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 e.d-i' < ng i /iad appealed ui t/ie 03
PondeAo-ia ptne ieedi'-cngi weAe additionally exposed to the standardized
*U. S. Environmental Protection Agency, Corvallis, Oregon.
635
-------
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
)
1— •
(O
QJ
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
^-1""1*S^ ~7 »r
^ ^^^^jrr-tnnt
• +
\ UP MELBOURNE
V\M:i "SYDNEY
x**^ - — K SY DNEY PEAK)
HC,
^Ox
3
2
1
*ob-\ 2 -1
20
0
STD.DEV
kg /miTh
20
0
RADIUS
8
miles
10
0
Figure 7. The HC and NO source intensity, their standard deviations
and HC/NO ratio Variation with city radius. Melbourne
and Sydney 24-hour data and Sydney a.m. peak results.
of 38% (28) in NO emissions for the average vehicle. The forecast reductions
that would be achieved over our Melbourne driving pattern can be seen in Figure 8.
Over the speed range 10 to 24 mph, reductions will probably be better than the
reduction achieved over the LA-4 cycle. Only at very low speeds, 2 to 4 mph,
will 75% of the forecast reduction be achieved. With emissions-controlled
vehicles, the emissions will tend to be highest at low speeds and hence worst
in the CBD area.
In the Organization for Economic Cooperation and Development (OECD)
countries, restricting traffic access to city centres has been carried out,
but it is doubtful if traffic reductions by 50% over a sufficiently large area
(10 ml2) could be achieved initially with public and economic acceptability (29).
EMISSIONS CONTROL BY TRAFFIC MANAGEMENT
Traffic management techniques, as a primary NO control strategy, do
appear to be practicable. Firstly, we examine in Figure 9 what reductions are
already achieved in urban driving. The results of Figure l(b) are replotted
showing emission points on each of the 10 Melbourne sections for which the
traffic flow was above 90"- of the saturated traffic density, as defined by
highway capacity manuals (17). By examining only saturated conditions, which
apply at the a.m. peak hour, we are able to observe emission rates from driving
that is restricted by road rather than restricted by driving habit, which is
more difficult to draw useful conclusions from. In Figure 9, it is clear
981
-------
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 t/5
0)
+-> 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— {/)
) E
•r- UJ
E
UJ
*
to CO
•O r—
S- i-
re 2:
•o
c s-
re co
4J 0.
tA
s= E
o re
*r~ &-
VI CD
to
•r- C
E 1-
UJ
>•}
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.
------- |