Too

ables in

W24
W24
SR06
NOX
HCoo
HOoo
HCoo

Model


HC30
W24 Hc30
SR06 W24 HC30
NOXoo SR06 W24 HC30
NOXoo SR06 NOX24 W24 HC30
NOXoo SR06 NOX24 Woo W24 HC30
     9          .3966        Too    HCoo   NOXoo  SR06   NOX24  W24  HC30


 Final Coefficient Signs    (+)     (+)    (-)   (+)     (+)   (-)   (+)


 The variables  in the above model have all been deemed significant at the .10
 significance level

 Note:  The modified forward selection technique, available in the
        Statistical.Analysis System (SAS)(7), was used in deriving
        significant variables and their signs.

ozone.  This is consistent with other observations (12,  13),  and probably
reflects the fairly high intercorrelation of temperature  with  solar radiation
(SRoo) (Table 4) as well as faster reaction rates at higher temperatures.
Wind speed (transport-layer) (W24) on the previous  day  is seen as  the second
most important variable and appears to be inversely related to ozone levels.
The inverse relationship may reflect the tendency for higher concentrations  to
occur during multiday stagnation episodes and may further reflect  the import-
ance of upwind emissions at rural sites.  Indeed, the inclusion of hydrocarbon
emissions early on the  previous morning (HC30) and  the  previous day's NOX
emissions (N0x24) as  significant variables positively affecting ozone levels
is  further evidence that upwind emissions are important.   Other significant
variables resulting from the stepwise regression are morning solar radiation
(SRoo) and local HC and NOx emissions (HCoo,  and N0xoo).   It is of interest
to note that the sign of the coefficient for local  NOx  emission is negative.
This is in contrast to  the positive coefficients for hydrocarbon emissions
and for upwind NOx emissions, perhaps reflecting the fact that NOX is emitted
primarily as nitric oxide (NO), which acts as a scavenger of ozone, before
subsequently being oxidized to nitrogen dioxide (N02),  an ozone precursor.

                                     911

-------
                                 CONCLUSIONS

     t   There  is considerable variability among individual sites with regard
         to  the correlations between ambient ozone levels and the emission and
         meteorological variables.

     •   Generally, ozone levels appear to be more highly correlated with
         meteorological variables  (particularly temperature) than with emis-
         sion variables.

     •   Contrasts between rural and urban data suggest that local precursor
         emissions do play a role  in ozone formation and that the relationship
         between hydrocarbon emissions and ozone levels is a positive one.
         Coefficients obtained in  the stepwise regression analysis suggest
         that NOx emissions may initially be inversely related to ozone
         levels, but the relationship subsequently becomes a positive one.
         This change in sign is consistent with the role of NO as a scavenger
         of  ozone and the role of  N02 as an ozone precursor.

     •   The prime utility of statistical analyses similar to the ones de-
         scribed herein may be to  qualitatively support relationships which
         have been hypothesized on a physical basis.  The statistical analysis
         is  limited by the crudeness of the data base (e.g., emissions) and
         analytical tools (e.g., trajectory model) as well as the peculiar-
         ities of individual sites.  In the stepwise regression analysis, only
         about 40% of the variance in afternoon ozone levels is explained by
         the statistically significant emissions and meteorological variables.
         Nevertheless, the variables selected as statistically significant and
         their signs are consistent with current theories concerning the
         formation and transport of ozone.


                               ACKNOWLEDGMENTS

     The authors would like to acknowledge the efforts  of the  Simulation
Support Section, Source Receptor Analysis  Branch,  Monitoring and Data Analysis
Division, Office of Air Quality Planning and  Standards,  U.  S.  EPA.   Through
the efforts of this  group,  computer generated emission  density maps  were made
possible.  This  work would  not have been feasible  without such maps.
                                  REFERENCES

 1.   Research Triangle Institute,  "Investigation of High Ozone Concentration
     in the Vicinity of Garrett County, Maryland,  and Preston County, West
     Virginia" EPA-R4-73-019 (Jan.  1973).

 2.   Research Triangle Institute,  "Investigation of Ozone and Ozone Precursor
     Concentrations at Non-urban Locations in the  Eastern United States", EPA
     450/3-74-034 (May 1974).

 3.   Research Triangle Institute,  "Investigation of Rural Oxidant Levels as
                                     912

-------
     Related to Urban Hydrocarbon Control Strategies", EPA 450/3-75-036 (March
     1975).

 4.  Stasiuk, W.  N. and P. E. Coffey, "Rural and Urban Ozone Relationships in
     New York State". JAPCA 24 564-368, (1974).

 5.  State of Wisconsin, Department of Natural Resources, "Ozone Monitoring-
     Wisconsin Summer 1974", Madison, Wisconsin (1974).

 6.  Heffter, J.  L. and A. D. Taylor, "A Regional-Continental Scale Transport,
     Diffusion and Deposition Model, Part I:  Trajectory Model"; NOAA Technical
     Memorandum ERL ARL-50, Silver Spring, Maryland (June 1975).

 7.  Service J.,  A Users Guide to the Statistical  Analysis System; North
     Carolina State University Press; Raleigh, N.  C.  (1972).

 8.  Stanford Research Institute, Monthly Progress Report No. 7, "The Relation
     of Oxidant Levels to Meteorological Features", Contract No. 68-02-2084 in
     progress, (April 1976).

 9.  Mancuso, R.  L. and F. L. Ludwig, "Users Manual for the APRAC-1A Urban
     Diffusion Model  Computer Program", Stanford Research Institute, (Sept.
     1972).

10.  Environmental Science and Engineering, Inc.;  "Residential  and Commercial
     Area Source  Emission Inventory Methodology of the Regional  Air Pollution
     Study", EPA-450/3-75-078, (Sept. 1975).

11.  Turner, D.  B., "Workbook of Atmospheric Dispersion Estimates", PHS Pub-
     lication 999-AP-26, (1969).


12.  Hartwell, T.  D.  and H.  L. Hamilton, Jr., "Examination of the Relation-
     ships Between Atmospheric Oxidant and Various Pollutant and Meteorologi-
     cal  Variables",  Research Triangle Institute,  Contract No.  68-021096,
     Research Triangle Park,  N.  C.  (1975).

13.  Stanford Research Institute, "The Relation of Oxidant Levels of Meteoro-
     logical Features",  EPA Contract No. 68-02-2084,  in  progress.
                                     913

-------
                       SESSION 19
OXIDANT-PRECURSOR RELATIONSHIPS AND THEIR INTERPRETATION
  IN TERMS OF OPTIMUM STRATEGY FOR OXIDANT CONTROL - II
                             E.L.  Meyer
             Environmental  Protection Agency
                           915

-------
                                                                          19-1
             REPORT ON OXIDANTS  AND  THEIR PRECURSORS IN CANADA

                                 L. Shenfeld*
ABSTRACT

     OxA.da.ntA  (ozone.] otic  biting mea&uAed at AeveAOt ioc.cuU.onA actioAA Canada.
At pficAcnt tke method OQ Aampling iA  by the ckejnuJtunu.neAce.nt pnA.nc4.ptn and
Apeci^icaltij ozone. iA beting monitored.   High conce.ntx.at4.onA kave been moAt
pnevatent in SoutkeAn Ontario.  TkeAe have occuAAed mainiy duAing the AummeA,
with uieatheA favourable faon. pko to chemical. fLeM.cti.onA in tke. atmoApheAe.  Higk
me.a6uAejme.ntA in remote tiuAai oAeai, indicate, evidence, o^ tke. tong diAtan.ee.
tAanApotit ofi ox.idantA OA weLt OA tkat ofa tke pticcusiAonA, nitAogen ozideA and
kydAocasibonA, wkick lead to tke ^onmation o& oxidantA.  Tke oiigin OjJ the,
poMLutantb pneAcnt in SoutkeAn Ontasiio  kaA been on a Sequent ba&iA tAaced
to the induAtAialA-zed OACOA o& the NoAtkeaAteAn United StateA.

                                INTRODUCTION


     Ozone  (Os) is presently being measured in most of the large cities across
Canada by chemiluminescent instruments.  The locations of these Cities that
are outside of the Province of Ontario are shown in Figure 1.  The highest
oxidant (ozone) measurements in Canada have been obtained in Southern Ontario
where monitoring began in  1969.  This area (shown hatch-marked in Figure 1)
extends southward to a latitude equivalent to Northern California.  The sam-
pling locations recording  oxidants  (ozone) that are discussed in detail in
this report are Windsor, Sarnia, Simcoe, Hamilton, Metropolitan Toronto,
Stouffville and Ottawa.  The locations of these are shown in Figure 2.  The
sampling sites are in urban areas, except for Simcoe and Stouffville which
are in rural farming areas.

                       CANADA'S AIR QUALITY OBJECTIVES
                             AND OXIDANT LEVELS
     The Canadian air quality  objectives  are given in Table 1.

     The Maximum Desirable  Level  defines  the long-term goal for air quality
and provides a basis for  anti-degradation policy for the unpolluted parts
of Canada and the continuing development  for control technology.  Maximum
Acceptable Levels are intended to provide adequate protection against any
*0ntario Ministry of Environment,  Canada.

                                      917

-------
Figure 1.   Location map.




          918

-------
                                                                I/I
                                                                cu
                                                                s_
                                                                4->

                                                                OJ
                                                                o
                                                                (O
                                                                .Q
                                                                S-
                                                                13
                                                                CD
                                                                C
                                                                 o.
                                                                 I
                                                                CO
                                                                "O
                                                                 (O
                                                                 C
                                                                O
                                                                 at
                                                                 o
                                                                C/}
                                                                CvJ

                                                                 OJ
919

-------
c/o
o
LU
CO
O
o
CQ
  •   0.
 (O  CU Q.
            -O CU
             Q.
               0> Q.
 x -Qo~>
 to  to  E
2: 4->\
    Q. CD
    CU  P-
    O
    o
               CU  E
               >  Q.
                  Q.
            E

            E
               OJ
            X -de
            (O  ra
              •i-  CD
               CO  P
            CD
            E
            CD CU

            2J
           (0
           o
           D-
                              *  *
                O
                O •
                CO
                                       o un
                                       tn i—
                                       C3 0
             C3 O
             C3 O -X
             C3 00
                   LO LO
               oo CNJ i—    o o un
               O O O    CNJ i— O
                           0 O O
                                       C3 O 0
O O O    000
ID LO OO    O O O
i—          «a- CM i—
                  LO
               LO r—
               o o

               o o
                           o o
                           o oo *
                                             oo
                                       *  * o
                                  o
                            *  *  10
                   CO
                S-  S-  S-
               -c-c  >,
                             C\l
                CO
                S_  S-
               -£=  >,
                                          CM
               CO
               -l->
               c
               
                                             CO
                                             •r—


                                             JD
                              CO
                              CU
                              o
                             ^

                             •K
                               920

-------
adverse effects of the pollutants.   Maximum Tolerable Levels denote con-
centrations of air pollutants that  require abatement without delay so as
to avoid further deterioration that could ultimately pose a substantial
risk to public health.

     No air quality objectives have as yet been set for hydrocarbons (HC).

     Levels of oxidants (ozone) are at times exceeding the maximum desirable
and maximum acceptable levels at almost all of the sampling locations in
Canada, the frequency of exceeding  these objectives being greatest in Southern
Ontario.  At some of the locations  the Proposed Maximum Tolerable Objective
is exceeded during several hours per year (Table 2.)

         TABLE  2.   NUMBER OF  HOURS  OZONE  LEVELS  EXCEEDED  OBJECTIVES
     LOCATIONMAXIMUMMAXIMUM         "   MAXIMUM
                       DESIRABLE          ACCEPTABLE           TOLERABLE
     	100ug/m3  (O.OSppm) 160ug/m3 (O.QSppm)   300ug/m3 (O.lSppm)
                     1974      1975     1974    1975        1974      1975
Windsor
Sarnia
Simcoe
Hami 1 ton
Toronto (Central)
Toronto (Suburban)
Ottawa
363
370
1004
396
619
325
*
482
618
1084
986
1175
685
541
59
80
135
37
146
43
*
155
132
141
317
386
181
85
0
0
0
0
1
0
*
14
4
0
6
11
0
0

 * Insufficient data
     The maximum concentration of oxidants recorded in Canada was 700 ug/m3
(0.35 ppm), occurring in August 1971 in a northwest suburb of Metropolitan
Toronto.  The measurement was of total oxidants obtained with an instrument
using a colourmetric method.   All monitoring instruments have since been
changed to those utilizing the chemiluminescent principle.  The highest
ozone level measured in Canada by this means was 550 ug/m3 (0.274 ppm).
This level occurred in August of 1975 at Windsor, Ontario.

                    RELATIONSHIP OF MAXIMUM OZONE LEVELS
              WITH LEVELS OF HYDROCARBONS AND NITROGEN OXIDES


     High levels of ozone at an urban location in Toronto and a rural loca-
tion, Simcoe, were divided into two ranges for analyses, 0.080-0.099 inclu-
sive and levels equal or greater than 0.10 ppm.  Table 3 gives the average
levels of total hydrocarbons, nitric oxide NO), nitrogen dioxide (M02), and
nitrogen oxides (NOx) during the morning hours 06:00-08:00 EST that the ozone
levels occurred in the two ranges.


                                     921

-------
                 TABLE 3.  RELATIONSHIP OF DAILY MAXIMUM OZONE
                 WITH HC AND NOX LEVELS FROM 06:00 to 09:00 am
LOCATION

TORONTO

SIMCOE

RANGE OF
MAXIMUM OZONE
ppm
0.080-0.099
* 0.10
0.080-0.099
£ 0.10
AVG. OZONE
LEVEL
PPm
0.093
0.117
0.084
0.106
AVG.
HC
ppm
2.28
2.47
1.95
2.07
AVG. AVG.
NO M02
ppm ppm
0.04 0.045
0.03 0.048
0.019
0.016
AVG.
NOX
ppm
0.076
0.082


     The data indicates a positive relationship existing between the daily
maximum levels of ozone and the average concentration of total hydrocarbons
measured from 6 and 9 a.m.   There is little if any relationship evident be-
tween the maximum ozone and levels of nitrogen oxides.  The data indicates
that at rural Simcoe location lower levels of hydrocarbons were associated
with the maximum ozone than in urban Toronto.   This may indicate that the
ozone is likely produced elsewhere and transported to the rural location.

                    AVERAGE DAILY MAXIMUM OZONE LEVELS AND
                             TIME OF OCCURRENCE
     The average of the maximum daily ozone concentration and the time that
the maximum occurred are given in Table 4.


                 TABLE 4.   AVERAGE DAILY MAXIMUM OZONE LEVELS
                           AND HOUR OCCURRED E.S.T.
Month
Avg.Max.
Cone. ppm
Hour Ending
Max. Cone.
April
0.037
1400
May
0.041
1500
June
0.053
1500
July
0.064
1500
Aug.
0.064
1500
Sept.
0.044
1400
     The table indicates the maximum ozone concentrations occurring in July
and August, with the time of occurrence during the afternoon hour ending at
1400 during April and September and at 1500 Eastern Standard Time (EST) from
May and August.

             RELATIONSHIP OF OZONE LEVELS AND WEATHER CONDITIONS


     An analysis of weather data during days of high ozone levels, (maximum
greater than 0.08 ppm) indicate that skies were predominantly sunny during
85% of the days.  About 10% of the high ozone days skies were reported cloudy,

                                    922

-------
but these days followed a sunny high maximum ozone day.  During these periods
the ozone formed during the previous day was not dispersed or reactively re-
duced by the following day.  During the remaining approximate 5% of the days
recording high ozone, thunderstorms were reported in the area of the sampling
stations.  Stratospheric contributions bringing ozone levels above 0.08 ppm
is rare.  However, on May 17, 1973, concentrations of oxidants in Hamilton
reached 0.10 ppm although conditions were not favourable for photochemical
reactions.    The Polar front was located along a northeast-southwest line
through the Appalachians with cyclogem'ses occurring east of the Lower Great
Lakes.  The 500 MB map discloses a deep north-south trough over the Great
Lakes.  This synoptic situation has been described by Reiter(l) as that asso-
ciated with the folding of the tropopause bringing ozone down from the lower
stratosphere to ground.  This likely explains the high ozone occurrence on
this day.  There have been one or two such occurrences of high ozone levels
each year that have no other explanation.

     The following table provides the percentage of days that the prevailing
wind was in the direction indicated on high ozone days:

                  TABLE 5.  RELATIONSHIP OF OZONE LEVELS WITH
                                 WIND DIRECTION
DIRECTION
% of Days with
N
2.2
NE E
3.3 12.2
SE
5.9
S
26.9
SW
32.8
W NW
14.2 2.2

     The above indicates that the favourable wind directions for days of high
ozone are from the south and southwest.  Winds in this direction occur with
high pressure areas located east to southeast of Southern Ontario.  Subsidence
inversions aloft would be associated with the anticyclone, resulting in lim-
ited mixing heights and poor dispersion.  Winds from the south and southwest
transport air from industrialized areas of the United States into Ontario.
Quickert et al.(2) has traced the origin of air-producing high ozone levels
in Ottawa back to Ohio.

     The high prevalence of days with southwest winds when ozone levels were
high at Simcoe would suggest the City of Cleveland, 200 km.  away, as the
principle source of precursor contaminants (Figure 2).  There are no other in-
dustrial or traffic sources in the immediate area of the sampling station in
the direction from Simcoe.   Almost 50% of the days that Stouffville, another
rural site, recorded high ozone levels, the winds were in the direction from
Toronto.  There have been many reports(3,4,5) of high rural  levels of ozone
in the United States, relating these to emissions from downwind urban centres.
The high ozone levels in rural areas are of major importance as documented
economic losses to agricultural crops of tobacco, white beans and tomatoes
due to oxidants have been reported(6,7,8,9).

     The daily average wind speed during high ozone days was generally light,
that is, less than 5 metres per second, for 90% of the days  recorded.   The
highest daily average wind speed during which 0.08 ppm was exceeded was
8.8 mps, occurring in Windsor July 13, 1974.


                                     923

-------
     The maximum temperature recorded during days of high ozone were above
25°C during 90% of the days recorded.  About 20°C may be considered the
lowest daily maximum that high ozone levels could be traced to photochemical
reactions.  During May 17, 1973, when the high ozone at Hamilton was indi-
cated to have a stratospheric origin, the maximum temperature was only 9°C.

     The influence of lake breezes on the development of high ozone levels
during the summer in the Toronto area has been noted by Anlauf et al.(10)
and Lusis et al.(ll).  Afternoon levels of ozone as high as 0.16 ppm was
measured on Toronto Island, which is just south of Toronto on Lake Ontario,
on August 7 and 8, 1974 during a time that a high pressure area was over
Southern Ontario.  Maximum downtown Toronto levels were 0.047 and 0.076 ppm
during the two days.  The primary pollutants emitted in the urban industrial
shoreline communities are considered to have been carried out over the lake
during the night by land breezes.  Photochemical reactions took place in the
stable air over the lake during the morning and ozone laden air was brought
in over the Island by the lake breezes thai: developed during the day.  Lower
concentrations measured in the downtown area were considered to be due to
the inhibitive action of higher levels of nitric oxide in the area.  There
is little automobile traffic on Toronto Island.  During the period, the levels
of ozone in the rural site of Stouffville northeast of Toronto were 0.168
and 0.133 ppm, respectively, similar to that on Toronto Island in the oppo-
site direction from the City.

                CONTROL TECHNOLOGY AND LEGISLATION IN CANADA


     On a nationwide basis, transportation sources have been attributed to
be the major emitters (65%) of nitrogen oxides and hydrocarbons.  However,
on regional or city-wide emission surveys, large fossil fuel power plants
have been shown to be major contributors of nitrogen oxides.

     There has been Ontario legislation on the control of hydrocarbons emit-
ted from automobiles sold in the Province as early as in 1968, that is, con-
trolling emissions from 1969 models.  In recent years the Canadian Government
has assumed jurisdiction and has passed legislation controlling emissions,
as indicated in Table 6.

              TABLE  6.   CANADIAN  AUTOMOBILE  EMISSION  CONTROLS

                                  Hydrocarbons               Nitrogen  Oxides
                                    gms/mi                      gms/mi

1973-74 Vehicles                       3.0                         3.1

1975-1980  Vehicles                     2.0                         3.1

Proposed for future                    0.41                        2.0
     With  respect to the control of stationary sources, most of the emphasis

                                    924

-------
has been involved in the control of hydrocarbons emitted.  Technology to con-
trol nitrogen oxide emissions are rarely enforced.  Evaporation losses of
hydrocarbons in storage tanks are controlled by the use of floating roofs.
Other abatement techniques employed by the petroleum and other industries
incorporate the use of scrubbers and afterburners.

     The success of present technology in the control  of emissions of the
precursor contaminants to prevent levels of ozone from exceeding 0.08 ppm
is doubtful.  With the presence of natural sources of hydrocarbons and the
long distance transport phenomena, stringent controls on a continental basis
would seem necessary.

                                CONCLUSIONS
     High oxidant (ozone) levels measured in Canada have been related prin-
cipally to photochemical reactions with only 5% of these high levels found
to be related to thunderstorms and downward transport of ozone from the
stratosphere.  Ozone levels over 0.08 ppm are normally associated with sunny
skies, afternoon temperatures reaching 25°C, and southerly winds less than
5 mps.  The transport of oxidants and their precursors from the industrial-
ized States to the south have caused injuries to agricultural crops in rural
Southwestern Ontario.  For the prevention of high levels of oxidants and
their adverse effects on man, more stringent controls of emissions of both
nitrogen oxides and hydrocarbons from stationary sources are likely to be
necessary.


                                 REFERENCES

1.   Reiter, E.R. Behaviour of Jet Streams and Potential Fallout Situations
     Archives of Meteorology, Geophysique, and Biocl imatology.  Vol.A17-1968
     pp 8-16.

2.   Quickert, N.,L. Dubois and B. Wallsworth. Characterization of an
     Episode with Elevated Ozone Concentrations.   Science of the Total
     Environment. Vol.5. 1976. pp 79-93.

3.   Gloria, H.R., G. Bradburn, R.F. Reinisch, J-N-  Pitts, J.V. Behar,
     L. Zafonte.  Airborne Survey of Major Air Basins  in California. Journal
     of the Air Pollution Control Assoc. Vol.24,  No.7, 1974. pp 645-52.

4.   Stasiuk, W.N., and P.E.  Coffey. Rural and Urban  Ozone Relationships in
     New York State. Journal  of Air Pollution Control Assoc. Vol.24, No.6.
     1974. pp 564-68.

5.   Ripperton, L.A., J.B. Tommerdahl, and J.J.B. Worth. Airborne Ozone
     Measurement Study.  Paper presented at the 67th  Annual Meeting of the
     Air Pollution Control Assoc. Denver, Colorado,  1974.

6.   Macdowall, F.D.H., L.S.  Vickery, V.C. Runeckles, and Z.A.  Patrick.
     Ozone Damage to Tobacco  in Canada.   Canadian Plant Disease Survey.
     Vol.43, #4.  Dec.1963.

7.   MuKammal, E.I.  Ozone as  a Cause of Tobacco Injury. Agricultural
     Meteorology: Vol.2, 1965.

                                     925

-------
8.    Linzon, S.N., Plant Damage in Southwestern Ontario.  Proceedinqs of the
     Joint Annual  Meeting of the American and Canadian Horticultural
     Societies.  Vol.10, No.5.  1975.  pp 494-5.

9.    Pearson, R.G., D.B. Drummond,W.D.  Mcllveen and S.N.  Linzon.  PAN-Type
     Injury to Tomato Crops  in Southwestern Ontario.   Plant Disease Reporter,
     Vol.58. No.12, 1974.

10.   Anlauf, K.G., M.A. Lusis, H.A.  Wiebe,  and R.D.S.  Stevens.  High Ozone
     Concentrations Measured in the  Vicinity of Toronto,  Canada.
     Atmospheric  Environment.  Vol.9.  1975.  pp 1137-9.

11.   Lusis, M.A.,  K.G.  Anlauf, Y.S.  Chung and H.A.  Wiebe.  Aircraft Ozone
     Measurements  in the Vicinity of Toronto, Canada.  Paper presented at the
     69th APCA Annual Meeting  in Portland,  Oregon,  1976.
                                    926

-------
                                                                            19-2
             PRECURSOR CONCENTRATION AND OXIDANT  FORMATION IN SYDNEY

              G.  H.  Allen, K.  Post, B. S. Haynes,  and R.  W.  Bilger*
ABSTRACT
     A-4 i&>  poAt -in the. Sydney Qtidant Study, the.  gAoup at the. Un-ive.AAity  o^
Sydney haA  commiAi>ione.d a Mobile. AiA AnalyAiA LaboAatoAy ^OA me.aAuAe.me.n-t ofa
ozone, and itA  pAe.cuAA OAA ••   nitAoge.n oxA.du>  and  non-methane. hydAocoAbonA.   ThiA
facility haA been uAe.d to de.ve.top impAove.d  catib>ia,tion pAoce.duAe^> faoA. ai.fi  pol-
lution monitoring inAtA.iune.ntA and to caAAy  oat  an  extent-cue pAogAomme. o& faie.ld
me.aAuAe.me.ntA.   The. {,-ie.ld pAogAomme. haA ide-ntifaizd  A agitable. AiteA ^01 the. mea-
Au/ieme.nt o&  pfie.cuAAou in the. Sydney ngion,  and the. mobile, laboiatofiy haA
enabled the.  de.ve.lopme.nt ofa p>ie.£uAAotL te.ve.tf> to  be  obAe.?ive.d -in aiA paA.c.e.lA  OA
the.y aAe. tAacke.d to fie.ce.ptoA OACOA.
      PAS.C.UAAOA c.once.ntAationA aAe. toweA than expected and not in agAe.e.me.nt
with  pAe.dictionA  of, A-impte. di^pe.uion mode.lA uAing cuAAe.nt e.miAA-ion inve.ntoA-
iu>.   Ozone.-pAe.c.uAAOA fie.iatJ.onA hip A aAe.,  howeuei,  conAiAte.nt with U.S.,  Jap-
anese,  and  EuAopean data, although much moAe. woAk.  iA Ae.quiAe.d to utabtiAh  a
          coAAe.lation {,OA Sydne.y.

                                   INTRODUCTION
      Pollution  from photochemical oxidants  has been recognized as  a  serious
problem  in  Sydney since November  1971,  when an episode was responsible  for
considerable damage to petunia seedlings.   The Sydney Oxidant Study  is  a  coop-
erative  program of field measurements  and  research coordinated and funded by
the  State Pollution Control Commission  (SPCC)  of New South Wales.  The  objec-
tives  and approach of the study have  been  outlined by Iverach (1975,  76).   A
review of existing information on the  relationship between oxidants  and their
hydrocarbon (HC) and nitrogen oxides  (NOX)  precursors had shown  that there  was
considerable conflict between correlations  developed by the Los  Angeles Air
Pollution Control District  (APCD),  the  U.S. Environmental Protection  Agency,
(EPA), the  Japanese Environment Agency, and smog chamber researchers.   There
was  thus a  great need to develop  data  for  Sydney so that control  programs
would  have  a realistic and  relevant basis.   Briefly, the roles of  the various
groups in the Study are as  follows:   University of Sydney, trajectory measure-
ments; Macquarie University, meteorology;  CSIRO Division of Process  Technol-
ogy,  HC "finger-printing" and smog  chamber  research; and SPCC, fixed site mon-
*The University of  Sydney,  Sydney 2006, New South Wales,  Australia.

                                      927

-------
itoring and airborne measurements.   At Sydney University we  are also engaged
in numerical modelling of photochemical  smog  formation (Allen  and Bilger,  1976)
under a project sponsored by the Australian Research Grants  Committee.   Fund-
ing of the Sydney Oxidant Study began in September 1974 and  is being continued
until September of 1977 at least.   It is believed that the interaction  on  a
continuing basis of these diverse groups is a very important ingredient in the
success of the Study.

     At the commencement of the Study, the only monitoring data being obtained
were from three fixed-site ozone (03) monitors operated by the SPCC, there
being almost no reliable information on the precursor NOx and  non methane
hydrocarbon (NMHC) (Iverach and Bilger, 1975).  A preliminary  study of high
oxidant days (Hawke and Iverach, 1974) showed the influence  of the land and
sea breezes during periods of general synoptic stagnation and  implicated
three main source areas:  an inland industrial source around Silverwater;  the
central business district (C.B.D.)  and the industrial region just to the south
of it; and a coastal industrial source at Kurnell near the mouth of Botany Bay
(Figure 1).  There was thus a great need to establish precursor concentrations,
correlate their occurrence with known emission sources, and  follow the devel-
opment of the polluted air masses as they pass over the region under prevail-
ing meteorological conditions.   The identification of suitable fixed
monitoring sites for both precursors and oxidants would be a further result
of this work.   These then were the first objectives of mobile  monitoring pro-
ject undertaken by the Sydney University group, and they have  in large part
been achieved.  Work is now aimed at quantifying the relationship between
emissions and precursor concentrations, correlating the downwind oxidant con-
centrations with morning precursor concentrations, and validating and develop-
ing a numerical trajectory model.  It is believed that the continuing
iterative interaction between field measurement and modelling  sharpens both
experimental and analytical techniques and yields new insights into the
problem, since "unexplained" factors of 4 and 2 and 1.5 between observations
                      Figure  1.   Sydney  location map.
                                     928

-------
arid predictions must, \>c investigated.   Here we report on the preliminary data
obtained and the first round of iteration with the modelling.  Earlier reports
may be found in Allen and Iverach (1976), Allen (1976), and Allen and Bilger
(1976).  Full details of the data are summarized in bimonthly and annual re-
ports to the SPCC (Post et al., 1976).  It is hoped that the data will attract
other modellers and interpreters.

     A considerable effort has gone into validation of the calibration tech-
niques used with the ozone, nitrogen oxides and hydrocarbons instruments used
in the Sydney University Mobile Air Analysis Laboratory.  This work is de-
scribed in the next section and it gives considerable support to the ozone
and NOx field observations.  Problems associated with suitable mobile moni-
toring sites free from local sources are also discussed in that section as
are current uncertainties inherent in continuous Flame lonization Detector
(FID) analyses of ambient hydrocarbon concentration.

     Non-methane hydrocarbon and NOx concentrations measured during the summer
oxidant season are considerably lower than those reported for winter condi-
tions by Allen (1975).   These observations, however, are in general agreement
with those projected from emission inventories by Iverach et al.  (1976) but
the relationship between ambient levels and upwind sources of different types
is not straightforward.  These results and the general features of the tra-
jectory measurements are reported in the Results section together with a dis-
cussion of attempts to quantitatively relate measured precursor concentrations
to emissions.  In the last section oxidant-precursor relationships are
examined.

                INSTRUMENTATION, CALIBRATION, AND PROCEDURES


     The Sydney University Mobile Air Analysis Laboratory used in this study
has a heat-insulated body shell mounted on a small truck chassis (Bedford
Elf 2000cc).  It has a 5kVA 240V 50Hz electrical power generator to keep the
instruments operating while mobile and a small air cooling unit to remove heat
dissipated by the instruments.  Sampling is via an all teflon sampling system
with an inlet 4 m above the ground.   Residence times in the sampling lines are
less than two seconds.   A full description of the laboratory may be found in
Allen (1976).

     An MSA model 11-2 dual FID is used to obtain total hydrocarbon and meth-
ane data.   This instrument is calibrated using NATA certified span gas mix-
tures supplied by the Commonwealth Industrial Gases Limited  (CIG).  Zero span
gas is generated by the instrument itself by passing the sample stream through
a catalyst bed containing Hopcalite at 375°C.  A gas mixture of 14.5 ppm meth-
ane was used to span both the total  hydrocarbon and methane burners and a mix-
ture containing 6.0 ppm methane and 2.5 ppm propane in air was used to deter-
mine the sensitivity of the total hydrocarbons burner to hydrocarbons other
than methane.  High purity hydrogen is used as the fuel and  it has been found
that the relative molar response (RMR) (Schoefield, 1974; Blades, 1973;
David, 1974) obtained for propane could be made to vary between 1.05 and 3.24.
Relative molar response factors are a function of burner configuration and
fuel.  Sample, and combustion air flow rates and the performance of this in-

                                      929

-------
strument with respect to its total  hydrocarbons function is still  under exam-
ination.  Throughout this period, however,  the instrument functioned under
conditions which gave a propane RMR of approximately 2.2.  All  hydrocarbon
data were obtained with the instrument operating in this manner.   No correc-
tion has been applied for non-theoretical  RMR behaviour.

     Ambient ozone concentrations are determined using a Norian model NCM-100
chemiluminescent analyser.   This instrument is calibrated against ozone mix-
tures generated by a Penray ultraviolet (UV) lamp incorporated in a flow dilu-
tion system.  The absolute ozone generation rate of this lamp is determined at
each calibration by either gas phase titration (GPT) or the neutral buffered
potassium iodide (NBKI) procedure.   Both analytical methods were consistent to
within ten percent (1.56 (+0.06;90%) x 10'7 moles 03 min'1 by NBKI, and
1.46 (+0.09;90%) x 10-7 moles 03 min"1 by GPT).  Post (1976) gives details of
these two methods.  An untraviolet absorption system is being developed to
give a further method of absolute calibration.

     Nitric oxide (NO) and total nitrogen oxides are determined using a Monitor
Labs model 8440 chemiluminescent analyser.   This instrument is calibrated using
a flow dilution system together with NATA certified gas mixtures containing
41 ppm and 185 ppm nitric oxide in nitrogen.  Calibrations were performed
within the range 4 to 400 pphm and instrument grade air  (CIG) was used as the
diluent.  Many inter!aboratory comparisons have been performed between the
University of Sydney and SPCC.  Both groups, using various combinations of
dilution systems and span gas mixture cylinders, have generated diluted mix-
tures which gave interlaboratory agreement to within ten percent.  Instrument
zero settings have been determined by switching off the ozone generator with-
in the instrument, thereby simulating a zero NOx condition.  This "zero ozone"
method is thought to be unsatisfactory and a considerable effort was made to
obtain a true zero NOx sample to get an independent measure of this zero set-
ting.  Ascarite (reg. T.M.) was found to be very effective in scrubbing NO
and nitrogen dioxide (NOa) from spiked air streams and was used to manufacture
"zero NOx" from ambient air.  This material gives a reading of +0.3 pphm on
the NO channel of an instrument zeroed by the "zero ozone" method.  The zero
N02 signal has been found to vary between - 0.5 and +0.2 pphm, but since this
signal is the result of the subtraction of the NO signal from the total NOx
signal, and is consequently dependent upon the sensitivity of both the MO and
NOx channels, its value from day to day is variable and  its absolute value
uncertain.  Tests using activated charcoal to scrub ozonized instrument grade
air have given similar results.

     In Figure 2 the quantity [Os][NO]/[N02] is shown as a function of the
time of day for March 19, 1976.  The smooth curve presented on this plot is
the clear sky theoretical [03][NO]/[NO;?] -vs- time relationship derived from
the solar zenith angle for Sydney in mid-March (calculated from the Nautical
Almanac (1976)), the rate constant data of Jones and Bayes (1972) for the
reaction N02 + hv ->• Os + NO, and the rate constant data  of Schofield (1976)
for the reaction NO + 03 > NO?. + 02 corrected for the temperature prevailing
at the specified time of day.  The relationships used in these calculations
are otherwise the same as used by Calvert (1976).
                                     930

-------
                      Cloud .
                 Figure  2.   Experiment  ratio [03][NO]/[N02]
                            for  March  19,  1976.
     The [03][NO]/[NO?] values calculated from the experimental data for this
and other days fall around the theoretical relationship and follow the pre-
dicted trend throughout the day.   The scatter in the experimental data is sim-
ilar to that obtained by Eschenroeder and Martinez (1970), Stedman and
Jackson (1975), Calvert (1976) and in this instance it can be attributed, in
part, to the lack of instantaneous 0-;, NO, and TI02 measurements.  The data
points in these figures were calculated from 3 minute averages which have been
corrected for the zero offsets.  The [Ch][NO]/[fl02] ratios calculated with
uncorrected N0>; data show no consistency and this supports the view that the
zero settings obtained by the "zero Ch" method are incorrect.  There is thus
considerable support for the accuracy of the corrected NO/ and Ch data and
their interpretation can be attempted with confidence.

                                      931

-------
     Wind speed and direction are measured using small  buoyant balloons that
are ballasted to give a rate of rise of 1  rn/s.   The balloon is followed from
its release point and the time and horizontal  distance  travelled (usually
about 50 m) is measured to give the wind velocity.   The heading after 100 s of
flight gives the mean wind direction.  The height of low level inversions can
be estimated from the time of rise to points where  a sudden deviation in drift
direction occurs.  Air temperature and relative humidity are measured using a
hand-held whirling hygrometer outside the laboratory.

     On account of the possibility of undispersed local emissions giving val-
ues which are not representative of the area or the air parcel being tracked,
the measurements were made at sites where these effects were considered to be
minimal.  The usual sites chosen have been near large areas of parkland with
few trees and with the laboratory downwind of the open  area.  Golf courses,
bowling clubs, and sports areas have been found to be ideal for this purpose.
Usually mains power was obtained for operating the instruments, but when mea-
surements had to be made with the generator on, the exhaust was directed down-
wind to avoid contamination of the ambient sample.

     "Oxidant Days" are days of predicted stagnant meteorological conditions
with all groups called into the field by the coordinator.  Measurements are
made with the mobile laboratory on Oxidant Days and on many other days when
rain or high winds had not been forecast.   At other times the van is garaged
in the Heat Laboratory in the Department of Mechanical  Engineering at the
University of Sydney and monitoring is continued from a sample point located
above the roof of the building.  Mobile measurement strategies included survey
work aimed at determining precursor sources, concentrations and suitable sam-
pling sites, and trajectory measurements aimed at tracking air masses from
source to receptor areas.

     Trajectory measurements have been made on days when the wind is light and
high oxidant formation expected.  The usual practice has been for high morning
concentrations to be located near a known source area and then followed when
the winds increased.  The trajectory is predicted from the wind speed and di-
rection in the lowest one-hundred metres of the atmosphere measured at each
successive site.  The mobile laboratory is shifted downwind by the distance
the air would move if it continued with the measured velocity.  Allowance is
made for the monitoring period and time to travel between points on the tra-
jectory.  It has been found that this practice has to be modified when trav-
elling in the sea-breeze front.  Otherwise one gets ahead of the front.
Starting points have been at positions downwind South and East of the C.B.D.
and East of the inland industrial source at Silverwater.  From the C.B.D. tra-
jectories were followed inland with the sea breeze.

                                 RESULTS
     The observations given here are taken from measurements made  from Oc-
tober  1975 to April 1976.  In this period the calibration of the ozone and   ~
oxides of nitrogen monitoring instruments is considered to be accurate to
within ten percent both  from month to month and in absolute terms.  A zero
correction of 0.3 pphm has been subtracted from NO concentrations  according  to


                                     932

-------
the findings above.  Hydrocarbon calibration  is  reported  in  a  consistent way
;jo pprn C fas methane), although no allowance  is  made  for  correction for the
relative molar response for hydrocarbon  components of the  ambient mixture.

     The distribution of measured half-hour average  concentrations of  NMHC
and NO/ for the morning period  0600  -  0900 at the  University of  Sydney and
other sites is given in Figure  3.  These data include measurements on  days
when either field measurements  were  made, NMHC  exceeded 0.5  ppm, maximum 0-
exceeded 5.0 pphm, or NOy  exceeded 5.0 pphm.   For  high concentration of NMHC
and NOx, the NMHC-to-NOx ratios vary between  4  and 12 (ppm C/ppm).   It was
considered that the NMHC/NOy  ratio would give an indication  of the source  of
precursors, since Australian  automobiles tested over the  U.S.  Federal  Driving
cycle give emissions approximately in  the ratio 5, industrial  combustion gives
a  ratio much lower than this, and refinery emissions, including  evaporation,
are much higher, being greater  than  10 ppm C/ppm.  The ambient levels  however,
are a mixture of emissions from many sources  and it  has not  yet  proved possi-
ble to isolate predominantly  automobile  emission from other  sources  using  the
NMHC/NOx ratio.  Because of the mixing of the emissions,  the ambient levels
are compared with predicted emission levels using  inventory  data.
               2.8
               2.4
              2.0
            I  1.6
            o
               1 2
o
| 0.8
              0.4
                                t
                              6 I , I5 0
                              * 2, 8 I
                              36. 70
                              2 9. 6 9

                         t POINTS NOT SHOWN ON ijR
                                                          o    8
                                          0°§°
                  O     9

                       O   o
                o        oo
                  o o o    *PO
           O 00%     -      °
                              o o
                                o  o o
                                °
     o
     o
)
   »   o o


° °°0° °   ?0
  o  o

1 o% °
                                 8   10   12    14   16   18   20   22
                                          NOX , pphm
                     Figure  3.   Half  hour  average  NMHC
                                and NOX  (0600 to 0900).
     The emissions inventory given by Iverach and Bilger  (1975) includes emis-
sions from auto exhaust, crankcase blowby, and evaporation, as well as  refin-
ery, distributed solvent evaporative, and stationary combustion emissions.
The automobile components are computed using factors estimated for Australian
motor vehicles and using traffic volumes and speed from a Sydney-wide traffic
survey.  An analysis of the November and December 1975 precursor observations
(Allen, 1976) shows that there is no statistically significant correlation  of
the observed ratio of NMHC to NOx with a ratio predicted  using upwind emission
from the inventory for each measurement site.  The median through the distri-
                                      933

-------
but ion  indicates that the predicted ratio  is about  thirty percent  higher  than
observed.  Recent emissions  laboratory  tests of Australian motor cars  show
that  the original automobile emissions  factors used in the inventory are  prob-
ably  too high for hydrocarbon emission  and too low  for NOx, and this fits with
the observed ratio.  Adjustments to stationary source emissions factors may
also  need to be taken into account.  The inventory  grid size, which in this
case  was two miles,  is  another factor which could influence the outcome of
comparisons between  measurements and inventory predictions, arid this needs
further  investigation.

      The measurements from March 19, 1976  have been chosen as a typical day
illustrating the relation between  concentrations predicted from the emissions
inventory and measured  values.  On this day there were putative health effects
(reported breathing  pains in children exercising),  peak ozone concentration
was recorded near noon  at 0.23 ppm, and visibility  was reduced to  a few hun-
dred  metres where the peak ozone was recorded.  The measurement record  (Fig-
ure 4)  shows ambient concentrations of  0^, NO, and  NO?, with wind  velocity
indicated at the top of the  figure.
                      30
                               6      12
                                  TIME (EST)
                     Figure  4.   Measurement  record  for
                                March  19,  1976.
     In these observations, the 0900 to 1200 readings were downwind of the
Silverwater source, 1200 to 1400 values were in sea breeze and were conse-
quently lower, and after 1400 the mobile laboratory was driven inland to
rejoin a trajectory computed from surface winds and high values were again
recorded.
                                     934

-------
     fhc- morning build up of ozone is plotted in Figure 5 and compared with
the predicted concentrations from the Eschenroeder and Martinez (1972) DIFKIN
program adapted for Sydney conditions (Allen and Bilger, 1976).  Following
Eschenroeder and Martinez, the predicted NO, emission has been reduced by a
factor of 4 and NMHC by 2 to better approximate the measured values.  Observed
ozone levels fell at the onset of the sea breeze and this is not shown by the
model, since the model presently cannot handle conditions of wind shear and
strong mixing.
                                                      z
                                                      o
                                                      CD
                                                      
-------
the source strengths from averaging the emissions inventory along the tra-
jectory are greater by a factor of two to four.   The source strengths fall
with time from 0900.  This may be due to chemical reaction, but is more
likely associated with dilution in a deepening surface layer.


           TABLE 1.  HOURLY AVERAGE NMHC AND NOX AND COMPUTED SOURCE
               STRENGTH DOWNWIND OF SILVERWATER, MARCH 19, 1976

Time


0900
1000
1100
1200
NMHC

ppmC
0.65
0.57
0.34
0.23
NOX

ppm
0.065
0.063
0.045
0.028
Computed NMHC
source
kg/4 sq. mile/3hr
1220
1080
640
430
Computed NOX
source
kg/4 sq . mile/3hr
400
390
280
170
  average source given by inventory        2400                  820
     All attempts to resolve the anomaly of the predicted concentrations from
emissions exceeding measured concentrations using the known mixing layer depth
have been unsuccessful.  On other days, downwind from the petrochemical plants
at Silverwater before 0700 ozone transients to 0.1 ppm have been observed fol-
lowing high transient NMHC concentration.  At these times high NMHC/NOx ratios
have been observed (greater than 50) and a purple brown haze was usual.  The
high concentrations were followed on December 12, 1976 about twenty kilometres
downwind and higher than normal NMHC, ozone, and nitrogen dioxide concentra-
tions were found.  The observation of high ozone concentrations within an hour
of sunrise is evidence for highly reactive hydrocarbon reactions.  This chem-
ical activity is not typical of photochemical smog production from auto ex-
haust where the reaction may take several hours to produce high ozone  levels,
and all of these observations have been traceable back to areas where  there
is a petrochemical industry.

     The identification of the C.B.D. as a major emission source is more dif-
ficult than for refinery and industrial areas, and there are two main  reasons
for this.  The distribution of the traffic is not over such a small area as
with industry, even in the C.B.D.  The ambient levels of NMHC and NOx  are con-
sequently lower even though the same mass of pollutants may have seen  emitted.
The second reason is that the peak traffic emissions occur at about, the same
time as the surface inversion breaks so that the ambient levels of the pollu-
tants are lower and not as easily distinguished from other emissions.  Ambient
levels from stationary sources are usually higher since they can be measured
before the inversion breaks and they generally originate from well-defined
areas.


                                      936

-------
                        OZONE PRECURSOR RELATIONSHIPS
      The development of ozone-precursor relationships from aerometric data
 requires a very large amount of data.  In the EPA study used to develop the
 correlations shown in document AP-84 (1971), data from some thousands of days
 were analysed.   These data were, however, for fixed site monitoring stations
 and the envelope correlations shown represent a very few points.  Because
 these points represent maximum ozone for minimum precursor concentrations, it
 is probable that they correspond to days when dilution due to burning off the
 ground-based inversion occurred early in the day so that the smog development
 was not affected by later dilution, and that the days were relatively calm so
 that the ozone was not transported from the monitoring sites.

     In the present study correlations will  be developed between downwind max-
imum ozone and near-to-source morning precursor concentrations.   Accordingly,
much less data will be needed to establish an envelope relationship.  Further-
more, by using the precursor concentrations  existing around 0900 hours, when
most of the dilution has already occurred, correlations which are comparable
to the EPA envelopes, the Japanese correlations, and smog chamber data should
be obtained.  The amount of reaction of NMHC and NOx that has occurred by this
time will be small.

     Table 2 shows data compiled in this way for six days over this last sum-
mer.  A perfect trajectory relationship between sites for the precursor and
the ozone have not been obtained in each case but they should be close.  The
data are plotted in Figures 6 and 7 to show correlation against NOx and NMHC
separately and then jointly.  It can be seen that the data lie well within the
envelope curve developed by the EPA for hydrocarbons and within the smog cham-
ber curve reported by Pitts (1976) for the effect of NOx.  The 0.49 ppm C
curve of Pitts lies close to the AP-84 envelope curve for correlation against
NOx.  On the joint correlation only the four points having ozone close to
14 pphm are plotted.  It can be seen that these four points show good cor-
relation with the Pitts (1976) data and the EPA AP-84 envelope curves.  How-
ever, they lie at quite a lot lower NOx and NMHC values than would be expected
from the Japanese data.  It should be reiterated that there are several ques-
tionable features about using this data in this way.  Also there are not
enough results at this stage to draw any conclusions with regard to control
strategy.  It is hoped that the accumulation of much more data in the
1976-77 season will make the picture clearer.
                                      937

-------
                     TABLE 2.   OZONE PRECURSOR RELATIONSHIPS

Date
30/10/75

27/11/75
1/12/75

NO
ppnm
7.3

2.5
4.4

NMHC
ppm
1.0

0.63
0.93

Site
Silvorwater

Concord
Leichhardt

T i me
0830-0850

0900-0920
0930-0940

Ozone
pphm
14

4.0
14.4

Site
Bel more to
Liverpool
Belmore
Canterbury

Time
1300-1400

1140-1240
1020-1100
                           Park

  30/1/76      7.4    0.56  Centennial    0900-0940   14     Lidcombe     1200-1300
                           Park

  27/2/76     10.1    0.67  Leichhardt    0820-0900    8     Bankstown    1400-1430
                           Park

  19/3/76      6.2    0.85  Marrickville  0910-0950   13.8   Liverpool    1430-1450
                           Golf Club
                    US EPA(1971)
                   =AP-8i fig C-2
             0.5         1 0

                NMHC - ppm
                                 I 5
                                                          • PITTS 1976
                                                            0 10
                                                                      0.15
Figure 6.  Downwind maximum  ozone  concentration as a function of  (a)  original
           upstream NMHC  concentration  and (b) original upstream  NOy
           concentration.
                                      938

-------
                         SMOG  ~
                       CHAMBER
 \.1fJ
 V  '
 \
  \
.4  \
'   \
                                                  \
                             i'r-USEPA(l97l)
                             •\  AP
                             U
                             •\

                              ''\.
   v>
    A:
    i  \
    ^-t'.
                                                     •10
                                  005
                                       NOX  ppm
                                             0.10
                                                        U 15
                   Figure 7.  Peak ozone concentration as
                              a function of upstream NMHC
                              and NOx concentration.  Data
                              shown is for Sydney, Oc-
                              tober 1975 to April 1976 and
                              the dashed lines show corre-
                              lations derived from both
                              ambient and smog chamber
                              data (see references).

                                 CONCLUSION
     Accurate measurements of ozone and precursors of photochemical smog have
been made from a mobile laboratory.  This method of operation for collecting
data on reacting species has the advantage that the same set of instruments
can follow a polluted air mass through an urban area, and sites can be chosen
to avoid any local source effects.  The chemiluminescent analysers proved to
be stable and reliable under these conditions; however, there still remain
problems with the hydrocarbons monitor operation.

     The concentrations of NOy and NMHC precursors have been found to be con-
siderably lower (by a factor of 4) than is predicted from emissions inventor-
ies and dispersion models.  This anomaly is still unresolved.

     In general the NMHC/NOy ratios are around 10, but under some conditions
the ratio groups are closer to 4.  It has not yet proved possible to correlate
the NMHC/NO\ ratio with apparent sources.  The inventories of source emissions
are not well-defined and for some sources, such as oil refineries, the esti-
mates of emissions could be too high.
                                      939

-------
     The precursor concentrations on high ozone days are in relatively low
ranges  (0.4 - 1.0 ppmC NMHC and 0.02 - 0.10 ppm NOX) compared with Los Angeles
experience, but are still consistent with available data on ozone formation.

     The first step in future work in the project must be the resolution of
the anomaly between measured precursors and the values predicted from invento-
ries.  More data will also be collected to establish the relationship between
precursors and ozone formation.
                                 REFERENCES

ALLEN, G.H.  (1975).   Sydney Oxidant Study - Trajectory measurement of ozone
     and precursors between November,  1974 and May,  1975.   Charles  tolling
     Research Laboratory.   Tech.  Note  El.   Univ.  Sydney.   June,  1975.

ALLEN, G.H.  (1976).   The  Use of Mobile Measurements of Oxidants and their Pre-
     cursors.  Proceedings Smog '76 Symposium, Clean Air Soc.  of Aust. and
     N.Z., February 17-19, Macquarie University,  Sydney.

ALLEN, G.H., and BILGER, R.W.  (1976).   Photochemical  Air Pollution Modelling
     in Sydney.  Proceedings of the Symposium on  Air Pollution Diffusion
     Modelling, Australian Environment Council, August 18-20,  at The
     Australian National University, Canberra.

ALLEN, G.H., and IVERACH,  D.  (1976).   Some Comments on a High Ozone Day.
     Clean Air, 10_, p. 23.

BLADES, A.T.   (1973).  The Flame lonization Detector.   J.  Chrom. Sci., 11,
     251 .

CALVERT, J.G.  (1976).  Test of the Theory of Ozone  Generation in Los  Angeles
     Atmosphere.  Env. Sci. Tech., K), (3), p. 248.

DAVID, D.J.  (1974).   Gas  Chromatographic Detectors.  J.  Wiley and  Sons, NY.

ESCHENROEDER, A.Q., and MARTINEZ, J.R.  (1970).  Analysis of Los Angeles
     Atmospheric Reaction  Data from 1968 and 1969.   General Research Corp.
     Final Report CR-1-170, CRC-APRAC Project No. CAPA-7-68 July, 1970.

ESCHENROEDER, A.Q., and MARTINEZ, J.R.  (1972).  Concepts and Applications
     of Photochemical Smog Models.  Advances in Chemistry Series, No.  113
     "Photochemical Smog and Ozone Reactions."

HAWKE, G.S., and IVERACH,  D.,  (1974).  A study of High Photochemical  Pollution
     Days  in Sydney,  N.S.W. Atm. Env.  8_, pp. 597-608.

IVERACH, D.    (1975).  Planning for Oxidant Control  in Sydney.  International
     Clean Air Conference, Rotorua, N.Z., Feb. 17-21, 1975.  (Clean Air Soc.
     Aust. and N.Z.).


                                      940

-------
IVERACH,  D.   (1976).   The Sydney Oxidant  Study.   Proceedings  of  the  Smog  '76
     Symposium,  Clean Air Soc.  of Aust. and  N.Z.,  February  17-19,  at Macquarie
     University, Sydney.

IVERACH,  D., and BILGER,  R.W.   (1975).  The  Status of Air  Pollution  and  Its
     Control in  Australia.   Advances  in Environmental  Science and  Technology
     edited by J.N.  Pitts and  R.L.  Metcalf,  John  Wiley and  Sons, Vol.  5,
     pp.  91-218.

IVERACH,  D., MONGAN,  T.R.,  NIELSEN, N.J., and FORMBY, J.R.   (1976).  Vehicle
     Related Air Pollution  in  Sydney.   J. Air.  Pollut. Control Ass.  26_,  39.

JAPANESE  GOVERNMENT,  A Case History of Oxidants and their  Precursors in  the
     Atmosphere of Japan.  National Report to O.E.C.D., Dec.   (1973).

JONES, I.T.N., and BAYES, K.O.   (1972) paper at Informal  Conference  on Photo-
     chemistry held at Oklahoma State University,  Stillwater, March  15-18
     1972.  Chem. Phys.  Lett.  VU 163, 1971.

PITTS, J.N. Jr.   (1976).   California's Experience in the Control of  Photo-
     chemical Smog.   Proceedings Smog '76 Symposium, Clean  Air.  Soc. of Aust.
     and  N.Z., Feb.  17-19,  Macquarie University,  Sydney.

POST, K.,  HAYNES, B.S.,  ALLEN,  G.H.,  AND  BILGER,  R.W., (1976).   Sydney Oxidant
     Study - Data Summaries for October to December, 1975.   Charles  Kolling
     Research Laboratory.  Tech. Note, ER-7.  University of Sydney,  February,
     1976.  See  also  ER-S,  ER-9 and ER-12.

POST, K.   (1976).  Notes  on Ambient Ozone Monitor Calibration Methods  used at
     the  University of Sydney.   Kolling Research  Laboratory.   Tech.  Note, ER-10.
     University  of Sydney,  May, 1976.

SCHOFIELD, K.  (1974).  Problems with Flame  lonization Detectors in  Automotive
     Hydrocarbon Measurements.   Env.  Sci. Tech.,  8. No. 9.   827.

STEDMAN.D.H. and JACKSON, J.O.  (1975). The  Photostationary State  in Photo-
     chemical Smog.  Int.   J. of Chemical  Kinetics Symposium 1 1975 Proc.  Symp.
     on Chemical Kinetics Data for the Upper and  Lower Atmosphere, Warrenton
     Virginia Sept.  1974 edited by S.W.  Benson, D.M. Golden and  J.R. Barker,
     J. Wiley and Sons 1975.

THE NAUTICAL ALMANAC  FOR THE YEAR (1976).  Her Majesty's Stationery  Office;
     N.P.   314-76.  (1974).

U.S. ENVIRONMENTAL PROTECTION  AGENCY, (1971).  Air Quality Criteria  for
     Nitrogen Oxides.  Pub. No. AP-84, Washington D.C.
                                     941

-------
                                                                           19-3
          SMOG-POTENTIAL  OF  AMBIENT  AIR, SAMPLED AT DELFT, NETHERLANDS:
                   THE  EFFECT  OF  INCREASING NO  CONCENTRATION
                                               X

                            J.  van Ham and H. Nieboer*

ABSTRACT

     Ejects Of) Mic.ie.cu>ing  pie.cauoi e.mi?>t>ionA on. the. Dalmatian o& photo-
che.micat ail pottution  have  ban  dnte.imine.d by e,xpe.lime.ntat A-mutation,  making
ate. oft ambient ail cotte.ct.ed at VeZfat,  Ne.theil.andt.  The. tmog-pote.ntiat  o^
e.aity-moining ail tampteA wat>  compaie.d with that o^ tampte^> o^ the. tame,  day
in which the. ni^tioge.n oxides conce.ntiati.on had be,e.n incie.ate.d 50% by  me.ant
of) addition ofa nitiic oxide.  In  oai e.xpeAime.ntt>, nitA.oge.n dioxide, maxima
on an average incAe.aA&d piopoit^ionatty with niJ:Aoge.n oxides pie.c.aA^oi con-
ce.n&Ldtion; faofi ozone, maxima a 10-15% inc.ie.aAe. wai> fiound &oi a 50% inc.ie.ate.
Of5 nitiogm oxides.  It iA  aigu.e.d that thue. ie.&uttA, which leJLate. to a
Coulee, aiound the. ^ampLing  bite. me.aAuling about 10f 000 km2, aie. indicative.
{,01 the. tie.nd o{, photoche.mic.at ail pottutA.on in the. NeJ:he.itandt>.

                                   INTRODUCTION

     Photochemical air  pollution  in the Netherlands has been documented
since 1966  (1).  There  have since been several incidents of high  concentra-
tions of ozone  (max. cone.  200 ppb)  and peroxyacetylnitrate (PAN)  (max.
cone. 16 ppb)  (2).   Recently,  peroxybenzoylnitrate (max. cone. 4.6  ppb)  has
also been identified as an  air pollutant in the Netherlands (3).  There  is
some concern about this type of air pollution because of resulting  material
and vegetation  damage,  including  crop decreases, and in view of  possible
health effects.

     In this respect it is  of  interest that the nitrogen oxides  (NO ) emission
density in  the  Netherlands  is  one of the highest in the world  (4).

     The Rijnmond  area, with its  extensive petrochemical activities,  is  at
the same time  a powerful  source area for hydrocarbons.  The high  traffic
density in  this part of Holland has been estimated to contribute  22%  and 30%
of, respectively,  N0x  and CxH   emissions (5).

     An increase  in  emissions  will influence  precursor  concentrations and
may result  in  a change  of the  average ratio of hydrocarbons to nitrogen
oxides.  The effect  on  photochemical air pollution of such  changes  is com-
plex; the question of  a relationship between  precursors and oxidants  has been
a matter of discussion  for many years (6), and it  still appears  to  be thought-
provoking  (7).
*TNO, Central Laboratory,  P.  0.  Box 217, Delft, Netherlands.

                                      943

-------
     In order to establish  relations  between  precursors  and  ozone
ing results as well  as  smog-chamber experiments  have  been  used  in
independent studies.   In monitoring
average oxidant concentrations  were
of total  hydrocarbons minus methane
though of some value  in urban areas
                               monitor-
                              several
studies (8), maximum daily one-hour
related to 6-9 a.m.  average concentrations
and/or total nitrogen oxides.   This approach,
 where automotive traffic is the main
source of precursors, has been criticized (8)  because it does  not account for
advection of precursors and oxidants after 9 a.m.

     Moreover, it appears that on account of calibration problems in NO
measurements, the relations obtained are probably in error (9).  Relations
between precursors and oxidants obtained from smog-chamber experiments are
most often based on  studies of systems containing mixtures of precursors,
oversimplified in comparison  to outdoor air.  An approach in which an emission
inventory is coupled with a chemical/transmission mathematical model appears
to be a promising tool for establishing the causality pattern of precursor/
ozone relationships  and  the consequences of various emission "scenarios".
Such an approach that, apart  from input data from the Netherlands, should
also take into account emission contributions from adjacent European countries,
is at present still  far  from  its realization.

     Another tool to predict  the influence of emission changes is experimental
simulation; it studies the effect of changing the precursor concentrations
or their ratio on the smog-potential of ambient air.  In 1973 we reported
on a technique for performing this  type of measurements  (10).  Measurements
of smog-potential of ambient  air have been carried out by several investiga-
tors  (11, 12, 13, 14, 15, 16).  As  far as we know, a systematic  investiga-
tion of the effect of increasing precursor concentrations on smog-potential
of ambient air has not been published.  Results obtained in a study on the
effect of a 50% increase in NO  emission on photochemical nitrogen dioxide
(N02) and ozone formation willxbe reported here.  Next,  the significance
of these results for the boundary layer will be discussed.
                                  EXPERIMENTAL
 SAMPLING
      Samples were collected  in Tedlar bags  (80 1.) mounted in PVC containers.
 Air was  taken  in at  1.50 m above ground level by evacuating the container
 through  a  critical orifice.  With the aid of a time switch, three 6-9 a.m.
 air samples of 65-70  litres  were obtained at a time.

 SELECTION

      Samples were irradiated only if one of the following concentration limits
 was exceeded:   50 ppb N0x, 5 ppb acetylene, or 5 ppb propylene.  When there
 had been rain  during  sampling, or in the preceding hours, samples were discarded.

 ADDITIONS

      Nitric oxide was added  by photolyzing N02 emitted by a permeation tube
                                     944

-------
in a slow stream of dry nitrogen.   Car-exhaust gas was added after dilution
with nitrogen.  The exhaust gases  are those produced by a BMW 2000 running a
driving cycle according to ECE 15.
IRRADIATION

     The air samples were irradiated in the Tedlar sampling bags in special
carts provided with TL-05 tubes (Philips).  The spacial light intensity
within each cart amounted to 100 W/m2, measured by the ONBA-method (17).
Owing to temperature fluctuations, variations in light intensity up t
percent were noted; they are insignificant in this comparative study.
irradiation in most experiments lasted 6 hours.
                                                                    to
5
The
TEDLAR REACTORS

     The reactors were conditioned as described earlier (10); this resulted
in the formation of minor amounts of ozone on irradiation of purified air.
Reactors were discarded when the blank ozone production exceeded lOppb/hour.
The ozone concentrations mentioned in this paper have been corrected for
                Decomposition of ozone at the walls of the reactor has been
              We assume that this loss of ozone is equal in bags of the same
             identical sampling histories.
"blank ozone
established.
age and with

MONITORING
     Nitric oxide and NO^ were measured by a Picos (manufactured by Hartmann
and Braun).  Ozone was measured according to Lindqvist (18).  C2-C5 hydrocarbons
were measured after having been concentrated (19).  Peroxyacetylnitrate was
measured as described earlier (2).

                                     RESULTS

     Between September 1972 and February 1974, 40 samples of ambient air,
selected after analysis, were—with and without addi Lions — irradiated in the
laboratory.

CHARACTERIZATION OF THE AIR SAMPLES

     The measurement of precursor concentrations in the samples revealed
that a division could be made into three source categories.  To this end,
NO  , ethylene (E), acetylene  (A) and propylene (P) were taken as source
indicators.

     In car-exhaust gas, prepared according to the European test cycle, we

found that the ratio 10 < [E'] + 'Al +  [P]
                             [NOX1
                                             = 4.8 ± 1.3.
^Average of six tests with a BMW 2000.  The average ratio C H  (in ppm Cfi)/N0
 in the exhaust gas amounted to 2.1 ± 0.1.  For the ratio C hr/NO  in exhaust*
 gas of seven cars of different makes and types, an averageValuexof 1.92
 was found.
                                      945

-------
 Under highway driving conditions, this ratio lies between 0.2 and 0.4.  More-
 over, city  traffic is characterized by a ratio [EJ  : [A] :  [P] =1:1: 0.25;
 the  [E]/[A]  ratio may vary, depending on driving conditions, between 0.7 and
 1.4.

     By  these characteristics,  samples  fell  into  three  categories.   One  was
of predominantly city-traffic  origin;  the  second  had important  industrial
hydrocarbon  contribution (  [EJ  / [A]  -2;  [PJ  /  [Aj  1);  the  third  category  was
rich  in  N0x, indicating  highway-traffic  emissions and/or contributions  from
heating  furnaces or power stations  as  the  predominant sources..  The  average
composition  of the samples  in  each  category  is  summarized in  Table  1.

              TABLE  1.   AVERAGE CHARACTERISTICS  OF  SAMPLES FROM
                         DIFFERENT SOURCE CATEGORIES

source
cate-
gory*


I
II
III
nt NOX [E] / [A]
ppb


11 159 1.0
14 92 4.6
12 197 1.4
10( [E] + [A] + [P] )
[N0x]


5.3
13.1
1 .6
Average
windspeed
during
sampl ing
m/sec.
1.85
1.8
3.0

   *  I  =  city traffic,   II  =  industrial,   III  = rich  in NOX
   t  is number of experiments
     The origin of the samples was also checked from the wind direction
during the sampling period;  the sample categories  matched  the various source
areas very well.  This is shown in Figure 1.   This finding enables us to
compare smog-potentials  of samples originating from different source areas.

SMOG-POTENTIAL

     A typical result of the irradiation of ambient air is reproduced in
Figure 2.

     The following data  of the oxidant components, e.g. ozone,  may serve as  a
measure for the smog-potential:

     •    the maximum concentration               -. [Q3]max (ppb)

     •    the irradiation time required to reach  maximum concentration
                                                         (03) (min)
          the integrated dosage (concentration x time)
                                                : n-hours-Oy-dosage  (ppb   hour)

                                     946

-------
 Figure  1.   Map  of  southwestern  part of  the province of South Holland; source
                 areas  of  precursors in  air sampled at delft.
          I  =  City Traffic  ;  II  =  Industrial  ;  III -  Rich in NO   .
                                                                 A
        400
        300
        200
                            DELFT ,  17-9 - 1973
                                               __	PAN
                                                             ppb

                                                              I
                                            NO;
20
15
10
           0       50      100      150     200      250     300
                                             	^-  t.min

Figure 2.   Smog profile of an ambient air sample irradiated in the  laboratory.

                                     947

-------
      The  figures obtained have been averaged per source category, differences
  in  ozone  data among the three categories appear from Table 2.


 TABLE 2.   AVERAGES OF PARAMETERS FOR OZONE  FORMATION IN THREE SOURCE CATEGORIES

Source


City Traffic
Industrial
Rich in NOX
N*


10
14
12
t (0.
max.
mm.

238
198
>295*
5/ [03] 3 hours
ppb

103
105
62
[03] max.
ppb

180
149
183*
3-hours
03-dosage
ppb hours
202
243
116
4-hours
03-dosage
ppb hours
396
351
241

n* is number of experiments
* In this category, several ozone maxima were not reached  within  the six hours
  of irradiation.   Averages were obtained by substituting  the  end-value of
  ozone or the total  irradiation time.   Therefore, the figure  in  the table
  is too low.

  EFFECT OF INCREASING NO  CONCENTRATION ON SMOG-POTENTIAL
                         A

       A typical  effect on the smog-potential of an addition  of NO is illustrated
  in Figure 3.
        200
        750
        700
                                 DELFT, 26-9- 1972
                           100      150      200     250     300
                                                                        25
                                                                        0
        Figure 3.   Formation  of NOp  and  0-,  in  ambient  air  irradiated  under
                               standard  conditions.
                                       948

-------
     The effects  have been expressed as  ratios  of corresponding quantities
of the spiked sample and of the control  sample.   These ratios  have been  aver-
aged.  Table 3 shows ratios for several  characteristics of the smog potential
on addition of NO, averaged per category with standard deviation.


 TABLE 3.  AVERAGE EFFECT (WITH STANDARD DEVIATION) ON FORMATION OF N07 AND 0
               WHEN NO -CONCENTRATION WAS INCREASE 45°,, IN THREE      ^
                      CATEGORIES OF AMBIENT-AIR SAMPLES

Source
Number [N02i
] max 4-hours-N02-dosage tpiax
[N02] max ° 4-hours-N02-dosage ° ^max
city
traffic
• industrial
rich
in NOX
Source

city
-
traffic
industrial
rich

in NO
X
10 1.
9 1.
8 1.
Number [0
[0
10 0.
9 1.
8 0.

44
32
54
J
3]
97
06
49

+ 0.09
+ 0.07
+ 0.09
3 hour
3 hour °
+ 0.07
^0. 11
+ 0.10

2.15 + 0.38 1.16
2.15 + 0.53 1.25
1.62 + 0.16 >
[03]
[03]
± 0-
+ 0.
1.24*
+
0
06
20

[03] max 4-hours-03-dosage
[03] max ° 4-hours-03-dosage
1 .09 + 0.04 0.92 + 0
1.14 + 0.04 0.91 + 0
>1.10* 0.60 + 0

.06
.07
.05

0





 *  In the spiked samples, the ozone maxima had not been reached at the end of
    the experiment; a higher average is probable.

      PAN  formation  has  been  measured  in  a  few  experiments  only.   For  the  "city
 traffic"  category,  some examples  of the  effect of  NO  addition  on  the  PAN
 maximum are  presented  in Table  4  and  compared  with 03 maxima  in  the same
 experiments.

                                    DISCUSSION

      The  smog-potentials of  ambient air  in the "city  traffic"  and "industrial"
 categories  (Table  2) were found to  be  such that ozone maxima  appeared on  the
 average after  3  to  4 hours of  irradiation.   The integrated light  dose after 4
 hours' irradiation  is  the equivalent  of  a  half summerday's sunshine in the
 Netherlands.
                                      949

-------
 TABLE 4.   PAN AND OZONE-MAXIMA IN SOME SAMPLES OF AMBIENT AIR AND THE EFFECT
                               OF NO-ADDITIONS
              Date        [PAN] max, ppb           [0,]  max,  ppb

21-11-'
20-12-'
19- 2-'
Control Spiked
sample sample
73 22 29
73 15 18
74 10 15
Control Spiked
sample sample
360
212 184
160 200

     The "rich in NO "  samples  need some  more  time.   It appears  that the emis-
sion situation, as far  as represented in  the  samples  under investigation, is
on the average favourable for the formation  of photochemical  smog.   In this
respect, the "rich in N0x" samples carry  less  weight  than the other categories
because they occur predominantly in the winter half-year.  At our latitude,
oxidant formation by photochemical processes  is then  negligible  owing to the
lower light intensity.   Samples from the  other two categories are more or less
distributed over the whole year.  Early-morning sampling is a necessary con-
dition for obtaining samples whose precursors  composition has not yet been
altered by chemical reaction.  It would be of interest to check  whether the
average composition of emissions which accumulate below the radiation inver-
sion during the night and early morning differs considerably from that of
a 24-hour day.  We have thus far confined ourselves to a crude inventory of
emissions; it indicates that two major sources will generally not be repre-
sented in early-morning samples:

     •    hydrocarbons  evaporation losses from storage tanks:  evaporation is
          promoted by insolation and occurs  mainly between 8 a.m. and 2 p.m.;

     •    nitrogenoxide emission from high sources:  the radiation inversion
          is often not dissolved until 9  a.m., thus preventing stack emissions
          to mix with the surface layer.

     This supports our assumption that the average ratio of hydrocarbons/nitro-
gen oxides in the samples does not differ much from the  ratio of the emissions
of a natural day.  Moreover, the samples  were collected  under a variety of
wind directions so that the resulting collection of samples covered widely
diverging hydrocarbons/nitrogen oxides ratios.  This  is  of importance, as it
justifies the assumption that the average effect of NO   additions on the smog-
potential of the early-morning samples equals the effect on a hypothetical
sample containing a proportional part of the average emissions during 24 hours,
     The most pronounced effects of future increase in N0x emission on smog-
potential are summarized in Tables 3 and 4, which give averages for a collec-
tion of experiments with 45% N0x increase.  These effects are:

                                     950

-------
     •    proportional  increase  of  NO-  maximum  and  NO;  dosage;

     •    increase in maximum ozone concentration;

     •    increase in maximum PAN  concentration:   the  effect  for  PAN  is  possi-
          bly more pronounced than  for  ozone;

     •    lengthening of time-period needed to  reach N02  maximum  and  ozone
          maximum;
          decrease of ozone formation after short irradiation  times,  corres-
          ponding with the lengthening of t    (03);  a marked  effect  in
                                           max
          [03]o ,      is only visible in the "rich in NO " category.
     In order to translate these effects in terms valid for the free atmosphere,
the role of factors that have not been incorporated in our experiments has
to be assessed.   In the boundary layer, wind and convection are important;
moreover, emission of precursors is a continuous process.   Of these factors,
continuous emission and dilution are counteracting and partly compensating
processes, and in our comparative experiments they have been left out of
account.

      Vertical transport,  which  is  primarily  responsible for  dilution, may have
 another effect  during  smog  episodes.   It  is  now  well  documented  (20)  that
 ozone  formed  in  smog episodes during  the  day may persist  overnight  above  a
 radiation-inversion  layer.   The next  day,  the  ozone  is mixed with  new precur-
 sors,  accelerating chemical  reactions  which  lead to  photochemical  smog.

      We wish  here  to point  out  that  our comparative  experiments  also  afford
 the  best practical  approximation  to  obtain the  average effect  of changes
 in  precursor  composition  on  the concentration  of ozone persisting  overnight.
 Most relevant in this  respect are  the  effects  on [03]    ,  which  were  tabula-
 ted  in  Table  3.                                       max

      We expect  that  the  retarding  effect  of  extra NO will  be  counteracted
 when more  ozone  from aloft,  also  resulting from the  extra  NO ,  is  available
 for  mixing.

      At the  present  moment,  a quantitative evaluation of  this  effect  is  not
 yet  possible, owing  to insufficient  knowledge  of atmospheric processes,
 expecially that  of mixing of air  parcels  after disappearance of  an  inversion
 layer.

      For this reason,  the effects  on  the  ozone-dosages in  ambient  air may
 ultimately turn  out  different from those  found in the smog chamber.   However,
 the  effects  on  the ozone  concentration and,  more specifically,  the  ozone
 maxima  indicate  the  trend of photochemical air pollution  for the chosen
 emission scenario.

      Horizontal  transport of air  parcels,  from which the  samples have been
 taken,  causes the  ozone  concentrations to reach  their maxima at  distances

                                      951

-------
from thp (.-minion cm;a that vary with  windspeed.   Consequently,  the  averaqe
effects found concern an area around the sampling  site of about  10,000 km-'
(based on the mean wind velocity during the night  before sampling).   The ozone
above a nocturnal inversion undoubtedly originates from emissions in a still
larger area.

     Summarizing, it may be noted that our technique is useful  especially
when a combined chemical-transmission model is not available, and when an
increase in precursor emissions is expected.  On the basis of an estimated
50  increase in NO  emission in the Delft area within the next 5-10 years,
a 10-15', increase in severity of photochemical smog is to be expected, at
least for those days when its formation is "favoured" by the weather.

                                ACKNOWLEDGEMENTS

     This work was  partly  supported by  a  grant of  the  Dutch  Ministry  of  Public
Health  and  Environmental Hygiene.   The  authors wish  to  express  their  apprecia-
tion  to  many  co-workers  at  the  Central  Laboratory  TNO;  in particular  to  Mr.
Th.  Lems, who carried  out  most  of  the  experimental work,  and to Mrs.  C.M.B.
Verwiel-Keijzer,  who  did the hydrocarbon  analyses.   Meteorological  data  were
provided by the  Ypenburg Airport department of the Royal  Dutch  Meteorological
Institute  (KNMI).   Car-exhaust  gas  was  obtained from the  Institute  for Road
Vehicles, TNO.

                                    REFERENCES

1.    a.    J.  G.  ten Houten,  Landbouwk.   Tijdschr.   78,  2  (1966).

      b.    J.  A.  Wisse  and  C. A.  Velds,  Atmospheric Environment  4,  79-85 (1970).

      c.    R.  Guicherit,  R.  Jeltes  and  F.  Lindqvist,  Environ. Pollut.  3,  91-110
           (1972).

2.    H.  Nieboer  and J.  van Ham,  Atmospheric Environment,  10, 115-20 (1976).

 3.    H.  Nieboer  and G.  M.  Meijer,  to  be published.

4.    J.  Doelman,  A. J.  Elshout,  R.  Guicherit, J.  van der Kooy,  De  Ingenieur,
      87, 894 (1975).

5.    N.  van Lookeren  Campagne,  J.  J.  Verbeek, M.  L.  Huisman, C. M.  Verheul,
      Papers presented at "Stacmap"  meeting, KIVI,  The Hague, March  1974.

6.    a.    J.  C.  Romanovsky,  R.  M.  Ingels, R.  J. Gordon, J.  Air  Poll.  Control
           Assoc.  ]]_,  454-9 (1967).

      b.    A.  P.  Altshuller,  S.  I.  Kopczynski, W.  Lonneman,  D.  Wilson, J. Air
           Poll.  Control  Assoc.  V7,  734-7 (1967).

      c.    W.  A.  Glasson and C.  S.  Tuesday, Environ.  Sci.  Techn. 4_,  37 (1970).

      d.    B.  Dimitriades,  Environ.  Sci. Techn.  6_  253-60 (1972).

                                      952

-------
     e.    A.  J.  Haagen-Smit  in  "Photochemical  Smog  and  Ozone  Reactions" Ad-
          vances in  Chemistry Series  No.  113,  Washington  1972.

     f.    J.  N.  Pitts,  Remarks  before the Subcommittee  on Public  Health and
          Environment of the Congress of  the  United States, September  1973.

7.    Workshop on Photochemical  Air Pollution,  Centraal  Laboratorium TNO,
     Delft, 18 May 1976.

8.    Air Quality Criteria for Nitrogen Oxides, EPA  Washington D.  C.  1971.

9.    Chemical Week 1972, July 19,  31.

10.  J. van Ham and H. Nieboer, Proc. 3d International  Clean  Air Congress,
     Dusseldorf 1973, C102.

11.  E. R. Stephens, F. R. Burleson, J. Air Poll. Contr.  Assoc. ]7_ (3) 147
     (1967).

12.  A. P. Altshuller, S. L. Kopczynxki, W. A. Lonneman,  F.  D. Sutterfield,
     Environ. Science Techn. 4- (6),  503  (1970).

13.  S. L. Kopczynski, W. A. Lonneman, F. D.  Sutterfield, P.  E. Darley, Environ.
     Science Techn. 6^ (4) 343, 1972.

14.  J. J. Bufalini, W. A. Lonneman, Environ. Letters ^  (2),  95 (1973).

15.  S. L. Kopczynski, W. A. Lonneman, T. Winfield, R.  Seita, J. Air Poll.
     Contr. Assoc. 25^ (3), 251 (1975).

16.  Japanese National  Case History  on Photochemical Air Pollution, presented
     at OECD-meeting, April  1974.

17.  J. N. Pitts, Jr.,  J. M. Vernon, J.K.S. Wan, Int. J.  Air Water Pollut.
     i, 595  (1965).

18.  F. Lindqvist, Analyst 9_7_, 549 (1972).

19.  H. Compaan,  presented at Orientatiedagen, "Milieumeting", October 1973.

20.  a.    S. Duckworth, R. W. McMullen,  68th Annual Meeting of APCA, Boston,
           June  1975.

     b.    H. van  Dop, R.  Guicherit and R. W.  Lanting, to be published.
                                     953

-------
                                                                           19-4
     A PRELIMINARY INVESTIGATION OF  EXPECTED  VISIBILITY IMPROVEMENTS
          IN THE LOS ANGELES BASIN FROM OXIDANT  PRECURSOR GASES
                   AND PARTICULATE EMISSION CONTROLS

            C. S. Burton, T. N. Oerskey,  and  S.  D.  Reynolds*


ABSTRACT
             one, pieAe.nte.d deAcsu.bi.ng  the. improvement!) in viAibi&ity,
faoute., on.Qa.yiid, ptumafiy pasvticulate.,  and total. ae.?io£ot conce-ntttationA that
can be. expected faon. d-i^eAe.nt  emi&A-ion control alt&tinativu -in the. Lot,
Ange£.e6 A  indicated,  which in tu.tin points to the.
imptLOveme.nt& -in vi^-ibiJLity and total &uApe.nde.d panticuJLatu with concomi-
tant fie.du.ctLo nt> in ozone. conce.ntn.ationi>.   Ike. importance, o^ &ie.cutinQ the.
totat aifi quality t>y£tm  (ga?>eA and pafiticalatu)  -in designing and e.vatu.-
ating viA-ibiLity and Aulfaate. control £t/iate.g-ieA -i!> empha&ize.d.

                               INTRODUCTION
     It is well recognized  that  in  addition to high oxidant concentrations,
another manifestation of  smog  in the  Los  Angeles Basin (LAB) is reduced
visibility.  Furthermore, it is  obvious  that visibility reduction is the
most generally recognizable form of air  quality degradation.  For cloud-free
conditions, the optical properties  of the atmosphere are in general deter-
mined by the light absorption  by gases and scattering by suspended particu-
lates.  In the Los Angeles  Basin however, recent field studies have often
shown that the measured particle volume  concentration in the range 0.1-1.0
pm (diameter) is correlated well  with the extinction due to scattering by
particles as measured with  an  integrating nephelometer(l).  This observation
is consistent with Mie scattering theory and, for the Los Angeles Basin at
least, indicates the relative  unimportance of light absorption by particles
and gases.  Additional measurements have also shown that the 0.1-1.0 ym size
range encompasses one mode  (mass mean diameter 0.4-0.8 ym) of a usually ob-
served bimodal volume distribution(l,2).   Furthermore, chemical analyses
have often shown that a dominant influence on both the aerosol mass concen-
tration of this mode and  visibility reduction is secondary aerosol--nonmari-
time sulfates, nitrates,  and organics formed in the atmosphere as a result of
a complex series of gas-to-particle conversion processes(l,3-6).  Of these
materials, accumulated evidence  suggests that the sulfate mass fraction of
the total aerosol mass is the  most  important single aerosol constituent af-
fecting visibility on a per unit mass basis(6).

*Systems Applications, Incorporated,  San  Rafael, California.

                                      955

-------
     The physical  and chemical  mechanisms  responsible  for  secondary  aerosol
formation are still  to be characterized in a  manner as detailed  as secondary
gas formation.   Nevertheless,  for the Los  Angeles  Basin, accumulated evidence
implicates the homogeneous gas  phase reactions  rather  than the heterogeneous
reactions as being the most important(l,7,8).   Correlations  between  sulfate
and organic aerosol  concentrations with ozone levels implicate various  por-
tions of the N0x-hydrocarbon oxidation reaction mechanism  as being the  most
important.

     It is qualitatively apparent from the foregoing,  therefore, that there
are a myriad of potential emissions control options for improving visibility,
including fine particle emissions control, S02  emissions control., and control
of oxidants via control of hydrocarbons and nitrogen oxides  precursor emis-
sions.  In particular, improvements in visibility  can  be expected in the
future from oxidant control programs already  underway.  However, the best
single or mixed control strategy has not been apparent, and  is in part  the
subject of the work reported here.

     In this paper it is our purpose to describe briefly the adaptation of
the Systems Applications, Inc.  (SAI) Urban Airshed Model to  include  secondary
aerosol formation and the prediction of visibility (visual range).   We  then
use this planning tool to examine the effects on visibility  improvements for
the Los Angeles Basin for various control  strategy options.   These options
include (a) continued oxidant control to meet the  ozone standard with and
without further reductions in sulfur dioxide  (S0?) and primary particulates
and (b) no further oxidant control from current levels, but  additional  S02
and primary particulate emission controls.  We  emphasize that the findings of
this part of the investigation are preliminary, in that further  verification
of the various elements of the analysis is necessary and for each control
scenario greater resolution in the degree of  control should be carried  out.
We also emphasize that the results are limited  to  the  Los  Angeles Basin.

                           ANALYTICAL APPROACH
     The analytical approach taken in this study consisted of three parts:

     •    Prediction of the mass concentration of the principal  aerosol com-
          ponents  (sulfate, organic, nitrate, and primary participates) in
          the Los Angeles Basin for a typical smoggy day using a numerical
          grid-type model(9).

     •    Calculation of the visibility in four-square-mile cells using a
          relationship between the visual range and the aerosol  scattering
          coefficient(lO), and a linear relationship between the aerosol
          scattering coefficient (bsca^) and the mass of each component of
          the aerosol derived from data collected in the California Aerosol
          Characterization Experiment (ACHEX) for the Los Angeles Basin(l).

     •    Application of the model to various control strategies including
          nitrogen oxides (NOX), hydrocarbon, S02, and primary particulate
          reductions.

                                     956

-------
     Application of this methodology,  which  will  be  described  in more  detail
shortly, is well suited for establishing the relationship  between  emissions
of primary pollutants and the expected levels of  deleterious products,  in
particular secondary pollutants.   It provides both spatial  and temporal  dis-
tributions of both the primary pollutants and their products,  and  the  effect
of controls on the concentration of primary  pollutant as well  as  secondary
pollutants.

     In this study 1975 was taken as the base year.   A day was selected with
meteorology typical of conditions that would occur on a smoggy day in  the
late summer or early autumn.  For the appropriate 1975 emissions,  visibilities
were calculated for each four-square-mile area.   Then reductions  in ambient
concentrations of S02, primary parti culates, and  ozone were made  so as to
meet various desired concentrations and the  visibilities for each  grid square
computed.  The desired reductions in ozone concentrations  led  to  reductions
in the nonmethane hydrocarbon and NOX concentrations.  It  should  be noted  that
the maximum ground-level concentration, which occurs in different places in
the Los Angeles Basin, was used to determine the  desired percentage reduction.

     In the following paragraphs our methodology  is described  in  greater
detail .

RELATING VISIBILITY TO MASS AND AEROSOL COMPOSITION

     Meteorological range or visibility is defined as the  distance at which a
standard object can just be distinguished from the background.  Koschmeider
(10) found that the visual range, Lv, could be expressed in terms of the ex-
tinction for light scattering by the relation

                         Lv = k (bscat) -1                                (1)

where bsca|- is  in units of 10~4 m"1.  Values of k have been determined theo-
retically and empirically.  For the former,  for a homogeneous atmosphere,  k
is 3.9 for an ideal black target, while for less  than ideal targets a value of
2.9 is appropriate.  Empirical values determined in California range from 2.27
to 3.37.   In this study we used k = 2.9.

     Of those field studies establishing relationships between bsca+- and aero-
sol mass,  the results of the ACHEX study are the most useful.   Statistical
analysis of two-hour average aerosol composition, relative humidity, and
corresponding bscat measurements provided a relationship between bscat and
the mass concentration of sulfate, nitrate, total aerosol  mass, and relative
humidity.  The  relationship is given by
  bscat(lO-lt m-1)- 0.02(MT - Ms - MN) + 0.062 Ms + (0.020 + 0.049y2)MN   (2)

where Mr, M^ and M$ are the total mass, nitrate mass, and sulfate mass, res-
pectively (all in units of yg m~3), and y denotes the fractional relative
humidity.   In using Equations 1 and 2 to relate the calculated mass concen-
trations of the constituents of the aerosol to the visual range, we:
                                     957

-------
     •    Utilized model  total  mass,  sulfate,  and nitrate results  averaged
          over a two-hour period and  a 2x2 mile spatial  average.

     •    Assumed a 50 percent relative humidity.

THE NUMERICAL MODEL

     The numerical model  used to predict the aerosol  species  concentrations
was a modified version of the grid-type Eulerian model  described  by Reynolds
et al.(9).  The model  solves the species transport equations
          3t
                       = v(Kvc.j) + R-j  + S1-,                               (3)
subject to boundary conditions which are described in Reynolds et al.(9).  In
Equation 3, c-j is the concentration of species  i,  u  denotes  tne wind vector,
K is the mass diffusivity, R-j  is the net rate of production  (possibly nega-
tive) of species i, and Sj denotes the source term.   In the  model's normal
operation, the concentration of each species or group of species is deter-
mined by the solution of the set of transport equations that are coupled
through the reaction terms.   For this study, the transport equations were
solved only for SOz, sulfates, organic aerosol, and  primary  particulates.  The
other species involved in the chemical reactions,  nonmethane hydrocarbons and
ozone, were obtained from measurements and were not  predicted.  By doing this
we avoided computational errors and reduced the cost of applying the model.
Nitrate aerosol concentrations were not predicted  by the model but were in-
cluded in the total mass by a method described  below.  Background aerosol
concentrations can be included in the model through  initial  and boundary
conditions; however, for this application background concentrations were
simply added to the total calculated mass prior to the calculation of the
visibility.

     Wind vectors and mixing depths are required inputs to the model for the
entire time period for which the model is run.   September 30, 1969 was chosen
from the days available because of the relatively  high pollutant concentra-
tions measured on that day and because the wind patterns were typical of many
late summer and early autumn days in Los Angeles.   The meteorology for that
day was used in all the computer simulations described in this report.

     Two basic assumptions were made regarding the formation of secondary
aerosols:

     t    Formation of each species of secondary aerosol occurs indepen-
          dently of the primary aerosol present and the rate of formation
          of the other species.

     •    Steady-state concentrations of condensable materials exist in
          the gas phase so that the rate of formation of condensable ma-
          terial in the gas phase is equal to the  rate of formation of con-
          densed material.
                                      958

-------
The validity of these assumptions  has  been discussed(ll ).   It  follows  immedi-
ately from these assumptions  that  all  condensable material  produced  in the
gas phase is converted to secondary aerosol.

     Of the possible paths available for sulfate aerosol  formation(12,13) ,
available field data for the Los Angeles Basin and elsewhere indicate the
significance of the indirect photooxidation mechanism for converting SOa to
sul fates (1 ,12).  For this mechanism the reactive species  HO.,  RO- ,  HOj, ROa
and 0(3P) are responsible for oxidizing S02 to sulfur trioxide (563), so that
assuming the reaction of S03 with  water vapor is sufficiently  rapid and the
concentration of water vapor is sufficiently constant, the rate of formation
of sulfate is given by

                    d[SULF]/dt = [S02]  z I^L"^]     ,                    (4)
                                        i

where Rj denotes the reactive species i and k^ is a rate  coefficient which  is
a function of the water vapor concentration.   Since these same reactive
species are related to the concentration of ozone, we assume that Equation  4
can be expressed as
                    d[SULF]/dt = k1[S02][03]Y     ,                        (5)

and will approximately represent the rate of formation of sulfate.   Roberts
(14) has used sulfate and S02 data from the Los Angeles Basin  to calculate
values of k, for Y=l .  The mean value obtained by Roberts, used herein, was
0.72 pprrf1 hr"1.  The mean molecular weight of sulfate of 125  g mole"1 was
used to calculate the mass concentration of sulfate(l).

     Accumulated laboratory and field observations indicate that ozone and
higher molecular weight olefinic materials produce traces of parti cul ate mat-
ter.  Neither the nature of the particulate materials nor the  reactive species
that are important is well established, although it is reasonably certain
that 1:1 adducts of ozone and an olefin are not important(7,13) .  Based on
what is known, an appropriate form of the reaction rate might be

                    d[ORG]/dt =  I  k-jjRHC-jllX.]     ,                   (6)
where k-jj is the reaction rate coefficient, [RHC-j] is concentration of non-
methane Hydrocarbon producing organic aerosol, and [Rj] denotes the concen-
tration of the transient reactants.  Following the same reasoning as used in
obtaining Equation 5 from Equation 4, we have

                    d[ORG]/dt = k2[NMHC]a[02]e    .                       (7)

Grosjean and Friedlander(7) provide data from which k2 can be estimated to be
0.2 ppm"1 hr'1 for a=B=l.  This value together with an average molecular
weight of 159 g mole"1 was used in calculating the mass concentration of
aerosol .
                                     959

-------
     There is virtually no quantitative  information  available on the rates of
conversion of NOX to nitrate in the  atmosphere, albeit many possible paths
exist.  In the ACHEX study, no  systematic,  useful  relation could be estab-
lished between observed nitrate concentrations and those of ozone, nitric
oxide (NO), N02,  sulfate,  relative humidity, or noncarbonate carbon(l).  The
method of estimating the nitrate concentration employed here was to assume
that it represented 8 percent of the aerosol mass  for the base  case year of
1975.  For other  years the nitrate level was rolled  back according to  reduc-
tions in NOV emissions.
           A

     Calculation  of the secondary aerosol  emissions  utilized 1975 emissions
for September and measured ozone and nonmethane hydrocarbon concentrations.
Primary aerosol emissions  were  calculated  from mobile carbon monoxide  (CO)
emissions and stationary S02 emissions  by  multiplying by the corresponding
emissions factors(15).  We assumed no change in these factors between  1973
(for which they were tabulated) and  1975  (for which  they were applied).  The
S02 emissions inventory was updated  to  September  1975 by contacting power
stations in the Los Angeles Basin.   Finally, following Trijonis et al.(16),
a two-hour average background concentration of 44 pg m~3 was assumed.

THE VARIOUS SCENARIOS

     Various scenarios were selected to compare the  1975 visibility levels
with those that would exist under conditions of reduced oxidant and S02 con-
centrations.  The method used to correct hydrocarbon, ozone, and nitrate
levels was based  on rollback using a modeling analysis developed by the Los
Angeles County Air Pollution Control District  (LAAPCD)(17,18).  Originally,
the following four scenarios were proposed:

     •    The 1975 base case;

     •    The 1975 base case, but with  S02 emissions rolled back to meet
          the 0.04 ppm California 24-hour standard;

     •    The 1980 oxidant, NOX, and nonmethane  hydrocarbon  concentrations
          predicted by the LAAPCD method for the  same meteorology  as  the
          1975 base case;

     •    The 1980 case above,  but with S02 emissions  rolled  back  to  meet
          the 0.04 ppm 24-hour standard.

After studying the results from the  above scenarios, two more  cases were
added:

     t    A yearly maximum oxidant level  of 0.16  ppm and  the  corresponding
          NOX and nonmethane hydrocarbon determined  by  the LAAPCD  method
          with no S02  rollback;

     •    The case above, but with S02 emissions  rolled  back  to meet  the
          0.04 ppm 24-hour standard.
                                     960

-------
SUMMARY OF APPROACH

     Ideally the visibility model  would include all  the  phenomena  and  all  the
inputs listed in Figure 1; however, because of the  limited level of  effort in
this study, not all of the inputs  or physical  phenomena  were  adequately
treated.  We have outlined the approach above; in Table  1  we  summarize our
assumptions.

                         RESULTS AND CONCLUSIONS
     Using the approach discussed in the previous  section we have examined
the effect of reducing S02 and ozone concentrations  on visibility within  the
Los Angeles Basin.  The results presented in this  section illustrate  the
following points:

     t    Although the sulfate aerosol  contributes more on a per weight basis
          to the visibility reduction,  sulfate,  in general, represents  less
          than 30 percent of the aerosol mass and  contributes less to visi-
          bility reduction than the other components.

     •    Since a reduction in oxidants reduces  the  concentrations of all of
          the secondary aerosols, greater improvements in visibility  can  be
          made by controlling oxidants  than by controlling S02.

     •    Controls on the large stationary sources (power plants and  refinery
          operations), which are generally regarded  as significant sources  of
          S02 emissions, appear to have a relatively minor effect on  sulfate
          concentrations.

     •    Our results suggest that the current California visibility  standard
          could not be met even by the implementation of controls that would
          result in achievement of both the Federal  oxidant standard  and
          the California 24-hour average S02 standard (even if extreme con-
          trols were placed on emissions of primary  particulates).

The information available from this modeling approach is extensive and cannot
be presented within the space allotted here.  Tables 2 and 3 summarize the
"worst case" visibilities for the various scenarios  considered in this
study.  Without controls on primary particulates,  the best visibility which
could be achieved was 6.4 miles with a yearly maximum one-hour concentration
of 0.09 ppm 03 and a maximum 24-hour average SO  concentration of 0.04 ppm.
The composition of the aerosol in this case is dominated by the primary par-
ticulates.  Table 3 shows the effect of reducing the primary particulate  con-
centration by one-third or two-thirds.   As expected, the most significant im-
provement in visibility occurs when the precursors of secondary aerosols  are
at their lowest concentrations.  In this case, the "worst case" visibility
increases from 6.4 miles to 7.8 miles with a two-thirds reduction in  primary
particulates.  Hence, controls that would reduce the yearly maximum one-hour
                                     961

-------
INPUT (All  as a function of loca-
tion and time)

Emissions
 Anthropogenic (Including pri-
  mary particulates)
 Natural
Pollutants—Hydrocarbons, NO,
 N02, S02
Meteorological Conditions
 Wind Field
 Inversion Behavior
 Diffusivities (as needed)
 Insolation
Initial  and Boundary  Conditions
PHENOMENA THAT MUST BE ACCOUNTED FOR
IN PHOTOCHEMICAL AIR QUALITY SIMULA-
TION MODELS
Transport
Dispersion
Inversion Behavior
Insolation
Natural Emissions
Anthropogenic
 Emissions
Photochemistry
Sulfur Chemistry
Nitrate Chemistry
Dry Deposition
Washout
Rainout
Gas-to-Particle
 Conversion
                        Grid-Based Photochemical
                         Air Quality Simulation
                      Model  (Scale:   Approx.  100
          km)
                             Gas-to-Particle
                         Conversion Relationship
                  bscat-Total Mass or bscat-Composition
                              Relationship
                      Visibility-b   ,  Relationship
                                  S C a L
           Figure 1.  Components of a visibility prediction model
                                    962

-------
               TABLE 1.   ASSUMPTIONS OF OUR APPROACH
(1)   The meteorological  variables  correspond  to  30 September  1969.

(2)   The rates  of formation of sulfate and  organic aerosol  from
     the gas phase are assumed to  be proportional  to  the  concentra-
     tions of S02 and nonmethane  hydrocarbons, respectively.   The
     rate constant in each case is assumed  to be proportional  to
     the ozone  concentration.

(3)   The nitrate aerosol  concentration is assumed  to  be 8 percent
     of the total aerosol  mass for the base case year [approximately
     the proportion found  in ACHEX (California Aerosol Characteri-
     zation Experiment)].   The gas-to-particle conversion of  this
     component  is not modeled.

(4)   Ozone and  nonmethane  hydrocarbon concentrations  are  not  pre-
     dicted by  the model;  they are interpolated  for the year  1969
     from APCD  measurements.  These 1969 concentrations were
     extended to 1975 using Los Angeles County APCD worst-case
     projections.

(5)   A linear rollback technique is applied to  the hydrocarbon,
     ozone, and nitrate concentrations using  hydrocarbon, NOX,
     and ozone  worst-case projections given by  the Los Angeles
     County APCD for the base case year (1975)  and other  scenarios.

(6)   This rollback is assumed to occur without a shift of the
     maximum ozone concentration in space or  in  time.

(7)   The formation of secondary aerosol and its  rate of  condensa-
     tion on particles is assumed to be independent of the compo-
     sition of the aerosol present.
                                963

-------
O
LU U_
                              CTl  CO

                              cxi  ro
o
o.
o
CO
cc: o
-
  co
 O =C
 LJ_ O >-
 o i— i—
        i — i
 co   «_ i
 I— 2: t-i
 <_) O OQ
 LU i — < I — I
 U_ I— CO
 1 1  i — i i — i
 LU CO >
 LU
 _1
 03
(X)  CXI

co  r^
                c
            S-  " fit
            o o o
           4-   C
                                         i—   O>
                                          ^ CTI c
                                          to O O
                                          OJ  • N
                                         c£ o o
                                                   c
                                                   o
                     CD
                     CO
        cr\  i—

        LT>  LO
   E o
l/) D-CJ
4-J CL
•—   OJ
13 <^O C
c/l i— O
(U  • N
c*: o o

•O CD E
OJ C =J
-M r- E
                                                 964

-------
      TABLE 3.  EFFECTS OF PRIMARY PARTICULATE CONTROLS ON VISIBILITY
                   AT THE LOCATION OF MINIMUM VISIBILITY

Scenario
1975 Ozone Concentrations
without S02 control :
1/3 prim*
2/3 prim+
with S02 control :
1/3 prim
2/3 prim
1980 Meeting 0.09 ppm Yearly
Maximum Ozone Concentration
without S02 control :
1/3 prim
2/3 prim
with S02 control :
1/3 prim
2/3 prim
1980 Meeting 0.16 ppm Yearly
Maximum Ozone Concentration
without S02 control :
1/3 prim
2/3 prim
with S02 control :
1/3 prim
2/3 prim
Aerosol
Primary
( yg m~ -1)

39
19

39
19


41
20

39
19

41
20

41
20
Mass
k
Total Mass scat,
(yg m-3) 10-ltm-i

251 7.8
231 7.4

234 6.8
214 6.4


148 4.7
127 4.3

136 4.0
116 3.6


169 5.5
148 5.1

156 4.7
135 4.3
Visibility
(miles)

3.6
3.8

4.1
4.4


5.9
6.5

7.0
7.8


5.1
5.5

5.9
6.5
* Primary particulates reduced by 1/3
+ Primary particulates reduced by 2/3
                                     965

-------
average ozone concentration to 0.09 ppm,  allow  the  California  S02  standard  to
be met, and reduce primary participates by  two-thirds,  would not result in  the
attainment of the present California visibility standard.

     These results also show that a more  significant reduction in  aerosol mass
and a corresponding improvement in visibility can be achieved  by reducing
ozone concentrations than by limiting emissions of  S02  from large  stationary
sources.  This result is observed even when the ozone is  reduced only enough
to meet a maximum hourly average concentration  of 0.16  ppm. Under this con-
dition, without S02 control the aerosol mass is reduced by 30  percent whereas
SO  control only reduces the mass by an additional  7 percent.

                                 ACKNOWLEDGMENTS
     In the course of this project we have benefitted greatly from discus-
sions and criticisms of several  colleagues.  In particular, we appreciate
the valuable comments of Drs. P. M. Roth and G. Z.  Whitten of SAI and Drs.
Attaway, Greenfield and Peyton of Greenfield, Attaway, and Tyler.  We also
wish to express our appreciation to P.  M.  Mundkur and J.  Ames for carrying
out the computer calculations.  This study was carried out under contract to
Greenfield, Attaway, & Tyler, San Rafael,  California, as  part of a larger
study to examine various pollutant control strategies in  the Los Angeles
Basin.
                                 REFERENCES

1.   Hidy, G. M. et al.   (1974), "Characterization of Aerosols in California.
     Volume  IV," Final Report from Rockwell International  Science Center
     to Air  Resources Board, State of California.

2.   Whitby, K. T., et al.   (1972), J. Colloid Interfac.  Sci., Vol.  39,
     p. 136.

3.   Junge,  C.  (1954),  J.  Meteor. , Vol. 11, p. 323.

4.   Dzubay, T. G., and R.  K.  Stevens (1976), J.  Environ.  Sci. (in press).

5.   Whitby, K. T. Proc. Gesellschaft Fur Aerosolforschung, Bad Soden, FRG,
     17 October 1974  (in press).

6.   Waggoner, A. P., et al.  (1976), Nature, Vol. 261, p. 120 and references
     therein.

7.   Grosjean, D., and S. K. Friedlander   (1975),  J. Air Poll. Control Assoc.
     Vol. 25, pp. 1038-1044.

8.   Kurosaka,' Donald (1976), "Sulfate Concentrations in the South Coast Air
     Basin," DTS-76-1, Division of Technical Services, State of California
     Air Resources Board.
                                      966

-------
 9.    Reynolds,  S.  D.,  P.  M.  Roth,  and  J.  H.  Seinfeld  (1973),  "Mathematical
      Modeling  of Photochemical  Air Pollution," Atmos.  Envi ron., Vol. 7,
      pp.  1033-1061.

10.    Koschmeider,  H.   (1924),  "Theorie der  Horizontalen  Sichtweite," Beitr.
      Phys.  Frei. Atmos.,  Vol.  12,  pp.  33-53,  171-181.

11.    Jerskey,  T. N.,  and  J.  H.  Seinfeld,  preprint  (1976).

12.    Wilson,  W.  E.,  et al.   (1976),  "Sulfates  in the  Atmosphere,"  69th Annual
      Meeting,  Air Pollution  Control  Association, Portland,  Oregon,  27 June-
      1  July 1976.

13.    Hidy,  G.  M.,  and C.  S.  Burton (1975),  "Atmospheric  Aerosol Formation  by
      Chemical  Reactions,"  Proceedings of the  Symposium  on  Chemical  Kin-
      etics  Data for  the Upper  and  Lower Atmosphere,  Airlie  House,  Warrenton,
      Virginia,  15-18 September 1974.

14.    Roberts,  P. T.  (1975),  Ph.D.   Thesis,  California Institute of Technology,
      Pasadena,  California.

15.    Birakos,  J. N.  (1974),  "1974  Profiles  of  Air  Pollution Control," Air
      Pollution Control District—County of  Los Angeles,  434 South  San Pedro
      Street,  Los Angeles, California.

16.    Trijonis,  J., et al.  (1975), "An Implementation Plan  for Suspended
      Particulate Matter in  the Los Angeles  Region,"  Final Report to EPA
      Region 9 on Contract 69-02-1384,  TRW,  Inc.,  El  Segundo,  California.

17.    Hamming,  W., R.  Chass,  J.  Dickinson, and  W.  MacBeth (1973),  "Motor
      Vehicle  Control  and Air Quality,  the Path to  Clean  Air for Los Angeles,"
      66th Annual Meeting, Air  Pollution Control  Association,  Chicago,
      Illinois, 24-28 June 1973.

18.    Chass, R., W. Hamming,  J.  Dickinson, and W.  MacBeth (1972),  "Los
      Angeles  Photochemical  Smog--Past, Present,  and  Future,"  International
      Conference on Automobile  Pollution, Toronto,  Canada, 26-28 June 1972.
                                     967

-------
                SESSION 20
CONTROL OF OXIDANT PRECURSOR EMISSIONS -  I

          Chcuxmcw:  R.W.  Bilger
      University of Sydney, Australia
                    969

-------
                                                                              20-1
            TRAFFIC MANAGEMENT AS A MEANS  OF  OXIDANT PRECURSOR CONTROL
                          FROM LOW-SPEED SATURATED TRAFFIC

                          R.  B. Hamilton and  H.  C. Watson*

ABSTRACT
     In laAge. CsUtieA, automob.tt contAA-bute. significant quantities o&  photo-
che.mi.cal  oK.ida.nt pAe.cuUoU,  and the. dsu.vi.ng 4equence  can ef^ec-t mo ton.  ve.hi.cie.
emiA&ionA .   ?Ae.viou& t>tudie..&  have, conce.ntAate.d on fte.du.cing caAbon monoxide.
by tAa^ic contAol.  ThiA  Ae.poAt. &kow&  that ave.Aage. Ape.e.d ti> the. de.teAmini.ng
facto i {^OA hydA.oc.an.bon  emiM>ioni>, wi.th  acce£e.Aation Aate.A loAgeJLy de.teAmintng
nitAoge.n  oxides
            c. de.n&-ity pno^il-ej, in i>ome.  lange. ci£iej> have, been ^ound to  be
           V&kicJte. miJie^  ofi tnaveJt. peA.  unit. ane.a ajie.  Iii.ghe.At in the. ce.ntA.al.
             &iict aA.e.a and fiaLt e.x.pone.ntial£y, beJ..ng 50% £&>-*> at. a 2 mi
Conve.nAeJLy,  aveAage. tJia^ic. 4peects i.ncAe.af>e. with incA.e.a&ing distance, faom
the.  ce.vitA.aL buAine^> du>tfiict.  AveAage. &pe.e.dt> and \)e.hicJie. hattA pe.A mite. oJte.
cio&e.ly Ae.tate.d.  Batud  on OUA. ve.hicie. e.miAt>i.on te^t data, city tAa^tc manage-
ment thAoagh impAove.d tta^ic ^tow can Ae.duce. hydA.ocaA.bon and nitAoge.n oxide.*
e.mi>iont> in ce.vitA.at ba^iviej>& diAtni-ct aA.e.a?> by 48%  and 66% AeJ>pe.ctive.ty.
Steady- Atate. dAtving, the. ide.a£, iMouid gi.ve. a nitA.oge.n oxtdu Adduction o^
appAoxi.mate.iy 70-90%.  A 501 Ae.diict.ion o{, hydA.ocoA.bon and nitA.oge.n oxides
in the. two-miJLe. AquaAe. aAound the. cc.ntAat bm&i.neJ>£ dif>tAA.ct ivotitd  reduce about
251  of, the. total automotive. PACCUUOA  e.mit>Ai.on& in the. g'z.eateA citij atea.
      The. upgrading  o^  manual faixe.d  cycle. tAa^ic signal *y&te.m6  can be. done.
at  a mode.&t cott and Ae.t>ult in oveAali community AavingA o& time.,  acci.de.nt!>,
e.ne.Agy,  and reduced photoche.mi.cal oxidant e.mi6*ionf>  in the. cA^itical ce.ntAal.
bui>i.ne.AA dit>tAi.ct   oA.e.a.

                                    INTRODUCTION

      At  the beginning  of automotive emissions control  it was recognized that
driving  sequences can  affect vehicle emissions.  Thus, emissions  have been
regulated according to measurements made over prescribed driving  schedules.
Rose et  al .  (1) demonstrated that carbon monoxide  (CO) and hydrocarbon (HC)
emissions were sensitive to the  average speed at which the vehicle is driven,
but they were unable  to define a speed influence on  nitrogen oxides (NO )
emission.                                                       .           x
*R.  B.  Hamilton, Shell  Australia, Melbourne, Australia.
 H.  C.  Watson, Melbourne University,  Melbourne, Australia,

                                        971

-------
     While recent works (2,  3,  4,  5)  have  quantified how changing  traffic
patterns influence CO emissions,  little effort appears  to have  been expended
in optimizing traffic management  to minimize the oxidant precursors HC and NO^.

     In this paper we enumerate how the vehicular emissions  source varies
in the city road network, show  which  areas give rise to greatest oxidant pre-
cursors, and propose control policies with simple traffic control  measures
rather than complex optimal  controls.  Projections of the effects  of traffic
management have been applied to traffic streams in which vehicles  are without
NO  emission control, since it  is  only with the implementation  on  July 1,
19^6 (Australian Design Rule 27A)  that N0x emission have been regulated to
approximate U.S. 1973 light vehicle levels, and thus the emission  charact-
eristics of controlled vehicles in Australia have become available to us only
recently.

     The forecasts made in this paper are founded on two earlier investigations:

     •    The development of an empirical  model which simulates the instan-
          taneous emissions from vehicles which are representative of the
          population average.
     •    A preliminary investigation into traffic patterns  in  Melbourne.

These investigations have been  combined with an analysis of traffic density
(VMT/m2) data and associated average vehicle speed for some  major cities.

                             VEHICLE EMISSIONS MODEL

     The emission model is an updated version of that developed by Watson (7)
and is available as a computer program called STEM-CA.   Its  essential features
are that for each of four driving modes, acceleration,  cruise,  deceleration
and idle (ACDI), the emissions  concentration, c, and exhaust flow rate, q,
are calculated  from the instantaneous velocity, v, and acceleration, a:
          q = q (v,a,m)

          cp = cp (v,a)

when m is the vehicle mass and p = CO, HC and NO .
                                                  (1)

                                                  (2)
     These expressions are combined with the pollutant density to calculate
the mass emissions rate:
e  =
                                                            (3)
     The functions q and c  are polynomial expressions in v and c whose
coefficients have been determined by regression analysis of emissions measure-
ments made in a baseline survey using U.S. 1970 Federal Test Procedure (FTP).
Other details of the procedure have been documented by Bulach (8).  Further
details of the validation of STEM-CA may be found in Reference 9.  Another
program, STEM-CB, is available to calculate instantaneous mass emission rates
from modal emission data by a constant volume sampler (CVS) method (US 1973
FTP), but requires further validation.
                                      972

-------
                              TRAFFIC PATTERN  STUDY

     In this preliminary study commenced in 1973 (10),  a single  route was
selected that contained central  business district (CBD), arterial,  sub-ar-
terial  and freeway driving.   The route was subdivided into 10 sections or  links
by major intersections, each section taking about 100 seconds of driving time.

     Standard techniques were used and multiple runs were made during peak
and offpeak hours to obtain a significant (10%) level of accuracy using the
chase car technique (11, 12).  All the data was manually digitized  for subse-
quent computer analysis.

                           EMISSIONS-SPEED VARIATIONS

     Input of the velocity (and acceleration by differention) records from
the traffic pattern survey yielded values of emissions rates for a  typical
vehicle.  Integration gave mass emissions on a distance or on a sectional
basis.   The HC and NO  emissions from the typical vehicle in gm/ml  as a function
of average speed for each pass along each section in peak/offpeak traffic  are
shown in Figure l(a) and l(b).  The HC emissions are strongly speed dependent
over the 1.5 to 43 mph average speed range encountered.  Regression analysis
of these results gives the expression:

          HC = 83.23/v + 3.19, where HC is in gm and v is in mph.

The correlation coefficient for this expression is R = 0.96.  The remainder
of the variation is the influence of changing accelerations from section to
section.

     Superimposed on these results, to establish credibility, are values
from several baseline emissions surveys according to European (13), Austra-
lian (14), Japanese, and American (15,  16) test procedures.  Where necessary,
results have been factored by 1.65 to convert non-dispersive infrared HC values
to flame ionization detector HC values, and adjusted to a common vehicle inertia
mass of 2,820 Ib wt. by assuming a linear  relation between emissions and mass.

     The corresponding variation of NO  with average speed is shown in Figure
l(b).  There is no simple dependence of" N0x emission on average speed.  Emis-
sion rates at constant average speed can vary by as much as 3:1, depending upon
acceleration rates.  This variation will be discussed  in more detail later
in the paper.

                URBAN TRAFFIC CHARACTERISTICS AFFECTING EMISSIONS

     The source strength along a roadway is the product of the emissions per
vehicle and the vehicle volume flow rate.  We have just shown that the emissions
rate is a function of speed and acceleration patterns.  In saturated peak hour
traffic, the volume flow rate is closely related to the average speed and is
a function of a large number of controlling factors which may be satisfactorily
simulated by 'queueing theory1 (17).  However, because of the complexity of
such simulation, we have resorted to published data on traffic volume measure-
ment to enable us to demonstrate the apparent generalities of urban traffic

                                      973

-------
flow and driving patterns.

TRAFFIC DENSITY A FUNCTION  OF  CITY  RADIUS

     Vehicle miles of travel  (VMT)  have  been  quantified  for  Los  Angeles  by
Roth et al.  (18) and by the Sydney  Transportation  Study  (19)  for each  grid
square of the urban areas,  using 2  ml  separated  orthogonal grid  lines.   The
highest VMT occurs in the CBD  and an  analysis of these  results to give VMT
as a function of radius from the CBD  are presented in  Figure 2,  together with
Melbourne data from Smeed (20).   The  sophistication of  the road  networks in
terms of greater or lesser proportions of freeways and  the effect on  road space
demand as influenced by city size have been  allowed for  by normalizing both
VMT and the radial distance, R:

          VMT* = MVT/VMT    and                              (5)
                        max
     30
     HC
      i
     25
     20
     15
     10
                   (a)   H_C
                                         KEY
                                        HC=83.23/v*3-19
                                    	MEDIAN
                                    ®   U.S. SURVEY (adjusted to 2,820 Ib) (16)
                                        AUST SURVEY BY A.D.R. 27 (13)
                                        EUROPEAN SURVEY BY £££.£(12)
                                        JJ.S. COMPACT CARS BY 7 MODE CYCLE(15)
                                        JAPANESE 10 MODE CYCLE (U)
                      10
                                                            35
        Figure la.
          15      20      25      30
            AVERAGE SPEffi  mi/h
HC and NO  emissions predicted for  Melbourne Route
        using STEM-CA program.
*Smeed (2) has shown that the city area,  A,  varies  as  A=P/500)°"7, whence  for

                    = fK = 1  ( p )0-35
                     WTT  V¥  500

                                     974
circular cities  R =/—

-------
      7
    NOx1
         v
                                    (b) NOx
                                        KEY
                                        AS!a)
                               .
                             \-
                              v
          \ ' .    •."'  • ^ ~N .  O •
          \          /        v
                      10      15      20      25      30
                                AVERAGE  SPEED    nri/h
                     35
        Figure  Ib.   HC  and NO   emissions  predicted  for Melbourne Route
                            usS'ng  STEM-CA program.
          R  = RM (1M) 0.35
                    (6)
where the maximum VMT, VMTmav and the population ratio of Melbourne  to  the
                          Hid X
other cities (PM/P) can be found from Table 1.


             TABLE 1.  CITY DATA - TRAFFIC VEHICLE MILES PER DAY
           P(xlO"6)
SYDNEY    LOS ANGELES   MELBOURNE

177,000*    359,000      126,000

  2.8         6.9          2.6
     *Sydney a.m. peak = 15,130/ml /hr

     We conclude from Figure 2 that in central areas there is considerable
similarity in the decrease in VMT with distance from the CBD with a 50% reduction
within two miles.  It will be noted also that the standard deviation between
cities and the time of day (Sydney 24-hour data and a.m. peak hour) remains
consistent.  A generalization extended to all cities cannot be justified on
the basis of the sample of three for which data was readily available,  but
warrants further study.
                                      975

-------
                                 *  SYDNEY
                                 +  SYDNEY A.M. PhAK
                                   LOS ANGELES
                                                             a1
                                                            STD.
                                                            DEV.
                                                            0.25
                                           6       8
                                   RADIUS   R*  miles

         Figure 2.  Normalized vehicle miles  travelled  (VM~)  against
                       normalized radius  for  3  cities.


AVERAGE SPEED INCREASES  WITH  RADIAL  DISTANCE

     The variation in  average traffic  speed with  increasing city  radius is
summarized in Figure 3 for Sydney (19), Melbourne  (21)  and London  (22).
There appears to be considerable  similarity in  the  rising speed with radial
distance for all three cities,  and only small shifts  in average speed occur
between peak and offpeak conditions.   The  large difference in standard devia-
tions at given radii between  London  and Sydney  can  be explained in  terms of
the different areas over which  the speeds  were  averaged:  city boroughs in
London (2 to 40 nil-") and 4 ml-1  grids  in Sydney.

TRAFFIC PATTERNS CHARACTERIZED  BY HALTS PER MILE

     The simplest way  of characterizing traffic patterns  is in terms of halts/
mile.  Data from U.S.  sources (23, 24), including  industry and government
driving cycles, are given in  Figure  4(a)  together  with  actual city  and suburban
driving measurements from the U.S. (12) and Australia,  largely for  Sydney  (25)
and our own for Melbourne (10).   Halt  frequency with  average  speed  is remarkably
consistent, except at  speeds  less than 10  mph where traffic flow  is very
congested.  Agreement  between the data from the same  sources  on idle time  as a
proportion of total trip time is  less  well-defined.
     The constraints of roadway design, signal  frequency driving codes,  ve-

                                     976

-------
           V
        mi/h
   —  MELBOURNE
   h- + SYDNEYAMPEAK
       SYDNEY
          20
          10
                                — -O LONDON AM &
                                     PMPE£K_
                                 3—7  ,

                               	O—•
                                          STD.

                                          cmt/h
                                                    :0
                    2       U       6       8      10"
                            RADIUS    miles

Figure 3.  Vehicle average speed V as a  function of  radius for 3 cities
IDLE %
 60
                                               HAUS/mi
 30
 15
    ©EPA.CITYIADR27A)
    @ E.PA HIGHWAY
    <3> FORD CITY
    ®GMCBH& SAE
            URBAN
it ©^-MELBOURNE
 v     REF. 8
          PELENSKY
          REF. 25
      (b)IDLE
                USTRAFFIC
                REF 11
                                  &
                                              ® GMSUBURBAN
                                              © SAE SUBURBAN
                                              ® GM HIGHWAY
                                              
-------
hide performance potential, and driver habits  appear to  give  consistency in
speed, halts frequency, and idle time, provided the averaging  time  is  greater
than 6 to 7 minutes.  This represents the residence time  of a  vehicle  in an
area of high precursor emissions, which later results in  an air mass  of high
oxidant potential.

SIGNIFICANT DIFFERENCES IN ACCELERATION RATES

     Significant differences in the acceleration characteristic of  traffic
flow appear to exist within the overall constraints, and  these can  be expected
to be reflected in differences in N0x emissions.  To illustrate this,  the
bivariate distributions of frequencyxversus acceleration, and  velocity of our
Melbourne measurements and the LA (E.P.A. City) cycle can be compared in
Figure 5.  There is only a small difference in average speed between the two
driving patterns (3.5 mph), yet the HC and N0x values predicted by  STEM-CA
for these schedules give higher NO  for the Melbourne pattern, but  lower HC.
The HC are contrary to the expected change on the basis of the median lines of
Figure l(a) and (b), as can be seen from Table 2.

           TABLE 2  COMPARISON OF EMISSIONS OVER LA-4 AND MELBOURNE
                               DRIVING PATTERNS

LA-4 cycle
Melbourne
Difference
Expected difference
Error
v(mph)
19.6
16.1
3.5
HC(gm/ml)
7.60
6.23
+1.37(18%)
-0.92(12%)
2.29(30%)
N0x(gm/ml)
2.58
3.19
-0.61(24%)
-0.30(12%)
0.31(12%)
     Further comparison between accelerations
City data  (26) can be made in Figure 6.
in Melbourne and the five U.S.
     The above examination of traffic patterns and emission consequences
emphasizes the need to account for local patterns and intercity variations in
the emissions.  Effort may only be justified in predicting these variations
when the modelling of the dispersion and chemical reaction components of
ambient air quality forecasting becomes sufficiently precise to make the likely
errors in source prediction unacceptable.

     In this  regard, it is likely that only a small section of the road network
system needs  detailed study using the concept ,of acceleration noise (11).
In major Australian cities, about 501- of VMT occur on the arterial road network,
i.e., approx  7; of the total urban network system (27).

                    EMISSIONS DISTRIBUTION WITHIN URBAN AREAS

     Accounting only for the median effects of average speed on emissions
(i.e., ignoring acceleration influences) for cars, it is possible to compute
the distribution of the vehicular emissions source and its variance, knowing the
                                      978

-------
         FREQUENCY15
         %at 10km/h,
         0.5 m/s*     10
         intervals
                      5
_ ^
' . . )
\\
^- 33.4%
**^me*m
X. i • ,
	E.PA.CITY&A.D.R.27A
     MELBOURNE SURVEY
       LIMIT OF
       ACCELERATION
       BOUNDARIES
                              ACCELERATION
        Figure  5.   Bivariate  frequency versus acceleration and velocity
            distribution  of E.P.A. city  (and Australian design rule
            27A)  and Melbourne survey  (weighted according to total
                   VMT  on  CBD road, arterial and freeways).
speed and VMT distributions  as  in Figures  2  and 3.   It can  be  seen  clearly
from Figure 7 that the peak  HC  and N0x  sources  from motor vehicles  arise  in
city centres.  The emissions loading  on the  environment falls  to  one-half
its maximum value at less than  a 2 ml  radius from the  city  centre.   This
is contrary to the expected  constant  HC and  rising  NO  , emissions from  the total
traffic stream that can be expected in  saturated traffic conditions as  the
speeds rise away from the city  centre  on a unit length of road (27).
                              fa1!
                                     979

-------
     ACCELERATION
     mi/hs
        6
        U/' AVERAGE
           AXEL.
                                       AVERAGE ACCEL.
                                       PLUS 2 STD. DEVS.
                                          .	.  SPEED
               10     20     30	4P	5Q.___J50 mi/h
                                      AVERAGE DKEL
                                      MINUS 2 STD. DEVS.
                                                              5 CITY SURVEY
— MELBOURNE
    SURVEY
         Figure 6.  Comparison of acceleration  velocity  relations  for
             Melbourne survey (weighted)  and Scott  Laboratories--
                                5-city results.

                CONTROL MEASURES TO REDUCE OXIDANT PRECURSORS

     The city centre  will  be  the  highest  source of  oxidant precursors from
the engine  with a  HC/NOx  ratio of about 2.7  to 3 for a non-emission controlled
vehicle population (seexFigure 7),  and this  is the  region to which control
strategies  for mobile sources  must  be  directed.

     Control can be effected  by the following,  separately or  in combination:
exhaust emission limitation,  traffic access  restriction, and  traffic management.

     The driving pattern  adopted  in the U.S.  and Australia for automotive
emissions control  (LA-4 cycle) has  an  average speed greater than that experienced
in the central business areas of  many  large  cities  at peak hour, (usually 8-15
mph, LA-4 cycle 19.6 mph), as well  as  other  differences  pointed out above.
Consequently, the  emissions reductions  achieved over the cycle may not be
achieved under these low  speed conditions.   The results  are of great practical
significance, for  the a.m. peak hour speeds  are lowest in and around the CBD
(27).  In Sydney during the a.m.  peak,  some  60"' of  the VMT is less than 15 mph,
with 69X less than 12 mph concentrated in the area  around the CBD.

     We have made  appropriate modifications  to  STEM-CA to reproduce the NO
emissions reductions  that can be  expected in Australia by the adoption of D.S.
1973 FTP (6).  Industry has forecasted that  this will bring about  a reduction
                                     980

-------

kg/mi2h
80

60
w \J
HC/NOx RATIO
^-
-------
60


50


40


30
      o  20
      a
      ^  10
      x
      o
          0
             '.V
                               38% REDUCTION
                       8
  16          24
AVERAGE SPEED  mi
                                                32
40
      Figure 8.  Predicted acheived NO   emissions reductions from average
            vehicle over traffic survey speed .range in Melbourne
             giving 38% reduction over  ADR  27A  (U.S. 1973 FTP).


that as average  speeds  increase, sections 4 and  6  have  significantly less
emissions  than sections  3, 5 and 7.  Also shown  in  Figure  9  is  the emission
rate that  would  be  achieved at steady-state driving, which would be impracticable
to maintain at low  speed.  The reductions actually  compared  to  those possible are
detailed in Table 3.
         TABLE 3:  N0y EMISSION  REDUCTIONS ACHIEVED IN DRIVING COMPARED
                      TO BEST POSSIBLE BY STEADY-STATE DRIVING
Velocity
Average
mph
2
4
8
12
16
20
26
36
Best Pos.
Reduction
%
65
79
89
90
74
76
68
42
Reduction
Achieved
from Med.
%
9
28
58
60
50
Redn.not
Achieved
of Poss.
%
86
42
35
34
40
Median
gm/ml
3.3
2.9
3.8
4.2
3.2
2.8
3.1
3.3
Mln.
Poss.
gm/ml
1.2
0.6
0.4
0.4
0.52
0.68
1.0
1.9
Min. Achieved
Obs. Reduction
gm/ml gm/ml
3.0
2.1
1.6
1.7
1.6
0.3
0.8
2.2
2.5
1.6
                                    982

-------
       N0x6
       gm/ini
                                            SECTION
                                            NUMBER
                                               3
                                        STEADY SPEED
                                         EMISSIONS
                                                                 gm/ni
                         10           20
                         . AVERAGE SPEED
        Figure 9.  NO  and HC emissions rates  from average  vehicle  in
             saturated traffic for CBD sections.   Also shown  are
                           steady-speed emissions.


     The cause of the NO  reduction  achieved on different sections may be
seen in Figure 10, wherexthe  reduced speed  range  of  Section 6 compared to 3
can be seen from the cumulative  frequency versus  speed diagram, and the reduced
frequency of high accelerations  (and decelerations)  can also be observed.  These
results are not unexpected.

                         STRATEGIES  FOR TRAFFIC CONTROL

     Some of the basic strategies  for the coordinated  control of a large
number of intersections include  minimum accidents, minimum  number of stops,
minimum overall delay, and maximum vehicle  miles  per hour.  A typical traffic
control system based on the  CBD  will cover  5%  of  the road network, 30% of the
total city VMT and 75% of traffic  signals.

     From the above findings  we  can  postulate  a simple traffic control policy
in which trip times are maintained but not  improved.   To control NO , the
acceleration rates and acceleration  noise should  be  minimized.  This could
be achieved by queueing vehicles in  the roadways  around the CBD area and then
releasing platoons of vehicles from  the queue, allowing them to pass through
the central area at approximately  a  steady  speed  close to the speed limit
by sequence control of traffic lights.  There  is,  in many cities, sufficient
excess road capacity around  the  CBD  to store vehicles  prior to CBD area access
under controlled conditions.
                                     983

-------
                                                    SPEED
                        ,0
IUU
60
60
h 3—^7 SPEED, CUMULATIVE
. ,' .' FREQUENCY .
1 X-6 <\
/ (SECT. NO.] ' i
/ K
/ !P i
i- {' ACCEL. 4

40

20
n
FREQUENCY
"
. .... ...._!.
- \\
_¥/ 3 ^ 	
                                                  8
                              ACCELERATION  mi/hs
       Figure 10.   Cumulative  frequency versus instantaneous speed and
             frequency versus  acceleration for sections 3 and 6
                         and saturated traffic flow.

     There are two basic  methods  of  control for a coordinated system of traffic
lights:   the simultaneous  and  the alternate systems.  However, deviations
from the fixed time traffic signals  can lead to difficulties with traffic
streams  of unequal volume, which  frequently occur in CBD areas.  Although
not used so far for oxidant precursor control stratagies, these difficulties
have not been experienced  in Sydney  where a comprehensive traffic control
system utilizing computer  and  TV  technology has been developed progressively
since 1971 (30) and now covers  340 traffic lights.  In Melbourne, a computer-
controlled system has  been  investigated.  It would be comparatively inexpen-
sive ($1-2M), have a net  cost  benefit, and improve speeds by 5-10% (31).  It
should be possible to  devise a  computer-controlled system which would opti-
mize traffic flow for  minimum  emissions in the high source CBD areas.  This
technique has been recognized  by  the OECD (32).
         Q
         LU
         ui
         n_
         CO
                        TYRCAL EXISTING  C.B.D. PATTERN
                           QUEUE &  CRUISE
                                                             TIME
        Figure 11.   Proposed  change  to CBD traffic patterns to reduce
                           NO   emissions source.
                             X
                                    984

-------
     The typical resultant changes in driving
11 and the changes in HC and N0x are given in
applies to the section overall; the reduction
much greater  (66% NO  and 48% HC).
                        patterns are depicted in Figure
                        Table 4.  The 50% reduction
                        in the CBD component will be
              TABLE 4 :  EFFECT OF TRAFFIC PATTERN MODIFICATION ON
                              PRECURSOR EMISSIONS
Section
Length
ml


ml/h
Queue*
Length
ml
HC
Before
gm/ml
After
gm/ml
Reduc.
%
NO
Before
gm/ml

xAfter
gm/ml

Reduc.
%
               8
0.5    11.6
10.3
11
3.8   1.87
50
       2 lanes per lane of C.B.D. flow.
                                   CONCLUSIONS

     A vehicle emissions model has been developed which has been used in
conjunction with a study of traffic patterns for four city driving types.
The driving patterns are most closely explained by the halts per mile and
average speed.  Hydrocarbon emissions are closely related to average speed.
The influence of acceleration/deceleration rates and frequency is a second
order effect which can cause a ±30% additional variation in HC emissions.
The dependence of NO  on speed is small and acceleration variations can cause
±70% variation in emissions at the same speed.  To establish N0x inventories,
it is necessary to take into account the dependence on acceleration.  This is
the subject of continuing research using the acceleration noise approach.

     There is a strong relationship between traffic density per unit area
and the radial distance from the CBD.  The highest concentratration of pre-
cursors occur at the city centre and fall to about half the maximum value at
a two-mile radius.  Some 45 percent of the total city traffic emission origi-
nate within a four-mile radius.  For uncontrolled vehicles the HC/NOx ratio
is 2.7-3.1 over the whole city.

     There are significant differences in the acceleration rates between
our Melbourne measurements and the LA (EPA) City Cycle.  The average speed
is also lower, which could effect future vehicle emission control policy in
a city with a different traffic driving cycle than the LA 4 cycle.
     Traffic management as a primary N0x control strategy appears practical.
Steady-state driving would reduce NO  emission by 90%.  Reductions of 60%
have been achieved in the CBD area wfiere traffic speeds are very low--8-12 ml/h.
A queue and cruise system involving simple traffic signalization could accom-
plish this.

     Greater reductions in oxidant precursor emissions should be achieved
by computerized optimization of traffic flow, which would tend to increase
average speeds, smooth flow, and reduce acceleration rates and acceleration
noise.  Hydrocarbon emission from an individual vehicle would fall 20:,- by
                                     985

-------
increasing the average speed from 12 to 20 mph.
     The control of photochemical  oxidant precursors  by traffic management
is attractive because of the potential  community savings in reduced pollution,
coupled with reduced direct energy consumption,  increased safety,  and reduced
travel time.  This has been recognized  by the OECD.   Further reductions in
model vehicle emission standards would  be borne  by the whole motoring community
irrespective of local air quality needs.
     The results
model builders.
                 of this work are directly applicable to the photochemical
     1.


     2.



     3.


     4.


     5.



     6.

     7.


     8.

     9.



    10.


    11.

    12.
                       REFERENCES

Rose, A.H., McMichael, W.,   and  Kruse,   R.   Comparison
exhaust emissions in two major  cities,   J.A.P.C.A.  Vol
pp 362-371, 1966.
Patterson,  R.M.,  and  Myer,  E.L.    An approach
traffic to ambient carbon monoxide concentrations
                                                                  of auto
                                                                 ,  15,  8,
                                                            for  relating
                                                            at signalised
          intersections
          1975.
          Dabberdt, W.F.,  Sandys,  R.C
          emissions/dispersion  model
          Meeting,  A.P.C.A.  No.  75-44
               68th  Annual   Meeting,   A.P.C.A.   No.    75-44.4,
                                         D.A.   ISMAP-A traffic/
                                         sources,   68th  Annual
                           ., and Buder,
                            for indirect
                           ,3, 1975.
Patterson, R.M., and Mahoney, J.R.   Traffic modern  controls as
an  emission  control  technique, 67th Annual  Meeting, A.P.C.A.
No. 74-3, 1974.
Watson,   H.C.,   and   Milkins,   E.E.    Prediction   of   CO
concentrations   in  street  canyons.   Atmospheric  dispersion
modelling conference,  Dept.   Environment, Canberra, Australia,
August, 1976.
Australian  Government  Department
Design Rule 27A for Motor Vehicles,
Watson,  H.C.   The  influence  of
localised urban emissions  source.
1973.
Bulach,  V.  Traffic  patterns  and
M.Eng.Sci. Thesis, University of Melbourne,
Watson, H.C., Hamilton, R.B., and  Boburka,
of motor vehicle speed on gross emissions
                                              of  Transport.    Australian
                                              1974.
                                              driving  patterns  on   the
                                              S.A.E.  preprint no.  730556,
                                               motor
                                            vehicle  emissions.
                                            1976.
                                            G.W.   The influence
                                          of photochemical  smog
          components in
          of Aust.  and
          Loughnan,  M.
          patterns   in
          Report.,  Mech
          Pignataro,  L
          Prentice  Hall
          Johnson,  T.M.
          Measurement
              Sydney.   Smog '76 Workshop no.5,   Clear  Air Soc.
              .Z.,  paper 5.3,  1976.
                and Walls,  N.   A preliminary  survey of  driving
              Melbourne.   Unpublished   Final    Year  Research
              Eng.Dept., University  of Melbourne, 1973.
              J.   Traffic   Engineering  Theory  and  Practice.
             , 1973.   502  pp
             , Formenti, D.L., Gray, R.F.,  and  Peterson, W.C.
             of  motor  vehicle  operation  pertinent  to  fuel
          economy.   SAE preprint no.  760003,  197b.
                                     986

-------
13.   Swedish  Ministry of Transport.  (Communications  Group).    Air
      Pollution from  Motor  Vehicles.  Final report, Stencil 1971:1,
      1971.
14.   Committee on Motor Vehicles  Emissions.   Australian  Transport
      Advisory  Council.   Air Pollution and  the  motor  vehicle  in
      Australia - 1974 review, 1974.
15.   Kruse,  R.E.,  and Hill, D.M.  Exhaust emissions  from  compact
      cars.  S.A.E. preprint no. 670688, 1967.
16.   Huls, T.A.  Evaluation of the Federal Light Duty Mass Emissions
      Regulations.  S.A. E. preprint No. 730544, 1973.
17.   Dobinson, K.W.  Traffic flow - fact and fallacy, J.S.A.E.  Aust.
      July - August, 1975.
18.   Roth, P.M., Roberts, P.J.W., Liu, M.K., Reynolds, S,D., and
      Seinfeld, J.H. Mathematical inventory of pollution emissions.
      Atmospheric Environment.  Vol. 8 pp 97-130, 1974.
19.   Ministry of Transport N.S.W. Sydney area transportation study.
      Vol.1, 1974.
20.   Smeed, R.J. The capacity of urban road networks.  Aust, Road
      Research Board.  Proc. 5th Conference.  Vol.5, pt.l pp 10-28,
      1970.
21.   Owens, J.M. Traffic delays in urban areas.  Aust. Road
      Research Board.  Proc. 4th Conference Vol.4, pt.l, pp 453-474,
      1968.
22.   Munt, P.W. Slow grind: results of speed studies.  Greater
      London Intelligence Quarterly.  No.29, pp 38-47, 1974.
23.   Austin, T.C., and Hellman, K.H. Passenger car fuel economy as
      influence by trip length.  S.A.E. preprint no. 750004, 1975.
24.   S.A.E. Fuel Economy Measurement Procedures Task Force.  The
      devlopment of the new SAE Motor Vehicle fuel economy measure-
      ment procedures,  S.A.E.  preprint no.  750006, 1975.
25.   Pelensky, K. The Cost of Urban Car Travel.  Australian Road
      Research Board.  Report No.5, 1970.
26.   Vehicle Operations Survey Scott Research Laboratories Inc.
      San Bernado, California, December 17th, 1971.
27.   Hamilton, R.B. Some aspects of speed  and energy utilisation in
      urban road transport.  Institute of Fuel (Aust. memb.) Biennial
      Conf. 1976.
28.   Roberts, G.W.  What's the Future for  Automotive Emission
      Control.  International  Clear Air Conference.  Clear Air Soc.
      Aust. and N.Z., Vol.1, pp 212-233, 1975.
29.   Better Towns With Less Traffic.  O.E.C.D.  Paris, April, 1975.
30.   Camkin, H.L., and Sims,  A.G. Co-ordinated Traffic Signals.
      Traffic Engineering Branch.  Department of Motor Transport
      N.S.W., 1974.
31.   Victorian State Government.  The Development of a Metropolitan
      Traffic Signal System for Melbourne.   Latrobe State  Library,
      Juno, 1969'.
32.   Photochemical Oxidant Air Pollution.   A report of the Air
      Management Sector Group on The Problems of Photochemical
      Oxidants and their Precursors in the  Atmosphere, OECD Paris
      1975.
                               987

-------
                                                                               20-3
                      CONTROL OF VEHICLE REFUELING  EMISSIONS

                       A.  M. HOCHHAUSER AND L. S. BERNSTEIN*
ABSTRACT

      A fauA-day te^t o& the. Exxon  Company U.S.A.  mani.falde.d vapoA  balance.
        fan. the. control oft ve.hi.de.  fie.faili.nQ' e.miA&ionA wat> caAAie.d  out at a
         station in  Ltnden, New) 3&ue.y.   l/apo/i ncoveAy nozzles weAe, uAe.d on
ati  &ix active. diApe.nAe.u at the. t>tati.onA.  POUA  o&  the. Aix diApe.nAeAA
int>tAume.nte.d to caAfiy out. the, le.ak-Aate. and baseline. method* o& me.oAuAi.ng
Ae. j{ ueJting
           the. combined ie.ak- note.- bat, eJU.ne. method, we. obtaine.d vaLid
me.at>uA.e.me.ntA on 45S  o& the. 764 cau  on whtch testing MAU atte.mpte.d.   On 77
COAA,  no &e.at wa& cb.ftu.ned at the. nozz£e-^xX£.necfe ^.n^eA^ace; S oft  th&Ae.
couid  be. fiueJLe.d onty M^Lth hand-heJtd  opeAation.  On  an add^tionai  14 c.au,  a
Ae.al WCL6 obta4,ne.d, but hand-heZd operation wai> fie.qaJJie.d.  Fofity-Yiine. teJ>&>
had to be. du>c.a>ide.d  because 0)J tnc-omplzte. data, and on 179 caA-6 a  &e.al waj>
obta4.ne.d,  but e.mibAi.on me.aAutie.me.ntA  could not be. made, because the.  dntveJi ne.-
       to coopeAate. \tiLtk QUA tej>t.
       The. majox. n.e.aboni>  fan. ^a^ituAe.  to make, a Ae.at  at the. nozzte.-^-Lttne.c.k
           uieAe.'-  ob&tAucJtion by the.  Li(ie.n&e. plate, oti bumpeA, no ^LnneA Lip
on the. &-lltne.ck, too  t>haAp a bend i.n the. ^xX£,nec.fe,  and COAA on whic.h the.
te.chni.CA.an am unawoAe. that a t>e.al had not been made.   We estimate, that
appAoxAjnately  tm-thiAdt> o{) the. cati on which a 4ea£ u)O4 not made, could
have,  been Abated i.fi a no Ae.al-no  loM nozzle
      Tfie baAe2A.ne.  pontion o^ the.  te^t pfioaduAe. i,n.di.cate.d average uncontrolled
           ^nom the. ve.hi.cle. sample. o£ 4.J5 g/gal, i.n good agAe.e.me.nt with otheA
btu.di.eA.   Without  the. no &e.al-no  falow nozz£e, we me.aAuAe.d 87% e.miAAionA con-
tAot i& all COAA weAe. c.onAideAe.d,  and S&% contAol  i.& caAA fLe.quiAi.yiQ hand-heJ-d
ope.fuzti.on weAe. excluded according to the. EnviAonme.ntal VAote.cti.on Agency '4
pfiopo&e.d AuleA.  We e^^xma^e that uAe. o& a no &e.al-no falow nozzle, would have.
pAovide.d 93% contAol o
                                    INTRODUCTION

 VEHICLE REFUELING EMISSIONS

	When  the  fuel  tank of an automobile  is  partially filled with gasoline,
*Exxon  Research and Engineering Company,  Linden, New  Jersey.

                                        989

-------
the vapor space above the liquid contains a mixture of hydrocarbons (HC) and
air.  During the refueling operation, as liquid is pumped into the tank,
vapor is forced out in approximately equal  volumes.  These vapors, referred
to as vehicle refueling emissions, have a magnitude of approximately 3 to 5 g
HC/gal.  of dispensed fuel.  Exact emission rate is a complex function of the
temperatures and Reid Vapor Pressures (RVP) of the dispensed fuel  and fuel in
the vehicle tank.  Earlier work(l) at Exxon Research showed that the ratio of
vapor generated to liquid dispensed could be calculated:

          ^.  =  1  +  (AT)(1.02  x  10"2  -2.06  x 10-\ -  8.88 x lO'"  RVP)
           L
where V = Volume of vapor generated,

      L = Volume of liquid dispensed,

     AT = Tt - Td, °F,

     T  = Tank fuel  temperature, °F,

     Tj = Dispensed fuel temperature, °F,

    RVP = Reid Vapor Pressure of the fuel.

Since engine heat generated during vehicle operation usually results in fuel
tank temperatures above dispensed fuel temperature, the volume of vapor gen-
erated will usually be smaller than the volume of liquid dispensed, a situa-
tion referred to as vapor shrinkage.  If dispensed fuel temperature is higher
than tank fuel temperature, the volume of vapor will be greater than the vol-
ume of liquid dispensed, a situation referred to as vapor growth.

VAPOR BALANCE SYSTEMS FOR THE
CONTROL OF VEHICLE REFUELING EMISSIONS

      Vapor balance systems  (Figure  1) are  the  simplest,  most  cost  effective
method of controlling vehicle refueling  emissions.  They  work  by  routing  the
vapors displaced  from the vehicle  tank through  a  vapor return  line  to  the
underground tank  from which the fuel was pumped.   These  systems require  a
tight seal at the nozzle-vehicle tank interface.  If a tight seal  is main-
tained,  and there is no vapor growth, vapor balance systems should give zero
emissions.  In cases of vapor growth, excess vapor volume can be vented
through the normal underground tank vent.

     There are two types of vapor balance systems, manifolded and unman-
ifolded systems.  In unmanifolded systems,  the Vapor from each dispenser is
routed back to the underground tank from which its fuel was dispensed.  In
manifolded systems,  vapors from all dispensers at a service station are mani-
folded before being routed to the underground tanks, which are also manifold-
ed.  Manifolded systems have the ability to balance vapor growth and vapor
shrinkage when they occur simultaneously.

MEASUREMENT OF VEHICLE REFUELING EMISSIONS
                                     990

-------
                            Liquid Fuel Line
            Vent
                                             Vapor Tight Nozzle
                                        -*- Underground Tank
                        Figure  1.   Vapor  balance  system.

     There are two sources of potential emissions from vapor balance systems:
emissions from the underground tank vent, and leaks at the nozzle-fillneck
interface.  Measurement of emissions at the tank vent is conceptually simple.
The emitted hydrocarbon can be collected on a carbon trap, or the volume and
HC concentration of vent emissions can be determined.  Measurement of leaks
at the nozzle-fillneck interface is more difficult because the leak rate is
a sensitive function of the pressure drop in the system.  Increasing or de-
creasing^pressure drop will change the vapor flow characteristics of the sys-
tem andmay lead to erroneous answers.  Three general approaches to measuring
leaks from vapor balance systems have been developed:  the leak-rate method(2),
the baseline method(3), and the vacuum collecting tube method(4).
      The leak-rate method  involves calibration of  the  leak  actually  occurring
during a refueling operation for each vehicle.  The  procedure  involves  two
steps:  calibration and measurement.

     •    Calibration:
          The dispensing nozzle is latched in place,
using appropriate valving, HC vapor is pumped into the
 at low pressure less than 1" of H?0.   If flow can be
        a leak is present and HC vapor is pumped into
                                                 leak-
          and
          tank
          maintained,
          the tank at three different pressures, allowing the
          rate" curve of volume of leak (in cfm) vs. AP (in inches
          of H?0) to be determined.

          Measurement:   During refueling, the pressure at the
          nozzle-fillneck interface and the concentration of HC
          in the vapor return line are recorded.  The volume of
          leak is calculated from the curve developed in the cali-
                                     991

-------
           bration  step.  Assuming that the concentration of HC es-
           caping through the  leak is the same as the concentration
           of HC  in the  vapor  return line, refueling emissions can
           be calculated as  follows:
                                      3
 Emissions,  g/gal  =
                    Volume  of  leak,  ft /min x HC concentration, g/t't
                              Fuel  dispensing rate, gal/min
 The  major  limitation  of  the  leak-rate method  is that it cannot be used when
 no  seal  is  made  at  the nozzle-fillneck  interface and gives poor results when
 large  leaks are  present.

     The baseline method determines  emissions as the difference between po-
 tential  emissions and the amount  of  vapor  recovered.  Prior to refueling,
 the  temperature  of  the fuel  in  the vehicle tank (T^) is measured.  The aver-
 age  temperature  of  the dispensed  fuel (T^j is determined during refueling.
 The  difference between the two  temperatures is defined as AT  (T^-Tj).  During
 refueling,  the amount of vapor  returned  is measured with a gas meter and HC
 analyzer.   An explosimeter or leak  test is used  to determine  whether Teaks
 are occurring at the  nozzle-fillneck interface.  Those vehicles for  which  no
 leaks  are  detected  are used  as  the  "baseline" population.  The amount of va-
 por returned with these  cars is plotted  against AT, and the resulting cor-
 relation is taken to  represent  the amount  of  vapor which would result if no
 control  were used.  The  emissions for cars  in which leaks are found  are cal-
 culated  as the difference between estimated uncontrolled emissions  (from the
 correlation with AT)  and the measured amount  of vapor returned.

     The vacuum  collecting tube method is currently under development by the
 Environmental Protection  Agency.  This procedure uses an open-ended sleeve
 that fits  loosely around  the  nozzle-fillneck interface.  Air  and any hydro-
 carbon leaking from the  interface are drawn through a series  of perforations
 on the inner  face of  the  sleeve at a rate of 2-5 cfm.   The volume and hydro-
 carbon content of the material  collected by the sleeve are determined and
 used to  calculate emission rate.  When fully developed, this method should
 provide  a  simple, more direct approach to measurement of these emissions.   We
 recommend  that EPA make  every effort to  replace the more complex methods des-
 cribed above  with this easier method.

                                  EXPERIMENTAL

       Vehicle refueling  emission  control technology is currently  in  a state
 of  flux.   New hardware and new  test  procedures are being introduced  at a
 rapid  rate.   We  therefore view  the test  described  in this paper as  indica-
 tive of  the state-of-the-art when the experiment was designed, rather than
 a definitive  measure  of  the  capabilities of the vapor balance system for
 control  of  refueling  emissions.

     The test was carried out using  the  Exxon  Company  U.S.A.  manifolded  vapor
balance system(5) at the  Research Exxon  Service  Station  located  at the  south-
west corner of the intersection  of U.S.  1 and  Park  Avenue  in  Linden,  New Jer-
sey.   This  station was chosen because of its  high  volume  (>100,DOO gal/month)
                                     992

-------
and its proximity to the Exxon Research Center and its facilities.   Emission
measurements we^e made using both the leak-rate and baseline methods.   We
learned of EPA's efforts to develop the vacuum collecting tube method  after
the completion of the test.  Vent emissions were collected in a carbon trap
and determined by weight.

     The test was run for  four consecutive days in April, 1976, from 6:00 a.m.
to 10:00 p.m.  Refueling, all test procedures, and data logging were carried
out by Exxon Research technicians.  The service station normally operates six
active dispensers, two each on leaded regular, leaded premium, and unleaded
gasoline.  All six dispensers were equipped with vapor recovery nozzles, but
only four of the six dispensers were equipped with the instrumentation neces-
sary for carrying out emission measurements.  One of the leaded premium and
one of the unleaded regular dispensers were left uninstrumented because of
manpower limitations.  The test as designed required 12 people on each of two
8 hr.  shifts, a total of 24 people.

      Measurements taken  included:

      •    Vehicle tank  fuel,  dispensed fuel,  and  returned vapor
          temperatures;

      •    Pressure  at the  nozzle-fillneck  interface and  ambient pressure;

      •    Volume  of:  (a)  vapor  flowing during  the  leak-rate  calibration,
          (b)  the vapor  returned  to the underground tank, and  (c) the
          fuel  dispensed;  and

      •    HC  concentration of the  returned  vapor.

We also  recorded  the date  and time of the  test, vehicle  identification  infor-
mation,  the  number  of miles  driven since the  start  of  the driver's trip,
whether  there  was_spill  or spitback, and whether  the fuel tank was partially
or totally  fueled.

      The measurement of  the  volume of vapor returned to  the underground tank
necessitated  placing a dry test meter in the  vapor  return line.  This raised
the pressure  drop between  the nozzle and the  underground tank  from 0.03 to
0.14  inches  of H20.  The results obtained  by  the  leak-rate method were  cor-
rected for  the increase  in pressure drop created  by the  dry test meter.
                                    RESULTS

LEAK-RATE METHOD

     Over the course of the four-day test, measurements were attempted on
764 light duty vehicles refueled at the station.   These vehicles can be
broken up into the following categories for the leak-rate portion of the
test procedure.


                                     993

-------
     •    No leak - A total  of 242 cars were determined to have
          no leak, based on  the fact that they showed vapor tight
          seals during the leak-rate calibration and a measurable
          amount of vapor was returned to the underground tank.

     •    Measurable leaks -• There were 203 cars on which the
          nozzle did not form a tight seal  and for which a leak
          was measured by the leak-rate procedure.

     »    Incomplete data -  Data on 49 cars, had to be discarded
          for reasons including instrument failure and technician
          error.

     •    No fit - There were 77 cars on which no seal was made
          at the fillneck and no vapor returned to the underground
          tank.  This category included tests in which the vehicle's
          driver refused to  participate in the test, but in which
          the technician noted that a seal  was not made.

     •    No latch - There were 14 cars on which refueling could
          be accomplished only by hand-holding the nozzle, but on
          which some vapor was returned to the underground tank.
          Data obtained on these cars could not be analyzed by the
          leak-rate method,  but could be analyzed by the baseline
          method.

     •    Refusals - 179 drovers refused to participate in the test.

      Emissions from the 445 no-leak and measurable-leak cars averaged 0.23
g/gal before correction for the increased pressure drop caused by the dry
test meter.   Correction for this increased pressure reduced emission to
0.15 g/gal.   The corrected figure will be used for all further discussion of
the data.
     The 77 no-fit cars could not be fit for the following reasons.

     •    21 cars had no inner lip on the fillneck.  The vapor
          recovery nozzle makes a seal by latching a ferrule under
          the lip of the fillneck and forcing the face of a rubber
          boot against the face of the fillneck.  With no inner
          lip, no seal can be made.  All of the cars in this
          category were improts manufactured by Nissan, Fiat,
          Peugeot, BLMC, and Volkswagen.

     •    10 cars had a sharp 90° bend in the fillneck which
          prevented the nozzle tip from entering far enough to
          engage the latch ferrule.  Ford vans were the most common
          vehicle with this problem, but the problem was also en-
          countered with a Dodge van, and cars manufactured by Fiat
          and Toyo Kogyo.

     •    34 cars had obstructions from either the bumper or the license

                                     994

-------
          plate preventing the nozzle from properly sealing.
          This category contained a wide variety of U.S. and imported
          cars.

     •    10 cars had no apparent leak during the leak-rate cali-
          bration but were classified as no-fit when it was discovered
          that no vapor was returned to the underground tank.

     •    1 car was reported to have too narrow a fillneck to be
          sealed by the nozzle.

     •    1 car had severe rear-end damage.

Hand-held operation was required on 8 of the 77 no-fit cars.

BASELINE METHOD

     Of the 242 no-leak tests reported in the leak-rate method, 125 could
be used for the baseline regression equation.  The rest of the tests had
missing data -- usually initial tank temperature-- which is an extremely
difficult measurement to make.  It was possible to determine baseline em-
issions on 142 of the 343 refuel ings on which vapor tight seals were not
obtained.  Here again the data most often missing was initial tank tempera-
ture.  No-fit and no-latch refuelings were included in the 343 cars with
leaks.
     The baseline regression was determined to be:

          g HC/gal  = 4.04 - 0.046 AT

This equation explained only  13% of the observed  variance and  had  a  standard
error of estimate of  1.10, not  significantly different  from  the  standard
deviation of the raw  data, 1.17.  Regressions against tank  fuel  temperature,
dispensed fuel temperature, and combinations of these parameters were also
attempted, but none were better than the correlation with   AT.

      Average AT for all  valid tests was -2.2°F with a standard deviation of
9.3°F.   Average potential  emissions was 4.15 g/gal.   Average leak for the
267 no-leak and measurable leak  cars on which valid baseline tests were ob-
tained was 0.15 + 0.21 g/gal.  The baseline correlation estimate of emissions
should be compared to the uncorrected leak-rate method estimate of emissions,
0.23 g/gal.  While it was possible to correct the leak-rate method for the
higher pressure drop  caused by the dry test meter, there was no way to correct
the baseline procedure.  Given the high degree of uncertainty around the base-
line estimate and the much smaller number of valid tests used in the baseline
estimate, the agreement between the two methods was good.
 COMBINATION OF LEAK-RATE AND BASELINE METHODS

     During the development of our test program we learned that EPA was
 considering a combined leak-rate/baseline method which would involve using


                                     995

-------
 the  leak-rate  results for  small leaks and baseline results for large Teaks".
 We also  analyzed  the data  with this approach using the following criteria.

      •     If the  emissions calculated by the leak-rate method
           are  less than  1.0 g HC/gal, use these emissions.

      •     If the  emissions calculated by the leak-rate test are
           more than 1.0  g  HC/gal, use the baseline emissions.

      a     For  cars in the  no-fit or no-latch category, use baseline
           emissions, if  available.

      »     If leak-rate data are missing, use baseline emissions, and
           if baseline emissions are missing, use the leak-rate method.

 Using these criteria, data were available on 458 cars.  Average emissions for
 these cars were 0.12 g/gal.

 UNDERGROUND TANK  VENT EMISSIONS

      Vent  emissions were measured daily and averaged 0.013 g/gal.

 TOTAL EMISSIONS USING THE
 DIFFERENT  ESTIMATING METHODS

      In  estimating total emissions, we are concerned not only with the vehicles
 for  which  valid and complete test data were obtained, but with the vehicles
 in the other categories:   incomplete data, no-fit, no-latch, and refusals.   It
 should be  remembered that  vehicles on which a tight fit could not be made were
 considered no-fits, whether or not the driver agreed to participate in the
 test.  Total emissions are the sum of emissions at the nozzle-filltank inter-
 face and the 0.013 g/gal emitted at the vent.

     For the leak-rate method,  the best estimate of the emissions from the
673 cars in the no-leak, measurable-leak,  incomplete data,  and refusal  cate-
gories, are the emissions from the 445 cars on which valid and complete data
were obtained.   Total  emissions for these cars averaged 0.16 g/gal.   The best
estimate for the  14 no-latch cars was  also determined in the baseline procedure
and found to be 0.58 g/gal.  The average of these values is 0.58 g/gal.  If we
exclude the 22 vehicles which required hand-held operation (8 of the no-fits
and  all  14 no-latch vehicles), average emission is estimated as 0.54 g/gal.
     For the baseline method, the best estimate of emissions for the 673 cars
in the no-leak, measurable-leak, incomplete data, and refusal  categories is
the 0.16 g/gal found on the 267 cars for which valid and complete data were
obtained.  The best estimate of emissions for the 77 no-fit vehicles is the
4.23 g/gal found on the six no-fit cars for which valid baseline estimates
were obtained.  The best estimate of emissions for the 14 no-latch cars is
the 0.58 g/gal found on the three no-latch cars for which valid baseline tests
were conducted.  The average for the baseline procedure is 0.58 g/gal.  Ex-
cluding hand-held operation from this calculation yields an average of 0.54
g/gal.
                                     996

-------
      For the combination method, the best estimate of emissions for the 673
cars in the no-leak, measurable-Teak, incomplete data, and refusal categories
is the 0.13 g/gal  found for the 458 cars on which valid and complete  test
data were available.  For the no-fit and no-latch cars, the baseline  values
of 4.23 and 0.58 g/gal  were used.  This results in an average of 0.55 g/gal
with hand-held operation and 0.51 g/gal  without hand-held operation.

     The bulk of the emissions in all cases comes from the 77 no-fit vehicles.
Since the estimates of emissions from these vehicles  (4.15 vs. 4.22 g/gal)
were very similar,  it is not surprising that the totals estimated by the three
different methods agree so well.

      If we assume  potential emissions  of  4.15 g/gal,  overall system efficiency
can be calculated to be 86.1-86.8%  if hand-held refuelings are included, and
87.0-87.7% if hand-held operations  are  excluded.  These values are quite close
to the 90% recovery envisioned in EPA's regulations.

     A comparison of the results obtained with the various test methods is
contained in Table  1.  We have also included an estimate of the potential con-
trol obtainable with a no seal-no flow  nozzle, that is, a nozzle with an in-
terlock that prevents fuel flow  unless  a seal is made  between the nozzle and
the fillneck.  The  basis for this estimate is presented in the next section of
the paper.
          TABLE 1.  TOTAL REFUELING EMISSIONS BY DIFFERENT METHODS
                 Method
   Estimated
Total Emissions*
^Control
  Leak-Rate

  Baseline

  Combination

  Estimated performance with a
    successful no seal-no flow nozzle
   0.58 g/gal

   0.58

   0.54


0.29-0.31
   86

   86

   87


   93
    Includes hand-held operation.
                                  DISCUSSION
IMPROVED SYSTEM EFFICIENCY
     Approximately 2% of the total emissions determined by the combined leak-
rate-baseline method were underground tank vent emissions, 257. were from cars
on which a seal could be made, and the remaining 72% were from cars on which
no seal was made--the no-fit cars.  Significant improvements in system effi-
                                     997

-------
ciency can be made only if the number of no-fits can be substantially reduced.

     The analysis of the reasons for no-fits presented above showed that all
but two of the 69 cars that were no-fits, and did not require hand-held re-
fueling, fit into one of four categories:  21 had no inner lip on the fillneck,
8 had too sharp a bend in the fillneck, on 7 the vapor line did not open,
even though it appeared that a good seal had been made, and on 31 the fill--
neck was obstructed by either the bumper or the license plate.  One of the
remaining two cars had rear-end damage, and on the other, the fillneck was
too narrow.   Below, we will consider the potential improvement in system ef-
"ic'ier.cy whicn could be afforded by a no seal-no flow nozzle.  We estimate
thai a satisfactory no seal-no flow nozzle could have provided a seal on 44
of the 69 no-fit cars not requiring hand-held operation in this test.  We
assume that all 21 of the cars with no inner lip on the fill neck, all 7 of
the cars on which the vapor return line did not open, and half (15) of the
31 cars on which the license plate or bumper wan an obstruction could have been
sealed in this fashion.  The cars with too sharp a bend in the fillneck and
the remainder of the cars with obstructions would probably not be sealed by
the no seal-no flow nozzle.  For these cars, some sort of manual  override of
the no seal-no flow device would have to be provided.

     Use of a no seal-no flow nozzle would mean that all of the 69 cars
which were previously no-fit cars not requiring hand-held refueling would
now require hand-held operation.  The cars on which a seal could be made
would require hand-held operation to maintain the seal, and the cars on which
a seal could not be made would require hand-held operation to override the
no-seal-no flow device.  If hand-held refuel ings were excluded from the data
analysis, all of these tests would be dropped from consideration.  Using this
criterion, the best estimate of total emissions from the remaining cars
would be the 0.13 g/gal found for the 458 rio-leak and measurable-leak cars,
using the combined leak-rate-baseline method.  This is equivalent to a
system efficiency of 96.8%.

     If hand-held refuel ings are included in the data analysis, emissions from
all  764 cars would have to be considered, and an additional  5 cars would have
to be added to the group sealed by the no seal-no flow nozzle.  The best
estimate of total emissions from the 673 no-leak, measurable-leak, incomplete
data, and refusal cars is the 0.13 g/gal found with the combined leak-rate-
baseline method.   The best estimate of the total emissions on the 49 cars for
which the no seal-no flow nozzle would require a seal to be made is between
0.13 and 0.58 g/gal.   The lower figure assumes that a seal as good as was
made on the cars sealed in this test can be made.  The higher figure assumes
that the seal will be no better than was made on the no-latch cars,,   The best
estimates of total emissions from the 14 no-latch cars and the 28 remaining
no-fit cars are 0.58 and 4.15 g/gal, respectively.

     Including hand-held refuelings and assuming total emissions of 0.13 g/gal
for cars on which the no seal-no flow nozzle requires a seal to be made re-
sults in average emissions of 0.29 g/gal, equivalent to a system efficiency
of 93.0%.  Assuming the higher value of emissions for these cars, 0.58 g/gal,
results in average emissions of 0.31 g/gal, equivalent to a system efficiency
of 92.5%.  In all of the cases we have considered, a successful no seal-no


                                      998

-------
flow nozzle could lower total  emissions from the system to below 0.4 g/gal  and
raise system efficiency to above 90%.

ADDITIONAL WORK

     Tests of prototype no seal-no flow nozzles and prototype hybrid vapor
recovery systems are planned for the near future.
                                   REFERENCES


   1.    Hochhauser,  A.  M.  and  Campion,  R.  J.    An  Experimental  Study
        of Vehicle  Refueling Emissions.   SAE Paper 760307,  February,  1976.

   2.    Listen,  E.  M.    Determination of  Hydrocarbon  Vapor  Losses  from
        Vehicle  Fuel  Tanks During  Refueling Using  a Leak-Rate  Procedure.
        API Report  CEA-22,  June, 1975.

   3.    Federal  Register.  40  (197):47666-47685,  October  9,  1975.

   4.    Principe, P.   Vapor Recovery Test Procedure.   Draft of EPA internal
        report,  May,  1976.

   5.    Comments of Exxon  Company,  U.S.A.  on Stage II Gasoline Vapor
        Recovery-Proposed  Regulations.   Submitted  to  the U.S.  EPA,
        November 24,  1975.
                                     999

-------
                SESSION 21
CONTROL OF OXIDANT PRECURSOR EMISSIONS - II

          Ckcuxman:  R.W.  Bilger
      University of Sydney, Australia
                   1001

-------
                                                                           21-1
                  NOV CONTROL TECHNOLOGY FOR STATIONARY SOURCES
                    A

                          G. B. Martin and J. S. Bowen*


ABSTRACT

     The. Envtn.onme.ntal.  Protection  Agency'A  Industrial Envtronme.ntal Rese.an.ch
Laboratory, Rese,arch Triangle.  Park,  Month Carolina,  has an e.zte.nslve. program
fior control ofi oxides o& nitrogen  ^rom stationary  combustion source^ by modi-
fication o£> equipment design and opiating  conditions.  This paper Mill briefly
describe the levels o{]  emission contAol that have  been achieved  on practical
combustion equipment.   It wWl than  describe -6eveAa€ ofa the. Env^Jionmantat
Ptiotaction Agenct/ pftOQfiam activ-itidA  fati de.vdopme.nt o& combustion tuchnotogy
capable, ofa Atgnsifitcantty enhanced  control ofa oxxcde-6  oft nitrogen.

                                   INTRODUCTION

     One of the major classes  of pollutants  emitted  into  the atmosphere from
man's activity is identified by the  term nitrogen  oxides  (N0x).   Approximately
99 percent of this NO   is generated  from combustion  of fuelsxin  a variety of
equipment types and tfie emissions  are concentrated in  population  centers.   The
evidence indicates that for most equipment  types,  95 percent of  the N0x is
emitted from the source as nitric  oxide (NO) with  the  balance  being nitrogen
dioxide (N02); however, recent results indicate  that for  certain  sources (e.g.,
gas turbines) the percentage of N0?  may be  significantly  higher.   In any
event, the total NO  emitted undergoes a complex scheme of atmospheric  photo-
chemical reactions with hydrocarbons  (HC) and sulfur oxides (SO  ).   These
reactions result in the formation  of  undesirable secondary species, such as
ozone (03) and nitrates, and in a  shift of  the  residual NO  toward N02.  The
recognition of the adverse effects of N02 and other  atmospheric  pollutants
on human health and welfare led to passage  of the  Clean Air Act  of 1970.  As
a result of this act the EPA was empowered:  (a)to establish a National Ambient
Air Quality Standard (NAAQS) for N0,;(b) to require  a  90  percent  reduction  of
NO  from the automobile; (c) to establish New Source  Performance  Standards
(N§PS) for stationary equipment; and(d) to  set  up  mechanisms to  insure  compliance.

     At the time the Clean Air Act was passed,  mobile  and stationary sources
each contributed approximately 50  percent of the N0x and  the major control
strategy was reduction  of mobile source emissions  by 90 percent  to 0.4  gm
of N0:,. per mile (0.25 mg of NO-, per  meter).  NSPS  were provided  to reduce
growth of NO  emissions from stationary sources.   Since that time a number
of factors have emerged which  appear  to require  some modification of this
*Environmental Protection Agency, Research Triangle Park, North  Carolina,


                                     1003

-------
approach to NO  control.   First, development of economic control  technology
to meet the automobile emission goal  has proved elusive.  Second, the energy
crisis has provided strong pressure for use of coal  in utility boilers with an
attendant increase in NO  emissions.   These factors  and others have resulted
in a greater emphasis onxNOx control  for stationary  sources and in a recom-
mendation for a Maximum Stationary Source Technology (MSST) program for control
technology development (1).

     There are a number of potential  approaches to control of MO  from sta-
tionary combustion sources:

     (1)  Combustion modification, which is based on alteration of combustion
          conditions to minimize formation of NO , is potentially applicable
          to all types of combustion  sources botft new and existing.

     (2)  Flue gas treatment, which involves an add-on device to remove the
          NO  from the flue  gas, is currently envisioned as a method to supple-
          ment combustion modification where a high  degree of control is required.

     (3)  Advanced alternate fuels and combustion processes may have properties
          leading to inherently low levels of N0x.

     (4)  Fluid bed combustion, which is predominantly an SO  control techni-
          que, has favorable characteristics for N0x reduction, especially
          with application of combustion modification principles.

The purpose of this paper is to describe the Combustion Modification program
as related to items (1) and  (3) above.  Flue gas treatment and fluid bed
combustion programs have been described elsewhere.

                                   BACKGROUND

     The combustion modification approach is based on the premise that N0x
can be reduced substantially by alteration of the conditions under which £he
fuel burns.  There are two predominant mechanisms for the formation of N0x
during combustion.  Thermal  NO  is formed by fixation of molecular nitrogen
from the combustion air through a series of reactions which are exponentially
dependent on temperature and slightly dependent on oxygen availability.  Fuel
NO  is formed by oxidation of organic nitrogen compounds contained in the fuel
through a series of reactions which are relatively independent of temperature
and strongly influenced by oxygen availability.  For any fuel containing bound
nitrogen both mechanisms contribute to the total NO   formation.

     Working from this knowledge of the NO  formation chemistry, a number of
control technology approaches have been formulated.   Techniques for control of
thermal N0x  are based on reducing the peak temperatures in the combustion zone,
and include staged combustion, low excess air, flue  gas recirculation, and
water injection.  Techniques for control of fuel NO. are based on reduction of
oxygen availability in the combustion zone, and inctude low excess air opera-
tion and staged combustion.   In addition, pilot scale combustion studies have
shown that changes in burner design can also significantly reduce the forma-
tion of both thermal and fuel N0v by  aerodynamically influencing local recircu-

                                    1004

-------
lation rates and/or oxygen availability in the flame.   The optimum level  of
control may require a combination of these approaches.   Although the control
techniques have shown good potential in experimental  systems and on some
practical equipment, the optimum levels achievable have yet to be established
and are the subject of this program.  In the application of N0x control  tech-
nology to practical equipment, care must be taken to minimize potential  adverse
side effects.  These include:  increases of other pollutants, loss of system
efficiency,  and system  operability  problems.  Based on  current  information,
emissions  of other pollutants  (e.g., carbon monoxide and  other  products of
incomplete  combustion)  can be  maintained  at low  levels  while  achieving signi-
ficant NO   reduction by  proper system  design.  Many of  the  NO   control techni-
ques,  such"  as  low  excess  air operation, even  offer potential  for  increases of
system efficiency.   Based on limited experience  with long  term  service it
appears  that operability  problems,  including  fireside  corrosion,  can be avoided
by  proper  design.

                          COMBUSTION MODIFICATION PROGRAM

      The NO  control technology  development program was initiated in 1969 at
the conclusion of  a systems  study  conducted under contract  by  Esso Research
and Engineering (2).  The system study  provided  an emission  inventory by source
category and fuel  usage;  in  addition,  it  reviewed the  available technology
and development potential  of various control  approaches and  concluded that the
combustion  modification  techniques  to  reduce  N0x formation  offered the most
cost effective potential.  Based on this  study and the  available  resources,  a
Combustion  Modification  R&D  program was structured to  develop  and apply con-
trol  technology for the  equipment  type  and  fuel  combinations  which constituted
the major  stationary sources of  NO  .   These major sources  are  shown in Table 1
based on a  recent  updated emissionxinventory  (3).  In  addition, the N0x con-
trol  technology developed was  to be capable of minimizing  undesirable side
effects  of  the technology, and maintaining  or improving process energy ef-
ficiency.   Ongoing developments  in  the  private sector  (e.g.,  gas- and oil-
fired utility  boilers)  were  considered  to insure complementary  activity where
appropriate  and to avoid  duplication where  significant  private  effort exists.
Details  of  the program  have  been documented in detail  in  two  recent papers
(4,  5).

      This  basic philosophy has been followed, although  in  the  intervening
years  the  program  emphasis has been expanded  and shifted  to  account for recent
trends.  As  an example,  in response to  the  recent increase  in  emphasis on
stationary  source  N0x control  and  the  Clean Air  Act requirement for demon-
strated  technology as the basis  for any NSPS, Industrial  Environmental Research
Laboratory-Research Triangle Park  (IERL-RTP)  has Aerotherm  working under
contract to  establish the requirements  for  a  Maximum Stationary Source
Technology  (MSST)  program.   The  inventory identified 137  combinations of
equipment  type and fuel  that account for  the  stationary source  N0x of which  38
contribute  90  percent (3).   While  this  appears to be a  formidablexnumber of
combinations,  there are  many common features  of  the systems  which allow
general  application of  the technology  and,  therefore,  a manageable development
effort.  A  report  containing the final  recommended program  and  funding levels
is  currently in preparation.
                                     1005

-------









CM
r^
CTl
' —
"* — •*

CO
UJ
01
ZD
O
co

CtL
LU
CO
— 1
	 1
UJ
— 1
u.

^»"
o
o;
u_

CO
^^
o
1 — 1
CO
CO
t— <
>
UJ

X
o
~zz.

_j

>
•J-;
re
^—
3
E

°
j_
O

0
O)
CO

>">

l/l
re
+_>
O
(-
—
're
0
•I-J

(4^
O

4_>
c
QJ
O
S-
O)
Q.
^^^<

c
o

4_>
o
3
-a
0
S-
c.

;
O


























14 —
o

c
O)
o

OJ
Q.
*^-^



r-~
•i—
0






r—
ns
O
0







t/)
fO
cu


<















s_
0
+*>
(J
Ol
CO








i — co r~- •=}• co co vo
(JO C^ CO CT> O CO kD
. .... . •
co r^ CTi r^~ LT> co cr» o
^3- tJDIOCCCTi CT. O~C






i — f^- CTi 1^^ CT) LO CO LO
^D r^ Cj" O O CO CM CO
cc cccMcor^ coi— co
^f •— r— O






CO r— CM «— O CO 1 1 O
in r^^ o "vj~ o CM i i •
..... LO
VO C\J i— CTi »d- r— CM





f^ r- LO r^~ i —
•* «^- CM *d- 1 1 >JD
II.. .11.
CM i i «a- o o r~-
CO ' CT






LO  c: to i — •!— en
i- •>- Q; -i- to c c
aj 4-> c o a) -r- o
i — (O -r- CD Q: +->•!- SZ
*'~ O J3 ^ — fO -*-1 O
O(/)Oi-i — I— CUC/l'r—
rn CJ i- 3 fO fO T" Z5 4-^ ^™-
C Q. 1 — •«— -r— jQ re CL)
^> •' — •" — i- o en to E i- 3
•M Cnot04-> S-C W O OJLl-
"r~ C QJ fl3 C/l U *r~ QJ CJ f*
f— ^|J^J fv fr^ •"j ^ -|__i {j | . ,» ^^
-(-><_> cocut-oc
o
i — c\j oo^- Lovor--i —

























































































f- — ^
^.
>^
— ^
cr.
•^
cr.
c

X
CM
CO
LO
o
C-
S-
^J
»^^
(/)
c
o
+J

ko
o

X
LO

10

r—
,

II

to
c
o

to
00

i
LlJ

re
a. >
o
,-t~
—


















1006

-------
     The general program approach involves five types of activities:

     fl    Field testing provides accurate emission information on the source
          class as normally operated in the field and evaluation of control
          possible through changes in operating conditions only.

     •    Process research and development evaluates the potential  for minor
          hardware changes for increased level of control by retrofit and,
          thereby, provides guidance for application of existing technology
          to new designs.

     f    Fuels research and development provides general development of various
          techniques for controlling N0x from each fuel type independent of
          equipment type, optimization of the technology for each fuel, and
          general guidance for application of the optimum technology to a
          variety of practical systems.

     •    Fundamental  studies of combustion chemistry and aerodynamics promote
          understanding of the combustion phenomena responsible for pollutant
          formation and, thereby, guide development and optimization of new
          technology.

     0    Environmental assessment documents the impact of the NO  control
          technology on total system performance and emissions to air, water
          and land.

                                 PROGRAM STATUS

     The results of the program to date fall into two general areas:   (a)
development and application of control technology to specific source categor-
ies, and (b) generation of technological approaches of potential applicability
to a wide range of sources.

     The development and application of technology for specific practical
equipment has concentrated on the major heat and steam generating sources of
NO .  These are utility, industrial and commercial boilers, and residential
heaters.  The significant results to date can be summarized as follows:
     (1)  Utility boiler field testing has established reliable emission
          factors for boilers fired with coal, oil and natural gas, and has
          indicated that significant levels of control are achievable through
          modification of operating conditions with the specific level of
          control dependent on fuel type and boiler design (6).  In addition,
          it has been shown that by minor hardware modifications to apply
          staged combustion, N0x levels can be reduced by 30-40 percent for
          coal fired boilers already operating at or slightly above the NSPS
          for that type of equipment.  Efficiency and fireside corrosion results
          are encouraging but show that more attention to these areas is
          necessary.

     (2)  Industrial and commercial boiler field testing has also established
          reliable emission factors for a variety of fuels; however, it has

                                     1007

-------
          indicated that the relatively simple designs of these boilers make
          them less amenable to N0x control  by possible changes in operating
          conditions.   The development and application studies, in which the
          emphasis has been on small  package type boiler fired with heavy oil
          and natural  gas, have shown the potential for NO  reductions of up
          to 50 percent for the heavy oil by application of staged combustion
          and over 60 percent for gas with flue gas recirculation (7, 8).

     (3)  Residential  furnace emission factors have also been established;
          however, changes in operating conditions, which consist of tuning,
          generally increase NO  slightly (9).  The development effort has
          produced an oil burner head design capable of reaucing N0x relative
          to conventional practice.  Further development of an integrated
          furnace to achieve a 50 to 75 percent reduction of NO  coupled
          with an efficiency increase is nearing completion (.10;.  Field
          application will be evaluated in the near future.

     The development of advanced technological approaches consists of investi-
gations in three general areas:  optimization of burner design and other
techniques to establish minimum NO  emissions achievable for conventional
fuels and system design; assessment of alternate fuels and equipment design
approaches for ultimate control potential through system redesign; and fund-
amental studies to provide understanding and guidance for the first two areas.
Selected significant results in each area are given below:

     (1)  Major effort for development of optimized technology has concentrated
          in the area of burner design alteration to give low N0x and efficient
           combustion  conditions.   Critical  parameters  have  been  established
          for natural gas and pulverized coal and progress has been made for
          residual oil.  Possibly the most significant result was the establish-
          ment of conditions capable of reducing NO  from a single pulverized
          coal burner from approximately 900 ppm toxapproximately 150 ppm (11).

     (2)  Advanced concept studies, which have been initiated, include alter-
          nate fuel combustion and catalytic combustion.  The only alternate
          fuel evaluated to date is methanol, which appears to have favorable
          combustion and emission characteristics (12).

     (3)  The fundamental studies have produced kinetic rate information
          for gas phase and fuel decomposition reactions, as well as basic
          aerodynamic mixing information for turbulent diffusion flames.
          This information provides input for development of engineering design
          models for application and optimization of the technology (13, 14).

                           PROGRAM DIRECION AND GOALS

     The program direction and goals are strongly dependent on both the level
of funding and the duration of the program.   Table 2 shows approximate current
emission levels for uncontrolled sources and the estimated level  of NO  control
achievable in 1980 and 1985 for the 10 year MSST.  The near term goalsxshould
be possible with hardware changes (e.g., burner design for conventional boiler
                                     1008

-------
design, while the long term goals may depend on significant system redesign
and/or application of advanced concepts.   Although the technology to achieve
these levels can almost certainly be developed by the year shown, the demon-
stration in practical equipment will depend on equipment lead time and other
factors, particularly for the larger systems (e.g., utility boilers).  For
less than a maximum program, lower priority sources will be dropped from
consideration.  Based on the current projected funding and a program duration
of 5 years, the probable achievable technology and demonstration goals are
indicated by notation on 1980 goals in Table 2.  The other 1980 goals are
achievable, but are not currently under intensive development.   This techno-
logy also has the potential for application to current equipment with a lower
probable degree of control.  The 1985 goals would be outgrowths of the tech-
nology with significant system redesign (i.e., combustion and heat exchanger
design matching) and/or utilization of advanced technology currently under
development (e.g., catalytic combustion).

     To illustrate the total approach, consider the overall development history
for wall-fired coal burning utility boilers shown in Table 3.  Given the
significant decrease available by relatively minor changes in hardware and
operating conditions coupled with the low level achievable on a small coal
burner at the International Flame Research Foundation, achievement of 200 ppm
in a field operating utility boiler by application of an optimum burner can
be projected with high probability.

     A second example can be taken for residential heating systems to illus-
trate not only NO  control with no increase of carbonaceous emissions, but
also improved efficiency.  The initial development was an optimum oil burner
head suitable for retrofit to at least 50 percent of existing residential
furnaces with an NO  reduction of 20 to 50 percent and no loss  in efficiency.
This burner head was then matched to a combustion chamber design with the goal
of a 75 percent NO  reduction and improved furnace efficiency.   The perfor-
mance of a prototype of this integrated furnace is shown in Figure 1 and is
compared to the performance of both the overall furnace population and the
unmodified stock furnace.  It can be seen that the NO  goal was almost achieved
at an operating excess air level that represents an efficiency gain.  Carbon-
aceous emissions were maintained at low levels (10).

     Finally, the exploration of advanced concepts and utilization of alter-
nate fuels may provide the basis for longer term (1985) technology to vir-
tually eliminate NO  emissions from some sources.  For bound-nitrogen-free
fossil and coal-derived fuels, catalytic combustion concepts may allow NO
to be controlled to the level of 10 ppm or less.  Control of NO  from nitrogen-
containing alternate fuels appears somewhat more complex; however, it appears
that this combustion technology should also be applicable (15).

                                   CONCLUSIONS

     The developing combustion modification technology can form :ie basis for
significant New Source Performance Standards in the future for several sta-
tionary combustion systems.  Implementation of a 10-year MSST strategy can
significantly broaden the degree of control and demonstrated source applica-
tion achievable with the technology.  It also appears that any adverse impacts


                                     1009

-------
of NO.  control  on cost, energy efficiency,  and other emission  can  be  minimized
or eliminated by proper technology application.
                                   REFERENCES

1.   Crenshaw, John and Allen Basala.   Analysis of Control  Strategies to
     Attain the National Ambient Air Quality Standard for Nitrogen Dioxide.
     Presented at the Washington Operation Research Council's Third Cost-
     Effectiveness Seminar, Gaithersburg,  Md.,  March 18-19, 1974.

2.   Bartok, W. et al.   Systems Study of Nitrogen Oxides Control  Methods for
     Stationary Sources, Vol. II.  National  Air Pollution Control  Administration,
     EPA No. APTD 1286, NTIS No. PB 192-789, November 1969.

3.   Mason, H. B. and A. B. Shimizu.  Definition of the Maximum Stationary
     Source Technology (MSST) Systems Program for No .   Preliminary Data from
     task in progress.

4.   Martin, G. B.  Overview of U.  S.  Environmental Protection Agency's
     Activities in NO  Control for Stationary Sources.   Presented at the Joint
                     X
     U. S.-Japan Symposium on Countemeasures for NO ,  Tokyo, Japan, June 28-29,
     1974.
X
5.   Lachapelle, D. G., J. S. Bowen and R. D.  Stern.   Overview of Environ-
     mental Protection Agency's NO  Control Technology for Stationary Com-
     bustion Sources.  Presented a! the 67th Annual AIChE Meeting, Washington,
     D. C., December 1974.

6.   Crawford, A. R., E. H. Manny, and W.  Bartok.   Field Testing:  Application
     of Combustion Modifications to Control NO  Emissions from Utility Boil-
     ers.  EPA-650/2-74-066, June 1974.       x

7.   Cato, G. A., L. J. Muzio and D.  E. Shore.  Field Testing:  Application of
     Combustion Modifications to Control of Pollutant Emissions from Indus-
     trial Boilers - Phase II.  EPA-600/2-76-086a, April 1976.

8.   Heap, M. P. et al.  Reduction of Nitrogen Oxide Emissions from Field
     Operating Package Boilers, Phase III  of III.   Report in review.

9.   Barrett, R. E., S. E. Miller and D. W. Locklin.   Field Investigation of
     Emissions from Combustion Equipment for Space Heating.  EPA-R2-73-084a,
     NTIS PB 223-148/AS, June 1973.

10.  Combs, L. P., and A.  S. Okuda.  Residential  Oil  Furnace Optimization -
     Phase II.  Final Report in Review.

11.  Heap, M. P., T. M. Lowes, R. Walmsley, and H. Bartelds.  Burner Design
     Principles for Minimum N0x Emissions.  EPA Coal  Combustion Seminar,
     Research Triangle Park, N* C., June 1973.


                                    1010

-------
TABLE 2. TECHNOLOGY  DEVELOPMENT  RESEARCH  GOALS—AVERAGE  NO (ppm 0 ?>%
                                                           A

Source
Utility
Gas
Oil
Coal
Industrial
Gas
Residual oil
Coal
Commercial
Gas
Distillate oil
Residual oil
Residential
Gas
Distillate oil
Current Technology 1980 Goal
15oja! 100
225^ ' 150

150 80
325 125, }
450 150lc;

100 50
125 70, v
300 100V ;

80 25, N
115 35Kj
Reciprocating Engines
Spark ignition—gas 3000 1200fdl
Compression ignition—oil 2500 1200^ '
Gas Turbines
Gas
Oil
(a) Current NSPS
(b) Estimated achievable
(c) Developed and field
(d) Developed technology
400[15o!^l 75^!
600[225l ;] 125^ '
with wet control technology
applied technology
from ongoing EPA program
1985 Goal
50
90
100

50
90
100

30
50
90

10
10
400
800
25
25

                                 1011

-------
C£
O
h- 00
OO C£
I—< LU
m _i
   *—«
I— O
z: CQ
LU
s: >-
D- I—
O »-i
LU
o
O HI
OL •z.
\- a:
Z. =3
O CQ
O
_l O
Q. O
X O
LU LU
   a:
ca
•o
CU
-p
o
OJ
•i — 5
o
i_
D-
T3
CU
CU
-C
o
=£
S-
-

fO
3
+J
(J
•a:
CM
O
<•— y;
cu ro
>
CU CE.
_J
X Q.
O 0.
•^f ^ 	 ,










4->
C
CU
>
LU
















r-~.
en
*~~







LO
r^
CTI
( n LO CM
cu r^ r^. r>-
S- CTi CTi CTi
D- i— i— r-

0 O 0 O 0
o LO in o LO
CTI LT> LD «a- i —



03
CU -H
CD C
c: cu
« E
-C •<-
0 CU S-
i— CU
C O) CL
O E C X
•i- CU -i- LU
•»-> 4J OO 1
-O tO CO 1
CU S- >, CU i_
i— CU OO r— CU
i— Q. (O c.
O O -D O S-
i- CU CO 3
4-> S- -r- CQ
C CU M- r—
O r— 00 -r- r— r—
O -r- Q- T3 fO "3
c o oo o E O
Z3 ca z: 2: oo o





i^~
r--
CTI
r~














o
CD
CM


0) r—
i— IO
cn+J
c c
••- cu
oo E
•r~
E i-
3 CU
E Q.
•r- X
-M LU
Q. 1
0 1
S-
OJ CU
r— c:
n3 s-
O 3
OO CO

r— ~ r —
i — fT3
3 O
Ll_ LJ





C
CO
CT.
*~














0
CD
CM







s-
cu
c
S-
3
CQ

E
3
E
•i —
-(->
Q.
O

>,

'o.
Q.

-------
      3.5
      3.0
      2.5
  flC

  OG
      2.0
      1.5
  X
  o
  CJ
  E
      1.0
      0.5
1       I      I       T
        1.0
i      i      i       r
                                                           — SMOKE >1
                                                          — SMOKE <1
            APPROXIMATELY 80% OF EXISTING OIL
            FURNACES PRODUCE CYCLE-AVERAGED
            EMISSIONS OF NOX IN THIS RANGE (REF.J
                   PROTOTYPE FURNACE WITH
                   RETENTION HEAD BURNER
                           y    /   /
                                  'PROTOTYPE LOW-EMISSION FURNACE
                     . TARGET LEVEL, CONTRACT 68-02-1819       	
                       NOMINAL OBJECTIVE OF FURTHER NOX
                       REDUCTION EFFORTS
               I       I      I       I      I      I       I      I       I
                          1.5

                  STOICHIOMETRIC RATIO
                          2.0
Figure  1.   Cycle-averaged  flue gas NO  concentrations, comparison of  the
          prototype  optimum furnace with  other existing units.
                                    1013

-------
12.
13.



14.


15.
Martin, G. B. and M. P. Heap.   Evaluation of N0x  Emission  Characteristics
of Alcohol Fuels for Use in Stationary Combustion Systems.   Presented at
the Symposium on Impact of Methanol  on Urban Air  Pollution,  80th  National
AIChE Meeting, Boston, Massachusetts,  September 7-10,  1975.
Axworthy, A. E., G. R. Schneider, M.  D.  Schuman,  and V.
istry of Fuel Nitrogen Conversion to  Nitrogen Oxides in
EPA 600/2-76-039, NTIS No.  PB250-373/AS, February 1976.
    H. Dayan.  Chem-
    Combustion.
Bowman, C. T. and L. S. Cohen.   Influence of Aerodynamic Phenomena on
Pollutant Formation in Combustion.   EPA-650/2-75-061a,  July 1975.
Martin, G. B.  NO  Considerations in Alternate Fuel
Symposium Proceedings:  Environmental Aspects of Fuel
II, EPA-600/2-76-149, June 1976.
Combustion.  In:
  Conversion Technology,
                                     1014

-------
                                                                         21-2
                 EMISSION ESTIMATES OF NO? AND ORGANIC COMPOUNDS
                     FROM COAL-FIRED FLUIDIZED BED COMBUSTION

                                 P.  F.  Fennelly*

ABSTRACT

      fluidized bud combustion *y*tem*  operate at significantly lower tempera-
ture* than conventional system*  (about 900°C versus 1500°C, respectively].
Question* have, ban raised concerning  the organic and nitrogen oxides
emissions which could result at these  lower operating temperatures.  Lining
available data ^rom bench scale on pilot plant fluidized bed combustion
experiments and/on Dimple thermodynamics and empirical correlations with
conventional combustion systems, estimates have been made o^ the concen-
trations o^ various compound!) in the &lue gas.  The&c estimate* are probably
good to within an order o& magnitude.

     The results indicate that with respect to formation ofa photochemical
ozidant precursors, no special environmental problem* should result farom
{,luidized bed combustion.  In &act, this combustion technique offers
significant advantages with respect to reductions in nitrogen oxides and
sulfur oxide* emission*.

                                  INTRODUCTION

     A fluidized bed boiler can be simply depicted as an enclosed Cavity
containing boiler tubes and a bed of granular solids, to which fuel is
added.  As shown in Figure 1, the solids are supported on a grid, at the
bottom of the boiler, through which combustion air is passed at high
velocities, typically 2 to 5 feet per  second (0.6-1.5 m/s).  The solids are
held  in suspension by the upward flow  of the air and a quasi-fluid is
created which contains many properties of a liquid.  The most important
liquid-like property is that the bed material is exceptionally well mixed
and flows throughout the system without mechanical agitation.

     Fluidized bed coal combustion systems for the production of steam
and/or electricity have several advantages over conventional combustion
systems.

     •  High heat transfer coefficients and volumetric heat release
        rates will reduce the boiler size by one-half to two-thirds
        or more compared to a conventional unit.
*GCA Corporation, Bedford, Massachusetts 01730.

                                     1015

-------
                                                        g
                                                        
-------
     •  Capital  costs will  be reduced due to the size reduction and
        the potential for shop fabrication instead of field construction.

     •  The use of limestone as bed material provides a means for i_n_
        situ sulfur dioxide (S02) removal.

     •  The high heat transfer coefficients permit lower operating
        temperatures (850 to 950°c) which can potentially decrease
        nitrogen oxides (N0x) emission.

     Questions have been raised concerning the organic emissions that could
result at these lower operating temperatures.  Using available data and/or
simple thermodynamics and empirical correlations, this oaper provides esti-
mates of the concentrations of various compounds that could result from
coal-fired fluidized bed combustion (FBC).  These estimates are probably
good to within an order of magnitude.

                            STABLE ORGANIC COMPOUNDS

     Potential organic pollutants are those compounds which could form from
the incomplete combustion of coal.  One can view the burning of a coal
particle as a sequential process.  The particle vaporizes or volatilizes;
these volatile compounds can react among themselves in a chemically reducing
atmosphere.  Next, they react with oxygen within the system in a diffusion
flame.  After devolatilization is completed, the char continues to burn.
In specifying the potential organic compounds which could form, the basic
question is what types of chemical species are produced during coal pyrolysis
and to what extent will they survive in the reactive environment of a
fluidized bed combustor.

     The chemical structure of coal can be viewed as a network of inter-
connected aromatic hydrocarbon compounds (1),  Small hydrocarbons (HC) can
form directly from cleavage of substituted alkyl groups.  Polynuclear
aromatic hydrocarbons (PAH) can form directly via bond cleavages in the
structural network or they can form via condensation reactions of various
HC decomposition products.   These condensation reactions even though generally
endothermic, based on thermodynamic estimating techniques, at FBC temperatures
of 1500°F probably proceed at a significant rate (2,3).  Branched or cyclic
HC are seldom found as products of coal pyrolysis carried out at similar
temperatures  (1); hence, only two classes of HC should be of any significance
in fluidized bed coal combustion: small HC (i.e., less than three carbon
atoms) and PAH.  Both could form and survive within the bed at temperatures
on the order of 1500°F.

     Similar arguments apply to organic nitrogen and sulfur compounds.
Species such as pyridine decompose at temperatures of 1000°F to form HCN and
small HC (4).  Thiophenes and mercaptans can also decompose to form small
HC and hydrogen sulfide (H2S).
                                     1017

-------
                  CONCENTRATION  ESTIMATES OF ORGANIC COMPOUNDS
     For a rough estimate of the concentrations  at which some of the small
HC and reduced sulfur and nitrogen compounds might exist, one can use
equilibrium calculations based on free energy minimization.   An upper limit
can be obtained from calculations performed in conjunction with coal
gasification experiments, where highly reducing, fuel-rich conditions  •
exist (5).  Even with only 60 per cent stoichiometric oxygen present, the
data in Table 1 indicate that HCN, COS, CS:,, etc.  are less than 10 parts per
million (ppm).  Extrapolation of the calculations  to typical operating
conditions such as 20 percent excess air, where S02 and NOX become the
predominant sulfur and nitrogen compounds, indicates that compounds such

           TABLE  1.  CALCULATED EQUILIBRIUM CONCENTRATION FOR SELECTED
                SPECIES  PRODUCED BY  INCOMPLETE COMBUSTION OF COAL
               Coal analysis:  C-68.5%, H-5.3%, 0-8.5%, N-l.4%,
                               S-4.1%
               Oxygen present:  59% of stoichiometric requirements
               Temperature:     1400°F
                                                  Mole
                         Species                fraction
                         N2                       0.36
                         CO                       0.26
                         H2                       0.15
                         H?0                      0.06
                         C0?                      0.10
                         CH,t                      0.06
                         H?S                      0.008
                         COS                      0.0005
                         NH3                      0.0007
                         C(g), HCN, CS?         < 10"5
                         C?H?, C7Hlf,  (CN)?
                         S?, S07, S03
                         N-,, NO,
                  Based  on  Reference 5
                                     1013

-------
as HvS, HCN, COS, (CN)>,  etc.  should be present in concentrations less than
1  pprn.   Free energy minimization calculations for the more complicated PAH
are impractical.   To estimate  the concentrations  at which these types  of
compounds might exist in  the FBC flue gas, one can use empirical  correlations
between benzo(a)pyrene and methane (CHj concentrations from measurements  in
conventional coal-fired combustion systems.   Figure 2 indicates the variation
of total  hydrocarbons (as CHt() as a function of oxygen (0-) concentration
from one set of FBC experiments.  Under normal operating conditions, about
3 per cent 02 in  the flue gas  (20 per cent excess air), the concentration
of HC (as CHi+ ) is about  100 ppm (vol ume/vol ume-V/V) (6).  Although emissions
can often vary between different fluidized bed systems, 100 ppm provides
a convenient reference value.   Previous measurements with conventional
coal-fired systems indicate that compounds such as benzo(a)pyrene are
     aiiy iQ-5 times less than the concentration of total HC as CH4(7).
         3000
        ex
        o.

       V)
         2500
       §2000
z
Ul
u 1500
O
UJ
z

jE  1000
UJ
2
       z
       o
       CD
       tr
       o
       tr
       o
           500
                             TEST  CONDITIONS
                      RE ACTOR'FLUIDIZED  BED MODULE
                      SUPERFICIAL  VELOCITY =12-14  fps
                      BED  TEMPERATURE = 1600-I800°F

                      BED DEPTH' 12"- 13"

                                    TYPE  OF COAL

                        O WASHED.OHIO^B  SEAM, 2.6%  S

                        X  UNWASHED, OHIO #8 SEAM, 4.5% S
                                                                    30
                                                                       0)
                                                                       o
                                                                    20
                                                             10
Q.


tt




V)
                                                                X
                                                                UJ
       Figure 2.
                  1.0        2.0        3.0        4.0        5.0
                 FLUE  GAS  OXYGEN  CONTENT, percent

           Variation in hydrocarbons concentration  with  flue
             gas  oxygen content  in the  FBM  (6).
                                    1019

-------
Using our reference value of 100 ppm (CH..),  this  implies  that in a
fluidized bed system, PAH could exist in the flue gas  at  concentrations  (V/V)
on the order of 1  part per billion (ppb).   Considering that flue gases  are
eventually diluted by a factor of a thousand when they are emitted  from
the stack (8), this implies that ambient concentrations of PAH near FBC
facilities would be on the order of 1 part per trillion.   This corresponds
to about 0.6 ng/tn3 which is roughly comparable to the  natural background
concentration ranges found in rural areas  (7). Accordingly, it seems that
PAH concentrations should not be high enough to cause  problems.

     Recently, some conventional coal-fired flue  gases have been tested for
the presence of polychlorinated biphenyls  (PCB) and trace concentrations
have been reported (8).  Experience in coal  combustion and coal pyrolysis
indicates that at temperatures similar to  that in coal-fired FBC, chlorine
(when present) predominatly exists as HC1  (1). Accordingly, one would  not
expect significant concentrations of PCB's In coal-fired  FBC.  If oresent
at all, they should be in concentrations less than those  of the unsubstituted
PAH (i.e., <1 ppb (V/V)).

                            NITROGEN OXIDES EMISSIONS

     There is considerably more experimental data available for NOX emissions
in FBC compared with that for organic compounds (9-12).  These emissions can
vary from system to system and work is underway to summarize the pertinent
fluidized bed parameters which influence NOX emissions.  Of special interest
are the relative contributions of thermal  NQX and fuel  NOX.

     At the temperatures encountered in FBC (1500 to 1600°F), equilibrium
calculations based on thermal fixation of  atmospheric  nitrogen indicate
nitric oxide (NO)  levels should be 100 to  200 ppm, depending on the amount
of 02 present.  In fact, NO levels in most coal-fired  fluidized bed experi-
ments conducted at atmospheric pressure are on the order  of 300 to  500  ppm
(9-12).  Two plausible explanations are (1)  the excess NO results from  hot
spots within the bed which are far in excess of the equilibrium temperature,
or (2) the higher NO levels result from substantial conversion of the nitrogen
in coal.  There is some interesting experimental  evidence supporting the
latter mechanism.

     Workers at Argonne Laboratories in a  set of  experiments on fluidized
bed coal combustion substituted argon for  the nitrogen in combustion air
and found that NOX levels did not change significantly.  In one experiment,
the initial NO concentration of 580 ppm did not vary during the argon sub-
stitution by more than +_ 20 ppm (9,10).

     NOX concentrations in fluidized bed systems, however, can be very
apparatus dependent, and significant concentration profiles have been found
in some reactors (11).  Accordingly, one must be  wary  of  the manner in  which
emission measurements are made.

     In a recent review, workers at Battelle compared  NOX data from a number
of laboratories and, in almost all cases,  the emissions are less than the
                                     1020

-------
EPA New Source Performance Standard which is  0.7 pound per million  Btu  heat
input,  maximum 2-hour average expressed as  N02  (equivalent to  approximately
550 ppm N02 at 3 per cent 0?  in the flue gas)  (13).

     At higher pressures  (6 to 8 atm),  experiments  indicate NOX  levels  are
reduced to values on the  order of 100  to 200  ppm, although the reason for
lower values is not yet known (14,15).

                             SULFUR OXIDES  EMISSIONS

     Only brief attention will be paid here to  SOX  emissions as  their role
in oxidant formation is considerably less well  defined than that of HC  or
NOX.  The emissions of SOX depend markedly  on the calcium/sulfur (Ca/S) mole
ratio which in turn is dependent on the amounts and types  of coal and lime-
stone used in the bed.  In general, however,  with Ca/S mole ratios  greater
than 2.0, 90 per cent control of S02 emission is achievable.  By way of
example, experimental results indicate that a bituminous coal  with  a heating
value of 13,700 Btu per pound and a 2.8 per cent sulfur content  when burned
in a fluidized bed at velocities of 2.1 feet  per second and a  Ca/S  ratio of
2.9 has an emission factor of 0.23 pound S02/106 Btu compared  with  EPA's New
Source Performance Standard of 1.2 pounds S02/106 Btu (14).
                                   CONCLUSIONS

     Concentration estimates for the various types of pollutants that could
form in coal-fired fluidized bed combustion are provided in Table 2 (16).   As
mentioned previously, there is only sparse experimental information available
for most of these compounds and these estimates which are based on simple
thermodynamic considerations and empirical correlations with conventional
combustion systems are probably good only to within an order of magnitude.
The only compounds generated in quantities high enough to be of concern in
oxidant formation are NO, S02, and carbon monoxide (CO).  The concentrations
of these compounds, however, are significantly less than those generated in
conventional combustion; hence, with respect to reducing oxidant formation,
fluidized bed combustion is a promising alternative to present boiler systems,

     Most of the existing data, however, has been gathered on bench scale
equipment or small pilot plant systems.  Final acceptance of FBC will prob-
ably await tests on larger facilities; but, presently, no emission problems
are anticipated in the scaling up of the smaller facilities.
                                     1021

-------
           TABLE 2.   ESTIMATED CONCENTRATION  RANGES  OF  POTENTIAL
          POLLUTANTS FROM COAL-FIRED FLUIDIZED  BED COMBUSTION  (16)
 Gas  phase

   Several  hundred parts/million:     CH,+ , CO, HC1, S02, NO

   Ten  parts  per million:             S03, C?H4, C?H6

   One  part per million:              HF, HCN, NH3, (CN)2, COS, H2S, h^SO^,
                                     HN03, F, Na

   One  part per billion:              Diolefins, aromatic hydrocarbons,
                                     phenols, azoarenes, As, Pb, Hg, Br, Cr,
                                     Ni, Se, Cd, U, Be

   One-tenth  (0.1) part per billion:  Carboxylic acids, sulfonic acids,
                                     polychlorinated biphenyls, alkynes,
                                     cyclic hydrocarbons, amines, pyridines,
                                     pyroles, furans, ethers, esters,
                                     epoxides, alcohols, ozone, aldehydes,
                                     ketones, thiophenes, mercaptans

 Solids

   One  part per million:              Al, Ca, Fe, K, Mg, Si, Ti, Cu, Zn, Ni,
                                     U, V

   One  part per billion:              Ba, Co, Mn, Rb, Sc, Sr, Cd, Sb, Se, Ca

   One-tenth  (0.1) part per billion:  Eu, Hf, La, Sn, Ta, Th
        —	"~  "  '            ACKNOWLEDGEMENTS

     This work was performed under contract to  the  U.S.  Environmental  Pro-
tection Agency under the guidance of Mr.  D.  Bruce Henchel  as  Project  Officer.
The opinions, findings,  and conclusions  are those of the author  and not
necessarily those of the Environmental  Protection Agency.
                                   REFERENCES

1.  Lowry, H.H. (ed.).   The Chemistry of Coal  Utilization.   John  Wiley and
    Sons, New York, 1963.

2.  Franklin, J.L.   Ind.  Eng.  Chem. ,  41:1070,  1949.
                                     1022

-------
 3.   Reid,  R.C.  and  T.K.  Sherwood.  The  Properties of Gases and Liquids -
     Their  Estimation  and Correlation.   McGraw-Hill, New York, 1958.

 4.   Sternling,  C.V. and  J.O.L.  Wendt.   Kinetic Mechanisms Governing the
     Fate of Chemically Bound  Sulfur  and Nitrogen  in Combustion.  EPA-
     650/2-74-017,  U.S. Environmental  Protection Agency.

 5.   Stinnett,  S.J., D.P. Harrison, and  R.W.  Pike.  Fuel Gasification.
     The Prediction  of Sulfur  Species  Distribution by Free Energy Minimiza-
     tion.   Environ. Sci. and  Technol . ,  8:441,  1974.

 6.   Robinson,  E.B., A.H. Bagnulo,  J.W.  Bishop, and S.  Ehrlich.  Characteri-
     zation and Control of Gaseous  Emissions  from  Coal-Fired  Fluidized Bed
     Boilers.   Pope, Evans, Robbins,  Inc.,  Alexandria,  Va.  Report  prepared
     for Division  of Process Control  Engineering,  National Air Pollution
     Control Administration (now U.S.  Environmental Protection Agency).
     1970.   p.  103.

 7.   Hangebrauck,  R.P., D.J. von Lehmden, and J.E. Meeker.  Emissions of
     Polynuclear Aromatic Hydrocarbons and  Other Pollutants from Heat
     Generation and  Incineration Processes.   J. Air Pollu. Control  Assoc.,
     14:267, 1964.   See also:   Hangebrauck,  R.P.,  D.J.  von Lehmden, and
     J.E. Meeker.   Sources of  Polynuclear Hydrocarbons  in the Atmosphere.
     PHS 999-AP-33,  U.S.  Department of Health,  Education and  Welfare, 1967.

 8.   Cowheard,  C.,  M.  Marcus,  C.M.  Guenther,  and J.L. Spigarelli. Hazardous
     Emission Characteristics  of Utility Boilers.   EPA-650/2-75-006, U. S.
     Environmental  Protection  Agency,  July  1975.

 9.   Jarry, R.L.,  L.J. Anastasia, E.L. Carls, A.A. Jonke, and G.J.  Vogel.
     Comparative Emissions of  Pollutants During Combustion of Natural Gas
     and Coal  in Fluidized Beds.   In:  Proceedings of Second  International
     Conference on  Fluidized Bed Combustion,  Hueston Woods, Ohio, 1970.
     U.  S.  Environmental  Protection Agency  Publication  AP-109.

10.   Jonke, A., et  al. Reduction of  Atmospheric Pollution by the Application
     of Fluidized  Bed  Combustion.  ANL/ES-CEN-1001 , Argonne National Labora-
     tory,  1969.

11.   Pereira, F.J.,  J.M.  Beer, B. Gibbs, and A.B.  Hedley.  NOX Emissions  from
     Fluidized Bed Coal Combustion.   Fifteenth Symposium  (International)  On
     Combustion, 1974, The Combustion Intitute, Pittsburgh, Pa.
                                    1023

-------
                                                                           21-3
     EMISSIONS ASSESSMENT OF THE  CHEMICALLY  ACTIVE FLUID BED (CAFB) PROCESS

  A. S. Werner, R. M. Bradway,  D.  F.  Durocher,  S.  L.  Rakes, and R. M. Statnick*

ABSTRACT
     Stack ga(> and poAticLilate.  miAtxionA  &Aom the. Ckz.m-ic.atly Active, fluid
Bed pilot plant at the, EMO  Rue.ax.ch dntAe.,  Abtngdon, MeA.e, me.at>uAe,d during
Q0i> location o^ ^u.e.1 oiJL  and b-ituman.   Qua.ntLU.eA me.aAuAe,d include.d total
paAticulate. , poAticulatz.  &i.ztAibutsionA ,  nitAoge^n oztdu, Aul&uA
and gaAe.ouA organic compounds.   Both paAti.caI.ate- and goxSeou/S e,m-i!>£
analyzed &OA oAgan-ic faunctsional gAou.pt,  uA-ing  Liquid chAomatogAaphy and tn
Ae.d ApzctAOAcopy.  In addition,  bulk and  AuA&ace. tAace. mutal compoA-itionA
paAticatate. weAe. date,Am^ne.d  &Aom &paAk  Bounce. mo44 t>pe.ctAot>copy and X-Aay
pkoto-eJL&ctAon £pe.ctAo-t>copy.  Result* aAe. pAZAnntzd uihtch indicate, that
^actoAy con&iol o& ga&couA Apnci
-------
sure the environmental acceptability of this process.   To attain these goals,
a systematic evaluation of all waste streams from the  CAFB was made and a
process emissions inventory was compiled.   These data  were derived from
engineering estimates and from an extensive pilot plant field sampling and
laboratory analysis program.   Emission rates determined for the pilot plant
were then used to predict pollutant loadings for the CAFB demonstration plant
and proposed commercial units and to make  recommendations for additional
control requirements.

     The experimental phase of the project consisted of two parts::  a field
test program carried out at the ERCA pilot plant during the period December 3 -
11, 1975, which consisted of on-site determinations of air emissions and
collection of gaseous and solid samples; and a laboratory analysis program to
identify and, where possible, quantify chemical  constituents of samples
collected at the pilot plant.

                                THE CAFB PROCESS

     Figure 1 is a schematic diagram of the CAFB pilot plant and Table 1
lists mass flow rates and temperatures of  internal and external streams.
Limestone and oil are fed continuously into the  gasifier at a calcium (Ca)/
sulfur  (S) (limestone/oil) molar ratio of  unity.  In the gasifier,, limestone
(CaC03) is rapidly converted to lime (CaO) and carbon  dioxide (C0?), and the
CaO is maintained in a fluidized state by  a preheated  air/flue gas mixture.
The air input rate is roughly 20 percent stoichiometric with respect to oil.
Fuel oil is consecutively vaporized, oxidized, cracked and reduced at 870°C
(1600°F) to produce a low Btu gas.  Over 80 percent of the input feed sulfur
is removed by the lime.  The gas travels from the gasifier through cyclones
for particulate removal and then into a boiler for combustion.  The boiler
flue gas encountered a knockout baffle and another cyclone before entering
the stack.  Lime is continuously cycled between  the gasifier and the regen-
erator where the roughly 7 percent of the  lime which is sulfided in the
gasifier is oxidized to CaO.   At the pilot plant, sulfur dioxide (S02) pro-
duced in the regenerator is fed to the boiler stack.  (In the demonstration
plant, S02 will be fed to another unit which will reduce the sulfur (S) to its
elemental form.)  Some spent CaO is continuously withdrawn from the regenera-
tor and retained for disposal.  To maintain S removal  efficiency, an equiva-
lent amount of CaCQ3 is continuously added to the gasifier.  Proper disposal
of spent CaCOs is a major environmental concern.

                               FIELD TEST  PROGRAM

     Coincident with the planning and pre-test site visit was the announce-
ment of the "multilevel phased approach" to source sampling and analysis by
the Process Measurement Branch of the Industrial Environmental Research
Laboratory, EPA (4).  Because this project was a preliminary environmental
assessment of a facility which heretofore  had not been subjected to a com-
prehensive emissions assessment, the decision was made to combine measurements
of the so-called criteria pollutants using the standard EPA methods with the
new procedures for broad identification of organic and inorganic emissions.

     The field test program was carried out at the ERCA pilot plant location


                                     1026

-------
                                                                E
                                                                (O

                                                                Q.


                                                                •!->
                                                                O

                                                                •I	 •»
                                                                Q-
                                                               c_>   S-
                                                                ro -W>
                                                                S-  W
                                                                cu
                                                                Q.    o
                                                               •r-  O
                                                                E  S-
                                                               => O.
                                                                O)
1027

-------
                  TABLE 1.   ERCA PILOT  PLANT  MASS  FLOW RATES

No.
Fig
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
on
. 1 Process stream
Oil feed to gasifier
Limestone feed to gasifier
Gasifier to regenerator stone transfer
Regenerator to gasifier stone transfer
Product gas to cyclone
Cyclone solids return to gasifier
N? gas to solids transfer lines
Product gas to boiler
Air to regenerator
Spent solids from regenerator
Regenerator off gas to cyclone
Regenerator off gas, cyclone to stack
Flue gas from boiler
Flue gas recirculated to gasifier
Flue gas to Tuyere Blower
Recycled flue gas from cyclone
Flue gas and air to gasifier
Flue gas to stack
Solids from boiler flue gas cyclone
Solids from recycled flue gas cyclone
Solids from regenerator off gas cyclone
Start up kerosene to gasifier
Stack emissions
Fuel injection air
Mass fl
kg/sec
0.04
0.003
0.11
0.11
0.16

0.0006
0.16
0.01
0.002
0.01
0.01
0.50
0.03
0.02
0.02
0.10
0.50



0.0005
0.50
0.01
ow rate,
(Ib/hr)
(228)
(25)
(860)
(850)
(1,279)

(4.5)
(1,279)
(65)
(14)
(63)
(63)
(4,000)
(250)
(125)
(125)
(800)
(4,000)



(4)
(4,000)
(45)
Temperature,
°C ' (°F)
88 (190)



850 (1,560)

850 (1,560)


1,050 (1,920)
1,050 (1,920)










43 (110)


during December, 1975.   During this period, two fuel  oils were fired:  bit-
umen (vacuum bottoms) and residual  oil  (atmospheric bottoms).   Samples were
collected during seven  separate runs:   four residual  oil  gasification runs;
two bitumen gasification runs; and one  combustion/startup bitumen run.  The
field measurement program consisted of  on-site quantification of S02> sulfur

                                     1028

-------
trioxide (S03),  nitrogen oxides (N0x),  hydrogen sulfide (HrS),  total  parti-
culate and participate size distributions.   Flue gas was sampled and  analyzed
for N0x by Method 7, S0.9/S03 by Method  8 and H?S by Method 11  (5).   An Orsat
analyzer was used to measure carbon monoxide (CO), C0? and oxygen (02).
Particulate sampling was accomplished using a standard RAC train constructed
according to the procedures outlined in EPA Method 5 (5).  Due  to the positions
of the installed ports, eight point traverses were taken on two diameters
120 degrees apart.  The train was modified slightly to allow for sampling of
gaseous organic species (see below).  Particulate size distribution measurements
were  taken with a University of Washington eight stage instack impactor using
ungreased foil substrates.  A  single point was sampled isokinetically for
sufficient time (15 to 30 minutes) to collect a weighable quantity on each
stage.

      In addition  to collection of particulate and inorganic gases, organic
vapors were collected  in a  specially designed adsorbent trap placed between
the filter and the  impingers on the Method 5 train.  The polymer used in
this  adsorbent cell, Tenax, reportedly retains all organic gases larger than
C5.
                         LABORATORY ANALYSIS PROCEDURES

     Three general types of analytical procedures were applied to oil, flue
gas, and particulate samples collected during the field test program:  organic
functional group identification; trace element quantification; and surface
element and inorganic compound quantification.

ORGANIC FUNCTIONAL GROUP ANALYSIS*

     In this procedure, sample extracts are separated into eight fractions
by  liquid chromatography (LC), evaporated to dryness, weighted, redissolved,
and analyzed by infrared spectroscopy  (IR) (6).  Methylene chloride was used
to  extract oil and particulate samples whereas pentane was used to extract
organic vapors adsorbed on the Tenax polymer.

     Liquid chromatographic separation into eight fractions is accomplished
by  transferring the extract to an LC column and eluting sequentially with a
series of solvent mixtures.  After collection from the LC column each frac-
tion is reduced to dryness using a Kuderna-Danish evaporator and air evapora-
tion and then weighed to determine the amount of organic material in each
fraction.

     The dried fractions are then redissolved in methylene chloride and
subjected to IR analysis.  The IR spectra are then scanned for functional
group peaks.
 *These  analyses were  performed  by  Battelle  Columbus  Laboratories  under  sub-
  contract  and  by  the  Process  Measurements Branch  of  EPA.

                                     1029

-------
TRACE ELEMENT ANALYSIS*

     Stack particulate was analyzed for bulk elemental composition using low
precision (±200 percent) spark source mass spectrometry (SSMS).   This techni-
que is sensitive to 70 elements.  To calibrate the SSMS results, some elements
were quantified by higher precision atomic absorption (AA) spectroscopy.
Interference of organic ions with low atomic weight elements is  well known
in SSMS as are losses of volatile compounds.  Thus uncertainties of values
derived for light elements such as fluorine, sodium and sulfur may be higher
than the indicated precision.

SURFACE ANALYSIS

     A number of particulate samples were investigated for surface elements
and inorganic compounds by X-ray photoelectron spectroscopy (XPS or ESCA).
All samples analyzed by ESCA in this study were first scanned over the entire
electron binding energy range (broadband scan) to identify those elements
present in concentrations greater than 0.1 percent.  These broadband spectra
were then analyzed to yield surface concentrations of all identifiable
elements.  The compound forms of surface vanadium and sulfur are also of in-
terest in this study.  They were investigated by scanning the binding energy
ranges corresponding to the ejection of the 2p electron of vanadium and the 2p
electron of sulfur.  To supplement the bulk SSMS analyses, high  energy argon
ions were used to etch away surface layers exposing strata 20 to 100 ft deep.
The exposed sample layers were then rescanned over the entire binding energy
range and the resultant elemental concentrations were compared with the surface
values and the SSMS analyses.

                               FIELD TEST RESULTS

     Table 2 summarizes the stack emission rates of N0x, S02, S03, H2S, and
total particulate.  The average NO concentration of 53*5 ppm is  considerably
lower than the low end of the concentration range found for conventional oil-
and gas-fired boilers.  The measured emission rate is also about one-fourth
of the New Source Performance Standard (NSPS) for oil-fired boilers and one-
third of the NSPS for gas-fired boilers (7).  Furthermore, the relative in-
variance of the three measurements suggests a low correlation between the NO
emission rate and temperature, excess oxygen, bed stone history, and fuel.

     Several factors may contribute to the low absolute NO  emission rate.
The reducing atmosphere in the gasifier may severely inhibft oxidation of
fuel nitrogen.  It has also been suggested that CaC03 might aid catalytically
in the decomposition of NO or react directly with NO (8).  The presence of
nitrogen on the surface of some of the smaller particulate (which was noted
in the ESCA analysis) is consistent with this latter mechanism.   It is also
possible that reduced nitrogen species, such as ammonia (NH3), formed in the
gasifier, pass through the boiler without reacting.  It is more  difficult to
explain away the apparently small amount of NO formed by thermal fixation in
*This work was performed by Battelle Columbus Laboratories and Aculabs under
 subcontract to GCA and by Northrop Services under contract to EPA.

                                     1030

-------
                    TABLE 2.   SUMMARY OF STACK EMISSIONS
N0/ SO; S03 H;S Total
particulate
Run
1
2
3
4
5
fi

7
Fuel ppm lb/10" Btu ppm lb/10 Btu ppm lb/10 Btu ppri lb/10 Btu ppn lb/10 Btu
Fuel Oil 53.5 0.0851 292 0.6431 8.3 0.023 <0.05
Fuel Oil 45.7 0.0671 305 0.6191 7.9 0.0201 <0.05
Fuel Oil 0.23 2.6xlO~'4
Fuel Oil
Bitumen 58.4 0.0791 828 1.5624 11.1 0.0263
Bitumen and
Stone Feeding
Bitumen
0.0971
0.0561
0.063
0.0921
0.101

0.1046
0.1921

the boiler.  The use of the thermal fixation reaction is strongly affected by
boiler design and firing characteristics which cannot easily be evaluated.

     Two uncertainties are thus apparent.  What is the fate of the fuel bound
nitrogen which is not converted to NO ?  How will the rate of the thermal
fixation reaction be affected by the particular characteristics of the boiler
to be used in conjunction with the demonstration plant?

     The S02 emission rates for residual oil gasification were about 20 percent
below NSPS for oil-fired steam generators and correspond to retention of 80
percent of the fuel sulfur in the gasifier lime (7).  The higher S02 rate
measured for bitumen gasification is due to two factors:  the sulfur content
of bitumen is roughly 50 percent higher than in fuel oil and, more importantly,
stone present in the gasifier during this run had been unregenerated and unre-
plenished for some time because of clogging in the gasifier-regenerator transfer
duct.

     The primary source of particulate emissions from the CAFB is gasifier
bed stone which passes through the two internal  cyclones, the knockout baffle,
and the stack cyclone.  The NSPS for particulate emissions from oil-fired
boilers is 0.1 lb/106 Btu (7).  Table 2 shows that during oil gasification
two of the four runs produced emissions only a few percent below the standard.
The NSPS was exceeded during bitumen gasification and combustion; the final
bitumen gasification run exceeded the particulate standard by a factor of two.

     Two factors must be invoked to understand the variation in and magni-
tude of particulate emissions:  cyclone efficiency and fresh stone feed.
The cyclones used at the pilot plant were not designed specifically for the
CAFB and operated at an efficiency of only 50 percent.  The high emission
rate from Run 7 can be attributed to an unusually high fresh stone feed rate.
Fresh stone undergoes attrition and calcination as it enters the gasifier.
High stone feed rate normally occurs during start up (combustion) but was
employed during Run 7 to compensate for the buildup of saturated stone due
to the clogging of the gasifier-regenerator stone transfer duct.  The parti-


                                     1031

-------
culate sizing distributions measured during Runs  3  and  5  indicate  that a
substantial fraction of the particulate emissions is in the respi^able range
and hence of primary concern.   The large respirable fraction is typical  of
conventional cyclones and may be expected in emissions  from the demonstration
plant which will also employ cyclones for particulate control.

                           LABORATORY ANALYSIS RESULTS

     The laboratory program consisted of organic  and inorganic  analyses  of
stack particulate and organic analyses of flue gas, bitumen, and spent regene-
ration stone.   Stack gas emissions collected during bitumen gasification were
found to contain aliphatic hydrocarbons, alcohols,  ketones, aromatics  and
carboxylic acid salts.   The total  gaseous organic emission rate (C>5)  was 0.022
lb/106 Btu.  Particulate from this same run contained aliphatic hydrocarbons,
carbonyl compounds, esters, aromatics, phthalates,  phenols and  alcohols  to-
taling 7 x TO'5 lb/106  Btu.

     Elemental  analyses of these stack particulates indicated that the most
abundant species are:  calcium, 14.1%; sulfur, 3.9%; vanadium,  1%; silicon,
0.5%; sodium, 0.4%; magnesium, 0.3%; nickel, 0.2%; and iron, 0.1%.  Compari-
son of these values with EPA's Multi-Media Environmental  Goals  (MEG's)(9),
which are derived from TLV's, showed that only the vanadium emission rate is
of potential concern.  Compound identification by ESCA showed that particulate
vanadium exists as a mixture of oxides, V205, V02 and V203.

     Particulate from fuel oil gasification was found to contain aliphatic
hydrocarbons, carbonyls, esters, ketones, aldehydes, alcohols and carboxy-
lates.  The total organic particulate emission rate was approximately equal
to that from bitumen gasification.  The elemental composition of these parti-
culates is similar to the bitumen particulates but contains half as much
sulfur and roughly 0.15% of chlorine and fluorine.

                                   CONCLUSIONS

     The CAFB process was developed to control sulfur oxide emissions from
combustion of high sulfur heavy fuel oil.  The results presented here show
that this goal  can be met under proper system operating conditions.  Under
the upset mode in which gasifier stone becomes saturated, control effective-
ness of the process is impeded.

     Particulate emissions present the most pressing air pollution problem.
Even under normal operation the particulate emission rate, caused by attrition
of gasifier bed stone, approaches the NSPS for oil-fired units.  Improved
particulate collection efficiency is required and is being developed, in the
form of more highly efficient cyclones, for the demonstration plant presently
in the final design state  (2).

     Nitrogen oxide emissions from the CAFB are very low and appear to be
independent of fuel, gasifier stone history, and detailed operating conditions,
The reasons for this observation are not understood completely and additional
studies should be undertaken to uncover the fate of fuel  nitrogen.
                                     1032

-------
     Gaseous and participate organic stack emission rates do not appear to
present any significant problem.  Vanadium is the only trace element whose
emission rate is potentially harmful.   However, control Of total participate
emissions will significantly reduce vanadium output.

                                 ACKNOWLEDGMENTS

     This work was supported by the U.S. EPA, Industrial Environmental Research
Laboratory-Research Triangle Park, under contract number 68-02-1316, T.O.  No.
14.  The authors thank Dr. Graham Johnes, Mr. Z. Kowszun and Dr. Gerry Moss
and their staff at Esso for their cooperation during the field measurement
program; Dr. Peter Jones of Battelle for the loan of the gas adsorption column;
and John Langley and Stephen Brenan of the GCA staff for their assistance
in the sampling effort.

                                   REFERENCES

1.    Craig, J.W.T., G.L. Johnes, Z. Kowszun, G. Moss, J.H. Taylor, and D.E.
     Tisdall.  Chemically Active Fluid-Bed Process for Sulphur Removal During
     Gasification of Heavy Fuel Oil - Second Phase.  EPA-650/2-74-109, U.S.
     Environmental Protection Agency, Research Triangle Park, N.C., November
     1974.

2.   Foster-Wheeler Energy Corporation.  Chemically Active Fluid Bed Process
     (CAFB) Preliminary Process Design Manual.  U.S. Environmental Protection
     Agency, Research Triangle  Park, N.C.  EPA Contract Number 68-02-2106.
     December 1975.

3.   Werner, A.S., C.W. Young,  M.I. Bornstein, R.M. Bradway, M.T. Mills and
     D.F. Durocher.  Preliminary Environmental Assessment of the CAFB.
     U.S. Environmental Protection Agency, Research Triangle Park, N.C.
     Contract Number 68-02-1316.  Task Order Number 14.  June 1976.

4.   Dorsey, J.A., C.H. Lochmuller, L.D. Johnson, and  R.M. Statnick.  Guide-
     lines  for  Environmental Assessment  Sampling and Analysis Programs.  U.S.
     Environmental Protection Agency, Research Triangle Park, N.C., March  1976.

 5.   Standards  of Performance  for  New  Stationary  Sources.   Code  of  Federal
      Regulations, 40  CFT,  Part  60,  May  23,  1975.

 6.   Jones, P.W., A.P.  Graffeo, R.  Detrick,  P.A.  Clarke,  and  R.J.  Jakobsen.
      Technical  Manual  for Analysis  of  Organic  Materials  in  Process  Streams.
      EPA-600/2-76-072,  Environmental  Protection  Agency,  Research Triangle
      Park,  N.C.,  March  1976.   81  pp.

 7.   Title  40-Protection  of the Environment.   Part  60  Standards of Performance
      for New Stationary Sources.   Federal  Register, 36(247):24876,  December 23,
      1971.

 8.   Jonke, A.  A.,  E.L.  Carls,  R.L.  Jarry,  M.  Haas, W.A.  Murphy, and  C.B.
      Schoffstoll.  Reduction of Atmospheric Pollution  by the  Application  of
      Fluidized-Bed Combustion.   ANL/ES-CEN-1001,  Argonne  National  Laboratory,
      Argonne,  Illinois, June 1969.   65 pp.

                                     1033

-------
9.   Eimutus, E.G. et al.   Source Assessment, Prioritization of Stationary
     Air Pollution Sources.  Model Description.   EPA 600/2-76-032'a,
     U.S. Environmental Protection Agency, Research Triangle Park, N.C.,
     February 1976.
                                     1034

-------
                                                                           21-4
                             NO  CONTROL BY ABSORPTION
                               X

                                    G.  Sakash*

ABSTRACT
     The, elimination  ofa  nitrogen oxides ^ume4 exhausted ^fiom a. uAanium
nccov&iy operation -U>  accompLti>kcd by the, che.rn.ical. conveAAi.on o£ nitrogen
oxides to nitrogen dioxide, uAing puAe oxygen, and a AubAequcnt ab&o?iption
0$ the, nitrogen  dioxide,  by ia.pi.dty neciAcuJLating waten to faofw nWtic acid.
An exce^-6 ofa pome oxygen fad ^fiom a cryogenic &y&tcm iA tued to compJL&tnly
convafut aUL o^ the, nitsiogun oxA.de, QOA to nitAogun dioxi.de,>  The, exce-ii  oxy-
gen it> the, onty  QO*> UiLtwateJiy dit>c.haSLge.d to the, atmoApheAe..  The, nitsiic. acid
{fO.vne.d by the, ab^oftption pioceA* /u ?ie.cinc,uJtate,d untit the. deAisind  c,on.ce,ntna-
tion u> obtained.  The, acid u> then, pumped back to the 4cAap ticcovcfiy &y&tcm
faon. tieube.

                                   INTRODUCTION

     The high toxic hazard rating associated with oxides of nitrogen  (NO  )
are of major concern  to  all manufacturing facilities where the potential
of NO  fume emission  exists.

      It  is  the  intent of this paper to describe a process which  has  success-
fully eliminated  the N0x fumes emitted from the dissolution  of uranium dioxide
in  nitric  acid.
      The  process  chemically converts NO  to nitrogen dioxide  (N02)  using
 pure  oxygen;  the  N02 is concurrently absorbed by rapidly  recirculated scrubbing
 water to  form nitric acid.  The process utilizes an excess  of pure  oxygen to
 completely  convert all  of the NO  gases to N02 which is very  soluble  in  water.
                                 X
      This  process  has been in use at the General Electric  Company  in  Wilming-
ton,  North Carolina,  for three years.  It has completely eliminated all  visual
N02  fumes  and  reduced the N02 emission level to less  than  5  ppm  which is the
current  Threshold  Limit Value (TLV) as published by ACGIH.

                                    DISCUSSION

      In  the reclamation of scrap uranium dioxide, the  scrap  material  is
dissolved  in nitric acid.  During this dissolution process,  large  quantities
of gaseous vapors, in the form of nitric oxide  (NO) and N02,  are evolved.
These  gases are  identified as N0x and are exhausted as a brownish  plume,
creating a very  undesirable atmospheric pollution condition.
*General Electric Company, Wilmington,  North Carolina.

                                      1035

-------
     The TLV for NO, as published by ACGIH, is listed at 25 ppm or 30 mg/m3
of air.  The N0x fumes exhausted during the uranium dioxide dissolution pro-
cess is far in excess of this; therefore, it was necessary to devise a process
which would reduce this emission to an acceptable level.  This was accomplished
by the utilization of a device (Figure 1) in which oxygen and water convert
the NO  fumes into nitric acid.  This device consists of a conversion chamber
      X
where the NO  are converted to N02 and an absorption chamber where the NO
                                                                         2
is converted to nitric acid.  An excess of pure oxygen, fed from a cryogenic
system, is used to completely convert all of the NO  gases to N02.   The ex-
cess oxygen is the only gas ultimately discharged to the atmosphere.

     The absorption column, constructed of 317 stainless steel, measures 20
feet high and 4 feet in diameter and is divided into three compartments.  The
top five-foot section is used as the conversion chamber where the chemical
conversion of the N0x to N02 takes place.  The oxidation reaction occurs
readily at room temperature without a catalyst.  The middle ten-feet of the
column is used as a mixing chamber.  This portion of the column is filled with
ceramic Intelox packing.  In this portion of the column, the water and N02
mix to form nitric acid.  The bottom five-feet of the column is the acid
accumulation chamber.

     The scrubbing water, flowing at a rate of 250 gpm, is continually re-
circulated through the absorption column until the acid concentration is
increased to a reusable product.  The heat of reaction is controlled by
passing the recirculation solution through a stainless steel heat exchanger.
Continuous monitoring of the acid concentration is done at a main control
panel.  After the required acid concentration is reached, the acid is pumped
back into the dissolvers for reuse.

     The off gas consisting mainly of excess oxygen is drawn up through a
high efficiency filter housing, then through a Buffalo gas absorber scrub-
ber, and finally through another bank of high efficiency filters.  A Leeds
and Northrup N0x emission monitor is installed in the off gas stack and
continuously monitors and records the visible NO  fumes being exhausted out
the stack.  An alarm automatically sounds if a malfunction in the absorption
system occurs.

     Because the system is used in the reclamation process of radioactive
uranium dioxide scrap, constant monitoring for radioactive material is re-
quired.  This is done both on the recirculation solution by periodic labora-
tory analysis and on the exhaust gases by continuous stack exhaust monitoring.

     Chemically, the major reactions which take place during the absorption
are as follows:

          3NO + 1% 02 	»  3N02

          3N02 + H20  	» 2HN03 + NO
          2ND + ~\h 02 + H20        2HN03
                                    1036

-------An error occurred while trying to OCR this image.

-------
     The process (Figure 1) occurs as follows:

1.   Uranium dioxide is dissolved in nitric acid in the dissolution chambers.

2.   NO  gas generated by dissolution leaves the top of the dissolution cham-
     bers and enters the absorption column at A.

3.   Pure oxygen from the cryogenic oxygen system enters the absorption cham-
     ber at B.  When additional oxygen is required, it is automatically
     added through the pressure sensing V2, C2, T2 system.

4.   A water-acid mixture recirculating through pump, P, and heat exchanger,
     E, enters the absorption chamber through a spray nozzle at C.  Roto-
     meter, R, indicates flow.

5.   The excess oxygen is exhausted as spent gas through D and goes through
     absolute filters prior to exhausting to the atmosphere.

6.   As the concentration of the nitric acid formed is increased, it is
     sensed through T3; and when the desired concentration is reached, it is
     automatically pumped through control valve V3 into the dissolvers.

7.   Make-up deionized water enters the acid storage tank at G.  A level is
     automatically maintained by a float controller at F.

8.   A Leeds and Northrup N0x emission monitor, I, senses visual emission and
     records it at a remote control panel.

9.   A monitor, H, continuously monitors the off gas for radioactive
     particulate.

     It is conceivable that this system could be adapted to any facility
where control of NO  emission is a problem.
                                     1038

-------
                                                                              21-5
                    DEVELOPMENT  OF  A LOW EMISSIONS PROCESS
                      FOR ETHYLENE  BICHLORIDE PRODUCTION

           W.S.  Amato, B. Bandyopadhyay, B.E. Kurtz,  and  R.H. Fitch*
ABSTRACT
     Thii>  wofik coveAS the initial phase ofa the. development of, a low- emissions
ethylenc  oxyhydAochloAA.nation pAocess faA pn.oda.cA.nQ  1,1 dichloAo ethane
{ethylene dichloAide) .

     Firstly,  experimental woA.k on. an existing pilot plant scale, once.-thsiou.gk
ph.oc.Ui> was  used to obtain basetine emission data.  in mass o{> hydJLOc.aA.bon plus
ethylene  dic.htotu.de. pen. mass  Ofj hydA.oge.n cktofcide.  ^ed aj> a function o^  ie.acJ:o>i
tmpeAotuJKi  and peAc.nnt exce^-6 zthylzne. to the. nz.act.on.; and to ie4c£ve  potan-
&jjJi pnobtwA  which may OAAAZ. -in a >ie.c.ycle. opoAation, viz., whether catibon
monoxide.    oxtdize-d to cafibon di.oxA.dn by the. n.ii-
nation rniakt be. ne.c.u&an.y -in  oideA to 4aue &thyle.ne.  uoltie in the. tie&ctot
Jifa a c.ah.bon  monoxx.de oxidation unit weAe. ne.c&>Aariy ,  and whether, the. Ae.actoA.
wouJLd oxidize, methane, and ethane., theAe.by p>ie.ve.nting tkeJJi btu/dup -en the.
     Secondly,  tke. e.xi^ting onc.e.-thnou.gh piJLot plant IMOA  convened to a le.-
cycle. opeAation which then {^u.nctione.d 4ucce4-6 fatly and  yi.ztde.d em-ibtxion data
in moi4 Oj$  hydA.ocoA.bon plu* e.thyle.ne. dichloAide. pejt mo64  o& hydA.oge.n chic Aide.
fae.d 06 a  function o{, n.o,actoA tempeA.atuA.e. and pe.Ace.nt e.xceA& eJ.hyle.ne. to the.
A&actoA.   In poAticalaA, the. pn.oje.ct objective, o^ A~e.du.cing by 90 peA.ce.nt the.
hydAocaAbon piuA ethyte.ne. dichloAide. emiA&iont, ^Aom an  ethyle.ne. oxyhydAo-
chloAination pAoceAA, thAough the. Ae.cycting o^ A&actoA  o^-gat>eA, hai> been
            demom,tAate.d.
      Thirdly, vaAioitt,  operating di^icaltie^  weAe. aA-t>ej>Ae.d which womld be.
 impontant fan. fautuAe, control appLication*  and 4co£e-up e.^faAtA,  viz.,  the.
 incAe.aAe.d sensitivity  o^ the. pAocus to upsets in filom, tempeAotuAu, and
 conce.ntAotA.ons .

      Lastly, economic  analyses aAe pAesented  to dmonstAote the  competitive
 position oft the improved pAocess.
*W.S. Amato,  B.E.  Kurtz and R.H.  Fitch, Allied Chemical  Corporation, Syracuse
 Technical  Center,  Solvay, New  York.
 B. Bandyopadhyay,  Allied Chemical  Corporation, Syracuse Works, Solvay,  New
 York.


                                       1039

-------
     ThLt> Monk wo6 p&i&ofme.d in fiulfi'illme.n.t o& Project Number R0AP 27 AXM  20,
Contract NmbeA 68-02-1835, by Alti&d Chmic.al Corporation, lnd.(U>&uaJL Chwi.-
eo£6 V• C?HtfCl2 + H?0, arid

   cracking of ethylene di chloride:

                            2 C,H4C12 -- * 2 C2H3C1 + 2 IIC1

The overall reaction in this balanced process is:

                2 C2H4 + C12 + 1/2 02 -- * 2 C?H3C1 + H?0.

The oxyhydrochlori nation step is the most technically difficult part of  the
overall process and has had a profound impact on the industry since its  de-
velopment.  The first commercial ethylene oxyhydrochlorination process appear-
ed in 1964 and within the next few years nearly 90 percent of the classical
process for vinyl chloride from acetylene was displaced.

EXISTING PROCESS

     There are currently two basic variants of ethylene oxyhydrochlorination
processes, those employing fixed catalytic beds and those employing fluid
catalytic beds.  Most of the major producers employ a fluid bed.  Ethylene
(CzHiJ, hydrogen chloride (HC1), and air are introduced into the bottom  of
the fluid bed where the oxyhydrochlorination reaction takes place.  The  heat
released by the reaction is substantial, so the reactors are equipped with
internal cooling coils which are submerged in the fluid bed.

     Downstream product recovery involves cooling of the reactor exit gases
by either direct quench or a heat exchanger, followed by condensation of
the ethylene dichloride (EDC) product and water by-product which are separat-
ed by decantation.  The remaining gases still contain 1 to 5 percent by
volume of EDC, so they are further processed in a secondary recovery system
employing either solvent absorption or a refrigerated condenser.  The off-
gas is then vented to the atmosphere.  (See Figure 1.)
                                    1040

-------
f
\

1
V


x o
O h-

                                                         to
                                                         S-
                                                         o
                                                         o
                                                         o
                                                         X
                                                         o

                                                         O)

                                                         O)

                                                        ^
                                                        _c:

                                                         0)

                                                         en
                                                         X
                                                         
-------
      The  vent  stream  contains 85  to  95 percent nitrogen from the air employ-
 ed  as  a source of  oxygen  for the  oxyhydrochlorination reaction.  It also
 inc"[udes_  a  variety of other compounds which originat_e__ejther iin_the feeds
to the reactor or as by-products of the oxyhydrochlorination,  such  as  carbon
dioxide, carbon monoxide, EDC,  C2H4, methane (CHi+), ethane (C2H6),  ethyl
chloride and absorber solvent.

     The aspect of most current ethylene oxyhydrochlorination  processes
which contributes the most to hydrocarbon (HC) emissions is the use of air
as the source of oxygen.  The inerts (nitrogen)  accompanying the air neces-
sitate a once-through process.   If pure oxygen were used it would be possible,
at least in theory, to recycle  the reactor off-gases.   Only one plant  cur-
rently employs pure oxygen, but it does not recycle the reactor off-gases.

PROPOSED PROCESS

     Recycle of reactor off-gases will  eliminate the need  for  a secondary
recovery system and the resultant losses of solvent which  typically amount
to 0.001 ton per ton of EDC.   It will also eliminate unreacted C2H4.  Some
means will have to he employed  to limit build-up of feed impurities (CHa
and C2H6)  and by-products (C02,  CO, and ethyl  chloride), probably a combina-
tion of removal by chemical or  physical  processing and  a purge to atmosphere.
The purge, however, will amount to only a small  fraction of the vent stream
which characterizes present processes and it is  certainly  reasonable to expect
a 90 percent reduction in HC emissions  over present processes.

     The detailed technical approach currently visualized  may  best  be  explain-
ed by reference to Figure 2.
     A fluid bed oxyhydrochlorination reactor employing a conventional copper
chloride on alumina catalyst is fed with C?Hi+, HC1, and recycle gas contain-
ing added oxygen.  The C?hK is in 5 percent excess over that required by the
reaction stoichiometry, the oxygen in 60 to 80 percent excess.  Near the top
of the fluid bed a small stream of chlorine, amounting to 2 to 5 percent of
the HC1 feed, is added to convert the unreacted C?H4 to additional EDC.

     The gases leaving this reactor contain the products of reaction, EDC and
water, and other substances originating as unreactive impurities in the feeds
or as by-products.  The gases pass to a quench column where they are cooled
by circulating water, then to a condenser where EDC and water are recovered.
These are separated in a decanter and the non-condensable gases flow through
a recuperative heat exchanger where they are re-heated.

     The hot gases pass to a fixed bed oxidation reactor (shown dashed) em-
ploying a platinum on alumina catalyst where the CO will be converted to
C02 by reaction with the excess oxygen from the oxyhydrochlorination reactor.
The gases exiting the reactor are cooled against the entering gases in the
recuperative heat exchanger and passed to a booster compressor.  It may be
possible to eliminate the separate CO oxidation step if it can be shown that
CO can be oxidized to C02 by recycle to the oxyhydrochlorinator.  If this is
possible, then the need for post-chlorination is obviated.


                                    1042

-------
LJ
>~
CJ
LU
cc:
                                                                                       o.
                                                                                       ex
                                                                                      o:
                                                                                      I—
                                                                                      oo
                                                                          > o
                                                                                   ca
                                                                                   C£
                                                                                   o
                                                                                   oo
                                                                                   D3
/
^
_


V
\
,1 	
                                                                                     oe.
                                                                                     eC
                                                                                     Q_
                                                                                     UJ
                                                                                     CO
                                                                                                                           cu
                                                                                                                           o
                                                                                                                           O
                                                                                                                           s_
                                                                                                                           CL
 o
•I—
-M
 ro
                                                                                                                           S-
                                                                                                                           o
                                                                                                                           o
                                                                                                                           o
                                                                                                                          -a
x
o
                                                                                                                           0)
                                                                                                                           c
                                                                                                                           OJ
                                                                                                                          CD

                                                                                                                          TD
                                                                                                                          OJ
                                                                                                                          I/)
                                                                                                                          O
                                                                                                                          Q_
                                                                                                                          O
                                                                                                                          s-
                                                                                                                          o
                                                                                                                          M-
                                                                                                                          (1)
                                                                                                                          CU
                                                                                                                          CD
                                                                                                                          i-
                                                          1043

-------
     The majority of the discharge from the booster compressor is  recycled
to the oxyhydrochlorinator.   This recycle gas will  be predominately CO? with
the process as described here.   A small portion of the recycle gas is  purged
in order to limit the build-up  of unreactive impurities which  enter the sys-
tem with the reactor feeds,  for example:   CHit and C2He which  are  typical
impurities in the C2Hl+.

     This purge stream is first cooled in a recuperative heat exchanger and
then further cooled in a refrigerated vent condenser which removes most of
the EDC and ethyl chloride.   These are collected in a receiver from which
the purge gases, consisting of 80-90 percent C02 and 10 percent CH4 and C2H6,
are discharged through the recuperative heat exchanger and may be disposed
of via incineration.

     The net effect of this process will  be at least a 90 percent reduction
in HC emissions compared to typical present processes.

OBJECTIVE AND SCOPE OF PROJECT

     The objective of the project was to demonstrate that emissions from the
production of EDC by oxyhydrochlorination of 02^ could be reduced in a modi-
fied process by at least 90 percent from the levels encountered with typical
existing processes.  The modified process was to employ recycle of reactor
exit gas, oxygen feed, and whatever additional processing steps were deter-
mined to be necessary to control build-up of by-products in the recycle stream
and was to be economically competitive with present-day processes.  The pro-
cess performance was to be evaluated on a laboratory scale and a  preliminary
study of technical and economic feasibility was to be carried out.

              PRELIMINARY EVALUATION OF ALTERNATIVE APPROACHES
     The preliminary evaluation of alternative approaches considered the
following process alternatives:  oxygen feed with vent gas recycle on the
existing process; catalytic oxidation of vent gas; and thermal  incineration
of the vent gas.  The results can be summarized as shown in Table 1.

     The results of the economic evaluations of these alternatives show
clearly that the oxygen addition with vent gas recycle on the existing
process is by far the most economic choice.   The estimates are  based on a
700 MM Ib/year EDC plant.

                            EXPERIMENTAL PROGRAM
     The main chemical  reactions which occur in fluid-bed oxyhydrochlorination
of CHt  are:
                                    1044

-------
                       TABLE  1.   PROCESS  ALTERNATIVES

Economic Consideration
Net Capital Increase
Net Increase in
Manufacturing Cost
Unit Cost Increase,
it/lb EDC

Oxy Feed -
Recycle
($1 ,300,000)*
$377,000/y
0.054
Process
Thermal
Incineration
$1,750,000
$695,000/y
0.099

Catalytic
Oxidation
$2,250,000
$985,000/y
0.141

  *  Figures  in  parentheses  refer  to  savings  in  capital.



Principal Reaction:

             C?H4 + 2 HC1 + 1/2 02 - *C2H4C12 + 2 H20         (1)
                                Cupric Chloride
                                   Catalyst
Main Side Reactions:
                       C2H,f + 2 02          » 2 CO + 2 H20            (2)


                       C2H^ + 3 02 -        » 2 C02 + 2 H?0           (3)


                        C2H4 + HCI^         »C2H5C1                  (4)


                         2 CO + Op          » 2 CO 2                   (5)


QUESTIONS NEEDING RESOLUTION BEFORE FINAL DEVELOPMENT OF A RECYCLE PROCESS

Question 1:  Will CO build-up in the recycle system reach levels which will
             cause explosive conditions?

             Will the ratio of C02/C0 remain at about 2 in the recycle system
             as it does in the once-through system?

             Can the reactor system effectively oxidize CO to C02?
                                    1045

-------
Question 2:  If CO is not oxidized in the reactor, but is oxidized in a
             separate unit, how do we avoid the loss of CoH^ value by oxida-
             tion?

Question 3:  Can post-chlorination effectively work on the exiting reactor
             gases to convert unreacted C^Hi4 to EDC?

Question 4:  What is the effect of the reactor system on HC impurities, such
             As CH^ and C2H6, which might occur in the C2H1+ feed?

     With these questions before us, the scope of an experimental program
was evolved as follows:

SCOPE OF THE EXPERIMENTAL PROGRAM

1.    Once-Through Runs

     (a)  Baseline runs at 230-240°C with 5, 10, 15 percent excess C2H,4 and
          60-70 percent excess 02 (as air).
     (b)  Runs with CO and C02 additions.
     (c)  Hydrocarbon additions:  CH^ and C2H6.
     (d)  Post-chlorination work.

2.    Recycle Runs (02 Feed)

     (a)  Runs at 230, 240, 250°C with 5-20 percent excess C^ and 50-115
          percent excess 02 with vent rates of less than 10 percent of total
          off-gases.

RESULTS

     The results of correlation of emissions data from the once-through pro-
cess are:

   (Mean - 231°C) EM0 = 2.47(lQ-?) + 1.35(10-3) XSETH                 (6)

   (Mean - 240°C) EM0 = 1.32(lQ-?) + 1.35(1Q-3) XSETH                 (7)

     Where

   EM0 = Emissions in    VQ ur-i—  f°r once-through operation.


XSETH = Percent excess ethylene to reactor.

     The results of correlation of emission data from the recycle process are:

   (Mean - 230°C) EMR = -1.08(10^) + 2.05(1Q-I+) XSETH                (8)

   (Mean - 242°C) EMR = -G.ASdQ-1*) + 2.05(10-4) XSETH                (9)

   (Mean - 251°C) EMR = -8.0300-14) + 2.05(1Q-'4) XSETH               (10)


                                    1046

-------
     Where
   CM    c •   •     •   kg(HC + EDC) f        -,        . .
   EMp = Emissions in -|/ — rrpi — - for recycle operation.
XSETH = Percent excess ethylene to reactor.

     Comparison of the emissions from the once-through and recycle operation
at the same reactor temperature and percent excess ethylene is shown in Table
2.  Table entries show % reductions in emissions -n of recycle process com-
pared with once-through process at similar values of % excess ethylene and
reactor temperature, T°C.
       TABLE 2.   COMPARISON OF EMISSIONS DATA, RECYCLE AND ONCE-THROUGH
                Table entries show % reductions in emissions -n
                of recycle process compared with once-through
                process at similar values of % excess ethylene
                and reactor temperature, T°C.
                       % Excess Ethylene   230°C   240°C
5
10
15
20
97.2
95.1
93.6
92.4
97.4
94.2
92.3
91.1

     This result is also shown in Figure 3.  Thus, it can be seen that the
original  objective of this investigation, viz., the reduction by 90 percent
of the emissions from the EDC process through implementation of a recycle
scheme has been positively demonstrated.

                                 CONCLUSIONS
EXPERIMENTAL RESULTS OF ONCE-THROUGH OPERATION

     •    Crude ethylene dichloride (EDC) purity varied inversely
          linearly with reactor temperature.

                                    1047

-------
    100
   CO
   52;
   o
   H
   CO
   CO
   w

   ^
   H

   55
   O

   £
   o
   t>
   o
   w
   w
   o
   f£
   w
     90
0
2
                FOR  R/R+V ^0.9

SUPERFICIAL VELOCITY       1.2  ft/sec

        EXCESS OXYGEN -  50fo TO  100^  4=

                R = RECYCLE  RATE

                V = VENT RATE
6
  8     10    32    1



% EXCESS  K'HIYLKNE
                            16    18
                                                                20
          Most  commercial plants  opcrtite tit  50-70%  excess

          air and 5-10% excess C2II.} .
Figure  3.  n - Percent reduction in  emissions recycle  compared to once-through

          vs. % excess ethylene with  reactor temperature as parameter.
           Kilograms of  HC plus EDC per  kg of HC1  emissions could be

           correlated as a linear function of excess  C2Hi+ to the re-

           actor with reactor temperature as parameter.

           Ethylene efficiency decreases with higher  operating tempera-

           ture due to increased oxidation of ethylene.
                                  1048

-------
      •    Addition  of CO  to  the  feed  gases  indicated  that  CO was
           oxidized  to C02; also  large additions  of C02  to  the feed
           gases  did not adversely affect the  reaction.
      •    Methane passed  through the  reactor  system unchanged and
           C2H6 was  significantly reduced (80  percent  in this case)
           over inlet values.
      •    Post-chlorination  of excess C2H4  in the  reactor  exit gases
           was feasible provided  chlorine was  injected into the gas
           phase  and not the  catalyst  itself.

 EXPERIMENTAL RESULTS OF RECYCLE  OPERATION

      •    Crude  EDC purity varied inversely linearly  with  reactor
           temperature, but was a few  tenths of a percentage point
           lower  than the  once-through purity  at  the same temperature.
      •    The C2H4  efficiency  for the recycle operation was higher
           than that for the  once-through process due  to the recycling
           of off-gases.
      •    Ethylene  efficiency  decreases  with  higher operating tempera-
           ture due  to increased  oxidation of  ethylene.
      •    All runs  of this report involved  the recycling of greater
           than 90 percent of the reactor off-gases.
      •    Kilograms of HC plus EDC per kg of  HC1 emissions could be
           correlated as a linear function of  excess C2H4 to the reactor
           with reactor temperature as parameter.
      •    The recycle operation  is more  sensitive  than  the once-through
           operation and requires closer  control  for successful  operation.

 COMPARISON

      Comparison  of  the emissions data for once-through  and recycle operation
 indicates  that the  percent reductions of HC plus EDC  emissions have exceeded
 90  percent for all  runs conducted in  this work.
      It was  possible  to develop  a  computer  simulation program,  based on
 actual experimental  data, which  could  duplicate  vent  gas compositions and
 superficial  velocities at given  reactor  temperatures  and pressures for
 given  excess CaH^/oxygen  values  to  the reactor.  Based  upon  the optimum expe-
 rimental conditions  from  recycle  operation, viz.,  230°C, 5-7  percent excess
 C2Hit  and 50  percent  excess oxygen,  together with this simulation program, it
 was found  that the  recycle process  was competitive  with  existing processes to
 within a tenth of a  percent  of total  incremental  capital and  manufacturing
 costs.

      The work of this  report has  revealed that the  newly-developed recycle
 process is much more  sensitive to  upsets  than  the conventional  once-through
 pjigcess.

     Before a commercial prototype of the proposed  process  can be  designed,
these sensitivity factors  must  be more fully explored along with the  estab-
lishment of well-defined requirements necessary for stable  operation.   Thus,
it is recommended that further  work be performed  on the newly-developed
recycle process  through the  use of computer  control technology.

                                     1049

-------
          SESSION  22
AIR QUALITY AND EMISSION  TRENDS

    ChaiVima.yi'  R.E.  Neligan
Environmental Protection  Agency
              1051

-------
                                                                             22-1
        THE  IMPACT  OF EMISSIONS CONTROL TECHNOLOGY  ON PASSENGER
               CAR HYDROCARBON EMISSION RATES  AND  PATTERNS

                                    F.  Black*

ABSTRACT
     A bnJLu-h kiAtony o& the. Ae.guJLation o& pat>£e.ngeA cent hydAocaAbon mi&&
tnciading the.  contAol. AtjAtzmA and de.vi.ceA utitize.d &OA e.mi^^i.on Ae.daction, AJ>
pAeAe.nte.d.  ThAne.  typeA Of} e.mit>Ai.oni> cum. diAcuAAe.d'-  cAankca&e. ve.ntsULati.on,
e.vapoAative. ga&otine.,  and combustion ex^iaaAi,  OnJLy e.x.hauAt and e.vapoAative.
          aAe.  o^ Asignifai-cance, and d&taitad kydAoc.an.bon patt&inA ^on. th&>£ oJiAi.on&  and the. degree o{, control
            hydA.ocaA.bont> OA gfioa.p& o^ hydAocaAbonb  JJ>  vafuabte. with the. contAot
         HydAocaAbon emiAb-ion. patteAnt, afte. pAeJ>e.nte.d at, a function ofi contAol
           a  va?u.&ty o£ catalyst and noncatalyAt  ve.hic£e^>.   1ncAe.cu,e. -in
      wth ve.hicte. ag
-------
                             EMISSION TRENDS

     In 1971, the National Petroleum Council projected decreases in annual
emissions from motor vehicles as shown in Figure 1 (1). These estimates were
based on emission rates predicted using the 1968 Federal Test Procedure and the
promulgated schedule of emissions reduction. The current test procedure would
increase these estimates, but the trends illustrated are valid.  The currently
utilized urban driving simulation and analytical procedures are considered more
accurate than those used previously.  One must be careful when evaluating
historical data.  Driving patterns typically used in vehicle testing have been
changed and these changes affect the emissions significantly.

     Table 1 illustrates the magnitude of control called for by the legislation.
                 TABLE 1.  EXHAUST EMISSION REDUCTION
                         Hydrocarbons,
                         g/mi.   g/km.
                    Carbon Monoxide,       NOx
                      g/mi.  g/km.    g/mi.  g/km.
Pre-1968 vehicles
(uncontrolled)*

Statutory levels
(highest degree of
control called for)
 9.1   5.7


0.41   0.25
  97     60.2     3.34   2.07
 3.4     2.11     0.4    0.25
Percent reduction
95.5
96.5
89.1
*low altitude, 49-state passenger cars (excluding California) (calender year
 1976 values) (2).

The emission rates for uncontrolled vehicles were derived from actual tests with
320 vehicles, 1963-1967 model years, using the current Federal Test Procedure
(2, 3).

     Manufacturers responded to this call for reduction in emissions with many
engine modifications and varied emission control packages.  They have been
effective in reducing emissions and, although not called for by law, many
production engine groups are already emitting exhaust hydrocarbons at rates less
than the statutory 0.41 g/mi level in low mileage vehicles.

     In pre-emission control automobiles, there were three major areas of
hydrocarbon emission:  combustion exhaust, crankcase blowby, and evaporative
loss from the fuel system.    Table 2 illustrates the relative contributions of
each.
                                     1054

-------
    cc
    <
    LU
    >
    CC
    Ul
    Q.
    CO
    o
    CO
    CO
    CO
    CO
    LU
              CARBON
             MONOXIDE
                               TOTAL VEHICLES IN USE
                                               HYDROCARBONS
            NITROGEN OXIDES

                           PARTICULATE MATTER
                  1960      1965
         10
1955
1980     1985
              Figure 1.  Mobile source annual  emission  rates  (1).
CONTROL OF HYDROCARBON  EMISSIONS

     The advent of emission  control has eliminated crankcase blowby emissions  J
low mileage cars,  and exhaust  hydrocarbons have been reduced greater than 95
percent for many vehicles.   Crankcase blowby has been eliminated with Positive
Crankcase Ventilation Systems.  These systems provide for circulation of air
through the crankcase and  for  drawing the circulated air and blowby gases into
                                    1055

-------
                  TABLE  2.   PASSENGER  CAR  HYDROCARBON  EMISSIONS
                                Exhaust        Blowby          Evaporative
                             g/mi.  g/km.    g/mi  g/km.       g/mi.   g/km.
  Pre-control  *               9.1     5.7     4.1   2.5        2.53     1.57
  Percentage  of  total        57.9            26.1              16.0
  *     Low  altitude, 49-state  passenger  cars  (2)

the intake manifold, where they are carried to the combustion chambers.  The
systems perform very well when maintained, but are subject to total failure if
the critical control valve becomes clogged.

     Engine modifications, catalytic converters, and combinations of both have
been used to reduce exhaust emissions.  Generally the engine modifications
relative to exhaust hydrocarbon emissions permit operation on the lean side of
the stoichiometric air-fuel ratio.  Lean operation is attractive from an emis-
sions standpoint as illustrated in Figure 2.  Emissions of hydrocarbons (HC), CO
and NO  are favored by lean operation.  Engine performance can suffer under lean
operation, however, and most of the modifications are designed to maintain
drivability throughout normal  driving patterns.  The catalyst systems are
designed to complete combustion of the hydrocarbons emitted from the engine to
C02 and H20 prior to their being emitted from the tailpipe.  Tim's permits
operations at air-fuel ratios  favoring performance (12.5 best power, 15.7 best
economy).  The variety of catalyst systems employed is numerous and no attempt
will be made to detail them.  Hydrocarbon emission patterns from several systems
are presented later in the discussion.

     Evaporative emissions have been reduced by using adsorption-regeneration
systems which use canisters of activated carbon to trap fuel tank and carburetor
vapors and hold them until such time as they can be directed into the induction
system for burning in the combustion chamber. These systems first appeared on
motor vehicles in 1971.  The standard of 2 g/test appeared to be easily met by
the control systems utilized.   The test procedure adopted by the federal certi-
fication laboratory (charcoal  canister technique) (5), however, has been found
to dramatically underestimate  these emissions.  More accurate procedures (Sealed
Housing for Evaporative Determination, SHED,  technique) (6) indicate actual
emissions from post-control vehicles in the field to be near 12.0 g/test with
about 12.1 g/day diurnal  emissions (7).  This equates to about 1.76 g/mi (1.09
g/km) (2).  Ambient conditions (temperature and pressure) dramatically effect
evaporative losses.  Tests using the SHED technique with post-control vehicles
in Denver gave 34.8 g/test with about 47.2 g/day diurnal (8).  This equates to
5.51 g/mi (3.42 g/km).  This is probably due  to the lower atmospheric pressure
and different fuel composition used in the Rocky Mountains.  Exhaust hydrocarbon
emissions are also somewhat higher at high altitudes.

                                      1056

-------
                                  14    16    18

                                 AIR-FUEL RATIO
        Figure 2.  Effects of air-fuel ratio on exhaust composition (4).

     Heywood effectively projected the significance of the higher than expected
evaporative emissions on outyear aggregate emission rates as illustrated in
Figure 3.   The model  used by Heywood assumes deterioration of the crankcase
emission control  systems as predicted by Voelz, et al.  (9). It also assumes
applicable exhaust standards are met during the life of the vehicle.  The
potential  importance of evaporative emissions in outyears is readily apparent.
The Environmental Protection Agency (EPA) is currently revising the evaporative
emissions  certification procedure to incorporate the SHED technique for the
1978 model year.   Curve B is thus the most accurate prediction.  It is based on
95% control in the 1977 model year, however, and this probably will not be
realized until the 1978 model year.

Patterns of Hydrocarbon Emissions

     Recognizing the relative contribution of exhaust,  evaporative, and crank-
case hydrocarbon emission, an examination of the detailed patterns of each
ensues.  The mass and pattern of hydrocarbons emitted from motor vehicles are
dependent  on many factors.  The fuel;  the emissions control system; the vehicle
history (wear, deposits, vibration, and maintenance); ambient conditions
(temperature, pressure, and humidity); and driving patterns are all significant
controlling factors in vehicle emissions. Therefore, the conclusions of any
reported study must be tied to the specific conditions  under which the study was
conducted.  The trends indicated are generally true, but the absolute  numerical
quantities reported are of value only  to indicate the trends.
                                     1057

-------
                  LLJ
                                          EVAPORATIVE
                      1960        1970        1980
                                      YEAR
1990
  Figure 3.   Normalized urban  aggregate  HC  emission  rate  from  in-use  auto-
  mobile population,  baseline  case.   Crankcase,  exhaust,  and evaporative HC
  emissions indicated separately.   Curve A-Conventional  assumption that
  evaporative HC are  95% controlled  in post-1970  vehicles.  Curve  B-Baseline-
  field data used to  estimate  HC evaporative  emissions from 1971-1976 model
  year vehicles, with 95% effective  control assumed  in 1977 and subsequent
  model years.   Curve C-Field  data based estimate of evaporative HC used for
  all  post-1970  vehicles.  (8)


     Tailpipe hydrocarbon emissions  have been examined in some detail by gas
chromatography (GC)  for many new and older  cars  over the  past  several years in
the EPA mobile source program at Research Triangle Park  (RTF).  The exhaust
patterns to be discussed were  determined for, generally,  low mileage  automobiles
(<10,000 miles)  tested with the  1975 Federal  Test Procedure (FTP)  (5).  This
procedure calls  for  driving the  vehicle  on  a  dynamometer  over  a defined urban
driving cycle.

     Table  3 presents the  relative exhaust flow and  hydrocarbon  concentration
of the modes under which vehicles  are normally operated  and the  approximate
percentages of  the FTP  driving cycle in each  of these modes.

     The hydrocarbon patterns reported below  were determined  using a  consistent
batch  of lead-free gasoline the properties  of which  are  shown  in  Table  4.
                                     1058

-------
          TABLE 3.   EFFECT OF VEHICLE MODE ON  HYDROCARBON  EMISSIONS
                                Exhaust (10)          FTP  Time  In  (8)
           Mode              Flow        HC Cone.          Mode,%
       Idle              very low          high             18.2

       Cruise
low speed
high speed
Acceleration
moderate
heavy
Deceleration
low
high
high
very high
very low
low
very low
low
moderate
very high
30.2*
27.7
23.9

      *    The average speed of the 1975 FTP is 19.7 mph.
       TABLE 4.  FUEL USED FOR EXHAUST HYDROCARBON EMISSIONS TESTING
       Research octane number                                  93.2
       Motor octane number                                     84.7
       Reid vapor pressure, psig                               10.2
       Distillation, ASTM D-86, °F
                   10%                                         123
                   50%                                         199
                   95%                                         325
                  100%                                         383

       FIA analysis, %
                  aromatic                                     24.0
                  olefins                                       8.3
                  parafins                                     67.7
This fuel is reasonably representative of typically marketed lead-free gasoline
The detailed hydrocarbon composition of the test fuel  has been previously
reported (11).

     Tables 5a and 5b present the hydrocarbon exhaust  emission patterns for a
wide variety of catalyst and non-catalyst automobiles.   Detailed  GC  procedures
(12) were used to produce all listed data.   The totals  were based on summations
of appropriate individual  hydrocarbon concentrations.

                                     1059

-------











oo
•z.
OL
LU
1—
1—
=c
D-

"^
O
1 — I
oo
oo
h- -H
LU
^^
UJ
Q
"^~~
re
LU
a
	 i
-
1C
I—
oo
ID
C-T-
*--l.
rc
X
LU
LU
	 I
i — i
CQ
O
•^_
O
h-
ID
>
^;
CO
-a
i —
•<
+j
o
i—
ai
-o
>>
-C
CU
-a
s-
o
^9- LA-
'S C
*\ 4_
C 0)
O r-
-Q O
S-
ro
o ai
0 C
s- a)
-O r—
>^ ^
-C ^
4->
i — LjJ
ro
-M
0 0
+J
M- ro
0 g
o
cu s-
01 c£
ro
i ^
c cu
cu c
o a)
S- Nl
cu c
o- a>
CQ

cu
c
cu
">>
+->
cu
o
<
M-
L(—
ro
ro
Q_
cu
ro
-C
0)

-t->
1/1 ^
r; c
ro o
-C _Q •
X S- •!-
LU ro E
o -~-
i— O cn
ro s-
+-> ~o
0 >,
h- 31

'o E
S_ 0)
-*-> 4-J
cr to
0 >,
CJ CO

OJ
U
J=
ai
>




0 0
0 0
r-^ i—




co ro
CM CO
ro LO

0 0
CD r~-
LO ro
CM ro



r--- LO
CT. «*
CM r^



0 <*•
ro o
C\J CO
CM i —


ro **•
^r o
CM CM




•* 0
O CM
i— LD
CO CM
LO LO
0 r-
«* "*


cn cn
CM cn
cn co





LO r^~
i— «*•
_ i 	





+->
in r- s_
>> ro o
r— E 4-)
i ro s- cj
c: +-> cu ro
o ro ^r cu
C U 4-> ^,

a
1 — ' >i / — ^
cj ro s- i-»
CM -^- ~o ro o
r^ CD r^ N 4-J
cn LO cn ro o o
r- CO r- S l_^f




o *^
CM LD
rM ro




0 0
•— LO
1

LO r-- ^
CM ^T 53-
r^ o o
i — CM CM



CT) O LO
CM LO r-~
LO t-~. 00



r-^ 01 co
CO ej- LO
oo cr> cn
i i r—

co ro ^j-
CO f~- <3-





LO LO CM
cn ro r^
r— ro CXJ
LO LO CT)
CO LO O
i— LO r-
LO LO LO


CO I — • =3-
ro ro ^j-
cn LO LO
r-




O^ LO LO
^ LO CXJ
•





c c " c:
O 4-J O 4-J Q_ O 4-J
•i— (/I T- m E -i— t/)
QJ4->>> QJ 4-J >, ;s CU 4-> >>
CT ~O n3 CTi~a ro Cr>~O n3
CT -r- +-J C -i— 4-J S- C T- 4-J
•-— X rQ -f— X ro T- -r- X 03
t^OU LOOUra LOOCJ
>i "0
S- S- "
Z3 Q O rO CD " O
U- I-H U_ ~O •— 1 (T3 I-H
un • LO c LO > n3
r^ >iCO r^ ro c\j r-. cp 0.0
cn r— ,— cns-o cn_c;Ei-n
i — cxco ( — oro i — o i— H ro
1060

-------











00
~*^.
fy*
; i i

1 —
1—
CL


^^
o
t— -H
OO
OO
t— H
^—
UJ
„ — ,
i i i
Q
^
i i i
Q
— 1
)
L__ 2C




1 	
0 E
S- OJ
-M 4-1
C >
C-J OO

(V
r—
O
.) —
r-
01
^





CTl
CM
i-^



0
0

CO




CO
t^J-

00
ro



IH
LO
1 —



OJ

00
OO


CTi
I--.

i —



CTl
00
r^

00
CO
t~^
ro



CO

^j-




co
00





"O
CD
«i —
'1
•r- CD '—
+-> cn c
n3 S_ ro
S- ra CD
-U -C r—
OO O * — '


O CD

-a co o o
C 0 S_
o ;> O-i—
n: cj ^ — cn





! 	 1
LO
CO



^
CD

f—
i —



CD
CO

cn
ro



LO
, —
OO



ro
CO

cn
i —


LO
O

00



cn
LO
o
1
^
<*£>
cn
CXI



S

*^j-




01
ro





o

c~
0
S-
4_)
U C C
CD re !-
i— cu r;
CD r— -Q
i_
OJ '— -Q
r- O 1— 1
LO -U C_)
>> O
!- S- 0
r" o cj-
!_> 	 -•*





CD
cn
CO



o
ro

00




CD
ro

ro
•—



LO
LO
ro




CO

CO



co
LO





ro
OO
^«

LO
i—
CO
r*-



CTl

^}-
CM



ro





i
c c:
•a o o +-1
CD T- T- LO

C-) ro i —
i— 3 "O ra

Z3 QJ X ro
-a i- o o

^ — - / — 1
-M O t— i
CD +-I CJ
C O
CJ S- S- 00
s: o a-oo
cr rr^oo





LO
00
ro



^
CTl






o
r—

1 — ^
•—



^
LO




O
CO

1 —
CM


r- —
^3-

i —



ro
^
^3-

cn
LO
CO
LO



cn

cn
1



CD





1
>> = C
ra O O -!->
3 T- T- LO

CD O ro i—
CD Z3 ~a ra
S- "O -^ 4->
-£Z CD X ra
4-> S- O U

,^^
0
o +->
> o
r- S_ O
OQ.ro
> 	 	 r —
1061

-------
     In general, vehicles employing oxidation catalysts for exhaust hydrocarbon
control have a lower percentage of unsaturated hydrocarbons than do vehicles
without exhaust hydrocarbon control.  Also, the methane percentage is increased.
The net hydrocarbon photochemical reactivity is thus reduced.   Of interest to
those who use acetylene as a tracer for auto exhaust hydrocarbons in ambient air
studies, the acetylene relative abundance in auto exhaust has  been reduced
significantly.

     Four vehicle configurations of potential significance in  outyear automobile
populations are listed in Table 5b.  The stratified charge engine, in which net
lean combustion is used for emission  control, yields a higher  percentage of
unsaturated hydrocarbons than previous non-catalyst cars.  The Chrysler Corpo-
ration "lean-burn" engine also yields a higher percentage of unsaturated hydro-
carbons.  Two control systems are listed which are leading candidates for
combined HC, CO, and NOX control.  Both the dual-bed and three-way catalyst
systems yield lower percentages of exhaust unsaturated hydrocarbons than non-
catalyst cars.  This combined with the higher percentage yields of methane
results in significantly lower net HC photochemical reactivity.

     Exhaust aldehyde data are also listed.   Here too, the percentage yields are
lowest for vehicles employing catalysts for exhaust emission control.  In
general, the aldehyde rates are an order-of-magnitude lower than the hydrocarbon
rates.

     The detailed patterns of evaporative hydrocarbon emissions are dependent to
a great extent on the gasoline used and ambient temperature and pressure.  As
previously mentioned, evaporative emissions are much greater in Denver than in
Los Angeles,due primarily to the lower pressure and temperature of Denver and
the attendant greater percentage of low molecular weight hydrocarbons in the
fuel.  This situation should be considered when attempting to  define automobile
sourced hydrocarbons in ambient air studies.  In cities of higher altitudes the
contribution of evaporative gasoline  is greater.  In the vehicle there are two
primarily sources of evaporative losses:  the fuel tank and the carburetor.  The
carburetor bowl potentially reaches higher temperatures and consequently the
evaporative losses from this source reaches higher average molecular weight.
The bowl total volume is also of significance to total mass of emission. The
temperature, vapor volume, and pressure of the fuel tank are all of significance
to tank losses.  Many parameters impact the composition of evaporative losses,
but in  general it can be stated that  the emissions are very much dominated by
the Ci+ to C6 paraffins.  Normal-butane and isopentane generally account for in
excess  of 50 percent of the total evaporative losses.  Table 6 is representative
of typical evaporative emissions.  The dominance of n-butane and isopentane, and
C^-Cg paraffins in general is readily apparent.

     Crankcase blowby has been essentially eliminated with the positive crank-
case ventilation systems.  Voelz et al. estimated the systems  to be about  98%
effective  (9).  This estimate presumes proper maintenance of the system.   In a
study of over 75,000 vehicles in 15 metropolitan areas,  17 percent of the
vehicles tested needed PCV system maintenance and 3.6% were emitting some
crankcase fumes into the environment (9). The hydrocarbon composition of crank-
case emissions is dominated by carbureted mixture  (^85%) and combustion gases
(^15%) (15,16,17).  The patterns are thus very  dependent on the fuel being  used.

                                      1062

-------
Lubricating oil does not contribute to crankcase emissions in a significant
manner (18).

               TABLE 6.   PATTERNS  OF  EVAPORATIVE  HYDROCARBONS
                                      Fraction  of  total  evaporated  hydrocarbons,
                                      vol.  %
                                      Carburetor  (13)             Fuel  tank  (14)
 Paraffins
       n-butane                          11.9                         30.5
       isopentane                        45.3                         26.4
       U-C6                              66.2                         76.0

 Olefins
       Cit-Ce                               8.5                         14.0

 Aromatics
       C6-Cy                               1.0                          1.3
     The pattern of the aggregate hydrocarbon emissions from catalyst equipped
vehicles is very much dominated by the paraffins.   As discussed,  this is  due to
the increased relative abundance of evaporative emissions with its paraffinic
dominance and the greater activity of the catalyst for the exhaust unsaturated
hydrocarbons.  Table 7 is illustrative of the relative abundance  of the various
hydrocarbon classes.

     For the purpose of illustration the 1975 Chevrolet Impala exhaust class
weight percentages previously discussed were assumed to be typical of the
model year.  Also, it was assumed that 55% of the total evaporative emissions
were from the carburetor and 45% from the fuel tank.  About 75% of the aggre-
gate emissions are paraffin, 9% aromatic, 15% olefin, and 1% acetylene.  These
estimated are for the 1975 model year.  Outyear estimates vary as the relative
abundance of exhaust to evaporative emissions changes.  The 1975 model year
emission factors suggest that the aggregate is composed of 41% exhaust, 59%
evaporative  (2).  For the 1978 model year, the aggregate is estimated to be
35% exhaust, 65% evaporative (assumes exhaust standard is reduced to  .41  g/mi).

Emission Control Systems Deterioration

     Although mentioned, but not emphasized, deterioration factors are very
important when attempting to assess the importance of the automobile to atmos-
pheric pollution.  Most of the data found in the literature is for low mileage
(<10,000 miles) vehicles.  With deterioration of emissions control systems, both
hydrocarbon mass and patterns will change.  Figure 4 is indicative of typical
exhaust control deterioration factors.  These factors are based on linear
deterioration at the rate of 1  percent per year for the pre-1968 model years, 10

                                     1063

-------
percent per year for the 1968-1974 model  years,  and 20 percent per year for the
post 1974 model  years (2).  No inspection/maintenance program is assumed.   Such
programs would decrease the deterioration significantly.   The factors for post-
1974 model years were based on tests of prototype vehicles.   The actual dete-
rioriation factors for 1975 and later model  year production  vehicles should
likewise be higher than 1968-74 model due to the increased sophistication of the
control system,  i.e. catalysts, air pumps etc.

     The actual  emissions increases experienced  with post-1974 in-use vehicles
on aging may be  significantly different from that indicated  by the published
deterioration factors.   The deterioration factors reported were determined with
specific vehicles chosen for the durability  fleet.   They  were maintained  on
owner's manual prescribed maintenance schedules  and were  not abused as consumer-
owned vehicles might1be.  For example, these vehicles do  not have their catalyst
exposed to leaded gasoline, their carburetors are not tampered with to achieve
better drivability, the spark advance is not set variant from manufacturers'
specifications,  and their catalysts are not physically removed from the exhaust
system.

     Ongoing surveillance programs are carried out by EPA under which consumer
automobiles are tested and actual emissions from in-use vehicles are determined
(19).  Catalysts 'having been on automobiles for only two years, it is difficult
to accurately project outyear emissions increases.

                                  CONCLUSIONS

     In conclusion, the advent of emission control  has dramatically decreased
the rate of emission of hydrocarbons into the environment from motor vehicles.
Crankcase emissions have essentially been eliminated.  Exhaust emissions have
been reduced as much as 95% in some vehicles relative to uncontrolled vehicles.
Vehicles using catalysts have seen a significant shift in the pattern of emis-
sions with unsaturated hydrocarbons being reduced to a greater extent than
saturated.  The advanced catalyst systems also show significant increases in the
relative abundance  of methane in the exhaust.  Vehicles using lean-combustion
for emission control reduce saturated hydrocarbons to a greater extent than
unsaturated hydrocarbon.  Evaporative hydrocarbon emissions are dominated by the
Ct^-Cg paraffins with n-butane and isopentane being the most significant.

     In general, hydrocarbons emitted from outyear motor vehicles will show an
increased dominance of paraffins due to the impact of catalyst on exhaust
hydrocarbons and the increased relative abundance of evaporative hydrocarbons.
The relative abundance of acetylene will decrease.

     Ambient conditions significantly impact emissions.  Hydrocarbon emissions
are generally greater at high altitudes than at low altitudes, particularly
evaporative emissions. Emissions will in general increase with vehicle age,
potentially as much as a factor of three over ten years.

                                ACKNOWLEDGMENT

     The author wishes to acknowledge and express gratitude to Mr. Larry High
for the GC exhaust  hydrocarbon analysis and Mr.  Fred Stump and Mr. William Ray
for the aldehyde analysis.

                                     1064

-------
                                                00
                                                       o
                                                       DC
                                                       <
                                                       LU
                                                    co>
                                                    cc_
                                                    LU O
                                                    C3CC
                                                    NOISSIIAI3  I aV3A
JNOISSIIAI3 X aV3A
o
o
      aoiovd i\ioiivaoia3i3a
                                                                       I
                                                                      -l->
                                                                      ^r
                                                                      CT)
                                                                      S-
                                                                      o
                                                     co to
                                                     S- S_
                                                     O 13
                                                    +J O
                                                     u
                                                     03 a>
                                                    (+_ +j
                                                       to
                                                    •I-  I
                                                    4-> CTl
                                                    ra ^t
                                                    5_ * __ ^

                                                    O
                                                    •r- Ul
                                                    S- 0)
                                                    O) i—
                                                    •M U
                                                    >
                                                    -£= O)
                                                                       to i —
                                                                       Z3 O
                                                                       fO to
                                                                      _C ro
                                                                       X CD
                                                                       0)
                                    1065

-------
          TABLE 7.  ESTIMATE OF COMPOSITE HYDROCARBON EMISSIONS
                    FOR TYPICAL 1975 PASSENGER CAR *


Total Hydrocarbon, g/km (2)
Percent of total , wt.%
C^-Cg paraffins, wt.%
Olefin, wt.%
Aromatic, wt.%
Paraffin, wt.%
Acetylene, wt. %
Exhaust
.75
40.8
8.24
8.34
8.03
23.29
1.11
Evaporative
1.09
59.2
40.4
6.67
.66
51.9
-
Composite
1.84
100
48.6
15.0
8.7
75.2
1.1

*low altitude 49-state car
                              REFERENCES

1.   National Petroleum Council.   Environmental  Conservation:   The Oil
     and Gas Industries.  A Summary.   Vol.  1.   Washington,  D.C.,  1971.
     106 pp.

2.   U.S. Environmental Protection Agency.   Compilation of  Air Pollutant
     Emission Factors.  AP-42 part B.  2nd Edition,  third printing.
     OAQPS, Research Triangle Park, N.C. 27711,  February 1976.

3.   Fegraus, C.E., Domke, C.J.,  and  J.  Marzen.   Contribution  of  the
     Vehicle Population to Atmospheric Pollution.   SAE No.  730530.
     Society of Automotive Engineers,  Automobile Engineering Meeting,
     Detroit, Michigan, May 1973.

4.   Trayser, D.A.   A Study of the Influence of Fuel  Atomization,
     Vaporation, and Mixing Processes  on Pollutant  Emissions from Motor
     Vehicle Power Plants.  Battelle  Memorial  Institute. Columbus,
     Ohio.   April  1969. pp 16.

5.   Federal Register, Vol 37, No. 221,  Nov. 15, 1972, and   Vol.  39,  No. 101,
     May 23, 1974.

6.   SAE Recommended Practice, Measurement  of Fuel  Evaporative Emissions
     from Gasoline Powered Passenger  Cars and Light Trucks  Using  the
     Enclosure Technique - SAE J171a,  SAE. Handbook, SAE, Inc.  Two
     Pennsylvania  Plaza, New York, N.Y., 10001

                                   1066

-------
7.    Williams, M.E.,  White,  J.T.,  Platte,  L.A.  and  C.J. Domke.  Automobile
     Exhaust Emission Surveillance.   Analysis  FY-72 Program.  EPA,
     Ann Arbor, Michigan.Publication  EPA-460/2-74-001.  February, 1974.

8.    Heywood, J.B.  and M.K.  Martin.   Aggregate Emissions  from the Automotive
     Population^SAE No. 740536,  Society of Automotive  Engineers.   Combined
     Commercial Vehicle and  Fuels  and Lubricants  Meeting,  Chicago,  Illinois
     June 1974.

9.    Voelz, F.L., Onyon, E.W., Oust,  R.M.  Rusnack,  N.C.,  Segal, J.S.  and
     B.G. Gower.  Survey of  Nationwide  Automotive  Exhaust Emissions  and
     PCV Systems Conditions  - Summer  1970.  SAE 710834.   Society of
     Automotive Engineers, Combined Truck  Powerplant, and  Fuels and
     Lubricants Meeting, St. Louis, Mo., October  1971.

10.  Brehob, W.M.,  Control of Mobile  Sources,  Purdue University.  8th
     Conference on  Air Pollution Control.   Lafayette,  Ind.,  October 1969.

11.  Black. F.M., and R.L. Bradow, Patterns of Hydrocarbon Emissions  from
     1975 Production  Cars, SAE 750681, Society of Automotive Engineers,
     National Fuels & Lubricants Meeting,  Houston,  Texas,  June, 1975.

12.  Dimitriades, B., and D.E. Seizinger,  Environ.  Sci. Techno!., 5:223-
     229, 1973.

13.  Wade, D.T., Factor Influencing Vehicle Evaporative Emissions,  SAE
     670126, Society  of Automotive Engineers,  Automotive  Engineering
     Congress, Detroit, Michigan,  January  1967.

14.  McEwen, D.J.,  The Analysis of Gasoline Vapors  from Automotive  Fuel
     Tanks.  153rd  National  Meeting of the American Chemical Society,
     Miami Beach, Florida, April  1967.

15.  Bennett, P.A., Jackson, M.W., Murphy, C.K.,  and R.A.  Randall.
     Reduction of Air Pollution by Control of  Emissions from Automobile
     Crankcases.  Vehicle Emissions.   Part I.   SAE  Tech.  Prog. Sec.
     6:224-253, 1964.

16.  Patison, J.N.  and E.R.  Stephens.  The Composition of  Automotive
     Blowby Gases,   pp. 37-50.  Proceeding of  the 3rd Technical Meeting,
     West Coast Section, Air Pollution Control  Association.  Monterey,
     Call for., 1963.

17.  Rose, A.H., and  R.C.  Stahman.  The Role of Engine  Blowby in Air
     Pollution.  J. Air Pollut. Control  Associ.   11:114-117, 1961.

18.  Domke, C.J., Lindley, D.J., and  C.N.  Sechrist.  How  to  Study Effect
     of Blowby Gas.  Hydrocarbon Processing 45(9):303-306, 1966.

19.  White. J.T. Automobile  Exhaust Emission Surveillance.   FY-74 program:
     Contract No. 68-03-2183, 68-03-2184,  68-03-2185, 68-03-2229.
     FY-75 program:  Contract No.  68-03-2378,  68-03-2379,  Environmental
     Protection Agency, Office of  Mobile Source  Pollution  Control,
     Ann Arbor, Michigan 48105.
                                   1067

-------
                                                                             22-2
                          CONTROL OF OXIDANTS IN  SYDNEY

                           D.  Iverach and M.G. Mowle*
ABSTRACT

     The. YiatuAe.,  e.xte.nt and Ae.veAity o^ photochemical pollution hi Sydney aAe.
             Recently ob*eAve.d health e^ec-ts on school chiX.dn.e.n have,  empha-
-6-czed the. AeAiouAne.&A o^ the. problem.
     Vn.oQn.eAi> in. cnmbating the. problem i& n.e.viewed.   The. impoAtance.  o&  a. coop-
e.native. inve.Atigati.on known  oj> the. Sydne.y OxA.da.Yit Study i& &tAeJ>t>e.d,  The.
introduction o{, the. U.S.  1973  ve.hicle. e.n\ik&ion AtandaAdA JJ> atbo du>cuJ>&e.d.
Re.ce.nt te.AtA Ahow that thu>  AtandaAd MJJJ, not btung  about the. sie.du.ctLonA
ofu.gtnal^.y hope.d faofi.  Ton. e.x.amp£e., sieAtittA ofa ve.hicie. te.i>ti> tn Sydne.y  i>how
that e.x.hauj>t hydfioc.an.boYi £eveZ6 w^tt only be. n.e.du.ce.d appn.oxAjmate.ty 30%  com-
pan.e.d to  mon.e, than 60% pn.e.dtdte.d on the. baAtt, o&  U.S.  data,  A n.e.\)Lt>e.d
     tnve.nton.y L*> pn.ue.nte.d  taking the.be. n.e.4uItA  into account.
     \notheA dibtuAbing n-e.ce.nt finding AJ> that  poUute.d ain. ^n,om Sydne.y may be.
tnanAponte.d about 50 km to an  an.e.a whesie. con^ideAabte. population Qn.owth 
&on.e.caf>t to OCCUA.

     Fn.om  oJUL thu, it u> cie.an. that toweA e.x.hauAt  Atandan.d*> an.e. needed to
contAol the. pAobLem, but the. natuAe. o& the. vehicle. indut>tn.y in AuAtnaLLa wJUL
make. the. intAoduction ofa tow,  deteAioAation  ^ee emi^Aion te.chnology a chal-
lenging tat>k.


                                  INTRODUCTION

     With  three million people living in an  area of 1000 sq. km.,  Sydney is
Australia's largest city.  Located on the east  coast of the continent at lati-
tude 34°S,  the city lies across the east-facing mouth of the Parramatta River
basin.  Gently rising hills lie to the south, west, and north of the region.
Westward the land continues to rise and 90 km from the coast the hills are
about  1200  m high.

     In this region the use of 1.2 million motor vehicles combined with the
operations  of an extensive range of industries  result in the emission of more
than 400 tons/day of hydrocarbons (HC) and about 170 tons/day of nitrogen
*State Pollution  Control Commission,  Sydney, New South  Wales, Australia.
 Dr. Iverach  is  also presently  at the Sloan School  of Management, Massachu-
 setts Institute  of Technology,  Cambridge, Massachusetts.
                                     1069

-------
oxides (NOX).  It is not surprising therefore that photochemical oxidants form
in undesirable amounts whenever appropriate meteorological conditions prevail.

     The problem of photochemical pollution first came into prominence in
Sydney in November 1971 when widespread damage to petunia seedlings was identi-
fied as being caused by oxidants.  Full realization of the seriousness of the
problem came in March 1976 when thirteen school children were taken to the
hospital after suffering from chest pains and breathing difficulties while
exercizing. The symptoms have been attributed to oxidants; at the time of the
incident ozone (03) concentrations over 0.2' ppm were recorded in the vicinity
of the school.  Prior to this occurrence the main recorded effects of oxidants
had been vegetation damage, reduced visibility and accelerated rubber crack-
ing.   Few reports of eye irritation had been received.

     To put the problem in perspective it should be added that, the potential
for a Los Angeles scale problem does not exist in Sydney.  It is not likely
that hourly values as high as those measured in the Los Angeles basin (up to
0.7 ppm) could form in Sydney.  Perhaps more pertinent is the fact that mete-
orological conditions are such that Sydney should never experience the same
frequency of high levels as recorded in that city (150 days per year over 0.15
ppm).

     Typically in Sydney there are five to ten episodes per year, each lasting
one to five days when 03 concentrations in excess of 0.15 ppm (1 hr ave) are
measured.   Hourly values over 0.25 ppm and peaks over 0.4 ppm have been re-
corded infrequently.

     If nothing were to be done, 0.15 values would become 0.25 within a decade
and such values might occur on 20-30 days each year.  Although not compatible
to experiences in Los Angeles, such a situation would, on present day atti-
tudes, be considered a very serious one.  Indeed there is considerable concern
about the existing levels.  Fortunately the need for effective action has been
recognized.


                              PROGRESS TO DATE

     In 1971 when oxidant pollution was first recognized as a problem, very
few data existed on which to base a control strategy.  The extent of the
problem was not known; virtually nothing was known of the concentrations of
the precursors, reactive HC's and NOX; the important role of meteorology was
only suspected, and no emissions inventory was available.  Nor was there
provision in the Clean Air Act to regulate vehicle emissions.

     Progress in both data collection and analysis and in the legislative
areas  has  been substantial, but as is the case in the rest of the world, much
remains to be done.

     With  respect to data collection and analysis, perhaps the single most
important decision the commission has made was to enlist the aid of three ex-
ternal  organizations.   It is not common in Australia for a State Government
Department to contract outside organizations in the scientific area.  However,


                                    1070

-------
the study of oxidants in Sydney required a greater variety of measurements to
be made with more sophistication than had hitherto been attempted.  Contracts
have been entered into with specific groups within the University of Sydney,
Maquarie University and the Commonwealth Scientific and Industrial Research
Organization (CSIRO).

     At the same time, the Commission has increased its staff and developed
its facilities in order that it be still directly involved in the day to day
effort and not simply a contract manager or coordinator.  Approaching the
problem in this way allowed a much greater effort to be mounted more rapidly
than would have been possible by doing all the work in-house, but without
losing the essential information that derives from day to day involvement.

     The responsibilities and interests of the four groups are listed below.
The activities in parentheses are not financially supported by the Commission
but obviously contribute to the study.

     •   The State Pollution Control Commission - Coordination, development of
         control strategies, development of emission inventories, responsi-
         bility for fixed site monitoring and mobile monitoring, including use
         of ground, sea and air craft.

     •   The University of Sydney - Responsibility for ground based mobile
         monitoring along air mass trajectories, analyses of data and determi-
         nation of driving conditions in Sydney (development of p.c. smog
         models for Sydney and development of emission inventories).

     •   Maquarie University — Responsibility for meteorological measurements
         and analysis of wind data and air movements (development of models).

     •   CSIRO- Responsibility for detailed measurements of reactive HC's and
         development of a doubly compartmented smog chamber (measurement and
         analysis of aerosols).

Published results from the various groups are listed in the reference section.

     With respect to legislative developments, the most significant event was
the decision in November 1972 by the N.S.W.  Government to amend the Clean Air
Act to provide for the regulation of vehicle emissions.  Regulations were
gazetted in April, 1974 requiring, among other things, control of evaporative
emissions on new cars manufactured after 1 January 1975 and the U.S. 1973
exhaust emission standards to be met on new cars manufactured after 1 July
1976.   This action had been proposed in 1972.  Details of the development of
this and other vehicle related legislation and the anticipated effects may be
found in Murphy (1975) and Iverach and Bilger (1975).

     The important decision to adopt the U.S. 1973 standards and test pro-
cedures was based primarily on practical considerations.

     It has been pointed out that lack of data made it difficult to develop a
control strategy in the early years of the problem.   However certain qualita-
tive judgements could be (and were) made.   If the automobile were the major


                                   1071

-------
emitter of precursors in Los Angeles there were good reasons to suspect this
to be the case also in Sydney.  On a per capita basis vehicle ownership and
usage were somewhat comparable but there were fewer other major sources of
oxidant precursors (oil refineries and power stations) in the Sydney region.
Emissions of HC and NO  from vehicles had to be significant.  This was veri-
fied by emission inventory calculations which, when calculated on a two by two
mile grid square basis, also allowed non-methane hydrocarbons (NMHC) and NO
concentrations to be predicted. It was then possible to compare these predi£-
ted concentrations with the various oxidant NMHC NO  relationships that had
been developed.  These relationships had been used fn the U.S. and Japan to
determine whether HC or NO  should be controlled preferentially.,
                          X
     Iverach (1975) summarizes the situation.  Neither an HC only (U.S.
approach) nor an NO  only (Japanese approach at that time, 1973) strategy was
unmistakably indicated.  This suggested concentrating on HC control because
the technology for it was better developed both for mobile and stationary
sources.  Hence, adaptation of the U.S.  1973 standards was consistent with
both the strategy as it could be derived at the time and the philosophy of
using the "best practicable means" to control pollutants.

     Using a "best practicable means" approach to the problem rather than pro-
mulgating what may have been impossible, though justifiable, is typical of
pollution control in Australia and has been used since the first Clean Air
Acts were gazetted in the early 1960's.   Such an approach does not promise
that the lowest possible emissions will  be attained but neither does it
threaten to involve the Government in a "back-off" situation as has occurred
recently in the U.S. and Japan.  Perhaps the main disadvantage of the approach
is that it does not force, or at least encourage, the development of control
technology.


                   RECENT DEVELOPMENTS AND FUTURE OPTIONS

     A deal of data have been gathered since smog was first recognized as a
problem in Sydney and more will be obtained in the near future.  The most sig-
nificant developments are summarized below.

EMISSIONS  INVENTORY

     The emissions inventory for HC and NO  continues to become more accurate
as actual measurements of sources replace Emission estimates that were based
on the U.S. Environmental Protection Agency (EPA) emission factors.  Signifi-
cant differences have been found between the EPA factors and actual emissions
in Australia.   For example NO  emissions from stationary sources have been
found to be approximately one-half the EPA factors.  Preliminary measurements
of pre-1977 vehicle emissions have indicated average exhaust HC emission rates
of about one-half the pre-1968 U.S.  model  (uncontrolled) value.  NO  emission
rates are about 15-20% higher than pre-1968 vehicles in the U.S.   x

     Although  evaporative emissions have not been measured in Australia using
the SHED method the recent discovery in  the U.S.  of high evaporative emissions
from "controlled" vehicles will most likely carry over to Australia because


                                   1072

-------
identical control techniques have been used.  This has caused a further re-
vision in the estimates of the benefits of existing controls.

     Current estimates of emissions from vehicles in Sydney compared to U.S.
data and U.S. 73 standards are shown in Table 1.

     The present best estimates of Sydney's emissions are summarized in Table
2. Estimates for 1971 are shown in parentheses.   Most of these estimates
should be given 'B1  or 'C'  rating in terms of AP42 (EPA, 1975).

                     TABLE I.  VEHICLE  EXHAUST EMISSIONS

         Vehicle Fleet or Standard      Emissions      g/Km
                                       CO             HC        NOx

         Ave. of 81 Sydney Vehicles
tested in 1976
Uncontrolled U.S. (pre 68)
U.S. 1973 Standard
32
54
24.2
3.0
5.5
2.1
2.6
2.25
1.9

PRECURSOR MEASUREMENTS

     Much effort has gone into measuring early morning concentrations of
reactive HC and NO .  The detailed results may be found in Allen et al. (1976).
In general it has Been found that NMHC/NO  values are in reasonable agreement
with values predicted from the present emission inventory.  In addition, con-
centrations are rarely measured that are as high as the 'worst case1 predic-
tions made by Iverach et al. (1976).   High oxidant day precursor concentra-
tions are 0.02-0.15 ppm NO  and 0.2-1.5 ppm NMHC.  Perhaps most importantly,
precursor concentrations have been related to oxidant levels downwind on
several occasions, and fair agreement has been found with currently accepted
oxidant-precursor relationships.

AIR MASS TRAJECTORIES

     Hyde and Hawke (1976) describe the various kinds of meteorological condi-
tions that prevail on high oxidant days.  Their most important conclusion,
based on modelling air mass trajectories, is that on a high proportion of high
oxidant days photochemical oxidants may be transported some 50 km. to the
south west of Sydney — an area where considerable population growth is
anticipated.
                                    1073

-------
TABLE 2.   EMISSIONS OF NON-METHANE HYDROCARBONS AND
    NO  IN SYDNEY (TONNES/DAY), 1976 AND (1971)

Source
Light Duty
Vehicle-Exhaust
Vehicle-Evap.
Vehicle-C' case
Petrol Trucks,
all sources
Diesel Truck
Motor Cycles
Utility Engines
Air, Rail & Shipping
Total Mobile
Industrial Combustion
Incineration
Refineries
Storage, Transfer
of Gasoline
Solvent Evap.
Service Stns.
Dry Cleaning
Total Stationary
Grand Total
Hydrocarbons

110
63
10

18
1
14
16
3
235
8
9
53
21
67
14
10
173
418

(176)
(51)
(39)

(ID
(1)
(12)
(13)
(3)
(306)
(7)
(7)
(45)
(18)
(56)
(12)
(8)
(152)
(458)
N0x

103 (86)
-
-

8 (6)
13 (10)
-
-
4 (3)
128 (90)
33 (58)
1 (2)
7 (13)
_
-
-
-
41 (72)
170 (161)
                    1074

-------
                               FUTURE CONTROLS

     Future control will focus on vehicles, the refineries and associated
gasoline transfer operations that exist to serve the needs of motorists, and
on solvents.  This discussion presumes that relatively straight-forward steps
such as legislation of the Rule 66 type can be taken to reduce the importance
of solvent emissions.  Achieving significant  reductions in vehicle emissions,
however, is likely to be tougher.

     The total Australian market is relatively small -about 600,000 units/yr.
The N.S.W.  market is about 200,000 (170,000 cars and station wagons).  This
market is split between some twenty-five companies offering some 150 models.
The local manufacturers, General Motors-Hoi den, Ford (Aust), Chrysler Australia,
and Ley!and (small) have the largest share of the market but have lost ground
in recent years to Japanese imports (mainly Toyota, Datsun, Mazda and Honda).
The top ten companies achieve about 98% of sales in the light duty vehicle
market.

     Some automobiles are manufactured in Australia; some producers simply
import the assembled vehicle but may have to install seatbelts or, as happened
in NSW, evaporative emission controls, depending on local  regulations and the
ability of the overseas principal to comply with them.

     All companies rely heavily on their overseas principals for emission con-
trol system design.  Even local manufacturers have only limited design author-
ity. The importers and assemblers rely absolutely on their overseas companies
for virtually all emission control work.   And to reiterate:  often the
Australian (and N.S.W.) market accounts for only a small fraction of their
total sales. Consequently many of the companies are slow to adapt and react to
new standards. Few are actively pursuing new engine designs or modifications
in anticipation of future standards.   Indeed only the local manufacturers and
the local carburetor manufacturer (Bendix Corp.) have emission testing facil-
ities in Australia and these are located in their development laboratories.
Only Leyland of all the vehicle companies has an emission test facility in
N.S.W.

     Despite these difficulties this industry has to be persuaded, in some
manner, to adopt what preferably should be a new, low emission, technology.
It must be a new technology because,  even assuming evaporative emissions can
be properly controlled, the need for exhaust emissions substantially lower
than the current standards can be demonstrated.  And it is most important that
essentially non-deteriorating technology be used.  The use of technologies or
techniques that suffer the currently observed rates of deterioration (EPA,
1975 and Stork, 1976) will not allow even modest improvements to be made in
air quality.

     The only technologies that offer real hope in this regard are the diesel
and the stratified charge engine.  Both these technologies are new to the
mass-produced light duty vehicle scene.   However, influencing the adoption of
these and possibly other, yet to emerge,  engines is one of the great chal-
lenges facing governments in Australia today.
                                    1075

-------
                                  REFERENCES

 1.   Allen,  Bilger,  Post  and  Haynes  (1976), present proceedings.

 2.   EPA (1975),  A Compilation  of  Air  Pollutant Emission Factors, AP42 4th
     edition,  U.  S.  Environmental  Protection Agency.

 3.   Hyde and  Hawke  (1976), present  proceedings.

 4.   Iverach (1975),  Planning for  Oxidant Control in Sydney, Jnt'l Clean Air
     Conference,  Rotorua, N.  Z., Feb.  1975.

5.    Iverach and  Bilger (1975), The  Status of Air Pollution and  Its Control in
     Australia,  in Advances in  Environmental Science and Technology, Pitts &
     Metcalf,  Vol  5,  1975.

 6.   Iverach,  Mongan,  Neilson and  Formby  (1976), Vehicle Related Air Pollution
     in Sydney,  J.A.P.C.A., Jan.,  1976.

 7.   Murphy  (1975),  Motor Vehicle  Legislation in NSW,  Int'l Clean Air Conference,
     Rotorua N.Z., Feb.,  1975.

 8.   Stork (1976), Preliminary  Data  on Emission Performance of 1975 model year
     cars, internal  EPA memo, Feb. 25, 1976.
                                     1076

-------
                                                                            22-3
                       OXIDANT  AND  PRECURSOR TRENDS  IN
                     THE METROPOLITAN  LOS  ANGELES  REGION

                  J. Trijonis,  T. Peng,  G.  McRae,  L.  Lees*


ABSTRACT
          papeA. de^cribeA fte.ce.nt  historical. &ie.ndt>  in oxA.da.nt and precursors
in the. Lot, \ngeZu fis.gi.on.   Control strategies  and  basin-wide emission trends
^or nitrogen oxides and rcacti.vc  hydrocarbons are documented year by year ^rom
1965 to 1974.  Trends in the geographic diAtsu.buti.on  o^ emissions are illus-
trated by computing nit percentage  emission changes over the. de.ca.de. ^or indi-
vidual, counties.  The. changes in  emissions  are  compared to change.* in ambie.nt
psie.cuuosi concentrations and oxx.dant concentration^.   It i& ^ound that many o{,
the. changes in monLtotutd aifi quaJLity can be. e.x.p&iine.d by tn.e.ndt>  in total
e.mu>£ion and in the. bpatiaJt distribution o& zmiAAion*.


                                INTRODUCTION

     The control of photochemical oxidant is a  complex,  dynamic  process involv-
ing major technical uncertainties.   An  effective  way  to deal with the dynamic
and uncertain aspects of the problem is  to  adopt  a  "feedback" approach in
control strategy planning.   Progress  (or lack of  progress) in improving ambient
air quality should be continually reviewed.  Control  strategies  should be
periodically reexamined and, if necessary,  reformulated.   Control plans should
be flexible so that changes  can be  made  as  new  information is developed.

     Trend analysis constitutes a fundamental part  of the feedback approach to
air pollution control.  Air  quality and  emission  trends  should be major factors
in determining whether or not control strategies  need reformulation.   Trend
studies also can help to reduce technical uncertainties.   For instance, the
analysis of ambient precursor trends  can indicate the adequacies, or deficien-
cies, in emission inventories and emission  reduction  estimates.   Comparing
historical oxidant changes to precursor  emission  changes  can provide insight
into the relationship between emissions  and air quality.   This improved know-
ledge can be used in evaluating future  control  strategies.

     This paper documents trends  for oxidant and  the  precursors  (nitrogen
oxides (NOX) and reactive hydrocarbons  (HC)) in the Los  Angeles  region from
1965 to 1974. Changes in basinwide  precursor emissions and in the geographic
*J.  Trijonis, Technology Service  Corporation,  Santa  Monica,  California; T.  Peng,
 Jet Propulsion Laboratory, Pasadena,  California:  G.  McRae  and L.  Lees, Environ-
 mental Quality Laboratory, California  Institute  of  Technology,  Pasadena, Calif.


                                      1077

-------
distribution of emissions are compared to changes in ambient precursor concen-
trations and ambient oxidant levels.   Most of the results presented here are
based on a recent report of the Caltech Environmental Quality Laboratory (1).
The reader is referred to that report for detailed discussions of the methodol-
ogy and results.
                          PRECURSOR EMISSION TRENDS

     This section describes trends in control strategies and emissions over
the last decade for the two oxidant precursors, NOX and the reactive HC (RHC).
The trend estimates are based on much of the latest information available for
emission sources in the Los Angeles region.   Included are new reactivity
factors for HC emissions (2, 3), recent measurements of auto exhaust emission
and deterioration factors (4, 5, 6, 7, 8), test data on the unexpectedly large
contributions made by evaporative emissions  from light-duty vehicles (9, 10),
the latest stationary source NOX inventory (11), and recent data on traffic
(VMT) growth and fuel usage patterns (12, 13, 14, 15, 16).   A detailed descrip-
tion of the methodology for determining the  emission trends can be found in
Reference 1).

     Basin-wide emissions of each pollutant  are documented year by year from
1965 to 1974 in order to characterize overall emission trends and the changing
role of various source categories.  To illustrate the spatial distribution of
emission trends, net percentage changes in emissions over the decade are
determined for each of the six counties within (or partially within) the
Metropolitan Los Angeles Air Quality Control  Region (AQCR).  Figure 1 shows
the area covered by the Los Angeles air basin and the geography of the six
county sub-areas.
 TRENDS  IN  TOTAL  EMISSIONS

      The purpose of  documenting  basin-wide  trends  in  total  emissions  is  to
 help  explain  trends  in  average,  basin-wide  air  quality.   Total  emission  trends
 are also useful  for  illustrating changes  in  emission  levels  for various  source
 categories.   These changes  result from emission  control  actions in  competition
 with  the growth  of source  activity levels.   Since  different source  categories
 are associated with  different  rates of growth  and  degrees of control, the
 roles of various source categories have undergone  continual  alteration.

      Figure 2 presents  estimates of basin-wide  NOx emission  trends  over  the
 past  decade.  The basin-wide total  is  represented  by  the top line,  while the
 distances  between the other lines  illustrate the contributions  of various
 source  categories. Figure  2 shows  that total NOx emissions  in  the Metropolitan
 Los Angeles AQCR increased  by  about 36% from 1965  to  1974.   The major contri-
 butor to this increase  was  the rise in NOX  emissions  from light-duty  vehicles;
 total NOX  emissions  from light-duty vehicles increased by 75%  over  the decade.
 This  was partially due  to  a 41%  increase  in basin-wide VMT  and  partially due
 to the  large  increase in NOx emissions from  1966 — 1969  model year  vehicles.
 The control techniques  used to reduce  HC  and carbon monoxide (CO) in  1966 —
 1969  vehicles had the side  effects  of  raising MOX  emissions.   Mew car emission


                                     1078

-------
                                                     O
                                                     O)

                                                     'QJ
                                                     01
                                                     £Z
                                                     l/l
                                                     o
                                                      o
                                                      CL
                                                      O
                                                      
-------
 standards for NOX took effect in 1971, but as yet these standards have not had
 a great impact on basin-wide NOX emissions.
      YEARI Y  1000

      
-------
 an  even  greater rate  after 1975  because of the stringent new car NOX emission
 standards  which started in 1975.

      Figure  3 summarizes basin-wide trends in reactive HC emissions  from 1965
 to  1974.   The top  line represents  total RHC emissions, and the distances
 between  the  lines  represent the  contributions from various source categories.
 Here, the  definition  of reactive hydrocarbon emissions is based on a two-group
 reactivity scheme  proposed by Dimitriades (2) of the Environmental Protection
 Agency and on reactivity factors calculated in a recent TRW report (3).
Figure 3.   Total  reactive hydrocarbon emission trends in the Los Angeles region.

      The  net  change  in  basin-wide  reactive  hydrocarbon  emissions  over  the
 decade was  a  decrease of  18%.   Most  of  this  reduction was  due  to  controls  on
 light-duty  vehicles.  Light-duty vehicle  crankcase  emissions were  reduced  in
 the early and middle 1960's.   Slight reductions  in  evaporative  emissions were
 obtained  in the  1970's  from  the new  car evaporative controls.   Exhaust emission
 standards for new automobiles  resulted  in significant control  of  exhaust
 emissions.  The  net  change in  RHC  emissions  for  the average vehicle  over the
 decade was a  52% decrease.   In  the presence  of a 41% increase  in  VMT,  the  net
 reduction in  total RHC  emissions from light-duty vehicles  was  only  32%.
                                      1081

-------
      Gasoline-powered,  heavy-duty  vehicles  experienced  a  23%  reduction  of per-
 mile  RHC emissions  over nine years,  most  of the  reduction  coming  from crank-
 case  emission  control.   With the 41% growth in VMT,  the net change was  a 7%
 increase over  the period from  heavy-duty  vehicles.   Emissions  from other
 mobile  sources have increased  by 56%.

      The organic chemicals  category, the  largest contributor  of RHC  in  the
 stationary  source category, has been reduced by  approximately  33% from  1966  to
 1974  due to APCD Rule 66 in the late 1960's and  to substitution of water-based
 coatings for organic solvent coatings  in  the early 1970's.  Emissions from
 gasoline marketing,  which is a major source of organic  fuel emissions,  have
 increased by 26%, although  filling of  underground gasoline station tanks has
 been  partially controlled by local APCD rules.   An important  part of the
 organic fuels  and combustion category  is  the "geogenic  source" that  was re-
 cently  pointed out  by Mayrsohn and Crabtree (17).  It is  assumed  that this
 source  remained constant over  the  ten-year  period.


 GEOGRAPHICAL DISTRIBUTION OF EMISSION  TRENDS

      The spatial distribution  of emissions, as well  as  the total  amount of
 emissions,  is  important to  ambient air quality.   Accordingly,  trends in the
 spatial  distribution of emissions  should  be considered  in  analyzing  air qual-
 ity trends.  This section illustrates  large-scale trends  in the geographic
 distribution of emissions within the Los  Angeles  air basin by  documenting net
 emission changes over the decade on  a  county-by-county  basis.

      There  is  reason to expect that  the spatial  distribution  of emission
 trends  has  been very nonuniform in the Los  Angeles region.  Although similar
 emission control policies apply to each county,  growth  rates  vary considerably
 from  county to county.   Figure 4 illustrates the  spatial  pattern  of  VMT and
 population  growth from  1965 to 1974.   Growth rates for  each individual  county
 are presented  within the boundaries  of those counties,  while  basin-wide fig-
 ures  are given in the lower left-hand  corner of  the  figure.

      As  indicated in Figure 4, the average  VMT and population  growth rates for
 the entire  basin are 3.9% per  year and 1.1% per  year, respectively.  Los
 Angeles  County has  been growing at a rate considerably  slower  than the  other
 five  counties  within the Basin.  The fastest growing county,  Orange  County,
 has a VMT growth (7.5 percent  per  year) triple that  of  Los Angeles County  (2.8
 percent per year) and a population growth rate (4.3% per  year) more  than ten
 times that  of  Los Angeles county (0.3% per  year).  These  differential growth
rates, compounded year by year over  the decade, led  to  a spreading out  of
emissions away  from  Los  Angeles County to the outlying  counties..   Los Angeles
County accounted for about 75% of basin-wide emissions  of NOX and RHC in 1965,
but only  for about 65% in 1974.

      Figure 5  illustrates the  spatial  distribution of NOX emission changes
over  the decade.  The net percentage change  in NOX emissions over the decade
is presented for each of  the six counties  and for the entire basin.   Each
county experienced  an increase in NOX  emissions over the decade,  with the
basin-wide total increasing by 36%.  Following the spatial trends in popula-

                                       1082

-------
                                                      01
                                                      i—I
                                                       I
                                                      LD
                                                      i-D
                                                      (13
                                                      S-
                                                      -I-J
                                                      s
                                                      O

                                                      en
                                                      Zi
                                                      Q.
                                                      O
                                                      O.

                                                      -a
                                                      a
                                                      (T3
                                                      rd
                                                      O)
                                                      O)
                                                      CD
                                                      OJ
                                                      >
                                                     
-------
                                                CT>
                                                LO
                                                IT)
                                                cr>
                                                 o
                                                •I—

                                                 I/)

                                                 £
                                                 QJ

                                                  >
                                                O
                                                ^

                                                 c


                                                 l/l


                                                 O)
                                                 cn
1084

-------
 tion  and VMT  growth,  Los Angeles  County  NOX  emissions  grew by  the smallest
 amount, 25%.  NOX emissions  in  Orange  County rose  drastically,  an estimated
 89% increase  over the decade.

      Figure 6 presents  the county-by-county  distribution  of  RHC emission
 changes. Over the decade, basin-wide RHC emissions decreased 18%.   Following
 the order  of  population and  VMT growth rates,  Los  Angeles County showed the
 largest decrease, 24%, while  the  outlying  counties experienced  lesser  decreases,
 In fact, Orange  County evidently  underwent an  actual increase  in RHC emissions
 of six percent over the decade.


                         AMBIENT  AIR QUALITY TRENDS

      Having documented emission trends in  the  Los  Angeles region over  the past
 decade, the next logical step is  to compare  the emission  trends to  air quality
 trends.  This section presents  air aua1ity__trends  over the decade for  both
 oxidant and its precursors and  analyzes  these  trends in terms of the changes
 in total emissions and the spatial distribution of emissions.   Comparing
 ambient precursor trends to emission trends  provides a  test  of the adequacy of
 simple proportional (rollback) models for relating emissions and precursor
 concentrations.  This juxtaposition also serves as a partial check on  both the
 emission and  air quality data bases.   Comparing oxidant trends to precursor
 trends yields insight as to the oxidant/precursor relationship in various
 parts of the  Los Angeles region.

     The data base for determining air quality trends  in the Los Angeles
 region consisted of computer tapes of hourly average pollutant measurements
 obtained from the California Air  Resources Board (ARB).  These tapes contained
 data from  all stations in the Los Angeles region which  reported to the ARB
 during the years 1963 to 1974.  A search was conducted  among personnel of the
 ARB and the County APCD's to determine whether monitoring changes occurred
which could affect air quality  trends.    Data were included from each site only
 if they were  taken with the same  measurement method each year.   It was deter-
 mined that the locations of monitors in  a few  cities had been altered  during
 the period of interest; data from these  locations were  eliminated from the
 study.  Some  of the monitoring  sites that were included in the  trend analysis
 are shown  in  Figure 1.

     In order to compare air quality changes to emission changes over  the
 decade, net changes in ambient  concentrations were determined from  1965 to
 1974.   For each pollutant (NOX, nonmethane hydrocarbons (NMHC), and oxidant),
 a least-squares trend line was  fit to the yearly averages of daily maximum
 one-hour concentrations.  The percentage change in this trend line from 1965
 to 1974 was used as the measure of overall air quality  change for each site.
 Yearly averages of daily maxima were eliminated from the  analysis if the
 averages were based on a very limited number of sampling days.  The analysis
was performed only for stations with at  least  eight years of adequate  data.


 AMBIENT PRECURSOR TRENDS

     Figure 7 compares changes  in NOX air quality to changes in NOX emissions.

                                     1085

-------
                                                CTl
                                                 I
                                                un
                                                en
                                                 co
                                                 to
                                                 0)

                                                O
                                                 1/1
                                                T3
                                                 O)
                                                
-------
                                             I
                                            in
                                             03
                                             13
                                             CT

                                             S-


                                             fO
                                             o

                                             co
                                             O)
                                             c:
                                             O)
                                             en
1087

-------
 The  circled  numbers  in  Figure 7 represent the net percentage emission changes
 for  each  county  over the  decade (also  given  in  Figure  5).   For  instance,  Los
 Angeles County experienced  a 25% increase in NOX emissions  over the decade;
 Orange County experienced an 89% increase; Riverside underwent  a 69% increase,
 etc.  The net percentage  changes in ambient  NOX concentrations  over the decade
 are  also  plotted in  Figure  7 for eleven monitoring  sites.   For  instance,
 ambient NOX  levels increased by 8% at  downtown  Los  Angeles, 76% at Anaheim,
 47%  at San Bernandino,  etc.  Basin-wide averages in emission trends and ambient
 concentration trends  are  presented in  the lower left-hand corner of Figure 7.

      The  agreement between  calculated  NOX emissions changes and measured  NOX
 air  quality  changes  is  strikingly good.  Total  basin-wide NOX emissions in-
 creased by 36% over  the decade, while  average NOX concentrations at the eleven
 monitoring sites increased  by 35%.  Even the spatial distribution of emissions
 and  air quality  trends  agree on a county-by-county  basis.   NOX  emissions  grew
 by 89% in Orange County,  while the two Orange County stations showed an
 ambient NOX increase of 93%.  NOX emissions  increased by 25% in Los Angeles
 County, while the eight Los  Angeles  County stations  experienced a 19% increase
 in measured NOX  concentrations.

     Figure 8 compares  changes in NMHC concentrations to changes in RHC
 emissions.  The  circled numbers  again represent net  percentage emission changes,
 for each county  and for the  entire basin.   Only four monitoring sites  provide
 data sufficient  for determining  ambient HC trends.   The net percentage  changes
 in NMHC* levels  over the decade  are  plotted  at the  locations of these  four
 sites.

     It is not possible to draw  definitive conclusions  concerning ambient NMHC
 trends because of the sparsity of sites providing data  on changes in HC concen-
 trations.   However, it  is encouraging that the average  change in ambient NMHC
 concentrations among the four sites,  a 15% decrease, is close to the basin-wide
 change in RHC emissions, an  18%  decrease.   Also, the spatial pattern of concen-
 tration changes  seems to agree qualitatively with the spatial pattern  of
 emission changes (at least in the sense that Los Angeles, San Bernandino,  and
 Orange counties  fall  in the  same order for concentration trends and emission
 trends).  The only puzzling aspect of  the ambient trend  data for NMHC is the
wide disparity between trends at the  two sites in Los  Angeles County,  Downtown
 Los Angeles and Azusa.

 TRENDS IN OXIDANT AIR QUALITY

      Figure 9 illustrates percentage changes in oxidant concentrations over
 the  past  decade  and  compares these to  changes in RHC emissions.  The circled
 numbers represent RHC emission trends  (county by county and basin-wide) while
 the  other numbers represent trends in  ambient oxidant  levels at thirteen
 monitoring sites. Basin-wide, the agreement between RHC emission trends and
 oxidant trends is quite good; total  RHC emissions decreased by  18% over the
 decade, while the average improvement  in oxidant levels at thirteen stations
 was  19%.   Also,  the  spatial  distribution of air quality changes agrees quali-
*Nonmethane hydrocarbon trends are computed from total hydrocarbon trends
 using empirical formulas relating THC to NMHC (18).

                                      1088

-------
   BASIN'.HDE KHC EMISSION CHANGE:  (-18%
                            V	x

   AVERAGE NMKC CONCENTRATION CHANGE (4 STATIONS):   -15%
         Figure  8.   Trends  in  RHC  emissions and air quality, 1965-1974.

tatively with the spatial  distribution  of  emission changes, at least in  the
sense that Los  Angeles  County  experienced  the  greatest  improvement  in both  RHC
emissions and oxidant concentrations.

     The spatial pattern of oxidant  air quality  trends  can  be  even  better ex-
plained by considering  what has been  termed  the  "dual role" of NOX  emissions
in oxidant formation.   From the present understanding of photochemistry,  it
can be argued that,  at  the existing  ratio  of RHC to NOX, increases  in NOX
emissions would tend to decrease  oxidant concentrations  in  the source-
intensive areas of  the  Basin  (the  central  and  western-coastal  areas).  In
these areas, the highest oxidant  concentrations  occur early in the  day  (around
noon) before the pollution is  transportad  inland by the  daily  sea breeze.   The
most effective  strategy for reducing  oxidant near the source areas  early  in
the day is to inhibit the  reaction by reducing the RHC/NOX  ratio.   This  can  be
done by decreasing  RHC  emissions  and/or increasing NOX  emissions.   The  trans-
ported oxidant  problem  in  the  eastern-inland areas should depend more 9n  the
absolute levels of  RHC  and NOX than on  the ratio.   The  key  to  controlling the
transport problem is less understood, but  indications are that reductions are
needed in overall levels of RHC and  (possibly) NO .

     These observations help to explain  why  the  coastal  areas  of Los Angeles
County show the greatest improvement  in  oxidant  concentrations (a 30 to  40%
decrease from 1965  to 1974).   These source-intensive areas  have  experienced a
24% decrease in RHC emissions, and they  have been  aided  by  the NOX  emission
increase.  The apparent anomaly at Anaheim,  a  31% decrease  in  oxidant within a
county with a 6% increase in RHC emissions,  can  be partially accounted  for  by
the large NOX emission  increase (and  resulting RHC/NOX  ratio decrease)  at that
source-intensive location.   The lesser  rates of  improvement in the  eastern-
                                      1089

-------
                                                  r--
                                                  en
                                                  LO
                                                  l£>
                                                  CTi
                                                  ro
                                                  13
                                                  CT

                                                  S_
                                                  •i—
                                                  ro
                                                  ro
                                                  -o
                                                  •r-
                                                  X
                                                  O

                                                  T3
                                                  c:
                                                  ra

                                                  00
                                                  c
                                                  o

                                                  00
                                                  oo

                                                  E
                                                  CD
                                                  n:
                                                  o:
                                                  0-)
                                                 -a
                                                  c:
                                                  ai
                                                  S-.
1090

-------
inland portions of the basin may be explained by noting that basin-wide RHC
emissions have been reduced only slightly and that total NOX emissions have
increased.   In particular, Riverside may have experienced a 23% increase in
oxidant concentrations because it is downwind of Orange County, which had a 6%
increase in RHC emissions and a large increase (89%) in NOX emissions.


                                 CONCLUSIONS

     Basin-wide emissions of NCX have increased by 36% in the Los Angeles
region from 1965 to 1974, while basin-wide RHC emissions have decreased by
18%.  Controls on crankcase and exhaust emissions from light-duty vehicles
were largely responsible for the total RHC emission decrease.  VMT growth and
the increase in NOX exhaust emissions from 1966 to 1969 model year vehicles
were the major contributors to the total NOX increase.

     Projections of emissions made in the early 1970's  (19) have turned out to
be overly optimistic when compared to emission trends which we now determine
in retrospect.  The projections made in the early 1970's indicated a  20%
increase in NOX emissions and a 40% decrease in RHC emissions from 1965 to
1974.  The error in the earlier projections is caused partly by the fact that
test data for automobile3 do not show as large emission reductions as would
have been expected from the time-table of tightening emission standards.
Evaporative emissions, for example, are much higher than called for by federal
and state standards.  For RHC, the new reactivity scale we are using  now leads
to a more pessimistic picture because it gives greater weight to source cate-
gories that have been reduced little in the last ten years.  For NOX, the
shortage of natural gas for industrial interrupt]'bles and power plants and the
resultant unforeseen switch to fuel oil have produced higher emissions than
originally projected for stationary sources.  The occurrences of these unfore-
seen circumstances should serve as a warning in making  future emission projec-
tions.  A sensitivity analysis of air quality to the possible range of emissions
should be included as part of future studies.

     Emission trends reflect a competition between reductions from emission
control and increases from source growth.  Since the geographical distribution
of source growth, in particular traffic growth, has been very nonuniform over
the Los Angeles region, the geographical distribution of emission trends is
also very nonuniform.  Among the counties in the basin, Los Angeles County has
grown the slowest, while Orange County has expanded at  the greatest rate. This
nonuniform spatial distribution of emission trends is important in inter-
preting air quality trends.  What appear to be contradictory air quality
trends at different locations in the basin can often be explained by  the
geographical distribution of emission trends.  In analyzing  future strategies,
the geographical distribution of emission changes should be  accounted for.

     The basin-wide trend in ambient NOX concentrations  (about a 35%  increase)
agrees well with the basin-wide trend in NOX emissions  over  the past  decade  (a
36% increase).  The geographical distribution of NOX emission trends  and NOX
air quality trends on a county-by-county basis also agree.  Ambient data on HC
trends are very sparse.  However, the average trend in NMHC concentrations at
                                     1091

-------
four locations (a 15% decrease)  is  consistent  with  the  basin-wide change  in
RHC emissions (an 18% decrease).

     The basin-wide improvement  in  oxidant  air quality  (19% over the decade)
agrees well with the basin-wide  reductions  in  RHC emissions.  The geographical
trends in oxidant agree,  qualitatively,  with  the geographical trends in  RHC
emissions, especially for Los  Angeles  County.   If we  account  for the dual  role
of NOX we obtain an even  better  clarification  of the  geographical distribution
of oxidant trends.   The increase in NOX  emissions helped  to decrease the  RHC/
NOX ratio; this resulted  in substantial  oxidant reductions  in the source-
intensive, central-coastal  areas.   Little  or  no improvement in  transport
related oxidant occurred  in the  downwind,  eastern-inland  areas  because overall
RHC emissions were reduced  only  slightly and  (possibly) because NOX emissions
increased.
                                  REFERENCES

 1.   J.  Trijonis, T.  Peng,  G.  McRae,  and  L.  Lees,  "Emissions  and Air Quality
     Trends in the South Coast Air Basin,"  EQL Memorandum  No.  16,  Caltech
     Environmental Quality  Laboratory,  Pasadena, California,  January 1976.

 2.   B.  Dimitriades,  "The Concept of  Reactivity and  Its  Possible Applications
     in  Control," Proceedings  of the  Solvent Reactivity  Conference, EPA
     650/3-74-010, November 1974.

 3.   TRW Environmental  Services, "Impact  of Reactivity Criteria on Organic
     Emission Control  Strategies in Los Angeles,"  EPA Contract No,
     68-02-1735, Redondo Beach, California,  July 1975.


 4.   Automotive Environmental  Systems,  Inc.., "A Study of Emissions from
     Light-Duty Vehicles in Six Cities,"  EPA Contract No.  68-04-0042,
     Westminster, California,  June 1972.

 5.   Automotive Environmental  Systems,  Inc., "A Study of Emissions from
     1967-1974 Light-Duty Vehicles in Los Angeles  and St.  Louis,"  EPA
     Contract No. 68-03-0390,  Westminster,  California, October 1974.

 6.   California Air Resources  Board,  "Exhaust Emissions  from  Privately
     Owned 1966-72 California  Automobiles -  A Statistical  Evaluation of
     Surveillance Data," El  Monte, California, March 1973.

 7.   Environmental Protection  Agency, "Compilation of Air  Pollutant
     Emission Factors,"  AP-42,  Second Edition, April 1973.

 8.   Private  communication,  D.  Bratton, California Air Resources Board;
     "Interim Estimates  of  Emissions  from Mobile Sources in California,"
     unpublished draft  by the  California  Air Resources Board  staff, dated
     April  1975.
                                     1092

-------
 9.  C.  D. Paulsell,  "Mobile  Source  Evaporative  Emissions," an EPA dis-
     cussion paper,  June 1974.

10.  California Air Resources Board,  "Public  Hearings to Consider Fuel
     Evaporative Emission Regulations for Light-Duty Vehicles," No. 75-7-6,
     April 1975.

11.  D.  R. Bartz, et al., "Control of Oxides  of  Nitrogen from Stationary
     Sources in the  South Coast Air  Basin," prepared for the California
     Air Resources Board (ARB 2-1471),  Report No.  5800-179, KVB, Tustin,
     California, September 1974.

12.  P.  K. Mazaika,  "Trends of Energy Use in  California and the South
     Coast Air Basin,"  Environmental  Quality  Laboratory, California
     Institute of Technology, Pasadena,  California, May 1975.

13.  Los Angeles Regional Transportation Study,  "LARTS Base Year Report
     1960," California  Division of Highways,  Los Angeles, California,
     December 1963.

14.  Los Angeles Regional Transportation Study,  "LARTS Base Year Report
     1967 Origin Destination  Survey," California Division of Highways,
     Los Angeles, California, December 1971.

15.  Private communication, G.  Bennett,  Transportation Studies Analysis
     Group, LARTS branch, California  Department  of Transportation, Los
     Angeles, California, September  1975.

16.  Private communication, H.  Linnard,  California Air Resources Board,
     October 1974; "Motor Vehicle Population  and Model Year Distribution,"
     a California Air Resources Board working paper, February 1974.

17.  H.  Mayrsohn and J.  Crabtree, "Source Reconciliation of Atmospheric
     Hydrocarbons,"  Atmospheric Studies  Section, California Air Resources
     Board, El  Monte, California, March  1975.

18.  J.  Paskind and  J.  Kinosian, "Hydrocarbon, Oxides of Nitrogen, and
     Oxidant Pollutant  Relationships  in  the Atmosphere over California
     Cities," Presented at the  67th Annual Meeting of the Air Pollution
     Control Association, Denver, Colorado, June 1974.

19.  L.  Lees, et al., "SMOG:   A Report to the People," Environmental
     Quality Laboratory, California  Institute of Technology, Pasadena,
     California, 1972.
                                    1093

-------
                                                                           22-4
             COMPARISON  OF  PAST  AND PROJECTED TRENDS IN OXIDANT
                   CONCENTRATIONS AND HYDROCARBON EMISSIONS

              R. M. Angus,  E.  W.  Finke,  and J.  H.  Wilson, Jr.*


ABSTRACT
         qaoLity projection*  ^or photochemical oxx.dan.t6 have, been made fioti the.
year* 1980,  1990 and  2000.  Thus, projection*  con*-ider alternative level* o&
hydrocarbon  control, ^or mobile and *tatA.onary  *ource*.   In order to make. the.
projection*, a**wmption*  concerning background concentration* and the. hydro-
carbon /oxA.dant relation*hip Mere ne.ce.MOAy.  The6e a**umption* Mere te*ted
u*i.ng pa*t trend* -in  air.  qaotity and emi^*on*.   The. projection* indicate, that
{,or the. mo*t optani*tic level* o& con&iot,  txignifa-icant xjnp-tovemen^ti in CUA.
qaatity are.  pAOj'e.cte.d &OA many afie.au> within the. United States.  Howe.veA,  even
(tiJMi theJ>e. Jj/npAoveinentis many  o& the.-i>e. 4ajne  aA.eai may not attain the. Motional.
Ambient AxA  QaaLity Standard  &or oitidantb  by the. ye.aA. 2000.


                                 INTRODUCTION

     The U.  S. Environmental  Protection  Agency (EPA) makes projections of
photochemical oxidant concentrations in  order  to compare the effectiveness of
proposed hydrocarbon  (HC) control strategies.   Recently an interagency task
force on motor vehicle goals  beyond 1980 analyzed the effect on air quality of
different pollutant control strategies through the year 2000 (Tuerk 1976).
One of the effects studied by the task force was the expected change in future
year 0  quality.
      /\

     Oxidant air quality  projections were made utilizing two models relating
HC emissions and 0  concentrations.   These  air quality projection models
assume growth rates and future emission  levels for both mobile and stationary
sources. Uncertainties abound in the projections because of these assumptions.
For photochemical oxidant projections, there are two additional uncertainties
caused by inaccuracies in determining:  (1) the proper relationship between HC
emissions and ambient 0   concentrations, and (2) the appropriate background
concentration for typical urban areas.  These  uncertainties are assessed by
studying past trends  in 0 air quality and  estimated HC emissions.  Finally,
using the two air quality models, projections  are made to the year 2000.

*R. Angus, University of  Missouri,  Columbia, Missouri.   This paper was prepared
 while he was on leave with the Office of Air  Quality Planning and Standards,
 U.S.  Environmental Protection Agency, Durham,  North CarolinaA_
  E.  Finke,  formerly  of U. S.   Environmental   Protection  Agency.
  J. Wilson,  Office  of Air Quality Planning  and  Standards, Environmental
  Protection  Agency,  Durham,  North Carolina.
                                      1095

-------
                  TECHNIQUES FOR PROJECTING CONCENTRATIONS


AIR QUALITY MODELS

     The effects of projected changes in HC emissions on 0  concentrations
have been analyzed by using two different means of relating the HC emissions
and the resulting 0  concentrations.
                   X

     The first HC/0  relationship is based on a nonlinear upper limit curve
(Schuck and Papettix'1973).  The upper limit, curve enclosed the highest 0
values observed for a given HC concentration.   The relationship is appro$i-
mated by Equation 1:

         HC = 2.049 0  = 4.792 0 2 + 5.036 0 3                             (1)
                     XX           X

For a given base year 0  concentration, Equation 1 is used to calculate a
corresponding HC concentration (6-99 am).   Assuming a proportional relation-
ship between HC emissions and the HC concentration, a new HC concentration is
projected for the year of interest.   Then, the relationship implied by Equa-
tion 1 is used to estimate the 0  concentration corresponding to the new HC
concentration for the projection year.

     The second HC/0  relationship assumes that the 0  concentration is a
linear function of HC emissions.   Therefore, a change in HC emissions is
assumed to result in a proportional  change in 0  concentrations.
                                               X

     Both of the above calculation procedures assume a zero background concen-
tration of HC and 0 .
ASSUMPTIONS USED IN MODELING

     Projections of 0  air quality were made for 37 areas in the United States.
The assumptions on wh¥ch the 0  projections are based are summarized in Table
1.                            x

     Four assumptions are made which immediately affect the quantity of esti-
mated HC emissions from mobile sources.  First, for light-duty vehicles one of
several possible emission standards for HC is assumed to take effect in 1978.
Second, the replacement rate for light-duty vehicles is assumed to result in a
complete turnover in the vehicle population every 13 years.   Next, the deteri-
oration rate of control devices for mobile sources is reduced from the present
rate by assuming that inspection and maintenance plans will  be in operation.
Credits for inspection and maintenance are calculated using the assumptions
listed in AP-42, Supplement 5 (EPA 1976).   Lastly, the exhaust emissions for
light-duty trucks and gasoline and diesel-powered heavy-duty vehicles are pro-
jected based on the emission factors of AP-42.   It should be noted that the
emission reduction assumed for these latter vehicles is generally less than
that assumed for light-duty vehicles.
                                     1096

-------










GO
o
1 — 1
I—
UJ
o

Q_
>_
| —
i — i
	 1

=}
o-

1 — 1
et;

1 —
5p
Q
I — i
x
o

^
o
CO


0-3:
E 0
0 S-
<_) CJ3




0 S-
|Q
O CO
1- >-
re co
Q£ to
re
S- CO
o
4J O
O +J
re
1 i f™
UL_ t_
0
c *t~
O (O
•r— {/)
•}
s-
o
CD
Q}
4-)
(O
0

CO
0
S—
0

o
•r—
V)
°i
UJ











o
o
CM


O
en
r^



o
CO
Cn
1


o
CT^
,_



^J-
to
en

































CO
CO
r—
0

C"
O)
>
$
3
Q
^j
01
•t—
— 1


^5
O

CO





o r-^ CM oo to
CM •— f— O O



i — i*-- CM en to
CM r— r— O O




r— o co r~^ to
^f *^~ CO CO CO




o o o o o





00 00 00 00 CO
«d- ^j- ^j- ^j- ^}-













+
•4- *^
o LO en ^t~ > —
• • • • •
CM r—











to
^
re
CO
>>

to i —
CO TJ CO "O
II i— J. -0 C
•r- re o o
to E T3 E >»
C ^v, C CO
o to re o .a
•r- E •»-> 4J
 re co ~o
to S- -o c
T- 05 c co re
e o T-
UJ CO T- r— 00
• (/> Q. r^
o o to o_ en
r-- i — i- re r—
en E
r— UJ



^5
O
•
co




^c
r—



^J-
•— |




en
co
*



o
"




CO

j= cn
•*-> T-
O _l











LO
«-^



00
"— ;




LO
to
*



0
"




^~
I —


































O)

•r—
r—
O
to
re
o
£>
3
O

re
CO












co
CM



r-.
CM




fN^
CO
*



o
"





o




































f—
0)
to
CO
•r-
Q
-»->
3
O
1
>,
>^
re
O)




^^
CM
*
co




LO
CM



O
LO




LO
CO
*



O
1— J_
3
0
U
0

O (/)
• CO
c
re
.c
o

CO
>
o
-0
re

r—
•r-
4^
C
3

^

Q.
0.
re

(/)
•o
i.
re
TJ
c
re
4-)
to

C
to O
CO i-
o to
S- to
3 •!-
O E
CO CO
i. c
« £
0 J-
•<- 3
4J CJ
re -K
CO




















































•
T3
s_
re
-a
c:
re
4^ "O
co s_
re
C XJ
o c
•i- re
to 4->
to CO
•r—
E c
UJ O
•r-
E t/)
•f- IO
S- i-
CO E
4-> UJ
C
~ t
4-> O
^— 1 \
CO 3
to 4->
co re
i- 4->
Q. CO
+ +

1097

-------
     Growth rates have a major effect on HC emissions projected to future
years. A growth rate of three percent compounded annually was selected for all
mobile sources.  This growth rate is representative of the historical growth
in vehicle miles traveled in metropolitan areas.  Stationary source growth
rates were estimated from economic indicators from four groupings of sources:
electric generation units, industrial activities, area sources and miscellaneous
point sources.  The stationary sources of HC are expected to grow at a 3.2
percent compounded rate.


PAST TRENDS

     To help assess some of the assumptions used in the protective air quality
models a study has been made of past trends in 0  air quality and estimated HC
emissions.  For this comparison, the most effective sets of data come from
California.  In general, trends in air quality have followed the trends in HC
emissions.  For example, in Los Angeles HC emissions have been decreasing
steadily since the mid-1960's.  It was estimated that the average emissions of
HC have decreased by 15 to 18 percent between 1964 and 1974 in Los Angeles;
during this same period peak 0  concentrations for six locations in Los Angeles
have decreased an average of 20 percent (Trijonis et al.  1976).

     Figure 1 presents a summary of the observed changes in 0  concentrations
in the Los Angeles Air Quality Control Program (AQCR).  In addition, 0
concentration projections are shown.  The upper trend line represents £he
monitoring site with the least improvement in 0  concentration from 1964 to
1974, and the lower line represents the site with the most improvement.  The
middle trend line is the median of six sites in the AQCR.

     Noting that the projection lines are in agreement with the general trend
in air quality, it appears that the assumptions concerning the relationships
between HC emissions and 0  concentrations are reasonable.
                          X

                             PROJECTION RESULTS

     For the areas that were modeled, the base year concentration was the
second highest one-hour annual average for 1972-1974.  Table 2 shows the
results of the upper-limit curve projections for an HC emission standard of
0.4 grams/mile.  These calculations showed an average improvement in 0  air
quality of 15 percent by 1980, 42 percent by 1990 and 50 percent by 2o5o.  As
mentioned previously, these projections depend on the growth assumptions and
the assumption concerning implementation of the emission standard in 1978.

     In spite of these improvements in Oy air quality many areas "are not pro-
jected to meet the ambient standard for u  by 2000.  Table 2 shows that 26 of
the 37 areas modeled will have concentrations in excess of the 0.08 ppm stand-
ard in 2000.  In addition, none of these cities is projected to have 0
concentrations which are at least twice the standard.

     It should also be noted that there is a significant difference in the
results obtained for the two models used.   A comparison of the effect for
                                     1098

-------
luepixo UL




        1099

-------
                  TABLE 2.  PROJECTED OXIDANT AIR QUALITY*

BASE YEAR
CONCENTRATION

004 Birmingham
005 Mob.-Pensacola
013 Clark-Mohave
015 Phoenix-Tucson
024 Los Angeles
028 Sacramento Valley
029 San Diego
030 San Francisco
031 San Joaquin
033 SE Desert
036 Denver
043 NY-NJ-Conn.
045 Philadelphia
047 Nat. Capitol
067 Chicago
070 St. Louis
079 Cincinnati
105 S Lou-SE Texas
115 Baltimore
119 Boston
131 Minn-St. Paul
153 El Paso
160 Gnsse-Fngr Lks.
173 Dayton
193 Portland
197 S.W. Penna.
214 Corpus -Christi
215 Dallas-Ft. Wrth
216 Houston-Glvstn
229 Puget Sound
991 Region 1
992 Region 2
993 Region 3
994 Region 4
995 Region 5
996 Region 6
997 Region 7
(PPM)
.29
.29
.22
.19
.54
.28
.32
.33
.27
.32
.33
.36
.20
.38
.30
.39
.19
.32
.23
.21
.10
.13
.12
.16
.15
.17
.15
.19
.38
.13
.19
.19
.19
.17
.17
.20
.11
Average Percent Decrease
No. Cities Above Std.

1980
CONCENTRATION
(PPM)
.24
.24
.18
.15
.46
.25
.25
.29
.24
.28
.26
.30
.15
.30
.29
.36
.15
.28
.22
.17
.08
.09
.09
.13
.13
.14
.15
.16
.36
.09
.16
.16
.19
.16
.14
.19
.09
15.
36
1990
CONCENTRATION
(PPM)
.17
.17
.12
.10
.34
.18
.16
.22
.18
.21
.16
.20
.09
.19
.23
.28
.10
.20
.17
.11
.05
.05
.05
.08
.08
.09
.12
.10
.29
.05
.11
.11
.15
.11
.09
.15
.06
42.
30
2000
CONCENTRATION
(PPM)
.15
.,15
.10
.08
.30
.15
.15
.18
.15
.17
.15
.18
.08
.18
.18
.23
.09
.17
.14
.10
.04
.05
.04
.07
.07
.08
.09
.09
.24
.05
.09
.09
.12
.09
.08
.12
.05
50.
26
*Light Duty Vehicle Emission Standard = 0.4 grams/mile

Projected Cone:  Second highest 1-hour concentration in ppm
Base Year:  Base Year concentration is second highest 1-hour annual
            average recorded from 1971 to 1974
                                   1100

-------
various assumptions concerning HC emissions and 0  background concentrations
on the long-range projections to the year 2000 isxpresented in Table 3.   At
the statutory emission standard of 0.4 grams/mile the empirical  upper-limit
relation projects an average air quality improvement of 50 percent by the year
2000.   The linear assumption projects an average improvement of 61 percent if
background 0  concentrations are neglected and a 46 percent improvement if a
0.05 ppm background is assumed.

     These differences represent the average improvement for 37 areas.   Indi-
vidual cities may show improvements that differ from these average changes
because of the different magnitudes of base-year concentrations.   For example,
a background concentration will  have a smaller effect on the projected improve-
ment in cities with high 0  concentrations than for cities with low concentra-
tions.  In addition, there will  be less difference between the results  of the
two projective models for cities with low concentrations than for cities with
high concentrations.  These effects are illustrated by comparing the number of
cities which are projected to violate the ambient air quality standard in
2000.  With an emission standard of 0.4 grams/mile the upper-limit curve pro-
jects  that 70 percent of the cities considered will still  be in violation of
the air quality standard in 2000.  The linear assumption with zero background
projects that this number will be reduced to 49 percent.  However, when the
background of 0.05 ppm is used,  the number of cities in violation of the
standard is projected to be 86 percent.


                                   SUMMARY

     The methods used to predict 0  concentrations indicate that even under
stringent control measures the 0  standard will not be met in many cities.  It
is predicted, however, that the 6  concentrations can be reduced by as  much as
46 to 61 percent under the most optimistic assumptions.


                                 REFERENCES

 1.  Schuck, E.  A.  and R.  A.  Papetti.   Examination  of the  Photochemical  Air
    Pollution Problem in  the Southern  California Area,  Region  IX,  U. S.
    Environmental  Protection Agency,  San  Francisco,  California,  October 30,
    1973.

 2.  Trijom's, J.  C.;  T.  K.  Peng;  G.  J.  McRae;  and  L.  Lees.   Emissions and
    Air Quality Trends  in the South  Coast Air  Basin,  Environmental  Quality
    Laboratory, California  Institute  of Technology,  Pasadena, California,
    January 1976.

 3.  Tuerk,  E.  Air Quality,  Noise and  Health,  Report of a  Panel  of the
    Interagency Task  Force  on Motor  Vehicle  Goals  Beyond  1980, U.  S.
    Environmental  Protection Agency, Washington, D.  C., March 1976.

 4.  U.  S.  Environmental Protection Agency,  Compilation  of  Air Pollutant
    Emission Factors  (AP-42), Second  Edition,  Research  Triangle  Park,
    N.  C.,  February 1976.

                                     1101

-------






oo
I—
2^
-
1 —
t— H
	 1
— i
«=£
CO
UJ
— 1
CO
<
1 —














CO

r-.

vo

co *
0) LO
CO
(C
CJ

*d-


co



CM


•—














CO
c
o
T-
P
Q
E
3
CO
CO
<





+
+
<3~ CO LO CM
r-. _J
+
*± r- LO _l
r- ID
+
r— OO LO _1
r>« rs
+
«*• co LO _i
f>. 3


4.
cn oo LO _J
• r*. 3

+
*J- CO LO r—
r>» _j


+
LO CO LO _l
t-^ =3
r—

^- CO O — 1
LO ID








CO
i- fc« »S
^•"•H 1 1
O)
, 5
n)
-o o:
S-
(0 .E r- TJ
T3 -P O O
C 3 i- .C
(O O P P
P i- C • >• OO T-J
o
>>>>>> s-
& ff £ 0.
3 3 to
Q Q C P
o c
P -P -i- «
x: J= P XJ
O> CD  X
_1 _J CO O

VO
•*
LO
CM
LO

0
LO



VO
<3"

vo
«s-



5

Cf>
CM











•P
C
n
CO
>
2
CL
B. *
t-t *
O
CO O
cno
IO CM
•P
c o
CO P
o
i-o
cor^
o_ cr>
^_
CO
ens
gg
cot-
«a:
a>
•si-
CT>
LO
CM
vo

0
1^.



VO
r>»

vo
00



 O
CO C
a: o
o
t-
CO T3
cn c
i. 3
to p "o
_i £- c
cn 3
1^ ^: o
0 t-
•o  p
•i- i. S- LO S~
•P  • N
UJ C
IO -P JC .C
CO -Q i- -P 4->
r— S- E T- T-
-Q Z3 -r- 5 2
(0 —1
CO S- S-
O CO S- IO IO
CO CO CO CO
10 c: a. c c
a> o Q. i- T-
o: ZD _j _i
•a
•P CO II II II
CO CO
o 
-------
                                                                             22-5
                       TRENDS  IN  AMBIENT LEVELS OF OXIDANT
                  AND THEIR  POSSIBLE UNDERLYING EXPLANATIONS

                  E. L. Martinez,  N.  C.  Possiel, E. L. Meyer
                  L. G. Wayne,  K.  W.  Wilson, and C. L. Boyd*
ABSTRACT
           papeA deACAibu woAk  andeAway within the. Ofifii.ce.  ofi  kin. Quality Plan-
ning and StandoAdii, U. S. EnviAonme.ntal Protection Age.nct/,  to Ae.v-iew and eval-
uate, trend*  ofi ambient o aidant /ozone. concentration* at Aite*  -in. the nation
that have. suitable amount!, ofi h4AtoAA.cal data..  Among the.  tAend paAametcAA
being examined annually aAe-  (1}  the. second htghe^t hourly concentration;  (2)
the. average.  dajJLy maxAjnum hourly concentration; and  (3) the. peA.ce.nt o{, day-i,
Mith vatid kouAJLy A&adingA exceeding the. National. Ambie.nt  MJI Quatity StandaAd.
In aAe.aA having moAe. than one. A-ite.,  •dpe.ctal emphaAiA it> gi.ve.n to the. spatial
vaAiation. o
     TAe.ndi>  -in oxA.dant/ ozone. le.ve.lA  aAe. oJU,o beting e.xamine.d in Ae-lation to
hydAocoAbon  and nttAoge.n oxx.de-6  e.miAton friend*.  A majoA.  objective. o& evo€u-
ating the.&e.  AeJLation&hipA Lt> to  de^teAmtne. L^ tAe.nd anaJtyt>ti> can be o6ed to
e-valaate. the. e^ec^cvene^-6 o&  ox^.dant control pAogAam*.  Ambtent oxA.da.nt/ ozone.
tAe.ndi> and coAAe^ponding ern^ci-i-con  tAe.nd data have, been examined in Ae.veAal
CaLihonnJja. MA BOAJM.A; Ve.nveA, ColoAado;  Philadelphia, Pe.nnt>ylvanta;  and a
non-uAban  bite, -in wuteAn Maryland.
      In  oAe.oi,  wheAe. Atatii>tically ^tgni^-icant &ie.ndA in amb-ie.nt ox4.dantf ozone.
le.veJLi> have, been faound, theAe. WOA Ae.a&onable. coAAUpondcnce. to  tAendb -in
hydAoroAbon e.mu>AtonA.  Howe.veA,  the. combined hydAocaAbon  and oxtde-6 o^ nitAo-
gen e.rniA-!>i.on trundi, w&ie. fiound to be. moAe. Atgnifiicantly coAAclate.d to oxA.da.nt/
ozone. &ie.ndA than tho&e. oft hydAocaAbon e.miAA'ionA alone..
                          INTRODUCTION  AND BACKGROUND

     In  several  locations across  the United States, where  attainment of the
National  Ambient Air Quality Standards (NAAQS) for photochemical  oxidants (Ox)
(0.08 ppm — one  hour average)  is  unlikely, additional measures  to reduce
*E. L. Martinez  and N.  C. Possiel, U.S.  Environmental Protection  Agency,
 Research Triangle  Park, North Carolina;  on  assignment from the National
 Oceanic and Atmospheric Administration,  U.S.  Department of Commerce.
 E. L. Meyer,  U.S.  Environmental Protection  Agency, Research Triangle  Park,
 North Carolina.
 L. G. Wayne,  K.  W.  Wilson, and C. L.  Boyd,  Pacific Environmental  Services,
 Inc., Santa Monica,  California.


                                      1103

-------
precursor emissions are necessary.   Since these measures may sometimes be
expensive or difficult to implement, the question arises, "Have past emission
controls been effective in improving ambient concentrations of photochemical
Ox or ozone (Os)"?  Perhaps an effective means for answering such a question
is through ambient air quality trend analysis.

     Observed ambient, Ox/03 trends are likely to depend upon a number of
factors. Among these factors are population and industrial growth patterns,
trends in non-methane hydrocarbons (NMHC) emissions, changes in reactivity of
the organic mix, trends in emissions or ambient concentrations of oxides of
nitrogen (NOx) and annual variations in the frequency of meteorological con-
ditions conducive to pnotochemical  Ox formation.  It should be noted that in
the U.S., Ox/03 control strategies have stressed reductions in non-methane or
reactive NMHC or RHC emissions.

     Results of Ox/0o trend studies have been reported in the literature.
Notable among these have been those of Altshuller (1) and Trijonis et al. (2).

     Altshuller (1) reviewed data from six Continuous Air Monitoring Program
(CAMP) sites in the United States.   These were central city sites located in
the following U« S. cities:  Chicago, Cincinnati, Denver, Philadelphia, St.
Louis, and Washington, D,  C.  He noted significant reductions both in observed
levels of ambient Ox/03 and in the frequency of high levels between the two
periods studied, 1964-66 and 1971-73, at five of the six sites,.  The reduc-
tions in the frequency of the highest concentrations (>0.125 ppm) appeared to
be the most dramatic, ranging from 67 to 94 percent at the stations exhibiting
downward trends.  Perhaps a portion of these long-term trends could have
resulted from downward trends in NMHC emissions and upward trends in local
nitric oxide (NO) emissions from motor vehicles between the tv/o periods.
However, no concerted effort was made to relate ambient Ox trends to precursor
emission trends.

     A comprehensive analysis of air quality and the relation to emission
trends in the California South Coast Air Basin (i.e., Los Angeles area) was
prepared by Trijonis et al, (2) in a California Institute of Technology re-
port.  Table 1 provides a synopsis of emission and air quality trends pre-
sented in the Cal.  Tech.  report.  The analysis covered the 10-year period
1965-74.  An approximation of the prevailing wind direction can be thought of
as oriented from west to east or left to right in the table.   In general, the
large geographical  differences in HC and NOX emission trends were found to
agree reasonably well with the observed geographical variations in ambient 0
(average daily 1-hour maximum) and N02 (annual average) concentration trends?
In evaluating geographical agreement between ambient Ox and precursor emission
trends, consideration of local NO depletion and transport must be taken into
account.   Overall,  the Basin-wide trends indicated:

     (1) a 36 percent increase in NOX emissions was  accompanied by a 35 per-
         cent increase in nitrogen dioxide (N02) concentrations; and

     (2)  an 18 percent reduction in reactive HC emissions was accompanied by a
         19 percent reduction in ambient levels of Ox.   The Cal. Tech. analysis
         provides strong evidence that reduction of organic pollutants reduces


                                    1104

-------
                 TABLE 1.  EMISSIONS AND AIR QUALITY TRENDS  IN
                     THE SOUTH COAST AIR BASIN, 1965-1974
PREVAILING WIND DIRECTION i^
COUNTY
% CHANGE IN
RHC EMISSIONS
% CHANGE IN
NOX EMISSIONS
% CHANGE IN
COUNTY-WIDE
OXIDANT
(# OF STATIONS)
% CHANGE IN
COUNTY-WIDE
N02(#OF
STATIONS)
%OF BASIN
POPULATION
(1970)
SANTA
BARBARA
- 9%
+69%
NO
DATA
NO
DATA
2.6
VENTURA
- 2%
+42%
NO
DATA
NO
DATA
3.7
LOS
ANGELES
-24%
+25%
-31%
(8 SITES)
+19%
(8 SITES)
68.7
ORANGE
+ 6%
+89%
- 12%*
(2 SITES)
+93%
(2 SITES)
13.9
RIVERSIDE
- 8%
+69%
+12%
(2 SITES)
NO
DATA
4.5
SAN
BERNARDINO
-17%
+38%
+ 2%
(1SITE)
+47%
(1 SITE)
6.7
BASIN-WIDE
AVERAGE
-18%
+36%
-19%
(13 SITES)
+35%
(11 SITES)

 *ONE SITE REPORTED AN INCREASE OF +78%, THE OTHER A DECREASE OF -31%

  (FROM TRIJONIS ET AL. • REF 2)
         ambient levels of Ox.  However, an increase in NOx emissions  espe-
         cially in the vicinity of the Ox monitoring sites may  have  influenced
         the ambient 0  reductions.
                      X

     More recently, Pacific Environmental Services, Inc.  (PES),  in a study
done for the U. S.  Environmental Protection Agency  (EPA), performed  an analysis
of trends in ambient Ox/03 and precursor emission levels  for  areas where  these
trends have not previously been extensively evaluated.  The remainder  of  this
paper describes the PES Study and its principal findings  (3)  supplemented by
related trend analyses performed within EPA.


                         DESCRIPTION OF TRENDS  DATA

     In the PES Study, consideration was limited to locations other  than  the
Los Angeles Basin with 5 or more consecutive years of  rather  complete  data.
At least a 5-year period was considered to be necessary to minimize  possible
meteorological biases.  If the period of record is much shorter,  the meteoro-
logical influences on air quality trends are likely to be dominant.  Over a
sufficiently long period, variations due to meteorological anomalies are
likely to be identified and the effects of emission changes,  growth, etc.
probably will have a more consistent influence  on ambient trends.

     Initially, only the years 1970-74 were considered.   This period had  the
broadest common data base.  Yet nearly all the  suitable ambient  data sites
were located in California.  Changes in analytical method, missing data,
                                     1105

-------
changes in monitoring locations or, as was often the case, a brief history of
monitoring for Ox/03 limited the base of suitable trends data outside of
California.  Suitable sites were located in the San Francisco Bay Area (9
sites), San Joaquin Air Basin (Stockton, Modesto), Southeast Desert Air Basin
(Lancaster), Sacramento Valley Air Basin (Redding), North Central Coast Air
Basin  (Salinas), all in California; and in Denver, Colorado; and Philadelphia,
Pennsylvania.  In addition to the ambient data, HC and NOX emissions data for
the individual years 1970 through 1974 or information that would have indi-
cated  the relative trend during the period were obtained, as available, from
EPA and state/local air pollution control agencies.

     Ambient trend data were also examined for a non-urban site, Garrett
County, Maryland, located in the extreme western part of the state.  This non-
urban  site was the only one found to have a rather lengthy period of record (4
years).  However, measurements were taken only during the summer Ox season.
For this particular site, only ambient data and apparent trends, if any, were
to be  noted. No attempt was to be made to evaluate the relation to emission
trends since there is presently no reasonable, non-arbitrary means of assign-
ing an area of emissions that would significantly affect a remote non-urban
site.

     A map of the California Air Basins and station locations outside the Bay
Area Air Basin are shown in Figure 1.  Figure 2 shows station locations in the
Bay Area Air Basin.  Site locations outside of California are shown in Figure 3.


                      DATA ANALYSIS AND INTERPRETATION

     The following ambient 0 /03 trend parameters were analyzed at each urban
site annually, the:         x

     •   2nd high hourly concentration

     •   annual average daily maximum hourly concentration

     •   third quarter average daily maximum hourly concentration

     •   frequency of days with hourly concentrations greater than 0.08 ppm
         (NAAQS)

     •   frequency of days with hourly concentrations greater than 0.16 ppm
         (2 x NAAQS).

     Table 2 shows the indicated direction of the first four trend parameters
at all sites during 1970-74.   The fifth parameter, values greater than 0.16
ppm, occurred too infrequently to evaluate trends.  Note that different trend
parameters may indicate contrasting trend directions.  Only the trends of the
pooled values of the 2nd high concentration in the San Francisco Bay Area and
the average daily maximum in Denver proved to be statistically significant at
the 5 percent level.   The trends were downward in both cases.
                                     1106

-------
      STATIONS

 SF   SAN FRANCISCO
 RC   REDWOOD CITY
 SJ   SAN JOSE
 FR   FREMONT
 SL   SAN LEANDRO
 OK   OAKLAND
 Rl   RICHMOND
 PT   PITTSBURG
 LV   LIVERMORE
   Figure  1.  Bay area  air basin station locations.
TABLE  2.  TREND  DIRECTIONS FOR  VARIOUS OXIDANT INDICES,
       AT  INDICATED MONITORING  STATIONS, 1970-74

SECOND-HIGH
AVERAGE MAX.
3RDQTR. AVG. MAX.
FREQUENCY "• 0 08 ppm
SF ^
-
-
-
-
RC
-
+
,-
-
SJ
+
+
-t-
+
FR
-
+
-
-
SL
-
+
+
-
OK
-
+
-
-
Rl
-
-
-
-
PT
-
-
-
-
LV
-
+
+
+
RD
0
+
+
-t-
ST
+
4-
+
-f
MO
-
-
-
-
SA
-f
+
+
+
LN
+
+
+
+
DV
-
-
-
+
PH
+
+
+
-
SF   SAN FRANCISCO       SL
RC - REDWOOD CITY       OK
SJ   SAN JOSE           Rl
FR - FREMONT           PT
                           STATIONS

                 SAN LEANDRO      LV  LIVERMORE
                 OAKLAND         RD - REDDING
                 RICHMOND         ST  - STOCKTON
                 PITTSBURG         MO  MODESTO
SA - SALINAS
LN - LANCASTER
DV  DENVER
PH - PHILADELPHIA
                            1107

-------
          NORTH
          COAST
                      STATIONS
                   RD  REDDING
                   ST  STOCKTON
                   MO  MODESTO
                   SA  SALINAS
                   LN  LANCASTER
               SAN
            FRANCISCO
            BAY AREA
                    SA

                  NORTH
                 CENTRAL
                  COAST
                       SOUTH
                      CENTRAL
                       COAST
                                SOUTH
                                COAST
                                            SAN
                                            DIEGO
               Figure 2.  California  air basin station locations
                          (Outside  San  Francisco Bay area).
     Since the emission  data from the Denver vicinity were not available  in
sufficiently detailed  form,  evaluation of relation between ambient data and
emissions could  not be made.   Therefore, the San Francisco Bay Area was the
only area (outside of  Los  Angeles)  with sufficient information to analyze  for
the relation of  HC and N0v emission trends or control programs to definitive
0  trends.
 X
x
     The ambient Ox data  showing  significant trends in the Bay Area are shown
in Table 3.  Note the  large  improvement in western areas of the Basin  in  con-
trast to eastern areas  (e.g.  San  Jose  and Livermore).   Table 4 is indicative
of HC and NOX emissions trends  in the  Bay Area during the 1970-74 period  (in
box).  Table 4 shows that  overall emissions have not changed much in the  most
recent five years.  Only  HC  emissions  show a slight but gradual decrease.
                                     1108

-------
            Figure 3.  Station  locations outside  California.
TABLE 3.  OXIDANT LEVEL BY YEAR AND
TREND DIRECTION (2ND HIGHEST DAILY
MAX. HOURLY AVG., PPM)
TABLE 4.   ESTIMATED EMISSIONS* AND
COMBINED EMISSION RATIOS FOR HYDRO-
CARBON AND OXIDES OF NITROGEN IN
SAN FRANCISCO BAY AREA, 1962 -
1974

SAN FRANCISCO (SF)
REDWOOD CITY (RC)
SAN JOSE (SJ)
FREMONT (FR)
SAN LEANDRO (SL)
OAKLAND (OK)
RICHMOND (Rl)
PITTSBURG (PT)
LIVERMORE(LV)
'70
.14
.16
.16
.27
.21
.20
.18
.18
.29
71
.15
.16
.14
.24
.24
.21
.14
.18
.21
'72
.07
.17
.17
.29
.14
.09
.10
.18
.18
'73
.08
.12
.18
.18
20
.14
.10
.10
.20
'74
.06
.16
.26
.21
.17
.09
.08
.10
.25
TREND
-
-
4-
-
-
-
—
-
—
YEAR
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
HC
1300
1300
1400
1300
1300
1300
1300
1300
1140
1110
1130
1070
1030
NOX
490
500
550
550
620
620
670
700
740
740
790
820
720
HC/NOx
2.65
2.60
2.55
2.36
2.10
2.10
1.94
1.86
1.54
1.50
1.43
1.30
1.43
                                               *TONS PER DAY
                                  1109

-------
However, more detailed emissions data by county (not shown) and source type
(stationary or mobile) indicate that:  (1) significant HC emission reductions
have been achieved in the western part of the Bay Area; and, (2) mobile sources
NOX emissions have increased in most places.  Air quality and emissions trends
reflect new auto and stationary source emission control programs and differential
area growth patterns.  Notably, some eastern areas of the Basin have undergone
rapid growth, thus overshadowing beneficial effects of emission control pro-
grams .

     The correlation coefficient between pooled second highest Ox concentra-
tions in the Bay Area and regionwide HC emissions was only + 0.58, not statis-
tically greater than zero at the 5 percent  level.  The significance test
involves the Student _t test, as explained by Moroney (4).  The correlation was
greater when both HC and NOX emission trends were related to oxidant trends.
The relation between HC/NOX emission ratio and second highest Ox levels was
considerably higher at + 0.88 and statistically significant at the 5 percent
level.  The improved results are indicative of the important role of NOX
in the photochemical process.  In these results the variable effect of meteoro-
logical factors was not considered.

     A much clearer picture for the Bay Area emerges when 13 years of data are
examined and normalized by considering only days with photochemical Ox conducive
meteorological conditions.  Table 5 denotes the normalized Ox trends.  Over
the 13-year period for which data are available, the downward trend was found
to be highly significant (p<.001).   Thus, it is virtually certain (i.e. 99.9
percent) that improvement in the Bay Area oxidant levels is attributable to
some physical phenomena rather than chance.  The possibility that the improve-
ment was unduly influenced by meteorological factors was minimized by the long
period of record and consideration of ambient Ox levels  only for days with
similar meteorological conditions.  Table 4 shows basin-wide HC arid NOX
emission trends for  1962-74. Note  the much wider range of emissions occurring
during this  longer period.  When HC emissions (see Table 4) were correlated
with normalized Ox levels pooled for the six stations  haying complete records
over the 13-year period  (1962-74), a correlation coefficient of + 0.77, statis-
tically significant  at the  5 percent level, was found.   Moreover, the correla-
tion coefficient between Ox and HC/NOX emission ratio  trends was even more
highly significant, + 0.92, significant at  the 0.1 percent  level.

     The available 03 trend parameter data for the non-urban station at
Garrett County, Maryland, is given in Table 6.   No apparent trends were noted.
What seems evident, however, are considerably large fluctuations in some of
the statistics from year to year.   For instance, the percent of hours that the
hourly concentration exceeded the ambient standard varied by more than a
factor of 3.  Probably, these variations in the large-scale regional 0 /03
levels may be reflected in the levels experienced in urban areas.      x

     An obvious deficiency of this  paper was the lack of suitable trends
information outside of California.   Improvement in  the trend data base, how-
ever,  is forthcoming.  At least ten states  outside  of California now have
sites  with three consecutive years  of Ox/03 measurements.  These data will
eventually provide a data base for  assessing nationwide 0  trends  and their
probable causes.                                          x

                                     1110

-------
  TABLE 5.  AVERAGE HIGH-HOUR OXIDANT CONCENTRATIONS  FOR  DAYS WITH COMPARABLE
  TEMPERATURE AND INVERSION CONDITIONS  (APRIL  THROUGH OCTOBER OXIDANT SMOG
  SEASONS, 1962 - 1974)
AVERAGE HIGH-HOUR OXIDANT CONCENTRATION
(Kl PARTS PER MILLION)
MONITORING
STATIONS
SAN FRANCISCO (SF)
SAN LEANDRO (SL)
SAN JOSE (SJ)
REDWOOD CITY (RC)
WALNUT CREEK (WCI
SAN RAFAEL (SR)
BAAPCD
AVERAGE*
LIVERMORE**
'62
.14
.13
.11
.13
.10
08
.12
--
•63
.12
.16
17
.10
.11
.09
.12
--
'64
.15
.19
14
.10
10
.07
13
--
'65
.09
.19
.16
.14
.11
.08
.13
--
'66
.08
.14
.11
.10
.10
.07
.10
--
•67
.08
.12
.13
.09
.13
07
.10
.13
'68
.05
.11
.13
.08
.10
.06
.09
.18
•69
.04
.12
.13
.09
13
.07
.10
.18
70
.07
.12
.12
.08
.09
08
.09
13
71
.05
11
.08
.07
.09
07
.08
.11
'72
.03
10
.10
.08
09
05
.08
.09
'73
.04
.11
11
.07
.08
05
.08
.12
'74
.05
.10
.16
07
.08
.06
.09
.13
13
YEAR
.08
.13
.13
.09
10
.07
.10
.13
OXIDANT TREND
DIRECTION
1970-74
ONLY
-
-
+
-
-
—
—
+
ALL
DATA
-
-
—
—
—
—
—
—
   •FOR BENCHMARK STATIONS ABOVE, WITH 13 YEARS OF RECORD

   •STATION WITH 8 YEARS OF RECORD
                                                   	NO DATA
(SOURCE:   INFORMATION BULLETIN 3-25-75:  A STUDY OF OXIDANT  CONCENTRATION  TRENDS;
 TECHNICAL SERVICES DIVISION, BAY AREA AIR POLLUTION CONTROL DISTRICT)

                          TABLE  6.   OXIDANT/OZONE TREND
                          STATISTICS FOR  GARRETT CO.,
                          MARYLAND
STATISTIC
MAX HOURLY CONC. (ppm)
MEAN HOURLY CONC. (ppm)
% DAYS HOURLY CONC. > 0.08 ppm
% HOURS CONC. > 0.08 ppm
YEAR
'70
0.13
0.05
40
27
71*
--
--
--
--
'72
0.12
0.05
40
11
'73
0.15
0.07
78
37
'74
0.15
0.06
43
13
                 *NO SAMPLING PERFORMED IN 1971
                             REVIEW AND CONCLUSIONS

      The following statements summarize trends information as described  in
 this  report.   Long-term trends in oxidant in the Bay Area Air Basin  are  signi-
 ficantly related to trends in HC emissions.  Oxidant concentrations  in the  Bay
 Area  are more highly related to HC/NOX emission ratios.  This indicates  re-
 ductions in  NO  emissions may have a tendency to increase Ox levels  within  the
 Basin.   Gxidanr problems in the Bay Area, due to changing patterns of precur-
 sor emissions, are becoming more severe in inland locations than in  Bayside
 locations.   Monitoring data for a 5-year period may not, in general  be suffi-
 cient to establish statistically significant Ox trends in an area.   And  concen-
 trations of  Ox/03 in non-urban locations exhibit large fluctuations  from year
 to year that may have an influence on urban oxidant levels.
                                      1111

-------
                             REFERENCES

Altshuller, A. P.,  "Evaluation  of Oxidant  Results at CAMP sites in the
United States,"  JAPCA 25 (Jan.,  1975).

Trijonis, J. C.,  T.  K. Peng,  C. J.  McPae and L. Lees,  "Emission and Air
Quality Trends in the South  Coast Air  Basin,"  Environmental Quality
Laboratory, California Institue of Technology  (Jan., 1976).

Wayne, L. G., K.  W.  Wilson,  and C.  L.  Boyd,  "Detection and  Interpretation
of Trends in Oxidant Air Quality,"  DRAFT REPORT Prepared Pursuant to EPA
Contract 68-02-1890, Task Order 1  (July 1, 1976).

M. J. Moroney, "Facts from Figures," Third Edition, p. 311, Penquin Books,
Baltimore, Maryland  (1956).
                                1112

-------
                 SESSION 23
ON THE OZONE/OXIDANT CONTROL STRATEGY IN U.S.

                      A.B. Bromley
                OECD, France
                     1113

-------
                                                                           23-1
               TRENDS  IN PHOTOCHEMICAL OXIDANT  CONTROL  STRATEGY

                                  J.   Padgett*
ABSTRACT
     The national, ambient air quality *tandard  far ox.ic evidence ofi health and other adverse  e^ect* *upport thiA
*tandard although data Limitation* even today provide ample opportunity to
chatting?, it.  The ba*ic Environmental Protection Agency *trategy  ^or the,
control Of} oxiidant i& to contnot anthropogenic. e.mJJ>Aion& ofa organic compound*
that can react photochnmicatiy to ^orm ox.idantt>.  It 4eemi c£eoA that thi&
approach i& bound eve.n though much additional, research  -06 needed to answer
question* aAtociated with the e^ecti, o^ ox.idanti> and the be&t approach to
their control.  Oxidantb originally were o& concern only in large  citi&>,
but recent technical, finding* *hoio that high oxA,dant level* (exceeding the
standard] can occur over large geographical area* because ofa the transport
oft ox.idant* and their precursor* over a *everal day period.  Given enough
time, even low reactive organic* can faorm ox.idant* and  are o&  concern.
The emerging oxi.dant control *trat.egy .therefore place*  more empha*i* on
the control o{> ati organic* emitted within large geographical  area*.  In
parall.el with the continuing implementation o&  oxidant  control plan* in
many urban area*, the Environmental Protection  Agency i* developing  new
*ource regulation* to control organic cmi**ion* ^rom chemical,  plant*.
Guideline* and regulation* are al*o being developed to  effectively reduce
organic emi**ion* firom other new and ex.4*t^ng *ource*,  and State*  are
encouraged to apply reasonably available emi**ion contAol* over  broad
geographical, area* where emi**ioyi den*itie* are relatively high.

                                 INTRODUCTION

     To protect the public from the harmful  effects of  photochemical oxidant,
the U.S.  Environmental  Protection Agency (EPA) promulgated a national ambient
air quality standard in April, 1971(1).   This standard  placed a limit on the
amount of photochemical  oxidant that could be present in the air after July,
1975 (with some provisions for extending this deadline  for two years), based
on the best available scientific evidence of health and other adverse ef-
fects(2).   Many difficulties have been encountered as federal, state, and
local governments have attempted to implement air pollution control programs
which would ensure that the standard would be met.  Even today, there are
many places in the country that experience photochemical oxidant levels sub-
stantially in excess of the standard.
 *U.  S.  Environmental  Protection  Agency,  Research  Triangle Park,  North Carolina,

                                      1115

-------
     Our basic oxidant control strategy is to reduce anthropogenic emissions
of organic compounds that can react photochemically to form oxidants.  Even
though the rationale for this approach is well documented(S), questions still
are raised by community and industry leaders responsible for implementing it.
Typical questions are:  Why do we need to meet the oxidant standard?  Are
we concerned only with ozone  (which is the pollutant we monitor) or with all
oxidants?  Is the basic strategy of reducing hydrocarbons a valid one?  What
is the impact on control strategies of hydrocarbon reactivity, background
oxidant levels, and the transport of oxidants and their precursor into and
out of a community?

    There are as yet no simple non-controversial answers to many of the above
questions.  And in some cases, as you have heard at this conference, scienti-
fic opinion is sharply divided on the answers to these and other questions we
could ask.  But if we are to  succeed in our national goal of oxidant control,
it is important for the public, as well as the scientific community, to have
answers to these questions and to be convinced that the control strategies we
employ are needed and will be effective.  Conferences such as this are a
vital part of this process.   Also important is the support of independent
scientific organizations such as the National Academy of Sciences, which
stated in its comprehensive 1976 report on Ozone and Other Photochemical Qxi-
dants(4) the following:

          Despite uncertainties concerning the causative agents
     and their effects, we must proceed with the regulation of
     emissions that lead to the formation of photochemical smog.
     At the same time, research should continue on identifying
     the individual harmful agents in photochemical smog and
     their effects.

     This paper first will consider briefly the scientific support for the
oxidant standard.  We then will describe the current EPA control strategy
for oxidant control and how this strategy is evolving as a result of new re-
search findings.  We will not, however, attempt in this brief paper to try
to answer all of the questions we have posed.

                       SUPPORT FOR THE OXIDANT STANDARD

     The U. S. national ambient air quality standard for photochemical oxi-
dants is 0.08 parts per million (ppm).  This is an hourly average, not to
be exceeded more than once per year.  While few will disagree that high con-
centration levels of ozone and other oxidant cause serious adverse health
effects, some people question the need for such a restrictive standard, and
others question the basis on which the current standard was established.  A
summary of the health implications of oxidants is given in Reference 5.
Questions continue to arise because the data on human health effects from
exposure to low levels of photochemical oxidant still  are limited and hence
may never be completely settled even though many independent scientific
reviews by medical experts support the present standard.
                                      1116

-------
      One such study was commissioned in 1974 when (,on<)rtss asked trie National
 Academy of Sciences (NAS) to review and update the information on health
 effects from air pollution, including those effects caused by photochemical
 oxidant.  After reviewing and interpreting the available data, these highly
 qualified scientists from the various scientific and technical fields con-
 cluded:

           In general, the evidence that has been accumulated since
      the promulgation of the Federal ambient air quality standards
      by the EPA Administrator on April  30, 1971, supports those
      standards.  Hence on the balance,  the (NAS) panels found no
      substantial basis for changing the standards(6).

    Although  this  conclusion  applied to  all  national standards,  the  NAS,
 in  specifically addressing  photochemical oxidant,  noted that  even  with  the
 current standard,  the  risk  to  the  general  population may not  be  negligible,
 and that there  is  evidence  which suggests  that  there may not  be  a  completely
 safe  level  of this air pollutant for all people.   A similar conclusion  has
 more  recently been reached  in  a  report  by  the  World Health Organization
 (WHO),  which  states:

          It  is apparent that any primary protection standard  (for
     oxidant) between  .05 and  .10 ppm will  provide the narrowest margin
     of  safety against some possible detrimental effects in the more
     susceptible segments of the population.  There would seem to be
     little justification for exceeding  .10 ppm for a primary protection
     standard(7).

                           OXIDANT CONTROL STRATEGY

     Our understanding of the photochemical oxidant formation process has
grown over the years because of both improvements in the theory of photo-
chemical oxidant formation and continuing empirical research.   However,
from a regulatory point of view, the most significant fact remains that
oxidants are formed from two types of precursors:  nitrogen oxides and
organic  compounds  (of which hydrocarbons are a major component).  In most
urban areas, the ratio of the organic concentrations to the concentration
of nitrogen oxides is such that the nitrogen oxides are in surplus and
hence the organics are the limiting factor in the set of reactions which pro-
duce oxidants.  Therefore, hydrocarbon and other reactive organic emissions
must be  reduced in order to lower oxidant levels.

     The reductions in emissions of hydrocarbons and other organic compounds
are to be achieved through Federal  and State programs that have been formal-
ized in  regulations promulgated under the Federal Clean Air Act of 1970(8).
The Federal programs provide for the reduction in emissions nationwide
through  the Federal Motor Vehicle Control Program, the Federal program for
control of aircraft emissions, and the development of New Source Performance
Standards.  The State programs provide for addit-onal control  measures through
State Implementation Plans in those areas of the country where the Federal
programs will not be sufficient to meet the air quality standard for oxidants.
                                     1117

-------
     Because roughly half of the hydrocarbons emitted to the ambient air
in the United States currently is attributable to motor vehicles, a large
portion of hydrocarbon reduction is to be achieved through the Federal
Motor Vehicle Control Program.  This program has required progressively
stricter hydrocarbon controls on all light-duty motor vehicles since the
1968 model year.  In 1970, the Clean Air Act required that hydrocarbon
emissions from light-duty vehicles and engines be reduced by at least 90%
from emissions permissible in 1970 model year vehicles.  This will eventually
require emissions to be reduced to 0.41 grams per mile.  Interim standards
of 1.5 grams per mile (g/m) are now required on 1975 through 1977 model
year vehicles.  Regulations also exist to limit hydrocarbon emissions from
other types of vehicles such as motorcycles, trucks, buses, and aircraft.

     The Clean Air Act also authorizes EPA to promulgate "Standards of
Performance for New Stationary Sources".  These are standards for new sour-
ces which reflect the best demonstrated system of emission control, taking
the cost of emission reduction into account.  New sources are sources that
are constructed or modified after the standard has been proposed.

     When Federal programs are not sufficient to meet the oxidant. standard,
the State is required to develop plans for additional reduction in emission
of organic compounds in those areas that will not meet the oxidant standard.
The State Implementation Plans include regulations for existing stationary
sources as well as the implementation, where necessary, of Transportation
Control Plans.

     For the development of Implementation Plans, it is necessary to deter-
mine how much the emissions of hydrocarbons and other organics must be re-
duced.  It was realized that the relatively simple methods used for directly
emitted pollutants were not appropriate for the complex oxidant formation
processes.  Therefore, an attempt was made to quantify the amount, of hydro-
carbon reduction required to meet the oxidant standard.  In 1970, the data
to relate non-methane hydrocarbon to oxidant concentrations were available
only from Denver, Los Angeles, Philadelphia, and Washington, D.C.  The
6 to 9 a.m. average non-methane hydrocarbon concentrations were plotted
against the peak hourly oxidant concentrations observed later in the day
at the same measurement sites.  The points scattered because of the varia-
tions in the conditions that affect oxidant formation.  A curved line was
drawn to enclose the data points within an upper boundary.  This upper limit
curve depicted the highest oxidant levels observed for a given hydrocarbon
concentration.  From this curve the minimum hydrocarbon reductions required
to meet the oxidant standard could be calculated.  EPA published the result
of this calculation as Appendix J in the August 14, 1971 Federal Register.
Given a measured maximum oxidant level in an urban area, Appendix J is to
be used to determine the amount of hydrocarbon control required in that area.
In some cities the measured oxidant levels were above (greater than 0.28 ppm)
the applicable range of Appendix J.  These cities were permitted to reduce
hydrocarbon emissions proportional to the amount that oxidant levels exceeded
the standard.

    The August 14, 1971 Federal  Register also gives emission reductions that
are attainable through the application of reasonably available emission con-

                                    1118

-------
trol technology.  These emission limitations emphasize reduction of total
organic compound emissions, rather than substitution of less reactive organic
compounds, because of the evidence that few organic compounds are photochem-
ically unreactive.  For organic solvent usage, however, such as in surface
coatings and dry cleaning, the guidelines present a list of low reactivity
compounds that may be considered for exemption for control.

     Transportation Control Plans provide for reductions in hydrocarbon
emissions beyond the reductions achieved by the Federal Motor Vehicle Control
Program and stationary source regulations set forth in approved State Imple-
mentation Plans.  Transportation Control Plans typically include measures  to
decrease automobile travel, require inspection and maintenance of vehicles,
and limit vapor emissions during gasoline transfer.

                                RECENT FINDINGS

     Recent field investigations and laboratory studies of oxidant formation
and control have provided several new insights.  It is now apparent that
oxidants are a rural as well as an urban problem.  Oxidants can be formed
over long time periods during stagnant conditions in high pressure systems
or during transport of oxidants and precursors.  This implies that the long-
term behavior of oxidants and precursors is an important contributor to
oxidant concentrations.  It also implies that less reactive organic compounds,
as well as the more reactive compounds, can contribute to observed oxidant
levels.  There are also natural sources of oxidants which at times may con-
tribute to oxidant concentrations reaching levels near the oxidant standard.
The studies indicate, however, that man-made emissions of hydrocarbons are the
predominant source of the highest levels of oxidant.

     Non-methane hydrocarbon (NMHC) and nitrogen oxides (NO ) measurements
made in relatively remote, sparsely populated areas of the nation show ex-
tremely low NO  levels and NMHC/NOx ratios that are sufficiently high that
control of NMHC may be of little benefit in these areas.  For urban areas,
however, these ratios appear to be in the range where a hydrocarbon control
strategy is effective.

     Although the theory of a hydrocarbon control strategy is well establish-
ed and has been validated in smog chamber tests, only confirmation in an ur-
ban environment, which requires years of data because of the predominant
year-to-year influence of meteorological conditions, will be convincing to
many.  Emissions and air quality data are becoming available over a suffi-
cient number of years for a few locations—notably Los Angeles and San Fran-
cisco—to allow meaningful statistical analyses of the effectiveness of
hydrocarbon emission reductions in reducing oxidant levels.  Separate ana-
lyses recently completed for these cities do show good correlations between
reductions in hydrocarbon emissions and ambient oxidant concentrations.


                       IMPLICATIONS FOR OXIDANT CONTROL

     Based on findings to date, implications for oxidant control include
the following:


                                    1119

-------
     1.   Recent findings support the basic EPA strategy, which emphasizes
the reduction of organic emissions in urban areas.   In many urban areas,
the required reduction in emissions is far greater than can be achieved by
use of all reasonably available control technology and it is likely that
some cities will not achieve the oxidant standard for many years.  This
problem has been recognized by Congress as well as the EPA, and legislative
amendments have been proposed that grant extensions in time to meet the
standard if the city has itself committed to employ all reasonably available
control measures to reduce emissions.

     2.   EPA policy has strongly encouraged the positive reduction of non-
methane organics but still recognizes as a useful interim control measure
the substitution of low reactivity for high reactivity solvents.  Since
even a low reactivity organic may contribute in time to oxidant formation,
the use of solvent substitution, even as an interim measure, is now being
reexamined.

     There is growing recognition that organic emission controls often should
extend far beyond the boundaries of a defined urban area.  Statewide or even
multi-state control of organic emissions may be effective in reducing oxidants
in populated regions such as the northeastern United States.  At the same
time, however, emission control in remote rural regions of the nation, such
as portions of the northwestern United States, where NO  emissions are very
low, may be counter-productive.  Further study and analysis is needed to
better define this problem.

     4.   Our improved understanding of the mechanisms for formation of
oxidant and the role of transport emphasizes the long existing need for im-
proved predictive models to replace the present Appendix J curve that relates
peak oxidant concentrations to non-methane hydrocarbon emissions.  The use
of improved models will not necessarily affect our current control programs,
particularly in urban areas where required emission reductions are far greater
than can be achieved, but they will establish more accurately long-range
program goals and facilitate the development of more effective area wide con-
trol programs.

            ACTIVITIES IN SUPPORT OF EPA's OXIDANT CONTROL STRATEGY

     In addition to the extensive research programs needed to provide better
answers to basic questions such as those posed earlier in this paper, a
number of other activities associated with control  implementation are under-
way.  These include:

     1.   Model Regulations to limit organic emissions from existing plants,
based on reasonably available control technology, are being completed for the
paint and coating industries.  New regulations for inspection and maintenance
programs and the recovery of emissions during gasoline transfer also are in
preparation.   These will  provide guidance to states which may incorporate
such regulations in State Implementation Plans for oxidant control.

     2.   New Source Performance Standards.   A comprehensive study of the
synthetic organic chemical industry will result in federal standards to limit


                                    1120

-------
new plant emissions of many organic compounds now uncontrolled.  Guidance
documents for control measures for existing plants also will be provided for
use in State oxidant control plans.

     3.   Reactivity.  The role of photochemical reactivity in oxidant con-
trol strategies will be re-examined with the assistance of State and local
regulatory personnel, members of the technical community, and other concerned
groups.

     4.   Substitutes for the Appendix J Model are being developed to provide
a method for accounting for the effect of NO  as well as oxidant and precursor
transport.  These new calculation techniquesxshould provide more accurate
quantification of the required reduction in organic emissions needed to
achieve a desired oxidant concentration.

                                  CONCLUSIONS

     Recent technical findings indicate the need for oxidant control programs
that span much broader geographical regions than orginally believed.  Also,
because of the role of transport of both oxidants and their precursors over
time periods measured in days, even low reactive organics may have significant
time to form oxidants and should be controlled.   The EPA strategy, which
emphasizes the control of anthropogenic organic emissions, is basically sound
and programs are underway to establish more effective control of these emis-
sions.  The evolving oxidant control  strategy places increasing emphasis on
the control of emissions over broad geographical areas where emission densi-
ties are relatively high.  States are being encouraged to apply reasonably
available emission control  measures in these areas as needed to supplement
Federal standards for vehicle emissions and new stationary sources.

                                  REFERENCES

1.   Federal register, Vol. 36, No. 84, "National Primary and Secondary Ambient
     Air Quality Standards", April  30, 1971.

2.   "Air Quality Criteria for Photochemical Oxidants", U.S. Department of
     Health, Education, and Welfare,  Publication No. AP-63, March 1970.

3.   "Control of Photochemical Oxidants—Technical Basis and Implications of
     Recent Findings:, U.S. Environmental Protection Agency, Publication
     No. EPA-450/2-75-005, July 1975.

4.   "Ozone and Other Photochemical Oxidants", National Academy of Sciences,
     Washington, D.C., 1976.

5.   "The Health Implications of Photochemical Oxidant Air Pollution to Your
     Community", U.S. Environmental Protection Agency, August 1976.

6.   "Air Quality and Automobile Emissions  Control--A Report by the Coordinat-
     ing Committee on Air Quality Studies", Volume 2, National Academy of
     Sciences, Prepared for the Committee on Public Works, U.S.
                                     1121

-------
7.   "Environmental  Health Criteria  for  Photochemical Oxidants", World Health
     Organization, Draft Report EHE/EHC/WP/75.5, March  20,  1975.

8.   Clean Air Act (42 U.S.C.  1857 et. seq.),  Clean Air Act Amendments of  1970,
     P.L. 91-604,  December 31,  1970.
                                    1122

-------
                                                                            23-2
             PROBLEMS  WITH CONVERTING STATE-OF-THE-ART PHOTOCHEMISTRY
                         TO STATE LEVEL CONTROL STRATEGIES

                              W.  Bonta and J. Paisie*

ABSTRACT

     A bsiie.fi history  ofi  the, development ofi the. photochemical  oxidant  control
pfiOQfiam in the State,  ofi  Mafiyljand iA presented.  Included in thiA  hiAtony iA
a. diAcuss-ion ofi the. efifiect ofi changes -in Federal legiAlation, specifically,
the, 1967 Ciean AxA kct and 1970  Clean AXA kct ame.ndme.ntf> on the, photochemical
oxidant control.
     In addition to  thiA,  the e.&fie.ctA o& changes -in  the. & tote.- o&- the.- an£ photo
          aAe. cfx^ca44e,d i.n fioJLaution to the. de.veJLopme.nt ofa the. photoche.mic.ai
oxA.dant control. pfioQUtm -in the, State. o& Maftytand.  Spe.ci^icatty , attejnptA
on the. pant ofa the. State, to de.ve£op an alternative, to the. Fe,de.?ial£y piomuJL-
gate-d "Appendix J" one, alAo dii>ciit>&e.d.  Included in  tku, jj> a. di&cu&Aion o£
the. tieAditb ofa the. aie ofi aeAomztAsic data ancULy&iA and analytical, simulation
o£ smog cliambet e.x.pe.tume.ntt> .
     Finally, obseAvationi*  on the. e,^
-------
THE 1967 FEDERAL ACT

     The accumulated photochemical oxidant data prior to the Clean Air Act
of 1967 was compared to the existing California oxidant standard of 0.10
ppm to determine the severity of the Maryland problem.   Although sparse,
the data indicated that both Baltimore and Washington exceeded the standard
less than twenty times per year and peak values were thought to be around
0.15 ppm.  This problem when compared with that known to exist in Califor-
nia was given a second priority behind suspended particulates (TSP) and oxides
of sulfur (S0x).

     According to the 1967 Federal Clean Air Act amendments, the issuance
of criteria and control techniques documents by the Federal agency
triggered the development of a regulatory program within state governments.
In March of 1970, these documents were issued for photochemical oxidants
and for oxides of nitrogen (N0x).  Following this event, the States were to
establish ambient air quality standards protective of the public health and
the appropriate control programs required to implement these standards.
Throughout the summer of 1970, the available documents were reviewed and
a preliminary assessment of control technology was assembled.  The primary
reference point was that of the compiled information contained in both the
Federal criteria and control techniques documents; there had been no general
awareness within the State of the problems or technological solutions associa-
ted with photochemical oxidants until 1970.

     At the end of November in 1970 the State held public hearings at which
the proposed ambient air quality standard to photochemical oxidants of 0.10
ppm was presented.  The presentation report (1) also contained an assessment
of the current levels and a preliminary finding of the necessary control
program to implement these standards.  As there were no modelling data  shown
in the criteria documents that would apply to ambient levels as low as those
found in the Washington Region, the empirical diagrams showing the associa-
tion between 6-9 a.m. non-methane hydrocarbons (NMHC) and afternoon oxidant
peaks were used.  One of these diagrams, found in the hydrocarbons criteria
document (2), showed that peak oxidant values above 0.10 ppm were not observed
when 6-9 a.m. NMHC levels were not above 0.30 ppm.  Thus, a control program
was sketched out that would reduce the mean levels of NMHC such that the peak
value expected was less than 0.30 ppm NMHC.  The available NMHC data showed
this level was exceeded 97% of the time.  Therefore, it was concluded that a
major reduction in mean NMHC emissions was required in order to achieve the
oxidant standard.  The reduction of the mean 6-9 a.m. NMHC level from 0.65
ppm to 0.10 ppm equalled 85% control and was thought to be available through
controls on solvent usage, gasoline marketing losses, dry cleaning vapor
losses, and the adoption of the then proposed Federal new vehicle control
program which provided a 90% reduction from precontrolled autos.  The control
strategy considered for solvent controls was based on California Rule 66, and
was used simply because it was the only adopted regulation available at the
time.  The gasoline marketing controls were developed by the State as a system
for reclaiming fuel vapors from the retail station filling process.  The
vapors were collected in the tank trucks and when the trucks returned to the
gasoline terminals the processing took place.  Terminal controls were pat-
terned after those used in southern California and the controls were applied
only to newly installed tanks.

                                     1124

-------
     In summary, the State thought that it had the beginning of a viable con-
trol program for obtaining the oxidant goal established pursuant to the 1967
version of the Federal Clean Air Act.

THE 1970 FEDERAL ACT AMENDMENTS

     When the Federal Act was changed on the last day of 1970, the State
postponed adoption of its ambient standard for photochemical oxidants un-
til the Administrator of the newly created EPA could promulgate the nation-
al ambient air quality standards required by the amendments.  During the year
1971, the majority of State effort centered around the development of new
capabilities for measuring oxidants using a continuous, telemetered system
of monitors of the chemiluminescence type.  In addition, detailed inventories
were compiled of all sources of hydrocarbons and extensive studies of mobile
and stationary source technologies were begun.  Land use and transportation
studies were also begun.  Throughout the summer and fall of 1971, regulations
implementing the above controls were designed and presented at public hearings
for both the Washington and Baltimore Metropolitan areas.  The Implementation
Plans which were required by EPA to demonstrate how the State intended to meet
the photochemical oxidant standard were forwarded to EPA in December of 1971
and the regulations were adopted as law in March of 1972.  In the process of
generating the 1971 Implementation Plan for photochemical oxidants, the inven-
tory of hydrocarbons was discovered to be less controllable than previously
thought.  This discovery was due to the facts found in the detailed technology
review which showed more realistic reduction potentials than had been origi-
nally contemplated.  The air quality forecasting model had changed, however,
to a linear rollback of NMHC which was assumed to be proportional to peak
oxidant.  The change was based on the guideline included as Appendix J to the
Federal regulations prescribing methods for preparation, adoption and submittal
of implementation plans under the 1970 Act (3).  In the range lower than 0.16
ppm oxidant, the guideline was so nearly linear that linearity was assumed.
The rollback requirement for the Washington area changed from 85% based on
hydrocarbons found in the preliminary survey to 50% based on the oxidant
oriented Appendix J.  Quite coincidentally, the revised estimate of the effec-
tiveness of the planned hydrocarbon control program also showed a 50% emission
reduction potential.  The Baltimore oxidant measurements, however, based on
the 10 minute grab phenolphthalein method prescribed a 30% reduction using
Appendix J.  Due to the greater industrial HC emission component in the in-
ventory, the Federal new car program had a smaller impact on the Baltimore
inventory.  The proposed control program was predicted to net a 26% reduction
in emissions which was considered sufficient for that time.  Thus, the program
designed under the 1967 version of the Federal Act was confirmed under the
1970 amendments.  The State added Rule 66 type solvent controls and gasoline
marketing controls to its arsenal of regulations.  The drycleaning vapor
losses were ignored as insignificant and also the drycleaning industry had
been phasing out the use of photochemically reactive solvents.  In addition,
nitrogen oxides emissions were regulated but only for the sake of meeting
N02 health standards, then thought to be exceedec.  There was no consideration
of NO  emissions in the oxidant control program.
     X
     During 1972, the planning phase of the program started studies of the
long term impact of alternative land use and transportation profiles.  These
studies were thought of as maintenance measures at that time.   The overall

                                    1125

-------
program design parameters remained unchanged until  the new continuous chemi-
luminescence system demonstrated peak ozone levels  over 0.20 ppm.   These new
data plus the Natural Resources Defense Council  vs.  EPA court decision (4)
forced development of an oxidant plan revision in mid-1973.   This  revision
showed that a 70% reduction in 6-9 a.m. HC emissions was necessary to meet
the oxidant standard, still using the Appendix J guideline.   As a  result,
the State regenerated previously set aside plans for controlling drycleaning
solvent losses and the emissions from gasoline marketing at existing stations.
Regulations were adopted for this purpose in October, 1973.   Two months later
a regulation was added which was designed to prohibit new major HC sources
and freeze the inventory of emissions from existing major sources  in the
Baltimore and Washington metropolitan areas.  The most significant feature of
the December, 1973 regulations from the ozone formulation standpoint, was the
use of the term photochemically reactive organic material (PROM) in the regu-
lation.  For the first time the program directly considered reactivity in its
oxidant control program, and such reactivity controls were based on reducing
the production of ozone, not eye lacrimators, as was the case in California.

     During the same period in the winter of 1972-1973, the State  was con-
sidering the air quality impacts of the newly proposed Brandon Shores power
plant to be located in the Baltimore metropolitan area.  Since this plant
was to be a large emitter of N0x, the State wanted to determine the impact on
photochemical oxidant levels within the region.   Unfortunately, the Federally
provided guideline was incapable of evaluating anything but tine effect of NMHC
on oxidant levels.  As a result, the State launched an empirical exercise
aimed at evaluating the aerometric data from the then operational  continuous,
telemetered system.  The objective of this study was to develop a  statistical
relationship between oxides of nitrogen and ozone for the Baltimore-Washington
metropolitan areas.

     The first of two attempts at developing a suitable relationship used
an eigenvector analysis which attempted to isolate  the major determinants
in the production of ozone.  This analysis which was similar in concept to
that of Peterson (5) used data from the six AIRMON  stations for selected days
during the summer of 1972.  The ozone data for a given day was arranged in
a 6 x 24 matrix (6 stations x 24 hours).  From this matrix,  it was possible
to derive a set of empirical orthogonal functions which could then be analyzed
for possible physical significance.  The orthogonal  functions consisted of
two matrices.  One matrix was 6x6 and the vectors in this matrix were re-
lated to station located parameters.  The second matrix was 6 x 24 and con-
sisted of vectors which were related to time of day.

     This process was repeated for a number of days during the summer of 1972.
The first orthogonal function, in most cases, accounted for 85-95% of the
variance in the data.  The first vector which was related to time  of day
correlated well with temperature or solar angle  for the test day.   Since the
first orthogonal function accounted for most of  the variance, no other signi-
ficant features of the data could be isolated.  Based upon these rather dis-
appointing results, this technique was abandoned.
                                     1126

-------
     Since the first attempt resulted in no information of value relating to
NO , a second analysis using multiple correlation was tried (6).  The analysis
technique was patterned after an exercise performed by Merz, et al.  (7) on
Los Angeles data.

     The Maryland data used in this analysis was for May through August, 1972
and came from the six AIRMON stations located in the Baltimore-Washington
metropolitan areas.  Four of the stations were located in the Maryland suburbs
of Washington approximately 8 miles from the central business district.  The
other two stations were located in Baltimore City.  One station was  approxi-
mately one mile from the central business district.  The other station was
approximately two miles from the central business district.  Each of the
stations measured the following parameters; nitric oxide (NO), nitrogen dioxide
(N02), methane, total hydrocarbons and ozone.  Due to instrument malfunction
and non-valid data, the total amount of data available for this analysis was
only 350 station-days.  This amount of data was further reduced based on
considerations of station location.  Two stations were located near  major
roadways and the corresponding levels of oxides of nitrogen were significantly
higher than for the other four stations.  One of the remaining stations was
located in a field, far removed from sources of oxides of nitrogen.   This
station was removed from the analysis due to the fact of significantly lower
levels of oxides of nitrogen.  Thus, three stations remained, two in the
Washington suburbs and one in Baltimore City.  The resulting number  of station-
days of data was 159.  Meteorological data for the analysis came from the
National Weather Service office at Baltimore-Washington International Airport.

     Since none of the three stations had sufficient data to do a station
specific analysis, all the available data was used and analyzed collectively.
No consideration was given to atmospheric transport.  The correlations,
also, assumed that measured levels of precursors at a given station  were
related to maximum ozone values measured later in the day at the same station.

     This analysis yielded the following results.  The most significant
variable in predicting ozone concentrations was temperature.  The next most
significant variable was the 6-9 a.m. average N02 concentration.  The least
significant variable was 6-9 a.m. average NMHC.  The following equation
summarizes the best correlation which this analysis of data produced.

  In (peak (03)) - -16.772 + 0.23 In (N02)6_g a^^ + 0.061 In (NMHC)6_9 a>m_ +

                             3.36 In (Max. temp.)

  Where 03, N02 and NMHC - concentration in ppm
        Max. Temp. = Baltimore-Washington International Airport
                     maximum temperature in degrees Fahrenheit.

The multiple correlation coefficient was 0.673 and the percent of the variance
explained by this correlation was 45.1.

     Based upon the correlation analysis, the Bureau concluded that  NMHC
control would have little or no effect on afternoon ozone levels.  From
this analysis it appeared that NO^ emission control would have a greater


                                     1127

-------
effect on afternoon ozone levels.   However,  it was recognized that the cor-
relations did not explain a large portion of the variance and further correl-
ations would be attempted as more data become available.

     This work was forwarded to EPA's National Environmental  Research Center
(NERC) for review and comment with the intention that the relationship derived
might be used in program planning work.   The response was critical of the
conclusions drawn from the study; that is, NERC did not believe that the
study justified controlling N0x instead of NMHC as a means for reducing ozone
levels.  Similar critical review by the California Air Resources Board found
flaws in that the data base was limited, the State did not consider transport
and other problems.

     During the confusion that attended the preparation and submittal of
Implementation Plans for the entire State, the photochemical  aspects of
hydrocarbon emissions were once again examined.  Even though  the mid-1973
submission was based on Appendix J, the State was interested  in the relative
degree of control afforded by taking reactivity into account.  As a result,
during the fall of 1973, the State prepared an inventory of HC emissions
based on reactivity factors for ozone formation.  These factors were being
given preliminary consideration in mid-1972 by the atmospheric chemistry
group in NERC and had been used by EPA in the demonstration project for trans-
portation control strategy development for Los Angeles completed in December,
1972 (8).  Following submission to the Philadelphia regional  office of EPA,
that agency chose not to use reactivity in its analysis and promulgated
transportation control strategies based on the use of Appendix J and NMHC
emissions.  This action immediately followed the same action  in Texas where
the EPA promulgated a plan based on a reactive inventory and  linear rollback.

     In the beginning of 1974 the State challenged EPA over its promulgation
in the fourth circuit court of appeals.   Thus, while the judicial system was
engaged to sort out the program design, the State continued to implement the
non-controversial elements and initiated other studies on the photochemical
phenomenon.  Since the continuous, telemetered air monitoring system had
produced a sizeable quantity of information on the subject, it seemed that
more statistical analysis of aerometric data was in order.  Thus, during the
first half of 1974, the earlier multiple regression analysis  was duplicated
using the more extensive data.  No significant change in the  relationship
of the parameters resulted.

     In addition, a hypothesis was advanced that was intended to shed light
on the EPA's hydrocarbon-only control guideline, Appendix J.   The theory
was advanced that if automotive traffic was significantly less prevalent
on weekends than on weekdays, it should follow that significant reductions
in ozone would also be observed.  Such a study was carried out and the results
reported in the literature (9).  It was found that significantly lower levels
of NMHC were measured on weekends with no commensurate reduction in ozone.
Indeed, it was observed that weekend ozone levels may even be higher than
the weekday measurements, however, the analysis did not demonstrate this
point.
                                    1128

-------
     During the early summer of 1974, the State took an opportunity to analyze
new aerometric data gathered at a rural site in Southern Maryland.   Although
a limited amount was available, wind directions were correlated with ozone and
precursors along with corresponding data in Washington, Baltimore and the
Tidewater, Virginia area.  For the first time the state had first-hand exper-
ience with rural ozone levels and indications at that time were that transport
of ozone was occurring from areas outside the State boundaries.  This obser-
vation was qualitatively compared with work performed for EPA by the Research
Triangle Institute which had measured ozone and precursor levels at McHenry,
Maryland in the westernmost portion of the State (10).

     Following the above work, a careful scrutiny of Appendix J revealed a
flaw that had not been mentioned by others.  Regardless of the lack of a
cause-effect relationship, the so-called "observational" approach from which
Appendix J was taken still showed that for given levels of hydrocarbons,
many ozone levels were observed.  Thus, the fundamental equation between the
two variables was an inequality with an upper boundary observed.  Thus, the
fundamental equation between the two variables was an inequality with an upper
boundary observed on the empirical diagram.  This boundary, slightly modified,
was transformed linearly by subtracting 0.24 from each hydrocarbon level and
dividing the difference by that hydrocarbon level.   Thus, the percent reduc-
tion necessary to achieve the Federal 6-9 a.m. NMHC guideline was given as a
function of peak afternoon ozone measurements.  The mathematical flaw is that
an equality was presented as the Appendix J guideline where an inequality
should have followed the transformation.  Correctly presented, the label for
the abscissae should have read Minimum Reduction in Hydrocarbon Emissions
Required to Achieve National StandarcTfor Photochemical Oxidant.  Because
of this error, the State of Maryland (and possibly-a number of other States)
were led to believe that a lesser degree of NMHC control was necessary in
order to achieve the national standard for photochemical oxidants.   Behind
Appendix J, after this mistake was realized, the goal remained the same as
that used during the first exercise in 1970, that is, the National  guideline
of 0.24 ppm 6-9 a.m. NMHC.

Program Revision Beginning in 1974

     In mid-year 1974, the State decided that not enough effort was being
devoted to understanding the basic state-of-the-art photochemistry of air
pollution.  As a result, a staff reorganization was ordered and two scientists
were assigned to work full-time on the problem.  The first task was to insti-
tute a very extensive literature survey of both theoretical and practical
aspects of photochemistry literature with the overall objective to determine
if the majority of observations were consistent among each other.  Up until
that time the results of State efforts seemed contradictory to the path of
National policies.

     One year later, the literature survey was essentially completed with a
summary paper having been written.  This paper was intended as an educational
exercise for members of staff routinely dealing with hydrocarbon abatement
activities.  It also was to serve as a base which staff can update yearly
to make sure that the State remains well aware of all latest studies.
                                     1129

-------
      It was  discovered also that when  taken  in  perspective  the  State  was
 not yet ready for the analysis  technique  being  highly  touted  by some,  that  is,
 the general  photochemical  oxidant simulation model.  These  models  which
 incorporate  both atmospheric diffusion processes  and atmospheric photochemis-
 try require  more input data than the  State  can  adequately provide  at  this
 time.   Therefore, while these models  provide a  very  useful  learning exer-
 cise,  the quantitative results  of a detailed strategy  analysis  would  be  of
 uncertain accuracy.   This  opinion was  based  on  both  the  paucity of reliable
 input  data which the State could provide  and the  reported analysis of the
 performance  of the models.

      The most reasonable second generation  planning  model appeared to be
 the use of the smog chamber simulation data  once  demonstrations were  avail-
 able that were carried out over the full  range  of ambient precursor concen-
 trations (11).  In order that these chamber  experiments  could be fully uti-
 lized  by the State, however, the agency personnel  had  to feel fully confident
 that State air quality would behave similarly.   In February of 1976,  there-
 fore,  the University  of Maryland was  engaged to  provide a  smog chamber
 simulator.  The main objective  is to  provide the  simulation tool such that
 State  personnel can evaluate the reactivities and dynamics  of Maryland-based
 emissions inventories and hopefully gain  "hands-on"  experience  with photo-
 chemical reactions.   It was also desired  that the feasibility of the  hybrid-
 analog computer be evaluated for these simulations since many linked  equa-
 tions  appear in the set of chemical kinetics under study (12).   The objective
 is to  provide inexpensive simulation  from short run  times,  rapid turnaround
 time and low cost facilities.  Available  digital  programs will  be  utilized
 if the use of the hybrid does not perform adequately.   In the meanwhile,
 the State has begun to collect  a detailed emissions  inventory to use  for in-
 put.  Emission questionnaires have been sent to approximately 80 stationary
 sources of organic emissions.  These  sources account for approximately 25%
 of the total anthropogenic emissions  of organic compounds in  the Baltimore
 metropolitan area.  The information which the State  is requesting  includes
 operating schedules in terms of shifts per  day, days  per week,  weeks  per
 year and information on percent of yearly production by  month.   Emissions
 information  is also being provided by the stationary sources  in terms of
 composition  of organic emissions.  This information  is in addition to the
 existing, non-species specific  organic emission inventory.   This data will
 provide a basis for future modelling  efforts.  Naturally, since this  is  the
 first  attempt in the State at compiling such an inventory,  problems are
 occurring.  However, it is envisioned that  this will  be  a continuing  program
 and the quality and quantity of the information will  improve  as time  progresses.

Observations                                                            _

     With the changes initiated by Congress  in the 1970 Clean Air Act  amend-
ments and subsequently promulgated "guidelines", there began a series  of
rapidly fluctuating program design changes.   The major changes were from a
program design based on observations of non-methane hydrocarbons during the
6-9 a.m. period to an oxidant control  strategy based on observed levels of
ozone.   With  the newly discovered levels of ozone measured using EPA approved
methodology,  the program design changed once again.  Much more restrictive
program design requirements were expected since the 1970 amendments provided

                                     1130

-------
for an auto control program that was not phased in with the ambient air quality
implementation deadlines contained in the Act.  Program design persons were
puzzled when long-term abatement needs disappeared.  Within five years the
hydrocarbon control requirements had made a complete cycle and those abatement
needs are perhaps more restrictive now, than originally contemplated in 1970.
Such a loop was the direct result of a mathematical error contained in the
Federal Appendix J guideline used in evaluating State proposals and promul-
gating State Implementation Plans.

     Throughout those areas of the Nation which were plagued by elevated
ozone levels, there appeared to be different techniques utilized by the EPA
in evaluating proposed plans and promulgating rules to abate such levels.
California, Texas and Colorado were allowed to use an approach based on
linear rollback of a reactive inventory while others, including Maryland,
were not.  No apparent reason for this is known to the State of Maryland.
The use of such a reactive inventory in Maryland would have removed the re-
quirements for those onerous strategies promulgated by the EPA that were
largely the impetus for the State seeking relief from the Federal 4th circuit
court of appeals.

     From the extensive literature survey, it was found that almost all of
the observations reported in the literature are consistent with one another.
The conclusions of some papers, however, were found to have been drawn in-
correctly due to a lack of overall perspective of the state-of-the-art.  Such
was the case with investigations made prior to the end of 1974 by the State
of Maryland.

     An emerging issue appears to be that of the occurrence of ozone levels
far removed from urban areas.  It seems that the extent of the ozone prob-
lem proportionally follows the expansion of the ambient air measurement sys-
tem.  Although there are schools of thought on the origin of those ozone
levels, much controversy remains over whether second or third day reactions are
responsible or the transport of "mature" ozone or even the reaction of surplus
nitrogen oxides with rural emissions of both natural and man-made reactive
hydrocarbon species.

                                 CONCLUSIONS

     Throughout the six year period that Maryland has been developing an
oxidant abatement program, much confusion has attended the cause-effect
relationships which result in observations of elevated levels of ambient
ozone.  Most of the confusion was resolved when it was realized that part-
time efforts to understand the photochemistry of air pollution were not suf-
ficient to understand the state-of-the-art.  It is recommended that States
with ozone problems fund and develop at least one full-time position for an
expert in such matters.  It is also recommended that the Regional offices  of
EPA do likewise.

     In addition there appears to be an emerging phenomenon relating observed
levels of ozone in rural areas to man's activities.  There are several dif-
ferent theories behind the cause of the phenomenon although none of the hypo-
theses have been adequately tested.  From the State level viewpoint, it is


                                     1131

-------
difficult to perceive if a uniform National policy exists and in particular,
if the Congress has been advised to relax nitrogen oxide controls without
regard for the rural problem.  Is it possible that the same confusion which
exists in the State ranks also pervades our national legislators?  Hopefully,
this is not the case.

                                   REFERENCES

1.   Andrew, D. P., Report for the Presentation of Ambient Air Quality Stan-
     dards for:  Carbon Monoxide, Hydrocarbons, Photochemical Oxidants, for
     the National Capital Interstate Air Quality Control Region.  Maryland
     State Department of Health and Mental Hygiene, Division of Air Quality
     Control, October 20, 1970.

2.   Air Quality Criteria for Hydrocarbons.  U.S. Department of Health, Edu-
     cation and Welfare, Public Health Service, National Air Pollution Con-
     trol Administration, Washington, D. C., March 1970.

3.   Requirements for Preparation, Adoption and Submittal of Implementation
     Plans, Vol. 36, No. 158 (Part II), (August 14, 1971  and amendments
     dated October 20,  1971).

4.   Natural Resources  Defense Council, Inc., et al. v. Environmental Pro-
     tection Agency, 475 F 2nd 986 (1973).

5.   Peterson, J. T., Distribution of Sulfur Dioxide over Metropolitan St.
     Louis, as Described by Empirical Eigenvectors, and Its Relation to
     Meteorological Parameters.  Atmospheric Environment 4, 501-518  (1970).

6.   Lebron, F., Paisie, J., Bonta, W. K., Analysis and Evaluation of Balti-
     more-Washington Photochemical Oxidant Data - Summer 1972.  Maryland
     State Department of Health and Mental Hygiene, Bureau of Air Quality
     Control, June 4, 1973.

7.   Merz, F. H., Painter, L. J., Ryason, P. R., Aerometric Data Analysis -
     Time Series Analysis and Forecast and an Atmospheric Smog Diagram.
     Atmospheric Environment 6, 291-342 (1972).
8.   Transportation Control Strategy Development for the Metropolitan Los
     Angeles Region, APTD-1372, U.S. Environmental  Protection Agency, Re-
     search Triangle Park, North Carolina,  December, 1972.

9.   Lebron, F., An Evaluation of Weekend/Weekday Ozone Concentrations in
     the Baltimore-Washington Areas.  Atmospheric Environment 9,  861-863
     (1975).

10.  Investigation of High Ozone Concentration in the Vicinity of Garrett
     County, Maryland and Preston County,  West Virginia, EPA-R4-73-019,  U.S.
     Environmental Protection Agency,  Research Triangle Park, North  Carolina,
     January 1973.
                                    1132

-------
11.   Dimitriades, B., Effects of Hydrocarbon and Nitrogen Oxides  on Photo-
     chemical Smog Formation.  Environmental Science & Technology 6, 253-260
     (1972).

12.   Meet,  T. A.  , Seinfeld,  J.  H.  ,  Dodge,  M.  C.,  Further Development of
     Generalized  Kinetic Mechanism for Photochemical Smog.   Environmental
     Science  & Technology 8,  327-339  (1974).
                                     1133

-------
                                                                           23-3
            OXIDANT  CONTROL UNDER SECTION 110 OF THE CLEAN AIR ACT

                                 J.  L.  Pearson*

ABSTRACT

     Section  110  oi  the Clean Ax>i Act fiequiAeA AtatcA to Aubmit p£cuu that
contain meaAuJieA  designed to attain and maintain the. national. Atanda^td 30 A
photochemical  ox^idant.   The original Stat.e Implementation PlanA weAe Aub-
mitted ,01  1972 and nave been Au.bje.ct to Aevi^ion Aince that time.  The.
approach to oK4.da.nt  control oft the. Env4Aonme.nt.ajL Pnote.cti.on Agency OA well
OA that ofa mo At States  haA been directed towatid con&iol o^ organic compound
e.mtAA4,onA.  A  ^.ecen^: Aeu-tetv ^104 Akovon that many State. J.mple.me.ntat^.on PtanA
aAe .inadequate, to attain and matntatn the. national Atandofid.  WkJJLe. -it 4^>
dt^^icult -in many OACOA to attain and matntatn the Atandatid because o& the.
Lcm4jte.d control tecbu.Qu.e4 ava.itable,,  e.^otit^ aAe continuing to reduce
ambient concentiattonA  ofi photochemical ozidant with a uxLew towand nve.ntu.al
attainment o^  the national Atandatid.

                                   INTRODUCTION

     Section 110  of  the  Clean  Air Act  (1)  of 1970 required submittal  of plans
for the implementation  of control  measures  to reduce emissions of specified
air pollutants  in order  to attain National  Ambient Air Quality Standards,
which were to  be  established by the  U.  S.  Environmental  Protection Agency
(EPA).   The Act required attainment  of the  national  primary, (meaning health-
related), standards  within three years of plan approval  by EPA.  However,
in certain circumstances,  i.e., where  one or more emission sources were
unable to comply  with the  requirements of the implementation plan because
the necessary  technology or other alternatives were  not or would not be avail-
able within the three-year period,  the Administrator could grant a two-year
extension.  In  most  cases  the  date  established for attainment of the national
primary standards was May 31,  1975.

     In addition  to  State  Implementation -Plans (SIP) required by section 110,
the Clean Air  Act also  required a broader control strategy to reduce organic
emissions and  oxidant air quality.   The Federal  program has primarily focused
upon the control  of  new  motor  vehicle  emissions  and  new stationary sources
of organic compounds.   Mobile  sources  are a major contributor to anthropo-
qenic emissions of organic compounds  and thus the U. S.  Congress, in Title II--
Kmissions Standards  for  Moving Sources—of  the Clean Air Act, mandated control
*United States Environmental  Protection  Agency,  Research Triangle Park,
 North Carolina


                                      1135

-------
of new mobile sources.  Specifically, the Act required that by 1978 hydrocar-
bon emissions from light duty vehicles be reduced by 90 percent, based upon
1970 light-duty vehicle emission rates.  Additionally, the Act authorizes the
Administrator of EPA to prescribe standards applicable to emissions from any
class of motor vehicles or motor vehicle engines which in his judgment causes
or contributes to air pollution capable of endangering the public'health or
welfare.  As a result, the Administrator has proposed or promulgated emission
limitations for diesel engines, motorcycles, and aircraft engines.   The Act
also required EPA to consider new stationary sources of air pollution.  Section
111 of the Act requires the Administrator to establish Federal standards of
performance for new stationary sources.  These standards of performance would
apply regardless of ambient concentrations of the specified pollutant.  While
these programs are required under Sections of the Act other than Section 110,
the emission reductions obtained due to these programs are an integral part
to implementation plan development.

                           HISTORY OF PLAN DEVELOPMENT

     On April 30, 1971, the Administrator promulgated in the Code of Federal
Regulations National Ambient Air Quality Standards for six air pollutants.
Included in this promulgation was the national primary standard for photochem-
ical oxidant.  This standard was established as a maximum one-hour ambient
concentration of 160 pg/m3 not to be exceeded more than once per year.  Another
of the National  Ambient Air Quality  Standards promulgated was a national
standard for hydrocarbons.  This standard, 160 yg/m3 maximum three-hour con-
centration (6-9  a.m.) not to be exceeded more than once per year, was  intended
to be used only  as a guide in devising implementation plans to achieve the
oxidant standard.  While implementation plans for oxidant were required,  no
implementation plans, ambient monitoring,  or other actions were required  as
a result of the  hydrocarbon standard.

    The State initiated programs were  included in the SIPs submitted  to
EPA in  1972.  The plans submitted at that time included a  variety of
stationary source control measures.   Regulations for controlling organic
emissions from stationary sources have generally followed  one of two  pat-
terns.  The  two  basic  patterns  are  the Appendix B to the  Part 51 regulations
 (2) and the  Los  Angeles organic  emission  control regulations.  EPA has
recognized deficiencies in both  of  these  approaches  to stationary source
control or organics  and has  undertaken a  task  to examine,  industry by
 industry, the availability of  controls.

    When  the  implementation  plans were submitted  in  1972,  EPA  recognized
that many were deficient  in  their ability to  attain  the national standard
for photochemical oxidant.   As  a result,  in  1973, several  transportation_
control measures  were  promulgated as  part of  the  SIPs  to  help  correct this
deficiency.   Since  that time a  major  concern  to States  in  their  development
of State  Implementation Plans  has been organic emissions  from  mobile  sources.
 Control of emissions  from mobile and  mobile-related  sources  has  been  a two-
 pronged attack.   The  first of  these is designed to  reduce organic emissions
 by the  imposition of controls  on the  source.   Among  these measures  are included:
                                     1136

-------
     t    Inspection/maintenance programs,

     •    Medium and heavy-duty vehicle retrofit,

     •    Ship and barge controls, and

     •    Vapor controls for gasoline marketing.

     The second part of the control program is designed to promote the effi-
cient use of transportation sources.  The following four measures are examples
of these types of regulations:

     •    Transit improvements,

     •    Employer/employee incentives,

     •    Parking management/restrictions, and

     t    Traffic management/restrictions.

                            REVIEW OF EXISTING PLANS

     The Environmental Protection Agency conducted a review, from late 1975
to early 1976, of State Implementation Plans to determine if they were sub-
stantially inadequate to attain and maintain the national standards.   This
review of recent ambient air quality data has shown that a number of Air
Quality Control Regions (AQCR's) had not attained the national  standard for
photochemical oxidant by June 1975 and are not anticipated, even with full
implementation of existing regulations, to attain that standard for several
years hence.  As a result of this review, in July 1976 EPA notified the appro-
priate State Governors and responsible air pollution control agencies of
the need for revisions to the control strategy portion of their implementation
plan.  These notifications of the need for revision were issued for 61 indi-
vidual Air Quality Control Regions and eight entire states without identifica-
tion as to specific Air Quality Control Regions (Figure 1).  In the case of
photochemical oxidant, these notifications included at least one Air Quality
Control Region in 30 states, which compares to the total of 55  states and
territories required to submit plans under Section 110 of the Clean Air Act.
These calls for State Implementation Plan revision do not, however, include
those areas where the implementation plans have recently been revised or are
currently being revised.

     Revised plans must be submitted to EPA for review and approval by July
1977.  These plans must include all measures necessary to provide for attain-
ment of the standard, and where attainment, of the standard cannot be demon-
strated, the plan must include all achievable emission limitations.  The
term "achievable" as used by EPA is intended to require the imposition of
reasonable "technology forcing" control measures, rather than simply requiring
"off the shelf" technology.  All other measures required to attain and main-
tain the National Ambient Air Quality Standard must be submitted by July 1978.
These other measures are intended to include, bjt are not limited to, trans-
portation and land use measures and programs designed to reduce organic con-

                                    1137

-------
                                                                   
                                                                   03
                                                                  ' rs
                                                                   cr
                                                                   CD
                                                                  T3
                                                                   "3
                                                                   10
                                                                   05
                                                                   HI
                                                                  41
                                                                   O)
                                                                  T3
                                                                  c
                                                                  fO
                                                                  I—
                                                                  n_

                                                                  E
                                                                  O
                                                                  •i—
                                                                  +->
                                                                  ro
                                                                  4->
                                                                  C
                                                                  OJ

                                                                  O)

                                                                  CL
                                                                  QJ
                                                                  S-
                                                                  O)
                                                                  01
                                                                  s_
                                                                  OJ
                                          <   9
1138

-------
pound emissions.  It is anticipated that some areas will  be unable to demon-
strate attainment of national  standards.  In such cases,  the submitted plan
should contain a discussion of the State's program to continue to identify
and explore additional  control measures and to adopt such measures when they
are determined to be effective and workable.

                         IMPLEMENTATION PLAN DEVELOPMENT

     During the first stages in the development of an implementation plan
it must be determined which areas have ambient air quality levels that
exceed the national  standard of 160 vg/m3.  At the time of initial implemen-
tation plan development, it was felt that the photo-chemical oxidant problem
was limited to major metropolitan areas.  Since that time EPA and others have
conducted several field studies and determined that ambient concentrations of
photochemical oxidant greater than the standard may be observed even in rural
areas of the country.  While EPA recognizes the transport of photochemical
oxidant (3, 4) and its precursors, it appears that the urban area emissions
are the major contributor to ambient concentrations in excess of the standard
in the vicinity of that urban area.  One of the basic reasons for this is
that a major portion of organic compounds emitted, motor vehicle exhaust and
solvents, are highly reactive and form photochemical oxidant within a short
period of time.  As  a consequence, EPA has generally recommended to the States
that during the current implementation plan development process, control of
emission sources in  the urban area be emphasized.

     The Clean Air Act dictates an air quality management approach.  As a
result, implementation plans must specify the reduction in emissions neces-
sary to attain the National standard and how these reductions will be obtained.
To assess the reductions necessary, a correlation between precursors to photo-
chemical oxidant and maximum oxidant levels is required.   The correlation
generally used for the initial plan development in 1971-1972 was based upon
the available non-methane hydrocarbon and oxidant air quality date (5) from a
limited number of cities.  This relationship, commonly referred to simply as
Appendix "J" because it was published as Appendix "J" to the "Requirements
for Preparation, Adoption, and Submittal of Implementation Plan,"  (6), has
been severely criticized.  Among the criticisms are the following points:

     t    The relationship is dominant at the upper bounds by data from Los
          Angeles, California,

     •    The  relationship does  not account for  the  effects  of  oxides  of
          nitrogen,

     •    The relationship does not account for the effects  of  transport, and

     •    The effect of background levels of oxidant due to  natural sources
          is not included.

     In an effort to develop a more comprehensive relationship, the Environ-
mental Protection Agency is continuing to study the correlation of maximum
oxidant levels in relation to these and other factors.  The  EPA is currently
attempting to develop several alternatives  to the original correlation that


                                    1139

-------
can be used by State and local air pollution agencies to determine the neces-
sary organic emission reductions to be used.

     For many areas of the U. S. where the implementation plan has been
identified as inadequate to attain the standard, these correlations indicate
a requirement for organic emission control greater than is available with
current control technology.  As a result, many Air Quality Control Regions
in the U.S. may not be able to demonstrate attainment of the national photo-
chemical oxidant standard in the near future.  Where this is the case, the
submitted implementation plan must contain requirements to implement all rea-
sonably achievable control technology for both mobile and stationary sources
and must indicate areas where further investigation and study will be conducted
to determine other mechanisms by which the oxidant standard may be attained.
Further, EPA requires that State programs for control Of mobile sources be
coordinated with appropriate state and local planning and transportation agen-
cies to ensure that air quality measures are implemented as part of the trans-
portation planning process.  In addition, the State will be required to adopt,
as they become available and are needed, other regulations that are identified
as representing reasonably achievable control technology.  These procedures
will likely require that the implementation plans submitted in 1977 and 1978
be revised in the future until the degree of control required to attain and
maintain the national standard has been achieved.

     EPA recognizes the important role of oxides of nitrogen in the photo-
chemical process (3, 7, 8).  However, recent studies have indicated that,
except in certain circumstances, i.e., rural areas with low ambient levels
of organics, the most efficient strategy for reducing ambient levels of
photochemical oxidant is through reductions in organic compound emissions.
In addition, EPA recognizes that as ambient levels of photochemical oxidant
and organic compounds are reduced, it may be necessary in the future to consider
the application of controls on sources of oxides of nitrogen emission (3).
It is not anticipated that control of oxides of nitrogen, for the purpose of
attaining the national standard for phof'ochemical oxidant, will be required
in the immediate future.  Additional research is continuing in this area.

     Natural sources of organic emissions are another problem which will have
to be accounted for in the future.  While natural sources of organic emissions
may account for oxidant levels in the range of 0.04 to 0.05 ppm, it is believed
that the contribution of natural sources is generally small in urban areas
in relation to peak oxidant concentrations.  Because EPA has recommended that
implementation plans be developed for urban areas, the impact; of natural
sources is not a major concern in plan development at this time.  However,
various members of the scientific community and EPA are conducting research
to determine the impact of natural sources of organics.

     The Environmental Protection Agency recognizes the difficulties that
will be faced as the U. S. progresses toward attainment of the national photo-
chemical oxidant standard and remains committed to the basic goal of reducing
ambient levels of photochemical oxidant.  The long-range goal is to attain
the national standard by reducing emissions of all organic compounds (except
methane), since virtually all organic compounds eventually participate in the
formation of photochemical oxidant when given sufficient time and ultraviolet


                                     1140

-------
energy.  However, due to the fact that many areas of the country will f\ot be
able to attain the photochemical oxidant standard for several years, EPA
currently requires that regulations requiring substitution of organic compounds
classified as less reactive for those classified as highly reactive be included
in the implementation plan.  This measure will be used only as an interim
measure until sufficient positive control measures for mobile and stationary
sources are available to reduce emissions of organic compounds to a level so
that the national standard will be attained.  The regulations of this type
currently in effect in many areas of the country are generally a modification
of the Los Angeles Rule 66, which was based upon the hydrocarbon reactivity
data available at the time of development.  The EPA is currently assessing the
need to (a) revise this regulation, or (b) develop an entirely new regulation
based upon the most recent reactivity data available.  It should also be
noted that EPA is currently reassessing its policy on the use of reactivity
and its requirement for adoption of substitution regulations.  The Agency is
also continuing to evaluate controls for existing sources and to develop
standards of performance for new sources of organic compound emissions.

                                     SUMMARY

     EPA believes that the combination of motor vehicle controls, transpor-
tation control measures, definition and imposition of achievable control
technology, regulations for substitution of organic compounds, development
of standards of performance for new sources, and other programs will make up
the total package of measures needed to attain and maintain the national
photochemical oxidant standard.  In addition, the U.S. Congress is currently
considering amendment of the Clean Air Act to account for the need for consid-
eration of differing types of control measures, the need to revise the Federal
Motor Vehicle Control Program, and the need for additional time to implement
control measures.  The Congress is considering additional land use and trans-
portation type measures for the control of all auto-related pollutant emissions
and will likely require EPA to provide information on many new and innovative
measures for control of mobile sources and to promote the most efficient use
of transportation systems.  While attainment of the national standard for
photochemical oxidant  in some areas of the U.S. may be difficult, continuing
efforts by industry and local, State, and Federal air pollution control
agencies will ensure that  there will be continuing progress  toward eventual
attainment of the standard.  Even though attainment of the  national  stan-
dard for photochemical oxidant may be several years away, EPA recognizes
that there are many benefits, both to health and welfare, to be gained
by reducing  both maximum oxidant levels and the number of times the  stan-
dard is exceeded.
                                      1141

-------
                               REFERENCES

1.  Clean Air Act (42 U.S.C.  1857 et seq.)  includes  the  Clean  Air  Act  of
    1963 (P.O. 88-206), and amendments  made by  the Motor Vehicle Air
    Pollution Control Act - P.O.  89-272 (October  20,  1965,  the Clean Air
    Act Amendments of 1966 -  P.O. 89-675 (October 15,  1966),  the Air
    Quality Act of 1967 - P.O.  90-148 (November 21,  1967),  the Clean Air
    Act Amendments of 1970 -  P.L. 91-604 -  (December  31, 1970), the
    Comprehensive Health Manpower Training  Act  of 1971 - P.O.  92-157 -
    (November 18, 1971) and the Energy  Supply and Environmental Coordina-
    tion Act of 1974 - P.L. 93-319 - (June  22,  1974).

2.  Code of Federal  Regulations,  Title  40 - Protection of the  Environment,
    Part 51 - Requirements for Preparation, Adoption,  and Submittal of
    Implementation Plans, Appendix B -  Examples of Emission Limitations
    Attainable with Reasonably Available Technology.

3.  Control of Photochemical  Oxidants - Technical Basis  and Implications
    of Recent Findings, July, EPA-450/2-75-005.

4.  Angus, R.M. and E.G. Martinez, "Rural Oxidant and  Oxidant  Transport,"
    October 22, 1975.

5.  Air Quality Criteria for  Hydrocarbons,  March  1970, U.S.  Public Health
    Service, AP-64.

6.  Code of Federal  Regulations,  Title  40 - Protection of Environment, Part
    51 - Requirements for Preparation,  Adoption,  and  Submittal of  Implemen-
    tation Plans, Appendix J  - Required Hydrocarbon  Emission  Control as a
    Function of Photochemical Oxidant Concentration.

7.  Air Quality Criteria for  Nitrogen Oxides, January  1971, U.S.  Public
    Health Service,  AP-84~

8.  Air Quality Criteria for  Photochemical  Oxidants,  March  1970,  U.S.
    Public Health Service, AP-63.
                                   1142

-------
                                                                          23-4
                 OXIDANT  CONTROL  STRATEGY:   RECENT  DEVELOPMENTS

                                B.  Dimitriades*

ABSTRACT
     Recent /teieotch  fai.ncU.ngA -f.eXe.vant to  aMnoAph&iic ox.-ida.nt occuAAence
         and examined ^on. tmpticationA  AegaAdtng  optimum -ittategt/ fan.
ox.tda.nt zontsioi.  ConcJtuAtonA o^eAed  aAe  that the Ae.ce.ftt evidence dee/6
tnvaiidate the Appendix-! method  1301 calculating  ccnfio^ Aequ-ciementA but
doe-i not pAov.-c.de gAound-i ^OA que/sttonxng the  validity o{j kydAocaJibon
ai an approach to ox.tda.nt Aeduction.

                                   INTRODUCTION

     The first oxidant control  strategy was put into effect in the U.S.  in
1970.  Experiences accumulated  since then  by  the  enforcement agencies have
brought to surface some enforcement problems  that were almost invariably
blamed on the non-defensibility of this first control  strategy.   Furthermore,
new research, conducted and reported in the last  3-4 years, raised additional
questions on the validity of the  current strategy and introduced numerous
speculative -- but reasonable --  suggestions  for  revisions or for altogether
new strategies.  Because of the complexity of the problem of oxidant control
and because of the confusing nature of some of the new research  findings it
was judged useful to  review these latter findings and closely examine them
for implications regarding optimum strategy for oxidant control.

                        CURRENT OXIDANT CONTROL STRATEGY

     To help explain  the relevance of  and  implications from the  new research
findings, the current oxidant control  strategy will  first be briefly described
in terms of its key steps and assumptions.

     The basic assumption underlying the current  strategy concerns the phy-
sical and chemical mechanism of the overall process  that starts  with the
discharge of emissions and terminates  with the occurrence of oxidant at
problem-concentrations.  Thus,  it has  been assumed that emissions are dis-
charged in the source areas and react  in-situ to  form oxidant that afflicts
the "receptor" areas,  the receptor areas being the very same source areas.
Consistent with this  assumption,  aerometric data  were gathered from several
urban centers and were analyzed to derive  a quantitative relationship between
HJ.S.  Environmental  Protection Agency, Research Triangle Park, North Carolina
                                     1143

-------
maximum-daily-oxidant concentrations and early-in-the-morning (6-9 a.m.)
non-methane hydrocarbon (NMHC) concentrations; both the oxidarit  (Ox) and
NMHC measurements were made  in the same location, in the core areas of the
urban centers.  This relationship is known as the "upper limit"  0 -HC curve
(Figure  1)  (1).
c/o
O
X
     0.30
     0.25
     0.20
     0.15
     0.10
     0.05
       0

            APPROXIMATE UPPER LIMIT
               OBSERVED OXIDANT
         0
                    V      m
                .• • i  • :•.••: ••••
                                I
0.5         1.0         1.5
       NONMETHANE HC, ppm C
2.0
2.5
  Figure  1.   "Upper limit"  curve relating ambient maximum 1-hour oxidant to
             6-9 a.m.  concentration of non-methane hydrocarbon.
                                   1144

-------
     Additional assumptions made were that all of observed Ox (and NMHC)
arises from anthropogenic sources -- the natural background levels being
zero -- and that ambient NMHC concentrations are proportional to NMHC emis-
sion rates.  These assumptions and the "upper limit" curve were used to de-
rive the Appendix-J curve relating ambient Ox air quality to NMHC emission
control requirements (Figure 2).

     A final key element in the current oxidant control strategy is that the
emission control effort is focused on the urban centers.  The non-urban areas
are exempted from control but not totally; they are subject to the controls
enforced in the form of national emission and new source performance standards.

     The validity of this control strategy has been questioned both because
of the questionable assumptions used and because of suspected ineffective-
ness in reducing rural oxidant.

                 RECENT RESEARCH FINDINGS AND THEIR IMPLICATIONS

     Briefly, recent research and findings can be summarized as follows:

1.   Field studies have revealed the occurrence of a wide-spread Ox problem
     in rural areas (3).

2.   Triggered by the findings on rural oxidant, extensive research has been
     initiated that has focused specifically on:

     •    occurrence of short, mid-, and long range Ox transport (3-7)

     t    meteorology related to Ox transport (6)

     •    chemistry of photochemical oxidant formation under pollutant trans-
          port conditions (8)

     t    role and importance of natural emissions as contributors to tropo-
          spheric oxidant (9)

     •    stratospheric contribution to tropospheric oxidant (10, 11).

3.   Smog chamber studies have resulted in an alternative (to Appendix J)
     method for calculating control requirements for urban Ox reduction (12,
     13, 14).  Key advantages of this latter method are the cause-effect na-
     ture of the chamber-derived 0 -to-precursor relationships, and the abil-
     ity to consider the role of the NO  factor quantitatively; the method,
     however, also has some serious limitations.

4.   Smog chamber and field studies have indicated that the reactivity-related
     approach to control  is less effective than previously thought (15, 16).

     These findings, as well as the extent to which and the manner in which
they can be used to support any immediate or future changes in the control
strategy for Ox reduction, are discussed next.
                                     1145

-------
 E

 CD
 3

  ft
•z.
O
t—i

g

I—
1Z.
LU
O

o
    i±i

                                                                                                     !    O
                                                                                                         o

                                                                                                         o
                                                                                                         nr
                                                                                                         Q-

                                                                                                         o:
                                                                                                         ZD
                                                                                                         O
 co
-I-J
                                                                                                                OJ
                                                                                                                OJ
                                                                                                                S-
                                                                                                                CT
                                                                                                                O)
                                                                                                                S-
o   •
S-  C
•4->  O
C •!-
o •*->
O  fO
 o  c
JD  CD
 S-  (J
 n3  c:
 O  O
 O  O
 i-
T3 4->
 >> C
jr  to
   -a
 OVr-
 C  X
•r-  O
4->
 (O "O
•—  OJ
 a>  v
 S-  3
    I/)
 a>    cu
 s-  E
 X  X
•i-  ro
-O  E
 sr
 01
 Q.
 CL

-------
     The occurrence of elevated (>0.08 ppm) oxidant levels in rural areas
is a fact and need not be discussed further.  What needs to be examined,
however, is the relative magnitudes of the rural and urban oxidant problems
in terms of effects and population exposure.  Both the make-up of the rural
oxidant mixture and the population exposed to rural oxidant are different
relative to the urban case, and it could be that the rural problem is in
effect less severe than suggested by the relative occurrence data.  An appro-
priate assessment here certainly is relevant to the question of how adequate
the current control strategy is, and therefore, needs to be made.

     The studies on oxidant transport have shown conclusively that single-
day, short-range transport does occur and may cause the oxidant problem to
be more severe in the receptor areas (suburbs) than in the source areas
(downtown) (3-7).  The same studies have also provided indirect evidence that
multi-day long-range transport occurs and impacts distant rural (and urban)
areas downwind from the source areas.  This latter evidence, however, is only
qualitative, the degree to which transported pollutants account for the
oxidant observed in the receptor rural (and urban) areas has not been
established.

     The control implications of the findings on short-range transport are
relatively simple.  These findings do not raise questions on the validity
of the current strategy, except for the obvious deduction that unless such
transport is considered the calculated degree of control needed will be too
low.  The relative roles of the HC and NOX precursors under such short range
transport conditions are not substantially different than those under no-
transport conditions.

     Unlike the short-range transport, the control implications of the find-
ings on multi-day, long range transport are extremely complex.  Thus, under
long range transport conditions both the source-receptor relationship and the
relative roles of the HC and NOX precursors are different than those under no-
transport conditions.  The implication related to the source-receptor relation-
ship factor is an obvious one.  It simply means that in areas affected by
transported oxidant, the Appendix-J curve is not valid because the source-
receptor relationship is not the same as the one assumed in the derivation
of the Appendix-J curve.  The implications related to the oxidant-to-precursor
dependency factor are not obvious and cannot be expressed simply in terms of
an overall judgment.  These latter implications merit a more detailed analysis
presented next.

     Present understanding regarding the relative roles of the HC and NOX
precursors is based almost entirely on smog chamber evidence and briefly is
as follows.  The difference in oxidant-to-precursor relationship between
transport and no-transport systems arises from differences in irradiation
time, and in extents of dilution and of interaction with fresh precursors.
Thus, under long range transport conditions the HC-!MOX precursor mixture is
irradiated, diluted and interacted by fresh emissions for several days, as
compared to a few hours only for the no-transport systems.  The combined
effects of prolonged irradiation and dilution upon simulated emission-control-
led and non-controlled atmospheres are illustrated by the data from a 3-day
smog chamber test  (Figure 3) (8).  The pertinent observation to be made here


                                     1147

-------
                                     X
                                   o
                                      CXJ
                                   4-J  S-
                                   (13 •!-
                                   •r- 4-
                                   T3
                                   03  CD
                                   S-  C.
                                   S- T-
                                   fO
                                   •O 4->
                                    I   C
                                   O>  d)
                                   cu  o
                                   i-  S-
                                   JT  O)
                                   -*->  CL
                                   4-  >,
                                      -Q
                                   (S)
                                   C "O
                                   O   -O
                                   E
                                   QJ to
                                   O 03
                                   C 2
                                   o
                                   o cu
                                      S-
                                   O) 3
                                   c: +->
                                   o x
                                   M -i-
                                   o s:
                                   O)
                                   OV
uiddto
                  1148

-------
is that in spite of reaction and dilution, the reacted pollutant mixture was
capable of generating significant concentrations of Ox during the 2nd and 3rd
days of irradiation.
ant oxidant but less
 Obviously, the effect
than proportionately.
       of dilution was to dilute result-
     The effect of hydrocarbon control  is depicted in Figure 4  (17).  It can
be seen that control of hydrocarbons had a drastic effect upon  the first day
oxidant levels and a much  lesser effect upon the second day oxidant levels.
   0.50
   0.25
 C*J
 o
                                     HC   2.90
                  12
       18
  24
TIME OF DAY
12
18
          Figure 4.   Second day 03 formation from "non-controlled" and
                          "controlled" HC-NOX mixtures.
     The effect of fresh precursor influx into the reacting system has not
been determined; it can be speculated reasonably that it would be a complex
function of the absolute as well as the relative levels of the reacted and
fresh precursors.

     The clear deduction from all these observations and expectations is that
under long range pollutant transport conditions the quantitative relation-
ships between oxidant and precorsors are indeed different than those assumed
in the derivation of the Appendix-J method.   Therefore, for areas affected
by multi-day transported oxidant, Appendix-J is not the appropriate method for
calculating numerical control requirements.   Nevertheless, there is no indica-
tion that hydrocarbon control will not benefit such areas.  What is somewhat
in question is the effect of NOx control upon such areas.  This question is
examined in some detail next.
                                     1149

-------
      Results from smog chamber experiments  have  shown  that irradiation  of
 an urban-!ike HC-NO mixture for one  solar day  results  in  a mixture  which  is
 virtually depleted of the reactive organics and  has  a  much higher HC-to-
 NOX ratio (8).   Irradiation for a second day will  further raise  the ratio to
 as high as  100:1  or higher.  At these high  ratios,  the reacted mixtures still
 retain their ability to form oxidant upon irradiation, even though  the  react-
 ant HC is at only a few pphm and the NOX at few  ppb  (8).   Furthermore,  injec-
 tion of small amounts of NOX seems to increase oxidant formation, suggesting
 that the role of NOX in such "aged"  systems, i.e.,  in  urban air  masses
 transported into downwind rural (or  urban)  areas,  is contrary to the NOX
 role in the atmosphere above the source area.  These results may suggest
 at first glance that for reduction of rural oxidant  it will be necessary  to
 control NOX along with HC.   In actuality, however,  such a conclusion may  or
 may not be  justified.  The  reasons for this uncertainity  are as  follows.
 Based on presently available evidence, control of  NOX  will have  two opposing
 effects:

      (1)  It will cause increased oxidant levels during the first day --  a
           part of the increment to be carried  into  the downwind  rural  (or
           urban)  areas during subsequent days.

      (2)  It will result in an aged  mixture containing less NOX  and, there-
           fore, with less potential  for oxidant  formation in the downwind
           rural (or urban)  areas.
      The net result from these two effects is  neither known  or predictable,
 at present.


     In conclusion, once again, the findings from recent research on oxidant
transport, so far at least, have not provided grounds for invalidating HC
control as the optimum approach to oxidant reduction in the source (urban)
areas.  Nevertheless, these findings do substantiate some of the criticisms
against the quantitative bases of the current strategy (i.e., Appendix J),
and raise some questions in regards to the effectiveness of the current
strategy in reducing oxidant in rural areas.  The next logical question is
whether there are any options and what these options are, insofar as the
rural oxidant reduction and replacement of the  Appendix-J method are concern-
ed.  Replacement of the Appendix-J method is- discussed elsewhere (13, 14).
A brief discussion of the rural oxidant problem is given here, next.

     Rural oxidant, conceivably, can arise from several sources, namely:

     •    From natural 03 (from stratosphere and electrical discharge);

     •    From transported (urban) anthropogenic pollutants;

     •    From natural emissions acting alone or in mixture with anthropogenic
          ones;

     •    From local (anthropogenic) emissions  acting alone or in mixture
          with transported ones.


                                      1150

-------
Relevant evidence -- accepted or not -- reported lately on the relative
importance of these sources is summarized as follows:

     (1)  Stratospheric 03 can be transferred to the upper troposphere
          through a "tropopause folding" mechanism (10).   However, the degree
          to which such Q% reaches ground level  has not been established.
          Based on radioactivity measurements, 24-hour-average estimates of
          ground level concentrations of stratospheric 63 are well below the
          NAQS level (11).  No estimates were possible, however, of the maxi-
          mum 1-hour concentrations.

     (2)  Oxidant transport and its potential for rural oxidant buildup have
          been demonstrated repeatedly so that its occurrence is no longer
          in question.  What is in question, however,  is:  (a) the importance
          of this rural oxidant source relative  to the other sources, and (b)
          the dependence of such oxidant on the  transported (presumably urban)
          HC and NOX, and on the locally emitted (rural)  fresh NOX.

     (3)  Some natural hydrocarbons, namely terpenes,  were found to be oxi-
          dant producers when mixed with NOX at  HC-to-NOx ratios between 15:1
          and 25:1 (9).  Such low ratios do not  occur in  rural areas unless
          there are significant sources of anthropogenic  emissions; in this
          latter case, terpenes could conceivably form oxidant.  However,
          the importance of such oxidant in terms of concentration level,
          extent and frequency of occurrence is  far from having been estab-
          lished.  A similar function was proposed for methane also, with
          the suggestion that anthropogenic NOX  would be  necessary for rural
          oxidant reduction (18).  This latter evidence and conclusions,
          however, were rejected by others as erroneous (19, 20).

     (4)  Based on the recent understanding of transport phenomena, one could
          hypothesize reasonably that rural oxidant formation could be caused
          by local anthropogenic emissions in at least two meteorological
          situations:

          •    Under conditions of severe stagnation,  in  which case the local
               emissions alone will enter the photochemical oxidant formation
               process.  Such conditions may occur, e.g., within the center
               of a high pressure cell.

          •    Within the back side of a high pressure cell, in which case
               the observed oxidant would reflect the sum total of urban
               and rural emissions dispersed within the high pressure cell.
               In such a case, obviously, the rural emissions will account
               for a fraction of the observed oxidant.

     The statistics of severe stagnation cases coupled with elevated oxidant
levels have been studied to some extent, and the connection between this
stagnation factor and elevated oxidant concentrations  has indeed been estab-
lished (6).  However, one would still have to consider that the severe stag-
nation conditions occur in a relatively small area only,  within the high pres-
sure cell.  In the severe stagnation case, therefore,  one cannot make a case
for strong effects from rural anthropogenic emissions.

                                    1151

-------
     In the back side of a high pressure cell,  the rural  and urban (anthro-
pogenic) emissions should have relative effects roughly equal  to their relative
levels.  Emission control, as prescribed by the current control  strategy,
will reduce mobile and stationary source emissions in the urban  areas and
mobile source emissions in the rural  areas.  Therefore, insofar  as the rural
problem is concerned, the current control strategy is inadequate only to the
extent that it does not call for control of the stationary source emissions
in the rural areas -- an inadequacy which cannot be serious.

     From the preceding analysis of the rural  oxidant problem and its con-
nection to the current control strategy, it can be concluded that (a) at
present, the rural oxidant problem is not sufficiently well  understood, and,
therefore, it is not possible to define the optimum approach to  rural oxidant
reduction, and (b) the current control  strategy, while it may not represent the
optimum approach, should nevertheless have a beneficial impact upon rural  air
quality.

                                   REFERENCES

1.  U.S. Environmental Protection Agency.  Air Quality Criteria  for Nitrogen
    Oxides.  Air Pollution Control Office Publication No. AP-84, Washington,
    D.C. , January 1971.

2.  Federal Register, 36_,  15486, August 14, 1971.

3.  Research Triangle Institute, "Investigation of Rural  Oxidant Levels as
    Related to Urban Hydrocarbon Control Strategies,"  Environmental
    Protection Agency Publication No. EPA-450/3-74-034, March, (1975).
    Research Triangle Park, N.C.  27711.

4.  Blumenthal, L.L., W.H. White, R.L.  Peace, and T.B. Smith.   Determination
    of the Feasibility of the Long-Range Transport of Ozone or Ozone
    Precursors.  EPA-450/3-74-061, Nov. 1974.   U.S. Environmental Protection
    Agency, Research Triangle Park, N.C.  27711.

5.  Cox, R.A., A.E.J. Eggleton, R.G.  Derwent, J.E. Lovelock, and D.H.
    Pack.   "Long-Range Transport of Photochemical  Ozone in North-Western
    Europe,"  Nature, Vol. 255, No. 5504, pp.  118-121, May 8,  1975.

6.  Research Triangle Institute.  "Study of the Formation and Transport of
    Ambient Oxidants in the Western Gulf Coast  and North-Central and
    North-East Regions of the United  States."   RTI Report pursuant to
    USEPA Contract 68-02-2048, Sept.  1976.   U.S. Environmental Protection
    Agency, Research Triangle Park, N.C.  27711.

7.  Research Triangle Institute.  "Ambient Monitoring Aloft  of Ozone  and
    Precursors in the Vicinity and Downwind of  a Major City  Using a
    Balloon-Borne Platform and Aircraft."  RTI  report on  part  of Project
    DaVinci under USEPA Contract 68-02-2341, 1976.   U.S.  Environmental
    Protection Agency, Research Triangle Park,  N.C.  27711.
                                     1152

-------
 8.  Ripperton, L.A., W.C. Eaton, J.E. Sickles, II, "A Study of the Oxidant-
     Precursor Relations Under Pollutant Transoort Conditions,"  Final
     Report to Environmental Protection Agency, EPA Contract No.  68-02-
     1296, January, (1976).  U.S. Environmental Protection Agency, Research
     Triangle Park, N.C.  27711.

 9.  Westberg, H.H. and R.A. Rasmussen, Washington State University, Pullman,
     Washington.  "Relationships  Between Oxidant Yield and the HC/NOX Ratio
     in Smog Chamber Irradiations of Natural Hydrocarbons."  To be issued as
     an USEPA report, 1976.

10.  Mohnen, V.A., P.E. Coffey, and W.N.Stasiuk.  "Ozone in Rural and
     Urban Areas of New York State.  Part II."  Paper presented at Interna-
     tional Conference on Photochemical Oxidant Pollution and Its Control,
     Raleigh, N.C., September 13-17, 1976.   Proceedings to be published as
     USEPA report.

11.  Reiter, E.R., "The Transport of Radioactive Debris and Ozone from the
     Stratosphere to the Ground,"  subcontract to Stanford Research Inst.
     EPA Contract No. 68-02-2084, November  (1975).  U.S. Environmental
     Protection Agency, Research Triangle Park, N.C.  27711.

12.  Dimitriades, B. "Oxidant Control  Strategies.  Part I.   An Urban Oxidant
     Control Strategy Derived from Existing Smog Chamber Data."  Accepted
     for publication in Envir. Science & Technol.  1976.

13.  Dimitriades, B. "An Alternative to the Appendix-J Method for Calcu-
     lating Oxidant-and N02-Related Control Requirements."  Paper presented
     at the International Conference on Photochemical Oxidant Pollution and
     Its Control, Raleigh, N.C., September  13-17, 1976.  Proceedings to be
     published as USEPA report.

14.  Dodge, M. "The Combined Use of Modeling Techniques and Smog Chamber
     Data to Derive Ozone-Precursor Relationships."  Paper oresented at the
     International Conference on Photochemical Oxidant Pollution and Its
     Control, Raleigh, N.C., September 13-17, 1976.  Proceedings to be
     published at USEPA report.

15.  U.S. Environmental Protection Agency.   "Proceedings of the Solvent
     Reactivity Conference,"  EPA-650/3-74-101 , Nov. 1974, U.S. Environmental
     Protection Agency, Research Triangle Park, N.C. 27711.

16.  Dimitriades, B.  "Application of Reactivity Criteria in Oxidant-Related
     Emission Control  in USA."  Paper  presented at the International  Con-
     ference on Photochemical  Oxidant  Pollution and Its Control,  Raleigh,
     N.C., September 13-17, 1976.  Proceedings to be published as USEPA
     report.

17.  Jeffries, H.E., D.L.  Fox, R.M. Kamens, "Outdoor Smog Chamber Studies:
     Effects of Hydrocarbon Reduction  on Nitrogen Dioxide,"  Environmental
     Protection Agency Publication No.  EPA-650/3-75-011, (1975).   U.S.
     Environmental  Protection  Agency,  Research Triangle Park, N.C.  27711.

                                      1153

-------
18.  Chameides, W., and D.H.  Stedman.  "Ozone  Formation  from  NOX in Clean
     Air,"  Envir.  Science &  Technol. ,  10,  p.  150,  Feb.  1976.
19.   Dimitriades, B.,  M.  Dodge,  J.  Bufalini,  K.  Demerjian, and A.P. Altshuller,
     Envir.  Science &  Technol.,  (in press)  1976.

20.   Weinstock, W., and T.Y.  Chang.  "Methane  and Nonurban Ozone."  Paper
     presented at National  APCA  Mtg.  in  Portland, Oregon, June 29, 1976.
                                    1154

-------
                                                                          23-5
          CONTROL  REGULATIONS  FOR  STATIONARY  SOURCES OF HYDROCARBONS
                             IN  THE  UNITED STATES

                                R.  T.  Walsh*
ABSTRACT

     Stationary *ource*  aac.ou.nt faor over 60  peA.ce.nt o& the. volatLte organic
compound* released to the  atmo*phere o&  the  United State*.   While em-is*ion*
^rom motor ve.hicJLe* are.  beting re.duce.d, hydrocarbon* ^rom *tationary *ource*
one. *til£ increasing in  many kin. Quality Control. Region*.   Recently, the.
Environmental  Protection Agency placed added emphasis  on photochemical oxi-
dant* traceable to stationary source* and it i*  actively engaged i.n *everai
program* aimed at reducing the*e hydrocarbon emi**
-------
   iuAtninA, Auch  OA gaAotine. masike^ing.   Ton. AuAfiace. coating and -some typ&A
   oAgantc ch&nicat manu.{iactLin.t,  ani^oAm n.o.QUitationA may not be pAacticabte..
   ie conAtdeAation may nave,  to  be gac.e. coating opuna-
tiont>, batk gaAoLine. t&untnalA, &klp and baige. &iant>pont o& po.tA.otum product!*,
        AtattonA, pntn.otvjm fLe.^n&u.&> ,  gaAotine. and c^tade o-UL & towage, and
        otkeA. 
-------
         50
         40
 OO fO
 UO O)
 UJ S-
   
-------
fully acceptable, there often has been much disagreement over effectiveness,
cost, and energy requirements of control  technology.   In some instances in
the gasoline marketing industry, control  devices not  only prevent air pollu-
tion but also save product and money for the operator.   Where control mea-
sures don't save enough product to offset control costs, acceptance has been
slower.   At the far end of the spectrum are measures  that require removal  of
organic vapors from dilute exhaust streams, usually with no product recovery
and at a measurable energy penalty.

     While there are dissimilarities in exhaust gases from industrial and
commercial sources, they usually can be categorized by the character and
concentration of the organic(s), presence of other pollutants, and point(s)
of emission.  Major groupings are:

     •    Mixtures too rich to burn (10 to 40 volume  percent hydrocarbon),
          such as those released from gasoline storage and transfer opera-
          tions and some refinery and chemical processes.

     t    Lean mixtures (less than 0.5 volume percent hydrocarbon) often
          contaminated with particulate matter, such  as those released from
          spray booths, baking and curing ovens, and  many chemical manu-
          facturing operations.

     •    Streams containing particulates, carbon monoxide, partially oxi-
          dized hydrocarbons, and/or sulfur compounds, such as those released
          from refinery catalyst regenerators, many chemical manufacturing
          processes, and a few metallurgical furnaces.

     •    Streams containing non-flammable halogenated compounds, such as
          those released from dry cleaning and degreasing operations.

     •    Dispersed, multi-point sources, such as pumps and valves in
          refineries and chemical plants.

     §    Waste disposal sites where solvents or other organics--often mixed
          with solids--are landfilled or otherwise discarded.

     The applicability of add-on control  options—incineration, condensation,
absorption, and adsorption—is strongly influenced by stream characteristics.
Exhaust streams containing rich hydrocarbon mixtures  or halogenated solvents
tend to be the most economical to control, often yielding a dividend through
product recovery.  For example, floating roofs on gasoline storage tanks
(FRT's)  usually pay for themselves in a short period.  Vapor recovery at
truck loading terminals is less cost effective than FRT's but recovered prod-
uct can  offset much  or all of the control cost.  However, several terminal
operators have opted for less expensive incineration  systems wherein dis-
placed vapors are burned in flares with no product recovery.

     The simple displacement or vapor balance system  has been employed at
gasoline service stations and bulk plants.  It has also been used in unloading
bulk solvents.  The transferred liquid displaces a nearly equal volume of air/
organic  vapor to the dispensing vessel with any excess vapor venting to the


                                    1158

-------
 atmosphere.  Where  vapor  tight  connections  can  be  used," the system  is effi-
 cient  and  highly  cost  effective.

     Carbon  adsorption  can  be used advantageously  with  halogenated  solvents
 because  of the  value of recovered product as well  as  the non-flammable char-
 acter  of the solvent.   Carbon finds frequent use in recovering chlorinated
 solvents in  the dry cleaning industry  and to some  extent in degreasing.
 However, adsorption is  seldom used with  petroleum  solvents.

     The most costly exhaust gases to  control are  dilute streams where the
 hydrocarbon  level is too  low to  sustain  combustion, has  little recovery
 value, and is frequently  contaminated  with  sticky  particulate.  For these
 lean streams, incineration  at 1200 to  1500°F is usually  effective for air
 pollution  control purposes  but  requires  appreciable energy input in the
 form of  natural gas or  distillate fuel oil.  Condensation and absorption
 are ineffective because hydrocarbon levels  are  too dilute.  Adsorption can
 provide  high removal efficiency  but requires that  particulates be removed
 beforehand to avoid contamination of the adsorbent.


       In some industries—particularly surface coating—the most promising
means of curtailing organic emissions  are process changes that back out much
of the input solvent or sharply reduce the volume of exhaust gases.   Water-
borne, high  solids  and powder coatings have all  been employed to a  limited
degree in  the manufacture of consumer goods.  Where they are used,  organic
emissions  have  been lowered by 85 to 100 percent.   Unfortunately, these repre-
sent well  under 10  percent of the industrial surface coatings in use today.
Acceptance of low solvent paints by the  coating  industry and its customers
has been far less than was predicted a few years ago.   Manufacturers'  reluc-
tance to commit to  untested coatings isn't surprising since the products often
are subject  to severe conditions.  Applications  range from food and beverage
containers to structural and mechanical  products subject to marine  atmosphere.

      Architectural  coatings are one area where  low solvent content coatings
have been  highly successful.  Today, water-borne paints account for about
70 percent of interior and exterior applications.   While these coatings still
contain  about 20 percent organic solvent, they represent reductions of 75 to
90 percent over conventional oil-base paints.

      In light of the diverse control  possibilities for stationary  hydrocarbon
sources, EPA recently directed additional resources to the review and docu-
mentation  of applicable technology.   The aim is  to publish control  technology
guidelines for principal sources and where applicable to promulgate new source
performance  standards and/or national  emission standards for hazardous pollu-
tants.    At present, the bulk of this effort is directed at the control of
oxidant  forming non-methane hydrocarbons.  At the same time, we are looking at
the toxic  effects of organic emissions,  particularly those associated with
chemical manufacturing processes.  If and when it becomes necessary to apply
more stringent controls to specific toxic materials or to organic halogens
that impact  the upper atmosphere, we should be in possession of sufficient
knowledge  to proceed with an effective control  program.
                                     1159

-------
      With regard to toxic organic emissions, our greatest concerns are the
chronic toxicants associated with carcinogenesis, mutagenesis, arid terato-
genesis.  A compilation of available data is nearing completion by the MITRE
Corporation that identifies over 600 organic compounds that are produced in
large enough quantity, are volatile enough., and exhibit known or potential
toxicity characteristics sufficient to warrant further study.

      For oxidant control purposes, the program is aimed at reducing emissions
of all volatile organic pollutants regardless of their photochemical reactivity.
This course of action is based on review of smog chamber studies and appraisal
of commercial solvents.   These studies show very few organic compounds to be
of "low reactivity"; furthermore, most of the latter are unacceptable as com-
mercial solvents.  Thus, solvent substitution strategies based on photochemical
reactivity appear questionable at best.  Nonetheless, this subject is being
further reviewed to determine whether, in fact, some low reactivity solvents
can be employed and whether solvent substitution would suffice as an oxidant
control strategy for isolated areas of the nation where transport phenomena
do not adversely affect downwind receptors.

     EPA is assembling control technology information for a variety of
petroleum-related industries, chemical manufacturing processes, and end-use
solvent industries, as listed in Table 2.  In some instances, a good deal
of information has already been assembled.  Briefly, developments and ex-
pectations are as follows:

     •    Gasoline Storage Tanks - The floating roof continues to be the
          dominant control device for large tankage, but EPA is making a
          renewed effort to evaluate vapor recovery systems.  Most of the
          effort by EPA and the industry is directed at more positive mea-
          surement techniques for emissions from floating roof tanks.  The
          current program should result in more definitive specifications
          for vapor seals and other components of floating roofs that
          affect emissions.

     •    Bulk Gasoline Terminals - Comprehensive tests are being conducted
          to more fully evaluate the effectiveness of vapor recovery systems,
          particularly in areas where service station controls are in effect.
          These data will be compared with the already completed evaluation
          of incineration devices.   Test procedures are being standardized
          and the feasibility of various monitoring devices analyzed.

     •    Bulk Gasoline Plants - Contract studies will soon be completed to
          evaluate the feasibility of both vapor balance and secondary re-
          covery systems at bulk plants.   Of particular importance is the
          economic impact of control regulations on small operators.  Large
          bulk plants are expected to utilize the same types of vapor recovery
          and incineration equipment that are being used at bulk terminals.

     •    Service Stations - Control systems for underground storage tanks
          (Stage I) have been installed in 1976 in many areas of the nation.
          For the most part, they are simple displacement systems that appear
          to provide well over 90 percent vapor recovery when operated pro-


                                     1160

-------
perly.  The effectiveness of existing systems and operating prac-
tices will be evaluated to determine if hardware improvements and/
or operating restrictions are required.

     The vehicle refueling (Stage II) program has proceeded much
more slowly and there is still much uncertainty as to the effective-
ness of the competing vapor balance and vacuum assist systems.  EPA
will repropose Stage II regulations this month and will continue to
evaluate systems installed or being installed in California and a
few other areas of the nation.  Under the revised Stage II regula-
tion, the first EPA mandated control systems would be installed
in 1978.
Ships and Barges - Most of the effort to date has been directed at
the loading of gasoline into tankers in Gulf Coast ports.   In addi-
tion, there is a great deal of interest in crude unloading both on
the Gulf Coast and the Pacific Coast.  Studies conducted by EPA and
the petroleum industry indicate that emissions are not as great as
originally predicted because the ullage space in the cargo tanks is
well below saturation levels.   Nonetheless, operating practices can
greatly influence saturation and the magnitude of emissions during
loading, ballasting of tanks,  and during inerting and certain main-
tenance operations.

Crude Petroleum Production - Studies of offshore drilling operations
were recently contracted.   Initial efforts will  characterize emissions
and review the more obvious control technology for storage, transfer,
valves, pumps, and separators.  A large fraction of emissions from
production operations are believed to be methane, but heavier frac-
tions are also released.   In at least some instances, air pollution
control equipment will recover substantial quantities of LPG
feedstocks.

Petroleum Refining - We plan to complete a guidelines document in
mid-1977 covering blowdown systems, cooling towers, flares, valves,
pumps, and other miscellaneous sources.  Processing units are being
evaluated under a major research and development contract.  It is
anticipated that equipment specifications and operating practices
will be a large part of any process unit regulations.  For those
processes with discrete exhaust streams, emission limits can be
expected.

Chemical Manufacturing - Carbon black and a few other chemical
processes have been evaluated, but EPA's major contract in this
field will be a comprehensive  study to be awarded within the next
few months.  Control measures  in this area are expected to be ex-
tremely source specific.   We are looking to incineration, adsorp-
tion, and absorption techniques as well as innovative process
changes that can reduce the need for pollution control.  Manufac-
ture of the nine chemicals listed in Table 2 are important because
of their toxicity as well  as their impact on photochemical oxidants.
                           1161

-------
     •    Surface Coating - Principal contract studies in this area are nearing
          completion.  We expect the eventual  answer to solvent control to be
          a combination of improved coating materials and better application
          methods.  Incineration and possibly adsorption are expected to sup-
          plement control programs, with incinerator designs to be maximixed
          for purposes of heat recovery and energy economy.

     t    Degreasing - Improvements in design  and operation  are needed to
          minimize losses and improve capture  efficiency at  the degreaser.
          We also expect that significant improvements will  be achieved
          through better operating practices.   Adsorption will continue
          to be the dominant stack gas clean-up technique.

     •    Dry Cleaning - Carbon adsorption is  expected to become standard
          practice for most perchloroethylene  plants.  Even  without control
          regulations, plants using fluorinated solvents will continue to
          maximize vapor recovery because of the cost of the solvent.  The
          future of petroleum solvent dry cleaning is uncertain because of
          related problems of safety, economics, and energy  requirements of
          available control technologies.

     For all of these sources, consideration will be given to the promulgation
of new source performance standards (NSPS) under Section 111 of the Clean Air
Act or national emission standards for hazardous pollutants  under Section 112.
New source standards are the more likely vehicle.  Hazardous pollutant stan-
dards are specific to the pollutant and apply  to both new and existing installa-
tions.  If the Section 112 route is chosen, emission limits  will be established
for all significant sources of the pollutant.   To date, hazardous pollutant
standards have been proposed for only one organic pollutant, vinyl chloride.
The final regulations are expected to be promulgated shortly.

     Where we are dealing with large identifiable sources of oxidant-forming
organics, NSPS will be the choice.  Thus, bulk gasoline terminals, refinery
process units, major surface coating operations, and many chemical manufac-
turing operations are principal candidates for NSPS.  In these instances,
control technology, costs, and economic and environmental impact will be
documented and the information made available  to interested  parties.

     To cover existing sources, we intend to publish what we term "guidelines
documents".   The purpose of the guidelines is  to apprise State and local
jurisdictions of the effectiveness and costs of available control options.
The documents will recommend no specific emission limits but will leave the
choice of control level to the States.  We anticipate that this approach will
lead to source specific hydrocarbon regulations for many different industries.
The greatest specificity probably will be applied to the chemical manufac-
turing and surface coating industries where both economics and feasibility
can vary with the product and the service to which the product is subjected.
Conversely,  such industries as gasoline marketing, dry cleanirg, and degreas-
ing are well standardized; here we look for relatively un^fonr regulations,
although degrees of stringency are expected depending on tne desired hydro-
carbon reduction needed in the particular area.
                                     1162

-------
     The first guidelines document was prepared for five surface coating
industries — can manufacturing, fabric coating, paper coating, coil coating,
and automobile assembly plants.   A draft document was distributed in July
to control agencies, affected industries, and other interested parties.
The finalized document is scheduled for release in November.   It will re-
flect the large number of comments we have received on the draft document.

     In summary, to combat the oxidant problem and a growing  toxic organic
chemical problem, EPA plans to enact a number of new source performance
standards and a few hazardous pollutant standards for sources of organic
pollutants.   State regulations will require increased control of organic
emissions from existing stationary sources.  All of these are likely to be
in the form of positive reductions rather than substitution of less reactive
solvents.  Source specific guideline documents will provide information
needed by the States to make decisions on emission limits, costs, and eco-
nomic and environmental impact.

     It is clear that a nationwide cleanup of organics is beginning and that
it will have a major impact on almost all industries that use, handle, or
produce organic materials.

          TABLE  1.   ANTHROPOGENIC  SOURCES  OF  ORGANIC AIR  POLLUTION
                          IN  THE UNITED  STATES,  1975	

                                         Emissions,      Fraction  of Total

                                        10  tons/yr         Emissions,  cl


     Transportation                        11.7                37.6

     Fuel Combustion                       1.4                4.5

     Forest, Agricultural
        and  Other Open  Burning              1.1                3.5

     Solid  Waste Disposal                  0.9                2.9

     Petroleum  Refining                    0.9                2.9

     Oil  and Gas Production
        and  Marketing                       4.2                13.5

     Chemical Manufacturing                1.6                5.1

     Metals Manufacturing                  0.2                0.6

     Other  Manufacturing                   0.8                2.6

     Organic Solvent Usage                 8.3                26.7

                                           31.1                99.9
                                     1163

-------
           TABLE 1A.  CHARACTERISTICS OF INDUSTRIAL EXHAUST STREAMS
                    CONTAINING VOLATILE ORGANIC POLLUTANTS


•     Mixtures too rich to burn - gasoline storage and transfer,  petroleum
      and chemical  processes.

•     Mixtures too lean to burn - surface coating, baking and curing ovens,
      plastics and rubber product manufacture.

•     Streams containing particulates, CO, HC,  SOX, H2S and partially oxidized
      organics - refinery, chemical  and metallurgical  processes.

§     Streams containing halogenated organics - dry cleaning and  degreasing.

•     Multipoint Sources - pumps, valves, blowdown system, cooling towers
      in refineries and chemical  plants.

t     Waste disposal sites for solvents, chemical  and  petroleum plant wastes.
                                     1164

-------
TABLE 2.  SOURCES OF ORGANIC AIR POLLUTANTS BEING EVALUATED

PETROLEUM RELATED                      SOLVENT USE
STORAGE TANKS                          SURFACE COATING
BULK TERMINALS                         DRY CLEANING
BULK PLANTS                            DECREASING
SERVICE STATIONS                       RUBBER PRODUCT MFG.
  VEHICLE FUELING                      PHARMACEUTICAL MFG.
  AND STORAGE TANKS                    ADHESIVE MFG.
SHIP AND BARGE TRANSFER                WOOD PRODUCT MFG.
PETROLEUM PRODUCTION                   PAINT MFG.
PETROLEUM REFINING
  PROCESSING EQUIPMENT
  FUGITIVE EMISSIONS FROM
  PUMPS, VALVES, ETC.
                  CHEMICAL MANUFACTURING
                  BENZENE
                  ACRYLONITRILE
                  CARBON TETRACHLORIDE
                  FORMALDEHYDE
                  MALEIC ANHYDRIDE
                  NITROBENZENE
                  PERCHLOROETHYLENE
                  TRICHLOROETHYLENE
                  VINYLIDENE CHLORIDE
                         1165

-------
                                                                     ADDENDUM
Prepared Comments cm Japanese Photochemical Air Quality Standards and Control
Strategies, by Professor Raisaku Kiyoura, Tokyo Institute of Technology.

     Following are some comments of mine and questions on the subject standards.
In the opinion of several biologists concerning the relative toxicities of
oxidants and N02, the adverse effects of oxidants are 10 to 20 times greater
than those of N02.  Therefore, control of oxidants is essential in air pollu-
tion control strategies.  Until now, N02 has not been controlled as stringently
as the hydrocarbon precursor of photochemical oxidants.  However, some people
believe that N02 may also have to be stringently controlled to achieve the air
quality standard for oxidant.

     The oxidant standards in Japan and U.S. have been set at the same level:
0.08 ppm (1 hour)*.  The U.S. N02 standard was set at 0.05 ppm for annual
average (equivalent to a 24-hour average of 0.14 ppm), and some biologists
feel that one-hour-maximum standard for N02 is also necessary.  The Japanese
N02 standard of 0.02 ppm for 24-hour average is five to seven times stricter
than that of the U.S., but there is some uncertainty whether this difference
is justified by biomedical evidence.

     Very stringent control of N02 without similarly stringent control of non-
methane hydrocarbons could result in increased oxidant formation.  Therefore,
these pollutant emission levels must be kept in balance consistent both with
the biomedical justification of the N02 and oxidant standards and with the
relative roles of the NO  and hydrocarbon precursors in the photochemical
oxidant formation process.
     Next, I submit that the following four questions need to be answered
with respect to the air pollution situation in Japan:

     1.   Under such a low Japanese N0> standard, can Japan reduce oxidant
          formation by controlling HC?

     2.   If so, what is the ambient HC level in ppm that would need to be
          achieved?
*The Japanese standard for oxidant was originally set at 0.06 ppm but was
subsequently changed to 0.08 ppm for reasons related to the analytical mea-
surement method.

                                     1167

-------
     3.   What is the relationship between the HC level and the natural
          background level?

     4.   Is that HC level attainable?

     A guide level limiting nonmethane HC between 0.20 and 0.31 ppmC (from
6 to 9 a.m.) was recommended in August 1976 by an expert committee of the
Japanese Central Council for Control of Environmental Pollution.
                                    1168

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse bejore completing)
 1 REPORT NO
  EPA-600/3-77-001b
                                                            3 RECIPIENT'S ACCESSION-NO.
4 TITLE AND SUBTITLE
  INTERNATIONAL CONFERENCE  ON PHOTOCHEMICAL OXIDANT
  POLLUTION AND ITS CONTROL
  Proceedings:  Volume II                        	
                6. PERFORMING ORGANIZATION CODE
                5. REPORT DATE
                  January  1977
 7 AUTHOR(S)
                                                           8. PERFORMING ORGANIZATION REPORT NO.
  Basil  Dimj.triades, Editor
9 PERFORMING ORGANIZATION NAME AND ADDRESS
  Environmental Sciences Research Laboratory
  Office  of Research and Development
  U.S.  Environmental Protection Agency
  Research Triangle Park, N.C.   27711
                10. PROGRAM ELEMENT NO.
                  1AA603
                11. CONTRACT/GRANT NO.
 12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Sciences Research Laboratory
  Office  of Research and Development
  U.S.  Environmental Protection Agency
  Research Triangle Park, N.C.   27711
                13. TYPE OF REPORT AND PERIOD COVERED
                  In-house                	
                14. SPONSORING AGENCY CODE

                  EPA-ORD
 15. SUPPLEMENTARY NOTES
  This  Conference was sponsored by the US-Environmental Protection Agency
  with  the patronage of the  Organization for Economic Cooperation and Development.
 16 ABSTRACT

        The  proceedings consist  of 97 technical papers covering such areas  as
  analytical  methods for photochemical oxidants and precursors; causes of  urban,
  suburban, and non-urban oxidant; biological effects; oxidant control strategies;
  and  trends  in emissions and emission control technology.   The International
  Conference  was held in Raleigh, N.C. in September 1976.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS

  Mir pollution
  *Ozone
  * Photochemical reactions
                                              b.IDENTIFIERS/OPEN ENDED TERMS  C  COSATI Field/Group
                                13B
                                07B
                                07 E
13. DISTRIBUTION STATEMENT

  RELEASE  TO  PUBLIC
   19 SECURITY CLASS (This Report/
      UNCLASSIFIED
21. NO. OF PAGES
      620
   20. SECURITY CLASS (This page)
      UNCLASSIFIED
                              22 PRICE
EPA Form 2220-1 (9-73)
1169

-------An error occurred while trying to OCR this image.

-------