PROCEEDINGS
SEVENTH JAPAN-US CONFERENCE
ON
PHOTOCHEMICAL AIR POLLUTION
November 29-30, 1982
Tokyo, JAPAN
US DELEGATION
Dr. B. Dimitriades, Chairman
Environmental Sciences Research
Laboratory
USEPA
Dr. A. P. Altshuller
Environmental Sciences Research
Laboratory
USEPA
JAPANESE DELEGATION
Mr. Danjuro Miki, Chairman
Environment Agency
Mr. Saburo Kato
Environment Agency
Dr, Toshiichi Okita
National Institute for
Environmental Studies
Dr. Naoomi Yamaki
Saitama University
Dr. Hajime Akimoto
National Institute for
Environmental Studies
Mr, Tetsuhito Komeiji
Tokyo Metropolitan
Research Institute for
Environmental Protection
Dr. Haruo Tsuruta
Yokohama Research Institute for
Environmental Protection
Dr. Shinji Wakamatsu
National Institute for
Environmental Studies
COMPILED BY
AIR QUALITY BUREAU
ENVIRONMENT AGENCY
3-1-1, Kasumigaseki, Chiyoda-ku, Tokyo, JAPAN
-------
PROCEEDIN6S
SEVENTH JAPAN-US CONFERENCE
ON
PHOTOCHEMICAL AIR POLLUTION
November 29-30, 1982
Tokyo, JAPAN
US DELEGATION
Dr. B. Dimitriades, Chairman
Environmental Sciences Research
Laboratory
USEPA
Dr. A. P. Altshuller
Environmental Sciences Research
Laboratory
USEPA
JAPANESE DELEGATION
Mr. Danjuro Miki, Chairman
Environment Agency
Mr. Saburo Kato
Environment Agency
Dr, Toshiichi Okita
National Institute for
Environmental Studies
Dr. Naoomi Yamaki
Saitama University
Dr. Hajime Akimoto
National Institute for
Environmental Studies
Mr, Tetsuhito Komeiji
Tokyo Metropolitan
Research Institute for
Environmental Protection
Dr. Haruo Tsuruta
Yokohama Research Institute for
Environmental Protection
Dr. Shinji Wakamatsu
National Institute for
Environmental Studies
COMPILED BY
AIR QUALITY BUREAU
ENVIRONMENT AGENCY
3-1-1, Kasumigaseki, Chiyoda-ku, Tokyo, JAPAN
-------
Printed in January 1983
by the
JAPAN ENVIRONMENT AGENCY
3-1-1, Kasumigaseki, Chiyoda-ku
Tokyo, JAPAN
PROCEDINGS—PAGE j
-------
PREFACE
This conference is a part of the activities fostered under the
Japan-US Environmental Agreement negotiated between the two countries in
August, 1975. The purpose of the Environmental Agreement and associated
activities is to develop environmental awareness and to promote cooperation
between the Japan and US in effort to reduce air pollution. Cooperative
activities pertaining to photochemical air pollution were commenced in
June, 1973, with the conduct of the first joint Japan-US Conference on
Photochemical Air Pollution. As of todate such Conferences have been
held as follows:
First Conference: 1973, Tokyo, Japan
Second Conference: 1975, Tokyo, Japan
Third Conference: 1976, Raleigh, NC
Fourth Conference: 1978, Honolulu, Hawaii
Fifth Conference: I960, Tokyo, Japan
Sixth Conference: 1981, Research Triangle Park, KC
Seventh Conference: 1982, Tokyo, Japan
PROCEDINGS—PAGE II
-------
TABLE OF CONTENTS
Page
Introduction
VI
Agenda of Meeting »•
Joint Communique
Technical Papers
1. Countermeasures for Photochemical Air Pollution in Japan
( Miki ) ___
~ 1
20 Atmospheric Reaction Mechanisms for Photochemical Ozone/
Oxidants ( Dimitriades ) ^o
3. Smog Chamber Study of Photochemical Ozone Formation:
Reactivities of Hydrocarbon-NO Mixtures and Sampled
X
Ambient Air ( Akimoto ) „-,
[ Ref. ] Correlation of the Ozone Formation Rates with
Hydroxyl Radical Concentrations in the Propylene-
Nitrogen Oxide Dry Air System: Effective Ozone
Formation Rate Constant ^
4. Further Development and Validation of EKMA ( Dimitriades ) -,,
5. Acid Rain (Deposition) Chemistry and Physics ( Altshuller ) -,-,
60 Status of Recepter Models ( Altshuller )
7. Recent Development of Aerosol Studies in Japan ( Yamaki ) ^QR
Appendices
Agenda of Joint Meeting
Technical Papers
1. The Problem of Acid Rain in Japan ( Kato )
2. Progress in Photochemical Air Quality Simulation Modeling
( Demerjian )
3. Researches on Acid Rain in Japan ( Komeiji )
t Ref. ] A Numerical Model of Acidification of Cloud Water
4. Urban Ozone Modeling Developments in the U»S. ( Dimitriades )_ 2^1
5. A Numerical Simulation of Local Wind and Photochemical Air
Pollution ( Kimura )
Ł57
6. Transport and Transformation of Air Pollutants by Land and
Sea Breezes ( Tsuruta ) 2gy
7. Evaluation of Eight Linear Regional-scale Sulfur Models by
the Regional Modeling Subgroup of the United States/Canadian
WorkGroup 2 ( Demerjian )
PROCEDINGS—PAGE jv
-------
8. Field Studies on Photochemical Air Pollution in Japan
( Wakamatsu ) -,-,-,
[ Ref. ] A Lagrangian Observation of Polluted Air Mass
Using Aircraft
[ Ref. ] Distribution of Photochemical Pollutants and their
Three-Dimensional Behavior covering the Tokyo
Metropolitan Area
9. U.S. Studies on Stratospheric Ozone ( Wiser ) ,„,
10. Intrusion of Stratospheric Ozone into the Troposphere
( Muramatsu ) _.
11. The Vertical Distributions of CF2Cl2, CFC13 and N20
over Japan ( Hirota )
PROCEDINGS—PAGE v
-------
INTRODUCTION
Replacing Dr. Seigi Yoshizaki, Director of Air
Quality Bureau, Environment Agency, who could not attend
the meeting on urgent business, Mr. Danjuro Miki, Director
of Planning Division, Air Quality Bureau, Environment Agency,
welcomed the delegates. Dr. Basil Dimitriades, head of the
U.S. delegation, thanked the organizers of the Seventh
Conference.
PROCEDINGS—PAGE vi
-------
SEVENTH JAPAN-US CONFERENCE
ON
PHOTOCHEMICAL AIR POLLUTION
Conference Room
Restaurant Castle
2-1-1 Uchisaiwai-cho, Chiyoda-ku
Tokyo, Japan
November 29-30, 1982
AGENDA
Monday, November 29, 1982
Acting Chairman: Mr. D. Miki
10:00 - 10:40
10:40 - 11:00
Opening Remarks
Introduction of Participants
Election of Session Chairman
Approval of Conference Program
Refreshments
Dr. S. Yoshizaki
( Director of Air
Quality Bureau )
Session Chairman: Dr.RDimitriad.es
11:00 - 12:00
12:00 - 13:50
13:50 - 15:00
15:00 - 15:50
15:50 - 17:00
17:30 - 19:30
Countermeasures for Photochemical
Air Pollution in Japan
Lunch
Atmospheric Reaction Mechanisms
for Photochemical Ozone/Oxidants
Smog Chamber Study of Photochemical
Ozone Formation: Reactivities of
Hydrocarbon-N0x Mixtures and Sampled
Ambient Air
Refreshments
Further Development and Validation
of EKMA
Acid RainC deposition ) Chemistry
and Physics
Reception
Mr. D. Miki
Dr. B. Dimitriades
Dr. H. Akimoto
Dr. B. Dimitriadee
Dr. A. P. Altshuller
PROCEDINGS—PAGE
-------
Tuesday, November 30, 1982 ' Session Chairman: Mr. D. Miki
10:00 - 10:45 Status of Receptor Models Dr. A. P. Altshuller
10:45 - 11:15 Refreshments
11:15 - 12:00 Pecent Development of Aerosol
Studies in Japan Dr. N. Yamaki
12:00 - 13:30 Lunch
13:30 - 15:00 General Discussion
Plans for Future Activities
Preparation of Joint Communique
Conclusion of Meeting
Tokyo $ Tsukuba
PROCEDINGS—PAGE ix
-------
SEVENTH JAPAN-US CONFERENCE OF PHOTOCHEMICAL
AIR POLLUTION JOINT COMMUNIQUE
The seventh Japan-US Conference on Photochemical Air
Pollution was held in Tokyo, Japan, on November 29 - 30,
1982. The US delegation consisted of Dr. B. Dimitriades
(Head of delegation) and Dr. A. P. Altshuller; the Japanese
delegation consisted of Mr. D. Miki (Head of delegation),
Mr. S. Kato, Dr. T. Ohkita, Dr. N. Yamaki, Dr. H. Akimoto,
Dr. H. Tsuruta, Dr. S. Wakamatsu and Mr. T. Komeiji.
The two delegations discussed the following subjects:
- Countermeasures for photochemical air pollution in Japan
- Atmospheric reaction mechanisms
- EKMA Model
- Acid rain chemistry and physics (deposition)
- Studies on sampling, atmospheric chemistry/ removal, and
source apportionment of ambient aerosol
Particular interest on the part of the US delegation was
expressed in on-going and planned photochemical smog chamber
studies in Japan and in recent Japanese developments regarding
mechanism of photochemical ozone formation in the atmosphere.
The Japanese delegation was interested in the latest EKMA
Model for photochemical air pollution countermeasures, Source
Apportionment Model for aerosol countermeasures, and acid rain
studies in the US. Both delegations agreed to promote
exchanging information in such areas furthermore.
The two delegations tentatively agreed to hold the next
meeting of Photochemical Air Pollution Panel in the US.
The date and agenda of the next meeting will be coordinated
through future communications between the two Panel cochairmen,
Tokyo, November 30, 1982
Dr. Basil Dimitriades
Head of US delegation
Mr. Danjuro Miki
Head of Japanese delegation
PROCEDINGS — PAGE X
-------
COUMTERMEASURES FOR PHOTOCHEMICAL
AIR POLLUTION IN JAPAN
Presented by D. Miki
Japan Environment Agency
PROCEEDINGS—PAGE 1
-------
Report on Curbing Hydrocarbons Emissions from Stationary Sources
In addition to promoting emergency measures to prevent photochemical
air pollution, the Environment Agency has been striving to compile and
analyze relevant knowledge and reduce the substances that cause the
phenomenon.
Hydrocarbons are among the substances that trigger photochemical smog.
The importance of measures to curb their emissions from stationary sources
was stressed in a document on the "direction of measures against photo-
chemical smog," which was approved by a relevant conference on April 8,
1975, and in a report which was submitted by the Central Council for
Environmental Pollution Control on August 13, 1976.
The stationary sources discharging hydrocarbons and the way they are
discharged are greatly diversified, as is emission control technology.
Because of the need to take reduction measures according to cases, a
panel of experts was set up in November 1979, to grasp the realities of
emissions, appraise emission control technology, and consider other
relevant matters.
The experts, led by Naoomi Yamaki, professor at the Faculty of Engineering
of Saitama University, have recently presented a report. On the basis of
the report, the Environment Agency will announce its policy on measures
to curb hydrocarbons emissions from stationary sources and do its best
to promote them. The gist of the report follows below.
I. PRESENT STATE OF PHOTOCHEMICAL AIR POLLUTION AND NECESSITY FOR
CONTROLLING THE EMISSION OF HYDROCARBONS
1. Present state of photochemical air pollution
Danger to the health from photochemical air pollution became a
problem after the event at Rissho High School in July, 1970.
Since then serious studies have been made and preventive measures
have been taken one after another. Thus, the situation of photo-
chemical air pollution has tended to improve.
PROCEEDINGS—PAGE 3
-------
However, photochemical oxidant concentrations in most areas
exceeded the ambient air quality standard and in addition,
warnings based on the Article 23 of the Air Pollution Control
Law are still often issued. In these last two years, the number
of days when a warning was issued decreased remarkably (Fig. 1)
partially due to the influence of the cool summer. But in view
of the fact that many warnings were issued in May and in the
beginning of June this year, the present condition seems to allow
buildups of highly concentrated pollution in certain weather
conditions. Furthermore, warnings have been issued also in the
outer suburbs of large cities, and in a survey on the influence
of photochemical smog on plants, the influence on plants was
found to have spread. In these situations, photochemical air
pollution has become a problem covering a wide area.
As can be seen, photochemical air pollution still remains a
difficult problem, and in future the conditions of pollution must
be carefully watched, and proper control measures taken.
2. Generation mechanism of photochemical air pollution
Photochemical air pollution refers to a phenomenon in which air
containing hydrocarbons (HC) and nitrogen oxides (NOx) emitted
from various sources reacts with sunlight (ultraviolet rays),
affected by various weather conditions, to produce oxidants (Ox)
such as ozone (0^), and peroxyacylnitrates (PANs), nitrates,
aldehydes, etc. To clarify such a complicated generation mechanism
involves many difficulties. For this reason, research is being
made energetically in two different approach, that is, in chamber
studies to clarify the processes of reaction and in field studies
to mainly clarify the processes of transport and diffusion, and
information has been accumulating steadily.
PROCEEDINGS—PAGE 4
-------
(1) Chamber studies
Hitherto, various chambers have been used, to make studies
concerning the photooxidation reactions in the environment
and in an artificial atmosphere, and very accurate experiments
are being made in low concentrations close to environmental
conditions (Note 1). According to the information obtained
from these chamber studies (Note 2), in order to lower the
production of ozone in the region with excessive hydrocarbons
like in the normal atmosphere, it is surmised to be necessary
to take measures to (1) decrease the emission of nitrogen
oxides to suppress the possible amount of ozone production
by lowering the nitrogen oxide concentration, and (2) decrease
the emission of hydrocarbons to suppress the rate of ozone
production by lowering the non-methane hydrocarbon (NMHC)
concentration.
(2) Field studies
Since the generation mechanism of photochemical air pollution
in our country is closely linked with the land and sea breeze
generated in coastal areas, it is important to clarify the
processes of transport drift and diffusion as well as the
reaction mechanism, to determine the relationship between
oxidants and non-methane hydrocarbons. For this purpose,
field studies have been made mainly in the Kanto Area, and
as a result, precious information has been obtained. For
example, it has been found necessary to investigate photo-
chemical air pollution as a phenomenon over two or more
continuous days, in an area where the circulation of air
is poor.
Based on these results, efforts are being made to develop
a simulation technique to secure the quantitative relationship
between the emission of nitrogen oxides and hydrocarbons and
the concentration of photochemical oxidants under various
weather conditions in more detail.
PROCEEDINGS—PAGE 5
-------
3. Emergency urgent necessity for controlling stationary sources of
hydrocarbons
As has been already pointed out, to prevent the generation of
photochemical oxidants, it is necessary to decrease the emission
of nitrogen oxides and hydrocarbons which generate such oxidants,
and as mentioned in the previous section, the specific functions
of both these materials in producing photochemical oxidants are
gradually being elucidated. Of these, for nitrogen oxides, severe
legal regulations have been enforced step by step for emission
from stationary sources and motor vehicles.
For hydrocarbons, legal regulations for the emission of hydrocarbons
from motor vehicles have also been intensified step by step since
1970, and control measures have been enacted against stationary
sources by governments ordinances of local, etc. As a result, as
shown in Table 1, the non-methane hydrocarbon concentration in the
environment has tended to decrease a little.
However, in the present situations of photochemical oxidants,
measures for preventing production are still insufficient, and
efforts must be made successively to suppress the materials
causing such production. Above all, control of hydrocarbons
emitted from stationary sources is still lax compared with control
against motor vehicles (Table 2), and the intensification of
control measures is an urgent metter.
(Note 1)
In clarification of the processes of reaction of photochemical air
pollution by the chamber studies, attention is being paid to the genera-
tion of ozone as a main component (usually 80 to 90%) of photochemical
oxidants, for analysis. With regard to hydrocarbons as a primary
pollutant, the non-methane hydrocarbons (NMHC) excluding methane, the
reactivity of which can be neglected, are adopted as an index.
PROCEEDINGS—PAGE 6
-------
(Note 2)
According to studies concerning photo-oxidation reaction in the
system of propylene (C3H6), nitrogen oxides and air, using the
large chamber at the National Institute for Environmental Studies,
it has been clarified that if propylene concentration is excessive
(^Hglo/lNOxlo-tS), the maximum ozone concentration ([03] max)
caused by light irradiation continued until ozone has reached the
maximum concentration is proportional to the square root of the initial
nitrogen oxide concentration (/[N0x]0) irrespective of the initial
propylene concentration, that the maximum ozone production rate at
that time is proportional to the initial propylene concentration
([C3H6]0) and the maximum concentration of hydroxyl radicals ([OH] max)
existing in the reaction system, and so on.
Furthermore, it has been proven that these relations derived for
propylene are effective also for other hydrocarbons, and it has been
almost confirmed that the same applies also in studies concerning
photo-oxidation reaction in the environment using the Environment
Agency's movable chamber.
II. REALITIES OF HYDROCARBONS EMISSIONS FROM STATIONARY SOURCES AND
APPRAISAL OF EMISSION CONTROL TECHNOLOGY
It is an urgent problem to curb hydrocarbons emissions from stationary
sources. But consideration of emission control measures must be pre-
ceded by overall knowledge of the realities of emissions and a precise
appraisal of emission control technology.
1. Realities of Emissions
Hydrocarbons are used and handled in a great variety of fields in
the form of crude oil^ refined oil products, petrochemical products
and the like. The industries and facilities, which discharged
hydrocarbons, are also greatly diversified.
PROCEEDINGS—PAGE 7
-------
The amount of emissions, as surveyed by the Environment Agency, is
given in Table 3^. Emissions dropped from about 1.3 million tons
in 1973 to 1.14 million tons in 1978, a difference of about 160,000
tons. In both years, industries using solvents accounted for about
80 percent of the total. Particularly, the industries having to
do with coating accounted for about 50 percent of the total.
The survey dealt only with what seems to be principal emission
sources. It did not cover all the industries and facilities which
discharge hydrocarbons. The industries and facilities, which were
not covered, need to be surveyed hereafter.
To check on the effect which the amount of emissions has on con-
centrations in the environment, the experts carried out case studies
on the areas where data on emission sources and data on concentra-
tions in the environment are in order.
Since a diffusion model, which reproduces environmental concentra-
tions from the amount of emissions, has been developed, it now can
be expected that a national picture will emerge as detailed data
on emission sources and environmental concentrations become available.
2. Appraisal of Emission Control Technology
Technological measures to curb emissions are as varied as the
diversity of the modes of emissions. Principal ones are given below.
(1) Control through equipment
a) Evaporation control equipment
This approach is designed to prevent the discharge of
hydrocarbons from storage and other facilities by changing
their structure. The relevant devices include floating
roofs, inner floating roofs, and vapor return equipment.
PROCEEDINGS—PAGE 8
-------
b) Disposal equipment
This involves the use of devices to capture hydrocarbons
emitted from various kinds of facilities for recovery or
disposal by incineration. The devices come in five
different kinds. They are respectively based on an
absorption methods, an absorption method, a condensation
method, a direct incineration method, and a catalytic
oxidation method.
(2) Control through raw materials
Many of the conventional materials used for the production of
paints, printing ink and the like contain hydrocarbons in the
form of solvents and diluents. Consequently, hydrocarbons
evaporate when paints, printing ink and the like are put to
use. A good way to combat this is to switch from these
conventional materials to materials which contain no hydro-
carbons at all or only a little of them (referred to as low-
pollution materials in the report presented by the experts).
Other case-by-case approaches in accordance with the modes of
emissions are possible.
Control measures dealing with respective categories of emission
sources are scheduled to be discussed in the next issue.
3. Emission Control Measures by Sources
Reflecting the great diversity of processes whereby hydrocarbons
are discharged and the way they are discharged, emission control
measures that can be applied are also greatly diversified.
Under the circumstances, the Environment Agency's panel of experts
categorized emission sources by the kinds of facilities and industries
and studied and assessed the control technology that is applicable
to each category.
PROCEEDINGS—PAGE 9
-------
Their conclusions are shown in Tables ^ and IT, which list emission
control measures by facilites and industries. The details of these
measures were put together in the form of "Technological Guidelines
for the Control of Hydrocarbon Emissions from Stationary Sources,"
so that they will serve as a manual. The guidelines are attached
to the report from the panel.
III. DIRECTION OF EMISSION CONTROL MEASURES FOR STATIONARY SOURCES
1. Basic Direction of Control Measures
The rate of compliance for the environmental quality standard on
photochemical oxidants is low. On the other hand, damage attributed
to photochemical smog persists in the major cities and their environs,
chiefly in summer. Under the circumstances, it is urgently necessary
to curb hydrocarbon emissions from stationary sources as they are
among the agents that trigger the smog.
In addition, some of the numerous substances which come under the
category of hydrocarbons are feared to be toxic. For this reason
and from the viewpoint of conserving resources, emission controls
are also required.
Two approaches are conceivable. One is a realistic and technological
approach. This involves making the most of the currently available
technology to curb emissions and promoting technological development
at the same time to achieve a further limitation.
The other is a total volume control approach. With a view to
achieving the environmental quality standard on photochemical
oxidants and keeping the environmental concentrations of hydro-
carbons at levels that comply with the standard, this calls for
emission controls to be based on the findings of research on the
effects which individual emission sources have on the environmental
concentrations of hydrocarbons.
PROCEEDINGS—PAGE 10
-------
With regard to the total volume control approach, the quantitative
cause-and-effect relationship between photochemical oxidants and
hydrocarbons in the atmosphere and the effects hydrocarbon emissions
from respective sources have on the environmental concentrations
of hydrocarbons in the area where they are located have to be
cleared up. As has been stated earlier, these points are still
in the process of being clarified.
As for the realistic and technological approach, the Environment
Agency's panel of experts has drafted "Technological Guidelines
for the Control of Hydrocarbon Emissions from Stationary Sources."
The agency believes that this approach of making the most of the
currently available technology to curb hydrocarbon emissions should
be pursued for the time being.
There still remain aspects that require consideration before the
introduction of legal controls to implement this approach, since
data on individual factories and other business establishments are
scarce and there are facilities and industries, whose emissions are
still to be determined.
For the time being, therefore, the Environment Agency will promote
maximum emission control efforts in line with local conditions on.
the basis of the report and "Technological Guidelines" which are
attached to the report, particularly the control measures by
facilities and industries, and at the same time, it will do its
best to compile data on emission sources with the cooperation of
local governments to work out specific methods to be employed when
legal controls are introduced.
In addition, the agency will continue to compile knowledge to
determine the quantitative cause-and-effect relationship regarding
the mechanism whereby photochemical oxidants come into being and
the environmental concentrations of hydrocarbons. It will also do
its best to streamline the system to carry out anti-pollution
measures in emergencies affecting wide areas.
PROCEEDINGS—PAGE 11
-------
2. Actions for Emission Control
In order to promote the control of hydrocarbon emissions from
stationary sources on the basis of the basic direction discussed
above, the following requests will be made to quarters concerned
on improvement to be achieved through changes in equipment and
raw materials, for which Tables *J- and S serve as guidelines, and
other necessary emission control measures, and on the collection
of data.
(1) To local governments; 1 to exercise emission control
guidance on factories and other business establishments
discharging hydrocarbons in line with the report; 2 to
consider the introduction of regulatory measures or strengthen
those already in place in accordance with local conditions;
3 to check on the degree of improvement brought about by
emission controls, including measurement surveys; 4 to put
in order data on principal emission sources and report them
to the Environment Agency; 5 to adopt on a preferential
basis low-pollution materials for the use of public enter-
prises over which local authorities have jurisdiction and
give consideration to the timing of coating; 6 to strive
to build a monitoring network on the environmental concentra-
tions of non-methane hydrocarbons.
(2) To industrial organizations which have to do with factories
and other business establishments discharging hydrocarbons;
1 to advance emission control measures in the related
industries in line with the report; 2 to carry out research
and development and work for the introduction of emission
controls where their implementation is technically difficult
now; 3 to cooperate in the periodic surveys the Environment
Agency conducts to check on progress on the amount of emissions
and emission control measures.
PROCEEDINGS—PAGE 12
-------
(3) To the related central government departments: to adopt
positively low-pollution materials for public enterprises
under their jurisdiction and give attention to the timing
of coating. They will also be asked to cooperate by
exercising guidance on industries which have closely to
do with their operations and handle hydrocarbons and by
helping them with remedial actions.
PROCEEDINGS—PAGE 13
-------
Table 1 Change in the concentration of non-methane
hydrocarbons in the environment
(Simple mean values of 23 monitoring stations
measured continuously from 1976 to 1980)
Fiscal year
Annual mean
(ppmC)
1976
0.51
1977
0.53
1978.
0.45
1979
0.40
1980
0.44
Table 2 Emission ratio of hydrocarbons between years
(estimated)
1978 fiscal year/
1983 fiscal year
Mobile sources
Stationary sources
0.61
0.88
(Remarks) The figures for mobile sources were obtained
for motor vehicles in Tokyo and three other
prefectures (Chiba, Saitama, and Kanagawa)
(hydrocarbons), and the figures for stationary
sources were obtained over the whole country.
The ratio of "mobile sources" to "stationary
sources" in fiscal 1978 was estimated as 1 : 2
in Tokyo and the three prefectures.
PROCEEDINGS—PAGE 14
-------
Table 3 Hydrocarbons Emissions by Principal Stationary Sources
(Unit: ton)
Emission sources
Oil
Petro-
chemicals
Solvents
Coating
Printing
Other
Solvents
Plants
Refineries
and storage
facilities
Oil tanks
and storage
facilities
Gas stations
Plants
Storage
facilities
Paint
production
MO) Automobiles
(- M
• H CO .
w
-------
Table ^j-. Emission Control Measures by Facilities
Industrie* Discharging HC
Emission Control Measures
1. Storage facilities
Petroleum and coal products/Chemical
and allied products/Food and tobacco/
Electricity and gas/Mining/Wholeiele
and retail trade/Transport
(II Floating root tanks or inner floating roof tanks should be installed, with
light colors, such as white and silver white. Or treatment apparatuses employ-
ing the condensation or absorption method or the like should be installed.
(21 Particularly in the case of new large-scale facilities, tanks with floating
roofs or inner floating roofs should be installed.
(3) Where vehicles like tank cars and tank trucks are used for shipment.
vapor return equipment or treatment apparatuses should be installed. Fur-
ther studies are needed whether measures for loading into ships should be
taken or not.
(4) Vapor return equipment is applicable to facilities in certain conditions,
such as underground storage facilities at gas stations.
15) Such other measures as the cooling of material in storage and the use of
breather valves should be taken.
2. Transport facilities
(ships, tank trucks,
drum cans, and the
facilities)
Petroleum and coal products/Chemical
and allied products/Mining/Wtiolesale
and retail trade/Transport
A close link should be established with storage facilities or treatment
apparatuses through the use of vapor return equipment and the like.
. As for measures for ships, further studies are needed because there are
such problems as ascertaining the anti-pressure resistance of the hull and the
levels of oil in the holds.
3. Coating facilities
(including drying
facilities)
Lumber and wood products/Furniture
and fi xtures/Rubber products/Leather
tanning and leather products/Ceramic.
stone and day products/SteeVNonfer-
rous metals and products/Fabricated
metal products/General machinery/
Electrical machinery, equipment and
supplies/Transportation equipment/
Precision instruments and machinery/
Ordrujnce/Miscel laneous manufacturing
industries/Services/Construction
(1) Less pollutive coatings should be used for outdoor coating or similar
instances in which large machines and the like are coated indoors.
(2) In other instances of indoor coating, less pollutive coatings should be
used or treatment apparatuses employing the adsorption, catalytic incinera-
tion, or thermal incineration method or the like should be installed.
<3) Coating efficiency should be raised through improvement in the coating
methods, coating processes and the like.
4. Printing facilities
(including drying
facilities)
Publishing, printing and allied indus-
tries/ Pulp, paper, and paper products/
Lumber and wood products/Textile
mill products/ Ceramic, itone and clay
products/Fabricated metal products/
Nonferrous metal s and products/Mis-
cellaneous manufacturing inductries
(1) Less pollutive printing ink should be used.
(2) Treatment apparatuses employing the adsorption or catalytic incinera-
tion method or the like should be installed in the printing processes for
metal plate printing and the like.
5. Degreasing
facilities
Textile mill products/Furniture and
fixtures/Nonferrous metals and prod-
ucts/ Fabricated metal products/Gen-
eral machinery/Electrical machinery.
equipment and supplies/Transportation
equipment/Precision instruments and
machinery/Ordnance/Miscellaneous
manufacturing industries/ Services
(1) Treatment apparatuses employing the adsorption or condensation
method or the like should be installed.
(2) Alkaline, emulsive or low-volatile solvents should be used in the
degreasing process.
6. Adhesive coating
facilities
(including drying
facilities)
Textile mill products/Lumber and
wood products/Pulp, paper, and paper
products/Rubber products/Leather
tanning and leather products/Ceramic.
stone and clay products/Steel/Nonfar-
rous metals and products/Electrical
machinery, equipment and supplies/
Miscellaneous manufacturing indus-
tries/Construction
(11 Less pollutive solvent adhesive* should be used.
(2) Treatment apparatuses employing the adsorption or catalytic incinera-
tion method or the like should be installed.
7. Others
Hydrocarbons are also conceivably emitted from other sources, such as
dry distillation, mixing, kneading, and extraction facilities, although the
realities of emissions are not necessarily clear. To cope with this problem,
the realities of emissions should first be cleared up and then appropriate
ones should be selected from among the control technologies that are listed
in Chapter 3 of the report from the panel. This procedure will probably be
enough. In some cases, treatment apparatuses are already in place.
(by the Japan Industrial Standards)
PROCEEDINGS—PAGE 16
-------
Table 5 Emission Control Measures by Industries
^Facilities, Coun-
termeasures
Industries
Facilities
Discharging HC
Emission Control Measures
1. Petroleum industry
Refineries
Oil storage yard
Gas stations
Storage facilities
Transportation
facilities
Production facilities
Storage facilities
Transportation
facilities
Underground storage
facilities
(1) Floating roof tanks or inner floating roof tanks should be installed, with light
colors used for the tanks, such as white and silver white. Or treatment apparatuses
employing the condensation or absorption method or the like should be installed.
(2) Particularly in the case of new facilities, tanks with floating roofs or inner floating
roofs should be installed.
(1) Treatment apparatuses employing the condensation or absorption method or the
like should be installed.
(1) Strict maintenance checks and management.
(1) Floating roof tanks or inner floating roof tanks should be installed, with light
colors, such as white and silver white. Or treatment apparatuses employing the con-
densation or absorption method or the like should be installed.
(1) Treatment apparatuses employing the condensation or absorption method or the
like should be installed.
(1) Vapor return equipment or treatment apparatuses employing the condensation or
adsorption method or the like should be installed. Or both of them should be installed.
2. Chemical industry
Production facilities
Storage facilities
(1) Treatment apparatuses employing the thermal incineration or condensation method
or the like should be installed. Or hydrocarbons should be recovered within the
relevant processes.
(2) Strict maintenance checks and management.
(1) Floating roof tanks or inner floating roof tanks should be installed, with light
colors, such as white and silver white. Or treatment apparatuses employing the con-
densation or absorption method or the like should be installed.
3. Automobile
industry
Oegreasing facilities
Coating facilities
(including drying
facilities)
(1) Treatment apparatuses employing the activated charcoal adsorption method or
the like should be installed.
(1) Less pollutive coatings should be used.
(2) Coating eff iciecny should be raised.
(3) Treatment apparatuses employing the catalytic incineration or thermal incinera-
tion method or the like should be installed.
4. Shipbuilding
industry
Coating facilities
(1) Less pollutive coatings should be used.
(2) Coating efficiency should be raised.
5. Construction
industry
Coating facilities
Adhesive coating
facilities
(1) Less pollutive coatings should be used.
(2) Coating efficiency should be raised.
(1) Less pollutive solvent adhesive* should be used.
6. Printing industry
Printing facilities
(including drying
facilities)
(11 Less pollutive printing ink should be used.
(2) Treatment apparatuses employing the adsorption or catalytic incineration method
or the like should be installed.
7. Deaning business
Cleaning facilities
(including drying
facilities)
(1) Treatment apparatuses employing the adsorption or absorption method should be
installed.
8. Manufacture of
rubber products
Production facilities
(tires, rubber and
other footwear.
rubber belts, and the
like)
(1) Manufacturers should switch from solvent-containing paste to non-solvent gruel-
like paste, from solvent-type coatings to low pollutive coatings, and from solvent-
type model-removing agents to water-based agents.
(2) Treatment apparatuses employing the adsorption, catalytic incineration or thermal
incineration method or the like should be installed.
(3) Measures should be taken to raise the viscocity of rubber, seal facilities and con-
tainers, and improve operating methods.
PROCEEDINGS—PAGE 17
-------
c
•H
e
a)
u
§
•O
a; -a
1 S
3 0]
C M
•H
(days)
200
150
100
50
days on which warnings
were issued.
repoete
. of
erers.
(persons)
5,000
t-i ra
o -a i-,
V V
fH -P »H
<0 h V
l&fc
Is s
1977
'78
'79
'80
'8l
Fig. 1 Changes in the total number of days on which
warnings were issued and in the number of
reported sufferes (1977 to 1981)
PROCEEDINGS—PAGE 18
-------
ATMOSPHERIC REACTION MECHANISMS FOR
PHOTOCHEMICAL OZONE/OXIDANTS
presented by B. Dimitriades
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 19
-------
Current U.S. research in the area of atmospheric reaction mechanisms for
photochemical-03/0X formation is focussed on the following specific problems
or information gaps:
(1) Large uncertainties of predictions of current mechanistic models
for 63 yields from single-day photochemical HC/NOX systems.
(2) Lack of validated detailed mechanisms for use with regional 63
air quality models (RAQSM).
(3) Lack of validated, detailed mechanistic models capable of
predicting other-than-03 oxidants.
These problems/gaps are discussed, next, in some detail.
(1) Uncertainties in Current Mechanistic Model Predictions.
There are several mechanisms currently in existence in the U.S. explaining
the atmospheric formation of OyOx. These mechanisms include among
others the EKMA, Carbon Bond, Demerjian, and Cal Tech mechanisms, and
differ in several respects as shown in Table 1. Each of the four mechanisms
was developed and validated independently, and until recently we had not
intercompared them through parallel testing against a single set of data.
Such an intercomparison was done recently using one set of smog chamber
data and one set of field data, and results were presented and discussed
in a workshop on EKMA, conducted on December 15-16, 1981, at Research
Triangle Park, North Carolina (Proceedings Volume I has been sent to
Japan; Volume II is forthcoming.).
The smog chamber data, from irradiated auto exhaust mixtures of constant
HC composition, showed no large differences among the four mechanisms in
their ability to predict smog chamber 03 yields from given HC and NOX
concentrations. When the mechanisms, however, were coupled to a simple
dispersion model and were used to predict ambient 63 concentrations *rom
emissions and meteorology input data, the predictions disagreed with
observations substantially. Furthermore, the disagreement differed from
mechanism to mechanism. Also, when the mechanisms were used, through
isopleth diagrams, to compute HC control requirements for NAAQS-03
achievement, the results differed considerably from mechanism to mechanism.
The conclusions from these mechanism intercomparison studies were (a) that
most or all of the current mechanisms have substantial inaccuracies, and
(b) that the existing smog chamber and real atmosphere data are not
sufficiently diverse to permit a more informative evaluation of the
various mechanisms.
PROCEEDINGS—PAGE 21
-------
Mechanistic inaccuracies can be either in the description of the chemical
pathways or in the kinetic data used. Pathway errors occur either because
pathways are not known in detail (e.g. photooxidation of aromatic HCs) or
because they are used in an excessively condensed form (e.g. in lumped
mechanisms). In Table 1, the four mechanisms shown consist each of 36 to
76 reaction steps. Considering that in actuality there are hundreds of
operative steps, one may suspect that one cause of error in these four
mechanisms is the lack of sufficient detail, particularly, with respect
to the aromatic HC chemistry. It should be stressed here, however, that
increasing the number of steps is not without penalty; it raises the
computer capability demands of the model, a problem which can be prohibi-
tive in the case of AQSMs.
Errors associated with the kinetic data used also may be significant.
Table 2 shows the rate constant values used in the inorganic chemistry
sections of the four mechanisms. It can be seen that only in four (out
of 25 in all) cases the same value is used in the four mechanisms; in 20
cases, the values vary by factors of 1.1 to 6.3. A comparable situation
exists in connection with the organic chemistry sections of the mechanisms.
Our viewpoint on this problem is that several, different-type research
efforts need to be done before we can solve or minimize this problem.
First, additional mechanism intercomparison studies should be conducted
that would hopefully enable us to understand why the various mechanisms
show these differences in behavior. It may be crucially important, for
example, to have an in-depth understanding of the inaccuracies introduced
by the different types of "lumping" in the lumped mechanisms versus the
inaccuracies of the "surrogate" mechanisms. In these studies, it would
be useful to include as many as possible different mechanisms, and it is
for this reason that we would like to invite Japan to participate in the
effort by conducting a study to compare the Akimoto/Carter mechanism (for
propylene) with the other mechanisms in existence. In view of the much
greater number of reaction steps that compose the Akimoto/Carter mechanism,
inclusion of this mechanism in the mechanism intercomparison effort is
more than justified.
A mechanism intercomparison effort has already been initiated in the U.S.
and involves six mechanisms: Carbon Bond III (SAI), ELSTAR (ERT), LIRAQ,
Photochemical Box Model mechanism (Demerjian), and EKMA (Dodge). We
visualize a complementary effort to be conducted by Japan that will
include one or more of the above mechanisms as well as those of Akimoto/
Carter, of Derwent (England), and possibly others. If this is of interest
to the Japanese delegation, we would be happy to arrange for communications
between the Japanese and U.S. experts on the details of this cooperative
effort.
Additional efforts are needed, in our viewpoint, (a) to scrutinize existing
kinetic data and develop a standard set of such data to be used in all
mechanisms, and (b) to develop standard sets of smog chamber data against
which to evaluate all mechanisms. With respect to the latter need,
again, we would invite Japan to discuss with us consideration of
Dr. Akimoto1s smog chamber data as one of the requisite standard sets of
data.
PROCEEDINGS—PAGE 22
-------
(2) Lack of Validated Detailed Mechanisms for Use with Regional Air Quality
Models (RAQSM).
Relative to the mechanisms used with the urban models, the regional model
mechanisms must have two additional features: the ability to treat
terpene chemistry, and the ability to treat multi-day chemistry. In
response to this need there are laboratory and modeling studies on-going
within ESRL addressed to the chemistry of terpenes, and also smog chamber
and modeling studies addressed to multi-day chemistry. In the latter
studies, the most difficult problem is caused by chamber wall-related
interferences (e.g. adsorption/description of N02, HN03, PAN, ^65,
HCHO, etc.). Besides the mechanism of the multi-day chemical process, we
need also to elucidate the mechanism by which "aged" pollutant mixtures
affect the atmospheric chemistry of fresh pollutants. First indications
are that such effects, to a large extent, can be explained in terms of
effects of the aldehydes present in "aged" pollutant mixtures.
(3) Mechanistic Models for Non-Ozone Oxidants.
Recently, there has been an increasing concern in the U.S. about oxidants
other than 03, namely: N02, PAN, HNOs, aldehydes, organic nitrates, and
H202- While these species are products of the same atmospheric chemical
process that produces 03, it is not correct to assume that the same
mechanistic models that predict Oo can also predict (equally well) these
other species. Development of models specific for those species requires
an abundance of laboratory and field data on formation and fate of these
species, as in the case of 0^. Acquisition of these data requires, in
turn, availability of analytical methods for these species. ESRL is now
conducting smog chamber and modeling studies for those oxidant species,
and is also attempting to improve the analytical methodology required,
especially for ambient HN03, particulate nitrates, and
PROCEEDINGS—PAGE 23
-------
Table 1,
Some characteristics of existing mechanisms
No. of reactions:
(Inorganic):
(Organic):
Ekma
76
(23)
(53)
Carbon
Bond-III
75
(19)
(56)
Demerjian
37
(18)
(19)
Cal Tech
52
(25)
(27)
Representative organic
species used:
Fixed
Propylene
Butane
Formald.
Acetald.
Propionald.
Butyrald.
Variable
Ethylene
01ef. Bond
Para. Carbon
Arom. Carbon
Variable
C3-01efin
Cg-Paraffin
Cp-Aromatic
C-Aldehyde
Variable
Ethylene
Olefin
Aromatic
Paraffin
Formald.
Aldehyde
PROCEEDINGS—PAGE 24
-------
Reaction
1 «>2 - NO * 0
Z ' • * (02) » («) » 03
3 10 + 0, ~ «2 * 02
•« «0+9 -W2
5 W2 * 03 - "03 » Oj
t no2*o-w) + o2
7 IK>2 * 0 - »03
B OH « 03 * H02 * Q2
9 H02 * Oj * OH » 202
10 W « *02 • HNOj
11 OH * M - HOND
12 '1C » HO * (02) - N02 *
0,
13 W * CO - H02 + C02
14 MC + NC, - N02 + N02
15 nnŁ + H03 + H20 - 2HNO.
16 SC * HC2 - N02 + OH
17 «0 * N03 » N205
18 N205 . N02 + NOj
19 *205 * HŁ0 - 2HN03
20 HOj + NO + H20 - ZHONO
21 2HONO - NO + N02 (+H20
22 *)02 + H02 » H202 + 02
23 «-0- + hv - 20H
4. 4,
24 HOHO » OH.+ NO
25 «02 » H02 » HN02 + 02
26 »02 * H02 » HH04
27 W * HN02 * H02*-*»0
23 *04 -. H02 * IW2
29 03 + hv - Of3*)
3D 03 * O'D
31 c'o *'5' 0
32 010 + H20 - OH * OH
32 Total number of Sups
Carbon-Bond III
*1
•4.40 x 106
26.6
0.048
1.3 x 10*
100
2.40
1.60 x 104
9770
1.50 x 10"4
440
2. BO x 104
Heterogen
1.20 x 104
1.50 x 104
(=0.06 k,)C
•
(:10-3*,)"
4.44 K 1010
3.« x JO5
19
Dodg.
k,
4.2 x 106
25.0
0.045
1.3 x 10*
—
B4.0
2.4
0.8 x 104
3,000
1.3 x 10*
— —
1.2 x 103
5.6 x 103
22.0
5 x ID"2
2 x 10'5 (?)
1 x ID"3
D.84 x 104
0.0024 x k1
0.38 x k,
0.054 x t,
2.7 K 10"3
8.7 x 10*
19, V?
Berorjiar
*,
4.0 x 106
22.0
0.055
•
87.0
4.0
1.5 x 104
12,000
440
1.1 » 10*
68
CaJ. Tech T~
"l
4.4 x 106
23.5
3.9 x 103
0.047
1.3 x 10*
9
3.6 x 103
82.3
1.5
1.5 x 104
17,400
440
2.7 x 10
— ~-
1.2 x 104 i 1.2 x 10
1.0 x 104
0.19 ^
4.3
2.3 x 103
9.6 x 103
9.0
T\
— -
— —
Z3 J 18 '
3.9 x 10
6.9
29 x 10"
0.37 x 10'
0.0026k,
0.16 k,
1.7
1.7 x 103
9.6 x 103
4.4
0.072 k1
—
*
^ 1
Variation, ~
iqhest/loxest
--
1.1
i 1.2
--
1.2
1.0
•'
'1.2
9.7
2.0 .
5.8
—
1.0
2.5
~ ""
1.0
1.4
3.2
5.8
—
-
4.1
1.1
6.3
2.5
1.4
1.0
3.-.0
1.3
2.7
500.001
5.6
PROCEEDINGS—PAGE 25
-------
SMOG CHAMBER STUDY OF PHOTOCHEMICAL OZONE
FORMATION: REACTIVITIES OF HYDROCARBON-NOX
MIXTURES AND SAMPLED AMBIENT AIR
presented by H. Akimoto
National Institute for Environmental Studies
Japan EA
PROCEEDINGS—PAGE 27
-------
Smog Chamber Study of Photochemical Ozone Formation:
Reactivities of Hydrocarbon-NO Mixtures and Sampled Ambient Air
Hajime Akimoto and Fumio Sakamaki
Division of Atmospheric Environment
The National Institute for Environmental Studies
P.O. Tsukuba-gakuen, Ibaraki 305 Japan
Introduction
Quantitative charactrization of photochemical ozone forma-
tion in organics-NO -air mixtures is of critical importance in
X
planning ozone control strategy based on smog chamber data and
computer simulation. Particularly, when one challenges the
computer modeling of the photochemical processes of the ambient
air, "photochemical reactivity" of the polluted atmosphere
has to be defined by some means in -a quantitative manner.
It should be noted here that one can not a priori postulate
that the photochemical reactivity of the ambient air can be
predicted properly from the "known" pollutant composition at hand.
Since polluted ambient air contains numerous organic and inorganic
compounds all of 'which are not necessarily analyzed by conventional
analytical techniques,some unknown factor might affect the
reactivity of such systems. Thus, it is very interesting and
important to see if the photochemical ozone formation rate
PROCEEDINGS—PAGE 29
-------
in the ambient air can be predicted from the conventionally
analyzed hydrocarbon composition assuming the additivity of
the reactivity. By this reason, a new method for analyzing
the photochemical ozone formation rate in organics-NO -air
mixture has been developed and the reactivity of the sampled
ambient air is evaluated quantitatively in the present study.
As a rate parameter which could be representative of ozone
formation, NO oxidation rate has been recognized as a useful
measure of the photochemical reactivity of various hydrocarbons
(1). More recently, Darnall et al (2) has proposed an OH-hydrocar-
bon reaction rate constant as a measure of hydrocarbon reactivity.
While these parameters have been used as scales for the classifi-
cation of numerous hydrocarbons based on the reactivity, no
direct general relationship between these parameters and the
actual ozone formation rate observed in a photochemical run
of a selected organics-NO mixture has been proposed. The present
J\
study concerns with the method of analyzing the ozone formation
rate in organics-NO -air system and proposes a phenomenological
rate parameter which is useful for representing reactivity
of mixtures of organics. The analysis was first made on the
propylene-NO -air system for which both the systematic smog
J^
chamber data and a detailed reaction model for computer simulation
are available. Based on the analysis a phenomenological
rate parameter, "effective ozone formation rate constant" will
be proposed. Smog chamber runs for various single hydrocarbons as
well as hydrocarbon mixtures were next carried out, and the
effective ozone formation rate constants of various hydrocarbons
were determined and the additivity rule of the rate constants
was established. Finally, the smog chamber runs of sampled
PROCEEDINGS—PAGE 30
-------
ambient air were carrried out and the reactivity of the ambient
air was evaluated applying th.e method of analysis proposed
here.
Experimental
Experiments were carried out at 30°C using an evacuble and
bakable photochemical smog chamber (6m ) at NIES (3). The
light source consists of 19 of 1 kw high pressure Xe are lamps
with approapriate Pyrex filters. Experimental procedures for
the runs with single hydrocarbons and synthetic mixtures were
the same as reported previously (3,4). For the runs with the
sampled ambient air, a large air sampler which is made of two
plastic bellow-type bags (about 3.5 m each when inflated)
housed in a container was used. The air was sampled at 8:00-8:30
a.m. at a nearly city, Tsuchiura, carried to the Institute,
and introduced into the evacuated chamber.
Hydrocarbons and oxygen containing compounds were analyzed
by GC's after concentration. The DNPH-GC method was also used for
the analysis of formaldehyde and acetaldehyde. Carbon monoxide
was analyzed by a long-path FTIR (L = 221.5m).
Computer simulation were performed for the C,Hg-NO -dry
air runs using the detailed reaction model reported before
(5).
Results and Discussion
(1) C3Hg-NOx-dry air
Analysis was first made for propylene runs using two reaction
PROCEEDINGS—PAGE 31
-------
parameters, maximum ozone formation rate, (d[O,]/dt) , and
J H13.X
maximum OH concentration, [OH~I . Here the experimental maximum
lucLX
ozone formation rate (d[o^]/dt ) „ is defined as the maximum
J ulclX
slope of the plot of [0.,] vs. irradiation time for each run
and the experimental maximum OH concentration, [OH] , was
TOcLX
obtained from the decay rate of propylene by assuming that
the reacting species for C3Hg are only OH and O3 (6). Fig.l
shows the plot of (d[O.,]dt) vs. [OH] for the runs with
j max max
various [NOx] and k][ (NO2 photolysis rate) but with the same
initial concentration of C3Hg ([C3Hg]0 = 0.50 ppm). The linear
plot implies that (d[O.,]/dt) is linearly proportional to
j nicix
[OH]max. The proportionality was also confirmed by computer
simulation as depicted in Fig.l. A proportionality between
(d[03]/dt)fc and [OH]fc within a single run was next checked
by computer simulation. As shown in Fig. 2, (d[O.-]/dt) is
in general proportional to [OH]fc until d[O3]/dt and [OH] reach
their maximums except during the very early stage of photooxida-
tion. After d[O3]/dt and [OH] reach their maximums, the ozone
formation rate decreases faster than the decrease of [OH].
Fig. 3 shows the plot of (d[O,]/dt) vs. [OH]m rc_Hjn
j max max j o u
for the runs with defferent [C3Hg]0 and k,. The plots of both
series of runs fall on a single linear line going through the
origin verifying an approximate relationship.
- "e3"6 CC3H6
C H
where kg 3 6 is the effective rate constant of photochemical
ozone formation. The value of k C3H6 is obtained to be 6.0
e
4 _^ _^
x 10 ppm min . The relationship was confirmed by the computer
simulation as shown in Fig.3- In order to check the validity
PROCEEDINGS—PAGE 32
-------
/-» tl
of k 36, the experimental and calculated ratios (d[O..]/dt)/
(COH]max,[C3H6]0) were plotted as a function of [C3H6]
in Fig. 4. Solid curve is drawn through the calculated data
points. However, since it was found by the computer modeling
that at low [C_H,]n/[NO ]. ratio, time for (d[O,]/dt) and [OH]
J b U X 0 J
to reach their maximum does not coincide but the latter is
delayed, the ratio of (dCC/dt) to [OH]tmax [C3H6]0 is
also shown in a dashed curve. Here [OH]. is the OH radical
umax
concentration at the time when d[O,]/dt reaches the maximum.
From the above results, the effective ozone formation rate
C H
constant k 3 6 is found to be defined as an apparent constant
in the hydrocarbon excess region. More detailed discussion
will be found in our seperate paper (7).
(2) Single Hydrocarbon-NO -humid air
X
The effective ozone formation rate constants for various
hydrocarbons are next determined experimentally in the NO -humid
air mixture. Series of smog chamber experiments for five olefins,
twelve paraffins and nine aromatics have been carried out. As
anexample, the experimental results for toluene are depicted
in Figs . 5 and 6. As shown in Fig. 5, for the runs with constant
[NO ]Q and k. , ozone formation rate is proportional to [Toluene]Q
while the apparent first order decay rate of toluene which
Tol
is equivalent to k_u '[OH] is constant being independent
OH .IU3.X
of [Toluene] under our experimental conditions. This verifies
that the relationship similar to Eg. (I)/
can hold for the hydrocarbon-NO -air system in general and
X,
the specific k value for each hydrocarbon can be determined.
PROCEEDINGS—PAGE 33
-------
Fig.6 confirms that the similar dependence of kg on [HC]Q/[NO ]_
is seen for toluene as in the case of proplene shown in Fig.4.
Table I summarizes the experimentally determined k values
for various hydrocarbons studied. For comparison, literature
values (8) of the absolute rate constants of OH-hydrocarbon
reactions, kQH, are also cited in Table I. Fig. 7 shows the
correlation of experimentally determined k values with the
k values. It is interesting to note that the k values correlate
linearly with k^ and has the same order of magnitude as k__.
Un Uii
As demonstrated in Fig.7, the k values fall between the values
of kQH and 2k_H in most of the cases. Rationalization of this
finding has been discussed in our separate paper (7).
(3) Synthetic Hydrocarbon Mixture-NO -humid air
X
In order to confirm the additivity rule of the k value,
synthetic hydrocarbon mixture-NO -humid air runs were next
carried out. Fig. 8 shows the time profile of C,Hfi-toluene mixture
runs. In this series of runs, initial concentration of toluene
was varied keeping the initial concentration of C3Hg (0.5 ppm)
and NO (0.09 ppm) to be constant. Fig. 9 represents the maximum
Jt
ozone formation rate obtained from Fig. 8 and the maximum OH
radical concentration calculated from the decay of toluene .
From these data it is now possible to compare the experimental
k value obtained from
(d[03]/dt)max " * <[C3H6V[Toluene]0) COH] (II)
with the calculated k value,
koalc CC3H6]0^3H6 + [Tduene],,*^"*
[Toluene]0
PROCEEDINGS—PAGE 34
-------
where the values of k C3H6 and k Toluene are available in Table I
e e
4 4 -1 —1
as 4.64 x 10 and 1.12 x 10 ppra min , respectively. The
comparison of k and k is depiot/ed in Fig.10 demonstrating
a good agreement.
The additivity rule of the k value was further confirmed
by using the hydrocarbon mixture sample consisting of 23 components
simulating the ambient air sample which will be discussed below.
(4) Sampled Ambient Air
Finally, smog chamber runs for sampled ambient air have
been carried out and the data were treated according to the
above method of analysis. Table II shows the initial concentra-
tions of the organics and NO for the sample studied. In this
Jt
case, the effective ozone formation rate constant on C. base, k ,
defined by
= 'eXp "
(d[0 ]/dt) = k'[OH] [NMOGCppmC)] (IV)
o civ e civ u
was determined experimentally for each run. Here, the average
OH concentration was obtained from the decay of hydrocarbons and
M
LNMOGJg is the initial non-methane organics concentration
in ppmC unit as determined by a so called "non-methane hydrocarbon
monitor". The obtained value of k1 exp can be compared with
the average ozone formation rate constants calculated by
k =
E[OG(ppmC)]0
using the values of k obtained before for indivisual hydrocarbons
(Table I). For oxygen-containing compounds for which k values
have not been determined experimentally, k = 1.5 k „ was assumed
PROCEEDINGS—PAGE 35
-------
tentatively.
Table III summarizes the experimentally determined (d[0_]/
dt)av, [OH]av, k'eexp as well as calculated k'e according to
Eq.(V). Average was taken for 0-1 hour after the start of
irradiation since the maximum ozone formation rate was always
observed during this period for these runs. In table III,
k' (NMHC) is the value calculated using only the non-methane
hydrocarbons neglecting all oxygen-containing compounds and
CO, while k1 (NMOG) is the value calculated taking into account
all organics and CO analyzed in the present study. The results
showed that the sampled ambient air has-substantially higher
reactivity than expected from the additivity of k values for
hydrocarbons only, but the effective ozone formation rate constant
of the ampled ambient air can well be represented when oxygen-
containing compounds and CO, are taken into account. Among
the oxygen- containig compounds, aldehydes were found to contribute
predominantly thus demonstrating the importance of their analysis
in the ambient air for the prediction of photochemical air
pollution.
Fig. 11 shows the plot of (d[03]/dt)av vs. [OH]av,[NMOG]-M
The experimental points fall on a single straight line implying
that the Eq.(IV) holds for these systems and the effective
ozone formation rate constant does not differ too much for
these samples. The average ozone formation rate constant of
the particular ambient air investigated in this study was deter-
3 —1 —1
mined from the slope of the line as 5.42 x 10 ppmC min
PROCEEDINGS—PAGE 36
-------
Summary
1. In hydrocarbon excess region, the relationship
can hold in general and the phenomenological second order
at co sta t ke' ma^ ^e calle<^ fc^e effective ozone formation
rate constant.
2. The k values for various hydrocarbons of atmospheric importance
have been determined and their additivity rule was confirmed.
3. The photochemical reactivity of the ambient air as expressed by
the ozone formation rate constant was determined to be 5.4
x 10 ppmC min for the particular samples studied.
The reactivity was predicted successfully from the k values of
analyzed organics constituents including CO? and the substantial
importance of aldehydes to account for the contribution
to the ambient air reactivity has been noted.
PROCEEDINGS—PAGE 37
-------
References
(1) W.A. Glasson, C.S. Tuesday, Environ. Sci. Technol-, Ł,
37 (1970) .
(2) K.R. Darnall, A.C. Lloyd, A.M. Winer, J.N. Pitts, Environ. Sci,
Technol., ]jO, 693 (1976).
(3) H. Akimoto, F- Sakamaki, M. Hoshino. G. Inoue, M. Okuda,
Environ. Sci. Technol. 13, 53 (1979).
(4) F. Sakamaki, H. Akimoto, M. Okuda, Environ. Sci. Technol.,
Ijj, 665 (1981).
(5) F. Sakamaki, M. Okuda, H. Akimoto, Environ. Sci. Technol.,
1Ł, 45 (1982).
(6) H. Akimoto, F. Sakamaki, G. Inoue, M. Okuda, Environ. Sci.
Technol., 14.' 93 (1980).
(7) H. Akimoto, F. Sakamaki, "Correlation of the Ozone Formation
Rates with Hydroxyl Radical Concentrations in the Propene-
Nitrogen Oxide Dry Air System: Effective Ozone Formatin
Rate Constant, accepted for publication in Environ. Sci.
Technol.
(8) R. Atkinson, K.R. Darnall, A.C. .Lloyd, A.M. Winer, J.,N.
Pitts Jr., Adv. Photochem., 11, 375-488 (1979).
PROCEEDINGS—PAGE 38
-------
Table I Effective Ozone-Formation Rate Constants,
ke ;of Hydrocarbons as Compared with KOH
No.-
•o-i
0-2
0-3
0-4
0-5
P-l
P-2
P-3
P-4
P-5
P-6
P-7
P-8
P-9
P-10
P-ll
-P-12
A-l
A-2
A- 3
-A- 4
A- 5
A- 6
A- 7
A- 8
A-9
Hydrocarbon
Ethylene
Propyletie
•1-Butene
Isobutene
l-Pent=ne
n-Butsne
Isobutaire
n-Pentane
Isopentane
n-Hexane
2-Methylpent_ane
3-Methylpentane
2 , 2-Di_3et:hylbutane
2 , 3-Diaethylbutane
n-Hepteue
2-Methylhexane
3-Methylhexane
Benzene
Toluene
Ethylbeazene
o-Xylene
m-Xylene
p-Xylece
1,2,4 -Tri^ie thy Iben zene
1,3, 5-Tri_3ethylbenzene
p-Ethyltoluene
kef+rJxlO"3 kOHxld"3 (a
—1 —1 - —1 —1
(ppm min ) (ppm min • )
19+4
40+10
43+8
75
53+13
6.4+2.1
5.9
7 . 9+1 . 6
7.4
9.9+2.7
27
18
6.2
11
-5.8+1.4
15
20
2.0
IT. 2+1. 4
13
34
49+3
25
86
160
20
14.8
37.1
52.2
75.0
43.1
4.03
3.73
5.54
4.6
8.6
7.4
10.1
2.9
6.4
9.3
9.1
9.1
1.78
9.5
11.1
21.2
34.9
22.6
59.2
92.4
18.1
(a) R. Atkinson, K.R. Darnall, A.C. Lloyd, A.M. Winer
and J.N. Pitts, Jr., Adv. Photochem. 11, 375 (1979)
PROCEEDINGS—PAGE 39
-------
n
M
3
W
O
M
Tablell Initial Concentrations of the Constituents of the Sampled Ambient Air
Run
9
10
11
12
13
Data
1981
3/16
3/18
3/23
7/27
7/29
[NMOG]??
ppmC
0.26
0.40
0.44
0.37
0.40
[NOX]Q
ppm
0.032
0.052
0.063
0.055
0.032
CO
' ppm
0.79
1.19
1.33
0.42
0.43
R.H.
%
6
8
27
50
60
Hydrocarbon
Paraf .
44.7
43.8
39.6
37.6
34.3
Olef .
20.2
18.5
21.8
14.4
11.7
Composition
Arom.
18.1
20.3
19.5
37.6
46.8
Acet
14.5
15.1
17.3
9.0
6.0
(mole %)
Unknown
2.6
2.4
1.8
1.3
1.2
Z [HC]Q
ppb
43.1
66.6
80.3
25.6
40.7
-Continued
Run
9
10
11
12
13
Initial Concentrations
HCHO
11
14
13
39
24
CH3CHO
12.0
11.8
14.6
24.3
11.0
CH3OH C
5.6
8.3
4.3
26.9
17.3
2H50
3.1
6.9
1.7
8.8
3.4
(ppb v/v)
H CH3COCH3
4.8
4.9
4.3
45.6
11.8
1 [NMOG] 0
ppmC
0.20
0.32
0.37
0.31
0.28
E-[NMOG]Q
f [NMOGlo
0.82
0;80
0.84
0.84
0.70
Table III Experimental Results of the Sampled Ambient Air Runs
Run
9
10
11
12
13
(d[03J/dt)fXP
ppm min'1
2.60x 10~4
2.80
4.98
8.50
7.10
[OH]av
10~7 ppm i
2.09xlO"7
1.37
2.03
3.66
3.56
keexp
DPmC-lmin-l
4.8xl03
5.1
5.6
6.6
5.0
ke(NMHC)
ppmC-lmin-1
3.2 x 103
3.1
3.3
2.6
2.4
ke (NMOG)
ppmC^min"1
6.5x 103
5.6
5.8
6.2
5.0
-------
1.0 13 2.0
IOHjmox (ID"7 ppm)
Fig.l
plot of (dCo3l/dt)max vs.
OH) . Filled and open
symbols are for observed and
calculated values, respective-
ly. CC3Hg30=0.50 ppm.
Variable CNOv70 runs (A.,A),
k.j=0.16 min
-f
variable k,
runs (O,O)CNOx3Q=0.09 ppm.
'c
e 3
Q.
O.
Fig.2
Plot of (d[03J/dt) vs. [OH]t
in. three computer simulation
runs. Filled and open sympols
correspond to values before and
after the d[O3]/dt reaches' the
maximum. [C,Hg]0=0.5 ppm
[NO ] k!
A \J
A A 0.045 ppm 0.16 min
• O 0.090 0.16
• D 0.083 0.25
-1
0.5 to
(OH]t dO"7 ppm)
-i r
•Ł<
72 4
j
ffc'
az Q.I. as
10
HO"7 ppm2)
Fig. 3
Plot of
VS
~IIIU^t *S w - w
open symbols are for observed
and calculated values, respec-
tively. VariableCp3H6)0 runs
for t,NOx)0=0.04 ppm G3,D)
[NO ")Q=0.09 ppm (A,A),
k1=0.16 min'1, Variable k][
runs for t.C3HgJQ=0.50
[NOx)Q=0.09 ppm «
PROCEEDINGS—PAGE 41
-------
'j:
i7
Q.
>e 4
§ 3
•^ 1
O*
T?
0
6 10 12 K 16
Fig. 4
Plot of (d[03]/dt)max/[OH]max.[C3H6]0
vs. [C3H5]0/[NOx
dry air system.
in the C3H6~NOx~
Dashed line shows the ratio of
(d[03]/dt)nax/([OHltIIiax.[C3H6]0)
Filled symbols are for the experimental
and open symbols are for the calculated
;data points.
•jc
E
_
E
CL
Q.
OS
7
Fig.5.
Maximum ozone Formation rate (O) and
maximum hydrocarbon decay rate (A ) vs.
[Toluene]-, in the toluene-NO.,-humid
u ^
air system. [N0x]n = 0.09 ppm, k, =0.26
l
(ppm)
•c
I
a
a.
-j
•o
A
A
D
5 10
I Toluene JQ/!NOxl0
15
Fig.6.
Experimental values of ke as a function
of [Toluene]0/[NOx]Q in the toluene-NO -
Humid air system.
O : [Toluene]0-varied run
Q : [NOx]Q-varied run
A : k,-varied run
PROCEEDINGS—PAGE 42
-------
QJ
-X
en
o
/
/
./
/
OTO
/"*/
and k
OH'
Fig.7
Correlation of
i—
line A; ke=2kQH/B;
Oolefins, <> paraffins
A arromatics
Numbers correspond to those
in Table I.
Fig.8
Time profile of ozone formation
in the C3Hg-toluene-NOx-humid air
runs.
[C3Hg]0=0.5 ppm, [NOX]Q=0.09 ppm,
k^O.19 min .
Numerals in the figure are
[Toluene]Q in ppm.
Time
(hr)
Fig. 9
Maximum ozone formation rate and
maximum OH concentration vs.
(Toluene] 0 in the C-jHg-toluene-NO -
humid air runs.
Experimental conditions are the same
as in Fig.8.
(Toluene^
PROCEEDINGS—PAGE 43
-------
a. i
o
J'
01
je
,/
^
o/
2 3
pprfr-miri'')
Fig.10
Comparison of the observed and calculated
overall effective ozone formation rate
-constant in the C-,Hc-toluene-NO -humid
jo x
air runs.
[OH]av[NMOG]
M
Fig.11
(d[03]/dt)av vs.
in the sampled ambient air runs,
Average was taken for 0-1 hour
after irradiation. Numerals in
the figure correspond to run
numbers in Table II and III.
[OHyNMOGJM (107ppm.ppmC)
PROCEEDINGS—PAGE 44
-------
C Ref. ]
Correlation of the Ozone Formation Rates with Hydroxyl Radical
Concentrations in the Propylene-Nitrogen Oxide Dry Air System:
Effective Ozone Formation Rate Constant
Hajime Akimoto* and Fumio Sakamaki
Division of Atmospheric Environment
The National Institute for Environmental Studies
P.O. Tsukuba-gakuen, Ibaraki 305 Japan
PROCEEDINGS—PAGE 45
-------
Ozone formation rates obtained in smog chamber experi-
ments for the C_H,-NO -dry air system were analyzed with
j 0 X v
the aid of computer simulation using a detailed reaction
model. In the region of [C-jHg] 0/ fNOx^ 0~5' tne ozone formation
rate was found to be approximately proportional to the product
of the OH-radical concentration and the initial concentration
of C_H,. in the earlier stage of photooxidation until the
J o
d[0,]/dt reaches a maximum. The proportionality constant
was defined as an effective ozone formation rate constant
and is proposed to be a useful parameter to represent photo-
chemical reactivity of hydrocarbon mixtures based on the
ozone formation rate. The effective ozone formation rate
constant for C..H,. in the dry air-NO mixture was determined
-3D X
4 -1 -1
to be 6.2 + 1.1 x 10 ppm rain .
PROCEEDINGS—PAGE 47
-------
Quantitative characterization of photochemical ozone
formation in organics-nitrogen oxide-air mixtures is of criti-
cal importance in planning ozone control strategy based on
smog chamber data and computer simulation. From this view-
points, an "ozone formation potential" was proposed as one
of the generalized reaction parameter in our previous studies
(1-5). While the "ozone formation potential" governs the
maximum amount of ozone formed ultimately after prolonged
irradiation, the "ozone formation rate" is another even more
important parameter which characterizes the photochemical
ozone formation in organics-nitrogen oxide air mixtures.
The purpose of this study is to present a method for analyzing
the ozone formation rate and to propose a phenomenological
rate parameter which would be useful for representing reac-
tivity of mixtures of hydrocarbons and other organics whose
components are not necessarily known.
As for the rate parameters which could be representative
of ozone formation rate, the NO oxidation rate has been recog-
nized as a useful measure of the photochemical reactivity
of various hydrocarbons and most extensively studied by
Glasson and Tuesday (6). More recently, Darnall et al. (.7)
have proposed OH-hydrocarbon reaction rate constant as a
measure of hydrocarbon reactivity. However, although these
parameters has been used for classification of numerous
hydrocarbons based on a reactivity scale, no direct general
PROCEEDINGS—PAGE 48
-------
relationship between these parameters and the actual ozone
formation rate observed in a photochemical run of a selected
organics- nitrogen oxide mixture has been proposed on an
absolute rate basis.
This study presents an analysis of ozone formation rate
data for the propylene-nitrogen oxide-dry air system carried
out in an evacuable photochemical smog chamber. An approxi-
mately proportional relationship between the ozone formation
rate and the product of the maximum OH-radical concentration
and initial concentration of propylene was found to hold.
The relationship was further confirmed in terms of computer
simulation using a detailed kinetic reaction model.
Experimental and Computation
All the experimental data used in the present study
were obtained in C Hg-NO -dry air runs using the evacuable
and bakable photochemical smog chamber at NIES. Experimental
procedures and details of these runs have been reported previ-
ously (1). The experiments were carried out at 30°C. The
initial conditions and light intensity expressed as k1 (NO.,
photolysis rate) are given in Table I. Typical wall loss
rate of -0.04 ppm of ozone in the chamber was 0.07j^0.01 hr
The experimental maximum ozone formation rate, (d [O.J /dt)° 5
_j max
is defind as the maximum slope of the plot of ozone concentra-
tion vs. irradiation time. The experimental maximum OH-
PROCEEDINGS—PAGE 49
-------
radical concentration, [OH]° s was obtained from the maximum
lucLjC
slope of the semi-log plot of the decay of propylene after
subtracting the decay of propylene due to ozone reaction.
The rate constants used for the C-.H,-OH and C,H,-0., reactions
06 J o j
were 2.51 x 10~1:L(8) and 1.30 x 10~18(9) cm3molecule~1s~1,
respectively. The data reduction techniques to obtain the
maximum OH-radical concentration have been reported in our
earlier paper (10). The obtained tOH] for each run is
cited in Table I. The estimated error in tOH]max is +25%
for the runs with [C Hg]_=0.50 ppm. For lower tC3Hglg runs,
the error tends to get larger due to the larger scattering
error in the C.H,. concentration measurement and a larger
3 b
contribution of the correction term of C H,-O-, reaction.
JO J
Computer simulations were performed for the C3Hg-NOx-dry
air runs using the reaction model reported before (4). The
model consists of 158 chemical equations and 89 species,
and was used without modification. Wall loss rate of 0^,
0.058 hr"1 was used throughout the computation runs.
Calculated maximum ozone formation rate, (d[0-,l/dt) and
maximum OH concentration, [OH]Ł|XC were obtained directly
from the computer output of these values. The integration
program used was the same as described previously (4, 11).
The initial conditions of the computer runs were the
same as those of corresponding experimental runs. A few
supplemental computer runs were also performed under initial
conditions for which experimental runs were not available.
PROCEEDINGS—PAGE 50
-------
Results
Table I summarizes the observed and calculated maximum
ozone formation rates and maximum OH concentrations for all
the C-,H,.-NO -dry air runs studied in this work. The deviation
Jo x
of the calculated values from the observed ones is the largest
for the runs with high light intensity (Runs 23 and 24).
Although the agreement between the calculated and observed
values is thought to be satisfactory except for these runs,
it should be noted that the present reaction model predicts
much slower initial oxidation rate than observed experimental-
ly. For example, the simulation for Run 1 predicts the maxi-
mum ozone formation rate at 120 min, whereas the experimental
run gives the maximum rate at 80 min after the irradiation
began. Although this deviation would be due to the presence
of unknown radical sources as discussed by Carter et al.
(12, 13), this problem was not pursued further in the present
study since the following discussion on the relationship
between (dto-,]/dt) „ and tOHl av will not be affected by
.j max max
the delay of photooxidation.
Since it was generally found both experimentally and
by the computer simulation that the maximum ozone formation
rate and maximum OH concentration are observed at nearly
the same time after irradiation in .most of the runs it is
expected that the (d[0,]/dt) is correlated to [OH]
j max max
quantitatively. Figure 1 shows a plot of (d (O., ]/dt )max vs.
PROCEEDINGS—PAGE 51
-------
[OH] for the runs with various [NO ]_ and k, but with
max X U J.
the same initial concentration of propylene ([C^Hglg = 0.50
ppm). The linear plot implies that (d[O,]/dt) is linearly
o nictx
proportional to [OH] _ even though (d[O,]/dt) „ and [OHl „
nicLX j lucLX . lucLX
vary with the [NO ] _ and k,. The proportionality was also
confirmed by computer simulation as demonstrated in
Figure 1. Deviation from the linear line can be noted for
the run with a low [C,H,]_/[NO ]_ ratio. Thus, both the
o O U X U
experimental and calculated points for Run 15 are much lower
than the linear line. This is because the initial concentra-
tion ratio, [C-jHcln/tNO 1 n=1.7, for Run 15 does not fall
o O U X U
into the hydrocarbon excess region as will be discussed later.
A proportionality between (dto.,]/dt). and [OH], within
a single run was next checked by computer simulation. Figure
2 depicts examples for Runs 10, 11 and 25. (d[O3l/dt). is,
in general, proportional to [OH] until dtO^l/dt and [OH]
reach their maximums except during the very early stage of
photooxidation . After d[O.]/dt and [OH] reach their maximums,
the ozone formation rate decreases faster than the decrease
of [OH],
For the runs with different t^Hglg, a plot of (d[O3]/
dt) vs. [OH] •[C-.H.-]., was attempted. Figure 3 shows
max max 360
the plot for all the runs with different [G^lg, x 0
and k, . The plots of most of the runs except for those with
low [C-,H.-] _/[NO ]n ratio fall on a single linear line going
o o U X U
through the origin. Verifying an approximate proportionality
PROCEEDINGS—PAGE 52
-------
between (d [Ojl/dt )max and tOH]Mx.. K3H610, i.e.
(1,
max
where k is the effective rate constant of photochemical
ozone formation for propylene. From a least square fit to
C3H6
the all experimental points , the value of ke is obtained
4 -1 -1
to be 6.0 x 10 ppm mm as a slope of Fig. 3. The relation-
ship can be confirmed by the computer simulation as demon-
C3H6
strated in Figure 3 and the calculated k value agreed
with the experimental one within 10%.
C3H6
In order to check the validity of k , the experimental
and calculated ratios (d[0,]/dt) /( [OHl • tC^H.- ln ) for
j max max j b u
all runs were plotted as a function of [C_H,]n/tNO ] n in
J O U X U
Figure 4. Solid curve is drawn through the calculated data
points. However, since it was found by the computer modeling
that at low [C.H-]n/[NO ]„ ratio, time for (d[0-,]/dt) and
j b U X U , o
[OH], to reach their maximum does not coincide but the latter
is delayed, the ratio of (d[0,]/dt) to [OH]. •[C,H,-]n
3 max max " J b °
is also shown by a dashed curve. Here, [OH] is the OH
max
radical concentration at the time when d [O. ]/dt reaches the
maximum. The deviation of the two curves is apparent at
low [C3H,]0/[NO ]Q ratio- As shown in Figure 4, both ratios
decrease as the [C-,H^ ]„/ [NO ]n ratio decreases. Although
J D U X U
the experimental points scatters appreciably, they also tend
to decrease as the [C,Hg]Q/[NO IQ ratio decreses in accordance
with the prediction of the computer simulation. The region
PROCEEDINGS—PAGE 53
-------
where the ratio gives nearly a constant value, which can
C3H6
be defined as kŁ , can be called a hydrocarbon excess region
for the ozone formation rate. The hydrocarbon excess region
for the C3Hg-NOx-dry air system may be defined as [C3Hg]Q/
[N0xlgŁ5 from the plot of Fig. 4, if one takes the 90 % line
of the calculated limiting value. The average of experimental
data only in the region of [C-jHg] Q/ [NOx] Q55 gives the observed
k value of 6.2 j- 1.1 (2 ) ppm~ min .
Discussion
The importance of OH-radicals in the reaction of photo-
chemical air pollution is well recognized (14-21) and efforts
have been made to correlate OH-radical rate constants with
ozone formation both in experimental (23) and assessment
studies (24). However, no direct correlation between the
OH-radical concentration and ozone formation rate has been
verified experimentally. The results of the present study
shown in Figures 1-3 reveal that the ozone formation rate
and the OH-radical concentration can be correlated in the
hydrocarbon excess region by the equation,
C H
-). = ke3 6t°H]ftC3H6]0 (2)
C3H6
and an effective rate constant, k is defined as an
e •
apparent constant during the early stage of photooxidation
until the ozone formation rate reaches its maximum. Equation
PROCEEDINGS—PAGE 54
-------
(1) is a special case of Equation (2). Similarly, time-aver-
aged values, (d[O,]/dt) „ and [OHl r for a certain time
O a. V civ
interval may also be used to obtain the value of k when
these quantities changes slowly with irradiation time. It
is interesting to note that the ozone formation rate can
be expressed in terms of OH-radical concentration and hydro-
carbon concentration being independent of nitrogen oxides
concentration and light intensity. The latter two parameters
only affect the OH-radical concentration and do not appear
explicitly in Equation (2). Further, the phenomenological
i fi —d — i — i
effective rate constant, k =6.2 x 10 ppm~ min , has
the same order of magnitude as the OH-C-Hg elementary rate
C3H6 4 -I -I
constant, knu =3.64 x 10 ppm min (9).
Un
In order to check the validity of the proportionality
between the maximum ozone formation rate and the initial
concentration of C.Hg as represented by Eq.(l), numerical
evaluations of the terms contributing to the ozone formation
rate are made by means of the computer simulation. The ozone
formation rate and the NO decay rate in the photooxidation
processes can be written as,
dfO ]
-^T- = k2[0][02] - k3[NO][03] -ZL±(03) (3)
and
- ^p1= -k1[N02]+k3[NO] [03]+Zki[R02J [NO] +Skj [RO ] [NO] (-4)
where k,, k_ and k, are the rate constants of reactions,
N02+hv—» NO+0 (i), 0+O2+M -4 03+M (ii) and NO+O3 —> NO2+O2
PROCEEDINGS—PAGE 55
-------
(iii), respectively. The L.(O3) stands for every loss term
of O3 except the reaction (iii), RO_ and RO represent all
peroxy- and oxy- type radicals which can react with NO,
respectively. At the time when the maximum ozone formation
rate is observed, the inequalities, (d[03]/dt )» (d [N0]/dt )
and Ek.[RO] [NO]«k1[N02l , k^NOltOjl, and E^ [RO2 ] [NO ] hold,
which lead to the approximation, d[NO]/dt=0 and thus,
k1[N02]- k3[NO][03]=Eki[R02] [NO] (5)
Next the steady state approximations for [o] and [RO2 1 are
assumed
d[RO
k5)[03] -k2[0][02l=0 (6)
=Ł5i(R02) - Łki[R02] [N0]= 0 (7)
where k. and k- are the photolysis rate constants of 03/
O,+hv — ?• O_+O(3P) (iv) and 0,+hv — > O_-i-O( D) (v), respective-
J ^ J <«
ly, and S.(R02) stands for every source term of R02 . Oxygen
atom loss other than the reaction with O2, and RO2 loss due
to radical-radical reactions can be neglected at the earlier
stage of photooxidation and are not included in Eqs.(6) and
(7). From Eqs.(3)-(7), the ozone formation rate can be
expressed as,
d[0,]
) - 1^.(0) (8)
where Z'L.(O-) is the summation of net ozone loss terms ex-
cluding the terms of O- photolyses (reactions (iv) and (v) )
PROCEEDINGS—PAGE 56
-------
which mostly regenerate O.,. Further, each term of S.
and L. (O-,)" can be written as,.
ES:(RO ) = E S. (RO-)-t- E S. (RO7)+ E S.(RO-) (9)
1 ^ OH-CH,- RCHO 0-.-C..H,-
J ;O .. J- "J O .- . .
E'L (O ) = k [O ] FC H 1+k [NO ] [O ]+k [O 1 (10)
and all q.the.r;:terms are .found to be negligible. Here
E iS.. (ROrX and E S.(RO_) are the RO_ production
r\rr r1 13 i^i —/~* W
< K 1 1 6
rate in the QK+C^H,. reaction sequence and Q3-C.,Hg reaction
sequence, respectively. E S.(RO-,) is the RO_ production
RCHO
rate in -the-sequences ;of aldehydes (HCHO and CH3CHO) photoly-
sis>;,and 'OR-aldehydes reactions. Contributions of other
aldehydes have.been neglected. In Eq.(10), k-, k7, and k
;' - , O / Vr
are the. rate constant of the reactions O.+C,HC —> products
:- • - - J J D
(vi), NO2H"°3 ~^ .NO-HrO- (vii) and .0- wall decay.
Table II summarizes numerical values of the terms which
contributes; to ES.(RO_) and ^L^(O ) obtained in the computer
simulation ,for variable [-C.H,.3x. runs -( [NO • ] n=0.,04 ppm and
' 3 o u x u
k =0.16 min~ ) .; The ,given: values are for the time giving
the .maximum .rate ,of ozone formation for .each run. As shown
in :Table II,, ithe ozone formation rate calculated by Eq..(.8)
from each ±erms,j(dtO,J/dt)^i8), agrees well with (d [O ]
j UicxA J
/dtFalc. obtained directly in each computer runs within 10%
' IT13.X ;
for these runs. Figure 5 depicts the dependence of source
terms of RO2 on [C3Hv]Q and Figure 6 shows that of loss terms
of .0, -and (d[O, 1/dt) .obtained by the difference of ES..{RO,)
j j max -1- *•
and E'L-(O3)- Figure 6 also shows the concentration of OH
at the time of maximum ozone formation rate. It is seen in
PROCEEDINGS—PAGE 57
-------
Figure 5 that the RO. production rate due to OH-C..H,. and
e. JO
03-C_Hg reactions increases linearly with [C3Hg]Q while that
of aldehydes reaction is nearly constant being independent
of [C,H,_]-.. This is because the maximum ozone formation
J O U
rate appears at prolonged irradiation time as the [C3Hg]Q
decreases and the absolute concentration of aldehydes at
the time of maximum ozone formation are kept nearly constant
even at the lower [C-Hg]Q. The sum of all production rates
of R02 is approximately proportional to [C^Hglg as shown
in Fig.5. Figure 6 shows that a loss term of 0_ due to
0-.-C-H, reaction is a linear function of [C,HC] „ while that
J J O j a U
due to NO_-O3 reaction and wall loss of O« is nearly independ-
ent of [C,H,.]n and contribute only a minor extent except
jo u
at very low [C,H,]A. The maximum ozone formation rate result-
J D U
ant from the difference of ES.(RO-) and Z'L.(O3) is a linear
function of tC.,Hg]0 but has a small negative intercept as
shown in Fig.6. Thus, the approximate proportionality between
(d[0,]/dt) and [C-H,]^ as expressed by Eq.(l) and the
j max j o u
decrease of k value at the low ratio of [C3Hg]Q/[NO ]Q can
be rationalized by the analysis of computer simulation runs.
The reason that the effective ozone formation rate
constant is in the same order of magnitude as k..,, is also
UH
apparent from the analysis shown in Table II and Figs.5 and
6. The major term contributing to d[O3J/dt is the S(RO2>
due to OH-C,Ht reaction sequence which produces about twice
j o
as much RO2 radicals as reacted C3Hg. Contribution of all
PROCEEDINGS—PAGE 58
-------
other production and loss terms in Eq. (10) are a fraction
„
.-H,
3 6
of the term value of -„Ł „ S(RO0), yielding the k value
- (Jti-L.-H, 4 -1 e
between k-Ł and 2k_Ł .
(Jti (Jti
The above discussion suggests that the relationship
of Equation (2) may be applicable to any type of hydrocarbon-
N0x~air system in the hydrocarbon excess region during the
earlier stage of photooxidation
d[O.J
= ke [oHlt [HC]Q (ID
t
For a practical purpose the time averaged form of Eq. (11)
would be more useful, i.e.,
dfoj
The average should be taken for a time interval near the
maximum of d[0.,]/dt. Therefore, if the value of kg is
determined for each hydrocarbon and also for hydrocarbon
mixtures, it can offer a new useful scale of hydrocarbon
reactivity based on the ozone formation rate. It should
be noted that the k value determined in smog chamber
experiments is free from the error due to possible presence
of unknown radical sources in the chamber (10, 11), since
kg is defined by the ratio of the ozone formation rate to
the OH-radical concentration. Therefore, the quantitative
evaluation of the photochemical reactivity of ambient air
in terms of the ozone formation rate can be accomplished
by irradiating sampled air in a smog chamber. The determina-
PROCEEDINGS—PAGE 59
-------
tion of k for various types of hydrocarbons and for the
sampled ambient air is underway in our laboratory.
PROCEEDINGS—PAGE 60
-------
Literature Cited
(1) Akimoto, H., Sakamaki, F., Hoshino, M., Inoue, G.,
Okuda, M., Environ. Sci. Technol., 1979, 13, 53.
(2) Sakamaki, F., Akimoto, H., Okuda, M., Environ. Sci.
Technol., 1980, JL4, 985.
(3) Sakamaki, F., Akimoto, H., Okuda, M., Environ. Sci.
Technol., 1981, 15, 665.
(4) Sakamaki, F., Okuda, M., Akimoto, H., Environ. Sci.
Technol., 1982, JJ3, 45.
(5) Shibuya, K., Nagashima, T., Imai, S., Akimoto, H.,
Environ. Sci. Technol., 1981, 15, 661.
(6) Glasson, W.A., Tuesday, C.S., Environ. Sci. Technol.,
1970, 4, 37.
(7) Darnall, K.R., Lloyd, A.C., Winer, A.M., Pitts, J.N.,
Environ. Sci. Technol., 1976, 10, 693.
(8) Japer, S.M., Wu, C.H., Niki, H., J. Phys. Chem., 1974,
]8_, 2318.
(9) Atkinson, R., Pitts, J.N., Jr., J. Chem. Phys., 1975,
j[3, 3591.
(10) Akimoto, H., Sakamaki, F., Inoue, G., Okuda, M.,
Environ. Sci. Technol., 1980, 14, 93.
(11) Whitten, G.Z., Hogo, H., "Modeling of Simulated Photo-
chemical Smog with Kinetic Mechanisms Vol.2 CHEMK:
A Computer Modeling Scheme for Chemical Kinetics,
EPA-600/3-80-0286, Feburuary 1980.
PROCEEDINGS—PAGE 61
-------
(12) Carter, W.P.L., Lloyd, A.C., Sprung, J.L., Pitts,
J.N.Jr.,Int. J. Chem. Kinet., 1979, 11, 45.
(13) Carter, W.P.L., Atkinson, R., Winer, A.M., Pitts,
J.N.Jr., Int. J. Chem. Kinet., 1982, 14_, 813.
(14) Heicklen, J., Westberg, K., Cohen, N., "Chemical Reaction
in Urban Atmospheres", Tuesday, C.S., Ed., American
Elsevier Press, New York, 1971, p. 55.
(15) Niki, H., Daby, E.E., Weinstock, B., Adv. Chem. Ser.,
1972, 113, 16.
(16) Demerjian, K.L., Kerr, J.A., Calvert, J.G., Adv. Environ.
Sci. Technol., 1974, Ł, 1.
(17) Doyle, G.J., Lloyd, A.C., Darnall, K.R., Winer, A.M.,
Pitts, J.N.Jr., Environ. Sci. Technol., 1975, 9_' 237.
(18) Calvert, J.G., McQuigg, R.D., Int. J. Chem. Kinet.
Symp., 1975, 1, 113.
(19) Darnall, K.R., Lloyd, A.C., Winer, A.M., Pitts, J.N.Jr.,
Environ. Sci. Technol., 1976, 10, 692.
(20) Wu, C.H., Japar, S.M., Niki, H., J. Environ. Sci. Health,
Ser. A, 1976, 11, 191.
(21) Calvert, J.G., Environ. Sci. Technol., 1976, 10, 257.
(22) Wang, C.C., Davis, L.I.Jr., Wu, C.H., Japar, S.,
Niki, H., Weinstock, B., Science, 1975, 189, 797.
(23) Winer, A.A., Darnall, K.R., Atkinson, R., Pitts, J.N.Jr.,
Environ. Sci. Technol., 1979,-13_, 822.
(24) Singh, H.B., Martinez, J.R., Hendry, D.G., Jaffe, R.J.,
Johnson, W.V., Environ. Sci. Technol., 1981, 15, 113.
PROCEEDINGS—PAGE 62
-------
Figure Captions
Figure. 1. Plot of (d[ 0^] /dt )ntax vs. [OH]max. Filled and
open symbols are for observed and calculated
values, respectively. [C-H,]n =0.50 ppm.
J D U
Variable tNOx3n runs (*,A) , k. = 0.16 min ;
variable k, runs (9,Q) , [NO ]„ = 0.09 ppm.
Figure 2. Plot of calculated (d[O3]/dt)t vs. [OHJt in the
computer simulation for Runs 10 (A, A) , 11 (a,D)
and 25 (•,O) - Filled and open symbols correspond
to values before and after the d[O ]/dt reach the
maximum.
Figure 3. Plot of (d[O ,]/dt) v vs. [OH] • [C.,Hfi ]n. Filled
j nicLX max j o u
and open symbols are for observed and calculated
values, respectively. Variable [C^H^IQ runs for
[NOx]Q = 0.04 (1,0) and 0.09 ppm (4,0), kj_ = 0.16
min ; Variable [NOx]Q runs for [C-jHgJQ = 0.10
(A, A) and 0.50 ppm (T,V) ; Variable k, runs for
[C3H6]0 = 0>50' [N°x]0 ~ °-09 PPm (•'O).
Figure 4. Plot of (d [03 ]/dt )max/( [OH J^- [C^g] Q ) vs, [C3H6]
[NO ]Q. Symbols are the same as in Figure 1.
Dashed line shows the ratio of (d[O ]/dt) /
_j rti3X
([OH] • [C3H6]0) (see text) .
max
Figure 5. Production rate of RO. due to C3Hg-OH, C,Hg-0,
and RCHO-OH, hv reaction sequences, and the sum.
[NO ] = 0.04 ppm, k. = 0.16 min"1.
X \) JL
PROCEEDINGS—PAGE 63
-------
Figure 6. Loss rate of CU due to C-jHg-O.,, NO2-O, and wall
reaction, and the sum. (d[O^]/dt) calculated
j tllclX
from Eq.(8) and [OH] are also shown as a func-
max
tion of [C3Hgl0. fNOx^o ~ °-04
-1
PROCEEDINGS—PAGE 64
-------
Table 1 Experimental and Calculated Ozone Formation Rates and OH-Radical
Concentrations in the C3Hg-NOx-Dry Air System
Run
1
2
3
4
5
7
8
SI
S2
10
S3
14
15
16
6
17
18
19
20
S4
9
21
22
11
13
12
'Wo
ppm
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.50
0.50
0.50
0.50
0.50
0.50
O.C5
Q.10
0.15
0.20
0.30
0.40
0.50
0.10
0.20
0.33
0.50
0.50
0.50
ppm
0.009
0.020
0.026
0.034
0.036
0.052
0.063
0.010
0.020
0.045
0.150
0.187
0.291
0.038
0.043
0.039
0.040
0.039
0.039
0.040
0.086
0.086
0.091
0.090
0.090
0.089
kl
min
0.16a)
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
/d[03Mobs
V at /max
10~3ppm min"
0.12
0.30
0.33
0.37
0.41
0.38
0.35
—
—
2.31
—
3.40
2.48
0.13
0.35
0.71
1.00
1.40
1.74
—
0.19
0.78
2.10
2.97
2.59
2.50
/dio3]\calc
\ at /max
10 ppm min
0.38
0.43
0.42
0.41
0.39
0.38
0.33
0.84
1.46
2.37
2.78
2.58
1.84
0.16
0.38
0.64
0.92
1.45
1.90
2.22
0.30
0.77
1.73
2.86
2.65
2.60
10 ppm
0.56
0.79
0.73
0.91
0.78
0.95
—
—
0.84
—
1.11
1.25
0.42
0.88
0.71
—
0.69
0.70
0.99
0.96
0.97
0.74
—
[OH]CalC
max
10 ppm
0.64
0.80
0.83
0.87
0.86
0.90
0.91
0.28
0.48
0.80
1.10
1.10
0.96
1.07
0.88
0.86
0.89
0.86
0.80
0.75
0.96
0.88
0.99
1.06
1.01
1.00
23
24
25
26
27
0.50
0.50
0.50
0.50
0.50
0.085
0.090
0.083
0.088
0.089
0.37
0.31
0.25
0.19
0.13
7.27
6.25
4.38
3.50
2.82
5.26
4.60
3.94
3.22
2.47
2.50
1.97
1.40
1.23
0.80
1.91
1.69
1.42
1.18
0.91
(a) Typical uncertainty in k^ is 0.16 + 0.02 min
-1
PROCEEDINGS—PAGE 65
-------
Table II. Calculated R02 Formation Rates, 03 Loss Rates
at the time giving the Maximum Ozone Formation Rate(a).
[NOx]0 = °-09 PPm
Reaction Term \- C3Hg .(ppm)
RO, Formation
C,H,-OH Reaction Sequence
°2
CjHg+OH f CH_CH(O, )CH_OH
0 J 2 2
CH3CH(0)CH2OH — ^CH-jCHO+HCHO+HO
CHjCH (OH )CH2O — 2» CH3CHO+HCHO+HO
Z S.fRO,)
C3H6-OH *
Aldehyde Reaction Sequence
nrnn i on 2 . Itri ,,-,-, tll , ,
°5
CH CHO+OH — >• CH CO,+H O
O, 332
^H CO i PIT n i rn
°2
20, ' 2
HCHO+hV j TlO i CO
20, ' 2
CH3CHO+H ^— ^ CH302 + H02+CO
E S^RO )
RCHO
Ł3H6-03 Reaction Sequence
C3H6+O3 *" CH3CHO2+HCHO
»• CH-O,+CH,CHO
°2
CH,CHO, =— >• CH,O,+OH+CO
20, J 2
=-*• CH-O,+HO,+CO,
o2
=-* CH-O+CO+HO,
202 3 2
CH2°2 " C02+2H02
E Si(R02)
Z S.(RO )
all L 2
k^ » 0.16
0.
min"1
10
Rate (10~3ppm
0.
0.
2
2
0.
0.
0.
0.
0.
0.
. 0.
0.
0
0
0
0
0
0
0
0
099
053
095
051
30
038
058
023
046
012
014
22
.014
.014
.004
.005
.001
.001
.05
.56
0.
min
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
1
20
-1)
296
106
282
100
89
050
080
016
063
014
017
27
.046
.046
.012
.016
.003
.004
.15
.31
O, Loss Rate (10~ ppm
O3+C3Hg » products
03+N02 4- N03+02
O.j+wall »• O, Loss
Z'l^lOj)
o
"dT3" 5 J 'Si(Ro2)- -L^C
max all
d[o3JCalc
— 3t —
max
0
0
0
0
Formation
>3) 0.
0.
.056
.084
.060
.20
Rate
.36
.38
0.
0.
0.
0.
.182
.114
.073
,37
0.
0.
0.
0.
0.
1.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
1
.30
.496
159
.473
252
49
047
076
Oil
073
013
016
27
.069
.069
.019
.024
.005
.006
.22
.98
0
0
0
0.
0.
.40
.633
.267
.671
.336
1.99
0.044
0.073
0.008
0.087
0.
0.
0.
0
0
0
0
0
0
0
2
013
016
27
.094
.094
.025
.033
.006
.008
.30
.56
0
0.
0.
0.
.50
.825
.443
.7.83
0.417
2.47
0.045
0.
0.
0.
0.
0.
0.
P
0
0
0
0
0
0
3
074
007
106
014
017
26
.124
.124
.033
.043
.008
.011
.40
.12
min"1)
0
0
0
0
.278
.112
.066
.46
0
0
0
'o
.376
.114
.069
.56
0
0
0
0
.'496
.132
.066
.69
(10~ ppm min )
0.
0.
94
92
1.52
1.45
2
1
.00
.90
2
2
.43
.22
(a) Calculations for
= 0.10-0.50 ppm correspond to Runs 6, 18, 19, 20, and 21,
presented in Table 1, respectively so that [NO ]- value for each run varied slightly.
For the reactions which produces two RO2 radicals, the reaction rates were multiplied
by two to obtain ROj formation rates.
PROCEEDINGS — PAGE 66
-------
8
7
'c
E 6
Q.
ro
'O
x
O
E
5
4
0.5
[OH]
max
1.0 1.5
( 10"7 ppm )
2.0
2.5
Fig. 1
PROCEEDINGS—PAGE 67
-------
8
a
a
> .
'o 4
x
O
1 ~r——i r
*^00>V T
_j i
0
0.2
0.4
0.6 0.8
1.0
Pig. 3
I I
j I
CM
o
*~ Ł?
1<
o
co Z
ro
O
PROCEEDINGS—PAGE 68
-------
RCHO+OH
RCHOhP
0.5
Pig. 6
PROCEEDINGS—PAGE 69
-------
FURTHER DEVELOPMENT AND VALIDATION OF EKMA
presented by B. Dimitriades
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 71
-------
The discussion presented earlier on further development of atmospheric
reaction mechanisms applies, of course, to EKMA also. The discussion to
follow, therefore, will focus on the field validation component only of the
EKMA development effort in the U.S.
At first glance, it would appear that for a purely mechanistic scheme
such as EKMA, the only validation needed is against laboratory data. EKMA,
however, is used as a complete air quality predictive model, and as such,
according to an all-accepted rule, it should be validated against field data
also. Field validation of mechanistic models is extremely difficult because
of the: problem in separating out the meteorological influences. Nevertheless,
there; are some techniques that, in concept or partially at least, circumvent
this problem. These techniques, and some comments they received from the EKMA
Workshop experts are as follows:
(1) Trend Analysis of Historical Emissions and Air Quality Data (J. Trijonis)
Historical precursor trends are constructed for an urban area using
emissions data and ambient 6-9-am data for NMHC and NOX. The precursor
trends are entered into the standard EKMA model to predict historical
ozone trends. The predicted ozone trends are then compared to actual
ozone trends to test the EKMA model.
This method, while simple in concept, is complicated by the large uncer-
tainty in the precursor data. For example, emissions inventory figures
for the Los Angeles area during 1964-1978 show 29% decreases in reactive
hydrocarbons and 35% increases in oxides of nitrogen but corrections of
various inventorying errors over this same time period are at least of
this same order. Also, the model predictions are extremely sensitive to
the NMHC/NOX ratio and, hence, to the errors of the ratio estimates.
Results from one application of the method showed that the method under-
estimated observed ozone levels by as much as 35% for peak hourly ozone
levels at Azusa. The discrepancy was greatly alleviated by the use of
the more robust statistic, the 95th percentile ozone level. The disagree-
ment seemed to be comparable between the case of precursor-derived NMHC/NOX
ratios and that of emission-derived MHHC/NOX ratios. Although attempts
were made to outline the source areas the emissions of which had most
influence on each ozone monitoring station and to correct ozone formation
for meteorology, no account was made for day-to-day differences in initial
chemical composition, mixing depth composition at the upper boundary or
path differences over the emissions pattern.
One important conclusion from such studies was that the emissions changes
during relatively short periods (3-4 years) are not large enough to
provide an adequate test of the EKMA model. Also, the method requires an
accurate knowledge of the NMHC/NOX ratio as well as reliable emissions
and ambient air quality data. With improvements in these areas, the
trend analysis approach can be a useful one.
PROCEEDINGS—PAGE 73
-------
(2) Statistical Evaluation of EKMA (J. Martinez, C. Maxwell, H. S. Javitz,
and R. Bawol)
The method entails computation of maximum daily 03 from 6-9-am WHC and
NOX and statistical comparison of such predictions with daily maximum 63
observations. A central theme to this approach is the probability that
the ratio (R) of observed-to-estimated ozone lies within the range 0.8 to
1.2. The graphical representation of the output is represented as, accuracy
probability isopleths on the NMOC/NOX plane. The method was used'to
compare the Carbon Bond II, Demerjian and Dodge chemical mechanisms using
data from St. Louis, Houston, Philadelphia, Los Angeles, and Tulsa.
Results showed that qualitatively all three mechanisms give similar
shapes for the oyer-(R<0.8) and underprediction (R>1.2) regions, but the
Dodge mechanism is represented best of all of the mechanisms in the
0.8
-------
Also, the method suffers from the limitation that it does not use statis-
tical samples; its performance depends on selection of one worst-case
hour or day.
(5) Comparison of EKHA with AQSMs (6. Whitten)
This method for field-validating EKMA 1s an indirect one, based on compari-
son with other field-validated models e.g. the SAI airshed model. Results
from parallel applications of the EKMA and SAI models showed good'agreement
except when the NMHC/NOX ratio is defined differently (based on emissions
for SAI and on ambient concentrations for EKMA) and in situations in
which background pollutant concentrations are Important.
(6) Deriving EKMA Isopleths from Experimental Data: The Los Angeles Captive
Air Study (D. Grosjean, R. Countess, K. Fung, K. Ganesan, A. Lloyd and
F. Lurmann)
The method involves sampling and Irradiation of urban air inside Teflon
bags and comparing resultant 03 concentration to those estimated by EKMA.
An application of the method is under way in LA, but results are not
available yet. The method 1s realistic but, unless shown to the contrary,
it suffers from the usual uncertainty problems associated with chamber
wall effects.
Discussion
All the above techniques for field-validating EKMA are being considered
by USEPA as candidates for subjects of future research. At this time we have
no final choices but we do favor those of the existing techniques (or new
ones) that provide statistical measures of the model's accuracy. Furthermore,
we are interested in the "Captive Air Experiment" concept but not for precisely
the same reasons as those of Grosjean and co-workers. Since our reasons may
be of interest to Japan also, I wish to elaborate further on this concept and
to invite the Japanese delegation to consider this as another subject for
cooperative research with the U.S.
The captive air experiments we would propose would be for the purpose of
testing the common technique of simulating urban atmospheres using smog chambers
and synthetic pollutant mixtures. The main question posed here is whether
such simulations are indeed realistic or useful. More specifically, the
proposed experiments would be for the purpose of comparing the photochemical
behavior of real atmosphere pollutant mixtures with that of synthetic mixtures
presumably representative of the ambient mixtures. Briefly, these experiments
would entail parallel outdoor irradiation of bags containing ambient (urban)
air and bags containing synthesized "ambient" VOC/NOX mixtures. If the two
sets of bag mixtures display substantial differences in photochemical behavior,
we would have to reconsider all AQSM techniques that assume no such differences.
PROCEEDINGS—PAGE 75
-------
METHODS FOR VALIDATING EKMA AGAINST FIELD DATA
1, TREND ANALYSIS OF HISTORICAL EMISSIONS AND AIR QUALITY
DATA
2, STATISTICAL EVALUATION
3, COMPARISON OF OZONE FREQUENCY DISTRIBUTIONS
-4, SIMPLIFIED TRAJECTORY ANALYSIS-APPROACH
5. COMPARISON WITH AQSMS
6, COMPARISON WITH SMOG CHAMBER DATA FROM AMBIENT AIR
IRRADIATIONS
PROCEEDINGS—PAGE 76
-------
ACID RAIN (DEPOSITION) CHEMISTRY AND PHYSICS
presented by A.P. Altshuller
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 77
-------
Acid deposition often referred to as "acid rain" results from the combined
effects of wet scavenging by rain, snow, sleet, or hail, by fogs and dry depo-
sition of gases and particles. The substances usually considered of concern
are sulfur dioxide, sulfur aerosols, nitrogen dioxide and nitric acid and
hydrochloric acid.
The concern biologically is not only with the acidity of the deposited
substances, but with the total sulfur and nitrogen deposited. With respect to
the background pH level, this level often has been assumed in the past to be that
of natural rainwater containing only dissolved CO-. Such an aqueous solution has
a pH of 5.6. However, the pH of rain in remote areas has been measured below 5.6.
Charlson and Rodhe (1982) have discussed the concentration levels of sulfur com-
pounds, ammonia and carbon dioxide possible in cloud water in remote areas. They
conclude that pH values could range from 4.5 to 5.6 in remote locations. There-
fore, the pH values obtained in populated rural areas or cities should be evaluated
compared to this wider background pH range.
Natural and Anthropogenic Emissions for the United States
Estimates of the natural and anthropogenic emissions of sulfur compounds in
the United States are summarized in Table 1. In the eastern United States, the
—112 —1
natural emissions summing to about 0.2 Tg S yr (10 grams S yr~ ) are insigni-
ficant compared to the anthropogenic emissions of 12 to 13 Tg S yr (Robinson
and Homolya, 1982). The natural emissions appear to be more significant in the
western United States compared to anthropogenic sources. However, the method of
estimation is such as to possibly overestimate natural emissions -of sulfur from
the more arid and alkaline soils of the western United States.
For the eastern United States, the natural emissions of nitrogen compounds
based on two different methods of estimation (Robinson, 1982) sum to a lower
estimate of 0.24 Tg N yr or an upper estimate of 1.9 Tg N yr~ (Table 1).
The upper estimate of natural emissions of nitrogen compounds would constitute
about 18% of the sum of natural and anthropogenic emissions in 1978 for the
PROCEEDINGS—PAGE 79
-------
eastern United States. Natural emissions would have constituted a much larger
percentage of total emissions of nitrogen in earlier years because the anthro-
pogenic emissions of NO have increased rapidly with time. In the western
A
United States, natural emissions of nitrogen compounds may be substantial com-
pared to anthropogenic emissions. However, the range of uncertainty in esti-
mating natural emissions of nitrogen is so wide as to make the estimat.es of
natural emissions particularly in the western United States preliminary.
Coal-fired emission sources burning coals with substantial chlorine contents
are major contributors to the anthropogenic emissions of HC1 in the atmosphere
(Homolya, 1982). However, the natural emissions of chlorine appear to exceed
the anthropogenic emissions in the eastern United States (Table 1).
Sulfates or sulfur in particle form can arise either from primary emissions
of particle sulfur from fossil-fueled sources or be formed from sulfur dioxide
through atmospheric reactions. In some areas of the northeastern United States
primary particle sulfur is significant compared to secondary particle sulfur
particularly in the winter months (Homolya, 1982). These higher emissions of
particle sulfur are associated with oil-fired sources burning fuel oils under
operating conditions leading to higher percentages of particle sulfur than usual.
In parts of New England ant the mid-Atlantic states particle sulfur can constitute
5 to 10% of total sulfur emissions.
Emissions of sulfur oxides from coal-fired sources in the United States have
been high for the last century. However, there was a shift away from use of coal
after World War I for residential and commercial heating and for railroad travel.
Almost concurrently, there was a substantial growth in coal-fired power plants in
the midwestern and southeastern United States. As a result, anthropogenic
sulfur emissions today in the United States are emitted from many large sources
with tall stack heights in rural areas in addition to urban sulfur emissions.
PROCEEDINGS—PAGE 80
-------
Nitrogen oxide emissions in the United States increased about four-fold
between 1940 and 1978 from anthropogenic sources (Homolya, 1982). This in-
crease was associated with large increases in emissions from both highway
vehicles and from electric utilities. The relative contribution of highway
vehicles compared to electric utilities varies substantially from region to
region in the United States.
Air Quality Measurements
Sulfur oxide concentrations are highest in urban areas of the eastern
United States (Altshuller, 1982). However, the sulfur oxide concentrations have
decreased substantially because of reductions in the sulfur content of fuels
during the late 1960's and 1970's. Although sulfur dioxide is appreciably lower
in rural areas of the eastern United States than in urban areas, the differences
in concentrations between urban and rural areas has decreased.
Sulfate aerosols occur at substantial concentration levels compared to
sulfur dioxide both in rural and urban areas. These aerosols are mainly in the
submicron particle size range although lower concentrations of supermicron
sulfur particles can be measured. Sulfate concentrations have decreased in
eastern cities in the United States in the winter months, but not in the summer
months. In rural areas, sulfate concentrations have not decreased throughout
the year and they have increased in summer months. Sulfate areosols can contri-
bute one-third to one-half of the sulfur burden in rural areas 1n the summer.
The concentrations of nitrogen oxides are comparable to the concentration
of sulfur oxides in both urban and rural areas of the eastern United States
(Altshuller, 1982). However, the concentration of nitrogen in aerosol form
is substantially lower than that of sulfur in aerosol form in the eastern
United States.
Sulfate aerosols not only make a contribution to acid deposition in the
United States, but also to visibility reduction. Sulfate aerosols appear to
be the most important species contributing to visibility reduction in rural
areas in eastern North America.
PROCEEDINGS—PAGE 81
-------
Transport of Acid Compounds
The flow field in the planetary boundary layer (PEL) within which the
acid substances are emitted is affected by a broad spectrum of atmospheric
motions (Gillani, 1982). These factors include wind shear with height, strong
diurnal and seasonal effects, as well as complex flows in river valleys, on
mountain slopes, or on the shores of lakes or oceans.
Within the Midwestern United States prevailing winds, on the average are
from the southwestern quadrant in summer, but are more westerly in winter
(Gillani, 1982). The vertical transport layer from longer-range transport
varies from the ground up to 1 or 2 km in the summer and about half these
heights in the winter. Diurnal variability of the flow field is especially
large in the summer when a "nocturnal jet" with strong wind shear is a frequent
occurrence particularly in the midwestern United States. Plumes undergo sequences
of sheared stratification and distortion at night followed by rapid vertical
mixing during the day. As a result, emissions can be rapidly dispersed over
a regional scale.
The diurnal and seasonal variations in mixing height are important also
(Gillani, 1982). Because of substantially lower mixing heights in winter than
summer, a significant part of the emissions from tall stacks may remain elevated
and relatively undispersed to the surface over distances in excess of 500 km.
Such diurnal and seasonal aspects should have an important influence on the
atmospheric residence times, range of transport before deposition, of emissions
from elevated compared to near surface sources. For example, nitrogen oxide
emissions from tall stacks should have significantly longer residence times than
the nitrogen oxide emissions from highway vehicles. The emissions from highway
vehicles can undergo substantial dry deposition especially during the night-
time hours. Therefore, the relative contributions of various types of sources
to biologically sensitive areas downwind should vary substantially with season
of the year.
PROCEEDINGS—PAGE 82
-------
Transformation Processes
It is reasonably certain that OH radicals predominate in the gas-phase
reactions that transform sulfur dioxide to sulfate aerosols and nitrogen dioxide
to nitric-acid (Miller, 1982). The rate of transformation of nitrogen dioxide
to nitric acid is substantially faster than the rate of conversion of sulfur
dioxide to sulfate. The concentrations of OH radicals show large diurhal,
seasonal and geographical variations. The reactions associated with OH radicals
should be most significant during the daylight hours, warmer months of the year
and at lower latitudes. In polluted air, the concentration of OH is strongly
related to the concentrations of hydrocarbons, aldehydes and of nitrogen oxides.
The same sets of atmospheric reactions of importance in determing the formation of
ozone and other photochemical products are involved in determining the net OH
radical concentrations.
Various chemical reactions must be considered with respect to reaction in
cloud and rain drops (Hegg and Hobbs, 1982). However, available evidence indicates
that the primary source of nitric acid and hydrochloric acid in cloud and rain
drops are either the homogeneous atmospheric gas-phase reactions discussed above
or the direct emissions of hydrochloric acid. Production of sulfuric acid in
solution within raindrops or other hydrometeors can occur by several different
chemical reactions. The oxidation of sulfite by hydrogen peroxide appears to
be the single most important of these reactions producing sulfuric acid. However,
the amounts of hydrogen peroxide available in solution is not well characterized
at present.
Neutralization by ammonia absorption is an important process (Hegg and
Hoggs, 1982). However, the acidity of raindrops appears to remain higher than
the acidity of suspended particles in the atmosphere.
A wide range of rates of conversion of sulfur dioxide to sulfate have been
reported in urban plumes (0 to 32% hr"1) (Miller and Whitbeck, 1982). The rates
appear to be appreciable higher in urban plumes within or near cities than in
power plant plumes or in rural areas. Rates of conversion of nitrogen dixoide
PROCEEDINGS—PAGE 83
-------
ranging from 5 to 24% hr~ have been computed. Most of the products of reaction
of nitrogen dioxide in plumes are gaseous rather than aerosol species.
A wide range of rates of conversion of sulfur dioxide also have been
obtained from measurements within power plant plumes (Miller and Whitbeck,
1982) (0 to 15% hr ). However, the rates of conversion of sulfur dioxide
are moderate, 0 to 3% hr" , in a number of studies. The rates are dependent on
time of day, season of year and availability of "background" hydrocarbon con-
centrations to mix into plumes from ambient air. The latter factor may explain
the higher rates of conversion for sulfur dioxide frequently reported for plumes
in the eastern compared to the western United States. Where concurrent measure-
ments of sulfur dioxide and of nitrogen dioxide rates are made, the rates of
nitrogen dioxide conversion were several times faster than those for sulfur
dioxide conversion.
It is difficult to separate out the contributions of gas-phase and liquid
phase reactions in plumes (Gillani, 1982). However, based on measurements in
power plant plumes the following 24 hour average estimates of sulfur dioxide
conversion rates have been suggested: July, gas-phase 0.8 ^ 0.3% hr" , liquid
phase, 0.4 j^ 0.2% hr" ; January, gas phase, <0.1% hr" ; liquid phase, about 0.2%
hr~ . Similar parameter!zations are not yet available for nitrogen dioxide con-
versions in power plant plumes. Also, similar parameterization values have not
been suggested in urban plumes.
Precipitation Scavenging Processes
Precipitation scavenging may be defined as the composite process by which
airborne gases and particles attach to precipitation elements and subsequently
deposit on the Earth's surface (Hales, 1982). The processes involve many inter-
meshed pathways. Different storm types, differ in the significance of the pro-
cesses leading to deposition; nevertheless, there is a substantial overlap in
the characteristics of storm types. Important storm types include cyclonic or
"frontal" storm systems and convective storms.
PROCEEDINGS—PAGE 84
-------
The results of field studies on precipitation scavenging lead to the
following conclusions (Hales, 1982).
(1) Sulfur oxides and nitrogen oxides are removed with low efficiences
short distances downwind of power plant stacks.
(2) In urban plumes, moderate percentage removals of sulfur oxides
have been observed out to 100 km. Removal of nitrogen oxides from urban
plumes is less than sulfur oxides, but still significant.
(3) On a regional scale the removal of nitrogen oxides is greater in
proportion to its regional emission rate than the removal sulfur oxides.
Dry Deposition Processes
The rates of dry deposition from the air to surfaces depend on a large
number of chemical, physical and biological factors (Hicks, 1982). The relative
importance of these factors vary according to the nature of the surface, charac-
teristics of the chemical and the meteorological conditions. Despite such
complexities it is useful to consider a deposition velocity, Vj, for gases
and submicron particles which can be used along with airborne concentrations,
C, to estimate fluxes, F, by the expression F = Vd°C. For particles larger
than about 5 urn diameter deposition is primarily determined by Stokes' law,
while for submicron particles turbulent transfer tends to dominate. Deposition
of gases is often limited by diffusive properties close to the receptor surface.
Surface effects can be very important. Uptake rates of vegetation is often
determined by stomatal resistance. Surface wetness also is important on
biological and man-made surfaces.
For a specific surface such as a lake, air moving from land over .water
will equilibrate at a rate dependent on the stability regime involved. Henry's
law constants and chemical reactivity of substances are important to exchange
between air and water.
Wind tunnel experiments indicate very low deposition rates for sub-
micron particles. Over water, the role of waves can substantially increase
particle transfer by several possible processes.
PROCEEDINGS—PAGE 85
-------
Several micrometeorological measurement methods have been used to calculate
deposition velocities. These methods include eddy-correlation, concentration
gradient techniques, a "modified Bowen ratio" method and high-frequency variance
methods.
A substantial number of field investigations have been carried ouf to obtain
the deposition velocities for sulfur dioxide (Hicks, 1982). Deposition rates of
0.1 to 0.2 cm s" have been reported over snow. Measurements available over
pasture, wheat and soybeans result in deposition velocities usually ranging from
0.4 cm s" to 1.3 cm s . Most of these results are for daylight conditions.
One set of field experiments over a forested area (pine plantation) demonstrated
the wide variation possible in deposition velocities over the diurnal cycle.
Relatively few measurements are available for the deposition velocity of nitrogen
dioxide. One set of measurements over soybeans during daylight hours provided
values of Vd comparable to those of sulfur dioxide over field crops. The resistance
to deposition of nitric oxide, NO, is high. While no direct experiments are avail-
able for nitric acid a very low resistance to transfer is estimated with a sub-
stantial similarity to HF which has a high deposition velocity.
Deposition velocities for sulfate particles are greater than obtained for
other particles in the submicron range (Hicks, 1982). An explanation suggested
was the presence of sufficient amounts of larger sulfate particles to result in
higher deposition velocities. However, in recent dry deposition measurements sub-
micron particles were specifically considered in which sulfur dominated the
chemical composition. Deposition velocities of about 0.5 cm s were obtained.
Deposition Monitoring
Methods for monitoring dry deposition are inadequate (Hicks, 1982). Results
obtained from use of collection vessels and from surrogate surfaces are not easily
related to the actual behavior of natural surfaces with respect to deposition.
A wide variety of precipitation chemistry sampling networks have been
operated in the United States because of the ease of collecting rain and snow
(Stensland, 1982). Bulk samplers often have been used which do not permit
PROCEEDINGS—PAGE 86
-------
the separation of wet from dry deposition. Many networks in the past lacked
the spatial and temporal extent necessary to assessing national or even regional
trends and patterns. More recent networks in the United States and in Canada
use automatic samplers and have sufficient geographical extent to give useful
information over eastern North America. The highest concentrations of sulfate,
nitrate and hydrogen ions in precipitation occur in the midwestern United
States east of the Mississippi River, into the mid-Atlantic states and in parts
of southern Ontario and Quebec. These concentrations fall off the west, south
and east of this geographical area. The highest ammonium ion concentrations
in precipitation are in the midwest, west of the Mississippi River. The highest
calcium ion in precipitation are in approximately the same geographical area
as ammonium. Therefore, the depositions of alkaline species, ammonium and
calcium are displaced somewhat to the west of those for hydrogen ion, sulfate
and nitrate.
An increase of nitrate in precipitation in the United States since the
1950's can be demonstrated (Stensland, 1982). However, there are serious
questions as to the evidence to support a similar increase in acidity.
Air parcel trajectory analyses have been used to link precipitation
chemistry patterns to emission source regions (Stensland, 1982). While
general directional characteristics have been obtained in several investi-
gations, such techniques need to be further developed and verified with
field experiments.
Other aspects relating to deposition measurement townee direct impaction
of cloud water within mountains at cloud heights and fog water. Munger and
coworkers (1982) report at urban locations in California the pH of fog water to
be lower than rainwater (Liljestrand and Margan, 1978). The pH of a number of
samples of fog water was as low as the 2.0 to 3.0 range. This behavior in urban
areas appears different than in rural areas where the concentration of ions of
fog water has been reported as comparable to that in cloud and rainwater. The
composition of the fog water was such that nitrate ions exceeded sulfate ions
and ammonium ions usually exceeded hydrogen ions in concentration. The authors
concluded that "the chemistry of fog water in the samples they analyzed was
dominated by the composition of the haze-forming aerosol that proceeded it".
PROCEEDINGS—PAGE 87
-------
References
Charlson, R.J. and Rodhe, H. (1982) Factors controlling the acidity of
natural rainwater. Nature 295, 683-685.
Waldman, O.M., Hunger, J.W., Jacob, D.J., Flagan, R.C., Morgan, J.J. and
Hoffmann, M.R. (1982). Chemical composition of acid fog. Science 218.
677-680.
Liljestrand, H.M. and Morgan, J.J. (1978). Chemical Composition of Aoid
Precipitation in Pasadena, California. Environ. Sci. Techno!. 12, 1271-
1276.
From Critical Assessment Draft Document
The Acidic Deposition Phenomenon and its Effects Volume 1. Prepared for :
U.S. Environmental Protection Agency through North Carolina State Univ.
Acid Precipitation Program, October 1982.
Robinson, E., in Chapter 2 - Natural and Anthropogenic Emissions Sources
Homolya, J. in Chapter 2 - Natural and Anthropogenic Emissions Sources
Altshuller, A.P. in Chapter 3 - Atmospheric Concentrations and Distri-
butions of chemical Substances
Gillani, N. in Chapter 4 - Transport Processes
Miller, D in Chapter 5 - Transformation Processes
Hegg, D. and Hobbs, P. in Chapter 5 - Transformation Processes
Gillani, N. and Whitbeck, M.R. in Chapter 5 - Transformation Processes
Hales, J. in Chapter 6 - Precipitation Scavenging Processes
Hicks, B. in Chapter 7 - Dry Deposition Processes
Hicks, B. in Chapter 9 - Deposition Monitoring
Stensland, G. in Chapter 9 - Deposition Monitoring
PROCEEDINGS—PAGE 88
-------
Source
Anthropogenic (1978)
electric utility
Anthropogenic (1970)
electric utility
Natural Sources
biogenic onland
oceanic
volcanic
Anthropogenic (1978)
electric utility
Anthropogenic (1970)
electric utility
Natural Sources
NO biogenic onland
NO by lightening
A
NH2 biogenic onland
Anthropogenic (1974)
Natural Sources
Emissions, TgX yr
Element Contiguous U.S.
s
s
s
s
s
s
s
s
N
N
N
N
N
N
N
N
Cl
Cl
13
-
14
-
v5
Ł.23
.2
small
10.7
-
0
9.3
_
.84b, 6.9C
0.23, 2.5
0.01 - 0.1
0.6, 4.3
-
.
-1
U.S. East of Mississippi
River plus Texas
12
8
13
7.5
•v.,2
.07
.1
neg
8.9
2.9
7.9
2.4
.24b, 1.9C
0.04, 0.7
0.01
0.19, 1.2
>_.45
1.6
may be important in certain areas for short time periods
estimated from approach based on results of Adams and coworkers
See Robinson (1982)
estimated from approach based on results of Gal bally
See Robinson (1982)
PROCEEDINGS—PAGE 89
-------
STATUS OF RECEPTOR MODELS
presented by A.P. Altshuller
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 91
-------
Status of Receptor Models
Robert K. Stevens, Charles W. Lewis and A. P. Altshuller
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Introduction
The development of receptor-based models has progressed significantly in
the past few years (see Report of First Quail Roost Workshop, Watson et al.,
1981). Mathematical receptor models presently in use can be classified into
several categories: chemical mass balance (CMB), multivariate, microscopic,
and source-receptor hybrids. For the sake of brevity, we will only discuss
two types of receptor models, tracer (CMB) and factor analysis models
(multivariates) in detail. We also distinguish receptor models from
conventional dispersion modeling.
DISPERSION MODELS
Dispersion models are the conventional means of predicting the
environmental impact of an emission source on air quality. In general, the
dispersion model states that the contribution of source j to a receptor, S.,
is the product of an emission rate, E., and a dispersion factor, D., so that
J J
S. = D.E.
J J J
Dispersion formulae are usually classified by the geometric form of the
emission source, giving rise to expressions for point, line, and area sources.
A basic goal of air quality meteorology is to relate the dispersion factor for
a given source goemetry to known meteorological parameters such as wind speed
and direction and atmospheric stability. The available dispersion models have
been classified and described in various EPA reports* Current development of
dispersion models focuses on modeling dispersion in rough terrain, long-range
transport, wet and dry acidic deposition, and complex atmospheric chemistry.
RECEPTOR MODELS
In contrast to dispersion models, receptor models start with observed
ambient airborne particle concentrations /at a receptor ana seek to apportion
the observed concentrations between several source types based upon knowledge
of the composition of the source and receptor material. Receptor models and
their application have been reviewed by Gordon (1980).
The foundation of all receptor-based models of particulate source
assignment is a simple mass conservation argument. If a number of sources, p,
exists, and if there is no interaction between their aerosols that causes mass
removal or formation, the total airborne particulate mass measured at the
receptor, C, will be a linear sum of the contributions of the individual
sources S.:
P
C = IS. (1)
j=l J
*Dispersion models usually provide predictions involving a single pollutant such
as sulfur dioxide, carbon monoxide or ozone. The models are designed to satisfy
the emission, transport and transformation characteristics of the selected pollutant.
PROCEEDINGS—PAGE 93
-------
Similarly, the mass concentration of aerosol property i, C., will be
where a is the mass fraction of source contribution j possessing property i
at the receptor.
The Chemical Mass Balance Receptor Model
When the property i is a chemical property, Equation (2) represents a
chemical mass balance. If one measures n chemical properties of both source
and receptor, n equations of the form of Equation (2) exist. If the number of
source types contributing those properties is less than or equal to the number
of equations, that is, if p < n, then the source contributions S. can be
calculated by solving the overdetermined system of linear Equations (i) . Five
methods of performing this calculation have been applied: the tracer
property, linear programming, ordinary linear least-squares fitting,
effective variance least-squares fitting, and ridge regression. Effective
variance least-squares is the most widely used CMB method (Watson, 1979).
Tracer Method —
The tracer element method is the simplest and will be the only receptor
model discussed in detail. Tracer models assume that each aerosol source
type possesses a unique property that is common to no other source type.
Equation (2) then reduces to
S = * (3)
J ^T
for each source j with its own tracer t (using the same notation as in
Equation (2). It works well when the tracers meet the following requirements:
(1) a perceived at the receptor is well known and invariant between
source and receptor,
(2) C can be measured accurately and precisely in the ambient sample,
and
(3) the concentration of property t at the receptor comes only from
source type j .
These conditions cannot be completely met in practice, and limiting the
model to only one tracer property per source type means that valuable
information contained in the other aerosol properties is discarded. However,
in some instances, one can use a tracer approach as a first-approximation,
upper-limit estimate of source impact. Thus, solutions to the set of
Equations (2) have been developed to make use of the additional information
provided by more than one unique chemical property of a source type, and even
by properties not so unique. The other methods of solution must deal with the
case of a number of constraints, n, greater than the number of unknowns, p.
*examples of aerosol properties are chemical composition and particle size
distribution.
PROCEEDINGS—PAGE 94
-------
Multivariate Models
The CUB uses the aerosol property of chemical composition. If source and
ambient samples are taken in more than one size range, the size property can
be used to separate the contribution of one source from another. However, the
CMB solutions discussed above have no way to incorporate the variability of
ambient concentrations and source emissions. The multivariate methods can do
this. Multivariate methods include factor analysis, target transformation
factor analysis, extended Q-mode factor analysis, and multiple linear
regression, which is often used in conjunction with factor analysis. Only
factor analysis will be discussed here.
While the CMB receptor model is easily derivable from the source model
and the elements of its solution system are fairly easy to present,
multivariate receptor models are not as simple. Watson (1979) has carried
through the calculations of the source-receptor model relationship for the
correlation and principal component models.
The apparent mathematical complexity of these models does not remove the
requirement that every receptor model be representative of and derivable from
physical reality as represented by the source model. A statistical
relationship between the variability of one observable and another is
insufficient to define cause and effect unless this physical significance can
be established and is the same restriction as imposed on the CMB model.
The multivariate models (except the extended Q-mode model) deal with a
series of m measurements of aerosol component i during sampling period or at
sampling site k. From Equation (2),
P
C = I a S k=l...m (4)
1R -_i 1J JR
The multivariate models deal only with C., with the objective of predicting
the number of sources, p, and of predicting which a.. is associated with which
S. (or, more ambitiously, estimating a., and S.,).
J 1J JK
Factor Analysis --
Factor analysis, a special type of multivariate model, generally begins
with a cross-product matrix of the data; frequently, by calculating the
correlation coefficient:
rr . = 1 / /I / (5)
I* . V. . 7" . _
i j m-1 k=l
where m is the number of observations, C. and C. are the average values of C±
and C., and 0. and a. are the standard deviations of C. and C.. The
correlation coefficient is a measure of the extent that ambient concentrations
C. and C. vary in the same way.
PROCEEDINGS—PAGE 95
-------
The classical factor analysis model transforms Equation (2) to
C.f = Z a.. S.I + d. U..
lk j=i XJ Jk J 1J
where
and
(9)
where a (the "factor loading") is related to the source compositions through
Equation3 (9) and S s (the "factor score") is related to the source
contributions throughjtquation (8). The d. and U. . are unique factor loadings
and scores, respectively, and often left ont of tfei analysis. The result is a
principal component analysis:
s p s
c..S = I a..S..S (10)
»k j=1 !J Jk
The factor model treats the C., as components of a vector for each
chemical component i in an n-dimensional space. If Equation (4) is true, then
only p vectors, where p < n in a less complex p-dimensional space, are
required to produce the vectors of C. . This p-space is defined by the
eigenvectors of the correlation matrix of the C. . The p eigenvectors merely
define the space, however, and are not necessarily representative of the
sources, the S. vectors. They must be linearly combined to form the new
source vectorsr This is typically done by a procedure known as "VARIMAX"
rotation. The classical factor analysis can be used to screen a data set for
potential error identification as well as for source recognition.
MULTICOLLINEARITY IN SOURCE APPORTIONMENT
An important aspect of source apportionment concerns the handling of
sources with similar signatures, that is, sources whose aerosols are
chemically and physically similar. Large errors can result if source
apportionment by conventional regression or weighted regression analysis is
attempted when two or more sources with very similar signatures are included.
It is not uncommon for negative aerosol contributions with large magnitudes to
PROCEEDINGS—PAGE 96
-------
be estimated for some sources under these conditions. Another symptom is the
calculation of uncertainties that are larger than the calculated source
contributions themselves.
In statistical terms, the problem of similar souce signatures is one of
"multicollinearity." More generally, multicollinearity exists when one source
signature is nearly a linear combination of any subset of the other
signatures. Details of how one can perform source apportionment studies when
source signatures are similar are discussed in detail in the EPA Receptor
Model Validation Workshop Report (Quail Roost II).
Considerations in Design of Source Apportionment Studies: A group of the
participants headed by Dr. Glen Gordon, University of Maryland, examined and
evaluated the preliminary results presented at the Quail Roost II Workshop and
prepared reports that contained (1) an assessment of the value of the real and
simulated sets and (2) a series of recommendations which serve as a basis for
design of source apportionment studies which may be conducted by private or
governmental institutions. The section, prepared principally by Gordon with
input from Quail Roost II participants, will appear in an EPA report entitled
"Mathematical and Empirical Receptor Models: Quail Roost II". Some of the key
issues of this section dealt with the development of protocols to perform
different levels of source apportionment. For example, it was recommended
that source apportionment studies could be divided into three levels of
intensity based principally on resource availability.
Level I Minimal Resources Requirements. This effort uses existing source
signature libraries, existing ambient air quality data (mass and
elemental composition of fine and coarse particles), local emission
inventories, and literature values for soil composition. With this data
one can perform chemical mass balance and factor analysis to determine if
there are sources contributing to the receptor site which are not
accounted for in the emission inventory.
Level II Moderate Resources are available: This level of effort includes
all of the Level I work plus source elemental profiles from "grab"
samples of fugitive sources, soil and major primary particle emitters in
the airshed. These grab samples are resuspended and recollected in the
same manner and on the same filter material as used to collect particles
at the receptor site(s).
Level III Maximum Resources: This level III effort calls for (1)
collection of aerosols at receptor sites where maximum plume fumigation
may occur; (2) collection of particle emissions from the major emission
sources with a dilution sampler that is coupled with a sampler similar to
those used at the receptor sites; (3) both X-ray fluorescence (XRF), and
instrumental neutron activation analysis (INAA) are used to analyze
filters from the sources and receptor sites; (4) X-ray diffractron (XRD),
scanning electron microscopy (SEM) and optical microscopy are also used
to resolve sources with similar emission profiles.
Gordon included in his section a list (Table I) of aerosol sampling and
gas measurement methods that should be part of a complete receptor model
station. Figure 1 depicts the aerosol sampling analysis and data obtained for
a complete aerosol apportionment study. We note from this figure at least 4
PROCEEDINGS—PAGE 97
-------
different sampling procedures may be necessary to perform reliable source
apportionments. Gordon further recommended that multivariate and wind
trajectory receptor models be used whenever there is a need to obtain source
signatures not readily available by conventional source sampling procedures.
These models are capable of providing source signatures by their unique
treatment of air quality data. However, multivariate statistical models
require large data sets (40 or more observations) and wind trajectory methods
require short sampling periods (6 hours or less) and appropriate
meteorological measurements.
Recommendations for Future Research: Table II is a list of marker elements
used by the Quail Roost II participants in their various receptor models to
report the Houston emission sources contributing to ambient particles
impacting our Houston receptor site. The sulfate and nitrate listed in Table
II do not, in the absence of appropriate source signatures, have marker
elements which could reliably deduce their origins. At present, we can only
speculate as to the origin(s) of the sulfate and nitrate. Even analysis of
the Houston filters by SEM and optical microscopy failed to deduce their
origin. Parenthetically, ancillary analysis of the aerosol samples by SEM,
XRD and optical microscopy expands the number of sources beyond what can be
identified by CMB methods alone (see Figure 2). In most airsheds in the
United States sulfate and nitrate and their associated cations represent a
substantial portion of the particle mass below 10 pm. Procedures must be
developed to determine the origins of sulfate and nitrate that impact a given
receptor site if source apportionment methodology is to adequately address
urban and rural particulate problems.
At Quail Roost I and II there was no attempt by the participants to
incorporate within the receptor modeling framework procedures to deal with
questions of non-conserved species, particularly sulfates and nitrates. In
view of the dominance of these species in urban and rural aerosols, it seems
clear that this issue requires resolution.
Recent work by EPA scientists (1982) has been directed at addressing the
problem of source apportionment of the non-conserved species, namely secondary
products of SCL and NO. One solution they propose is the application of the
following general formula:
where:
T. = S. • A. • M (11)
T. = Mass Concentration of Secondary Particles (e.g., SO.) from
J Source J.
S. = Mass Concentration of Primary Fine Particles from Source j
J determined to impact at Receptor Site. This term is derived-
from CMB calculations.
A. = Ratio of SO to Fine Particle Emissions measured
at Source j.
M= Transformation Junction [e.g., rate of conversion
of SO to SOT enroute from Source j to Receptor
Site] Z *
PROCEEDINGS—PAGE 98
-------
Assumptions: 1. Transport mechanism for sulfate and fine particle
marker elements from Source J to Receptor Site
is nearly the same.
2. Back trajectory meteorological data available.
This general equation has several features which need further discussion.
For example, using the ratio of SO to fine particle emissions eliminates the
need to know absolute emission rates from source j. Also, the transformation
function which will be described in a future publication is decoupled from the
dilution of the primary emissions during transport. The only parameters we
need to know are the rate of conversion of the primary gas to secondary
particles, a relatively crude estimate for the additional deposition of SO
compared with the fine particles, and the time from source to receptor. The
time function can be obtained through back trajectory analysis. The longer
the distance from source to receptor the more uncertain will be the
apportionment. We have estimated that with this model it should be feasible
to estimate the impact of a coal or oil fired power plant on a receptor site
downwind as far away as 500 km under meteorological stability class D, with
the theoretical limitation being the dilution of the CMB marker species
reaching the receptor site to a concentration below the detection limit of
existing analytical methods.
We offer this novel approach to receptor measurements not as an ultimate
solution to handling the apportionment of the secondary particles but to
stimulate other investigators to examine this issue and put forth substantive
alternative models. Perhaps by the time we hold our Quail Roost III Receptor
Modeling Workshop, field experiments will have been conducted and new source
apportionment models will have been tested to resolve the sources of the
sulfate and nitrate particles.
REFERENCES
Gordon, G. (1980) "Receptor Models" Environ. Sci. Technol. 14:792.
Watson, J.G. Chemical Element Balance Receptor Model Methodology for
Assessing the Sources of Fine and Total Suspended Particulate Matter in
Portland, Oregon. Ph.D. Dissertation, Oregon Graduate Center,
Beaverton, Oregon. 1979.
Watson, J.G. Receptor Models Relating to Ambient Suspended Particulate
Matter to Sources. EPA-600/2-81-039, Research Triangle Park, NC. 1981.
Although the research described in this article has been funded wholly or in
part by the United States Environmental Protection Agency, it has not been
subjected to Agency review and therefore does not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
PROCEEDINGS—PAGE 99
-------
TABLE I.
TYPICAL SAMPLERS FOR COMPLETE RECEPTOR-MODEL STATION
Particulate Flow rate Filter
Sampler (1/min) Medium
Dichotomous 17 Teflon
Quartz
Species
Measured
Mass
Elements
Carbon
Method
0-gauging
XRF, INAA
Combustion
Ionic species Ion chromato-
(SO.2 , NO. , graphy/chemistry
etc?) J
Dichotomous 50
Hi-Vol 500-1000
(with cyclone)
2.5 |Jm cutpoint
Nuclepore
Teflon
Quartz
Individual
particles
Crystals,
minerals
Organic
compounds
SEM/XRF
XRD, LM
GC/MS.LLC, etc.
Gas sampler
Total sulfur
NO
Polyurethane
Gases measured
SO and other S-
containing gases
NO, N02
Organic vapors
Output
Real time, need data-logging
system
Real time, need data-logging
system
Real time, need data-logging
system
Later extraction and analysis
in laboratory
Meteorological
Wind Speed
Direction
Temperature
Humidity
Parameter
cm/sec
compass Heading
°C
Dew Point °C
Output
Data Logger-Real Time
Data Logger-Real Time
Data Logger-Real Time
Data Logger-Real Time
PROCEEDINGS—PAGE 100
-------
TABLE II. SOURCE CATEGORIES AND MARKER ELEMENTS FOR HOUSTON AEROSOL
Marker Elements
General Source
Categories
Specific Source
Categories
Na, Cl
Al, Si, K, Ca, Mn, Fe
Marine
Crustal
C, Br, Pb
Mn, Fe
Cu, Zn, As, Sm, Sb
Transportation
Steel
Nonferrous metals
Sulfate
Marine
Wind erosion of soil
Paved and unpaved roads
Construction
Limestone crushing and handling
Cement
Lime Kiln
Fly ash
Slag
Noncatalyst vehicles
Catalyst cars
Diesels
Iron and steel production
Steel finishing and handling
Carbonaceous
NO.,
Nitrate
Primary emissions in Houston
airshed
Conversion of SO in Houston
airshed
Regional background
Refineries
Botanical
Vegetative burning
Tire wear
Photochemical
PROCEEDINGS—PAGE 101
-------
F1g. 1
RECEPTOR MODELING
AMBIENT AIR SAMPLING PROTOCOL
SAMPLER NO. 1
<2.5 pm >2.5 pm
60 PAIRS:
AMBIENT PARTICLES
ON TEFLON
ANALYSIS:
MODELS:
XRF. MASS
NAA
XRD
CMB
MVA
FA
SAMPLER NO. 2
<2.5 pm >2.5 pm
60 PAIRS:
AMBIENT PARTICLES
ONNUCLEPORE.
0.4pm PORE SIZE
ANALYSIS: MASS
SEM; OPTICAL
MODELS: QCMB
SAMPLER NO. 3
<2.5pm >2.5 pm
60 PAIRS:
AMBIENT PARTICLES
ON QUARTZ,
MASS LOADINGS:
200 • 400 pg/cm2
ANALYSIS: CARBON
ANIONS
CATIONS
XRD
ORGANICS
MODELS:
MRA
CMB
SAMPLER NO. 4
<2.5pm >2.5pm
60 PAIRS:
AMBIENT PARTICLES
ON TEFLON,
MASS LOADING:
200 - 300 pg/cm2
ANALYSIS: MASS
XRD
SAMPLER NO. 5
DENUDER DIFFERENCE
^~CYCLONE
NYLON /
FILTER
ANALYSIS: HN03
FINE
PARTICLE
NITRATE
PROCEEDINGS—PAGE 102
-------
F1g. 2
FINE PARTICULATE SOURCE APPORTIONMENT RESULTS
HOUSTON AEROSOL DEVELOPED BY QUAIL ROOST II WORKSHOP
O
CD
E
o
o
UJ
5
3
eu
50
40
30
20
10
TOTAL 56.3 TOTAL 66.3
-
OTHER
15.1
CRUSTAL
2.5
—
STEEL 1.5
SULFATE AND
CATIONS 24.9
OTHER
CARBON 7.7
VEHICLE
EXHAUST 3.9
OTHER METALS
^- MARINE 0.6 — ,
> NITRATE <0.2**
OTHER FOSSIL
CARBON 1.4—*
DIESEL
EXHAUST 0.6-.
OTHER
15.1
CRUSTAL
2.5
STEEL 1.5
SULFATE AND
CATIONS 24.9
CONTEMPO-
RARY
CARBON 6.3
GASOLINE
EXHAUST 3.3
-
UNRESOLVED
S. A. M.
PROCEEDINGS—PAGE 103
-------
F1g. 2
COARSE PARTICULATE SOURCE APPORTIONMENT RESULTS
HOUSTON AEROSOL DEVELOPED BY QUAIL ROOST II WORKSHOP
«»u
30
»>
"
z*
O
m
E
S 20
ui
l-
0
oc
°-
10
0
—
—
—
TOTAL 38.4
OTHER
8.9
CRUSTAL
24.2
NITRATE 1.0
OTHER
CAR BON 1.9
UNRESOLVED
S. A. M.
CEMENT 0.4.
OILFLYASHO.S^S
. STEEL 0.2.
/MARINE 0.3C^
^SULFATE AND
^CATIONS!. 2 --*.
BOTANICAL 1.3
\
/GASOLINE 0.6
"A DIESEL 0.6 — _^
TOTAL 38.4
OTHER
7.9
OTHER
CRUSTAL
16.2
COAL FLY
ASH 2.2
GLASS SLAG
2.9
LIME KILN
Ca02.2
NITRATE 1.0
TIRE WEAR
1.6
RESOLVED
S. A. M.
—
—
—
PROCEEDINGS—PAGE 104
-------
RECENT DEVELOPMENT OF AEROSOL STUDIES IN JAPAN
Naoomi Yamaki
Is se i Iwamo to
Kazuhiko Sakamoto
Department of Environmental
Chemistry, Faculty of Engi-
neering, Saitama University
Presented by N. Yamaki
Saitama University
Japan
PROCEEDINGS—PAGE 105
-------
In the 1970's, aerosol studies have been considered
important as one of serious problems of air pollution and
recently a lot of studies have been actively conducted.
We present here briefly the aerosol studies conducted
in Japan for these 2-3 years, particularly with respect to
estimation of the contribution to atmospheric particulate
matter from various emission sources, secondary formation of
atmospheric particulate matter, and artifacts at filter
collection of atmospheric particulate matter.
I. Estimation of Contribution to Atmospheric Particulate
Matter in Urban Areas from Various Emission Sources
Friedlander first reported(1973) the CEB method to estimate
emission sources of particulate matter. Recently, a large
number of studies have been performed to estimate the contribu-
tion to atmospheric particulate matter in urban areas from
emission sources(Dzubay(1979), Cooper et al.(1980),Gordon(1980),
1) 2)
Matsuo et al.(1978) , Mizohata et al.(1980) ).
Contribution to Total Suspended Particulate Matter(TSP)
from Automobile Exhausts
Mizohata et al. tried to estimate the contribution to TSP
from each main source according to the CEB method, based on the
results of neutron activation analysis of TSP collected at Sakai
in Osaka(Table I).2* The change in the contribution to TSP
from the automobile exhaust during 5 years from 1975 to 1980
was also estimated with respect to TSP samples at Osaka monitor-
ing station of NASN.3* The sum of the percentage contributions
from the six main sources(soil, marine aerosol, iron and steel
industry, refuse incineration,fuel oil combustion,automobile)
with respect to the 10 index elements(Na,Al,K,Sc,V,Mn,Fe,2n,Br,
Pb) was calculated to be ca. 50%(Table 2). The percentage
contribution from the automobile was estimated as ca. 10% and
the percentage contribution from diesel engine vehicles to the
automobile increased from 88% to 95% year by year; most of
particulate matter from automobiles arises from the diesel
engine vehicles and is the most important primary man-made
source.
PROCEEDINGS—PAGE 107
-------
The results were obtained according to the CEB method by
choosing inert elements as the index element. Recently, the
introduction of decay factors for reactive materials to the
CEB method was reported by Friedlander(1980) and in Japan,
the possibility of introduction of PAH as the index material
is now discussed.
The .Contribution to TSP from Natural Sources
From estimation by Mizohata et al.(1979)(Table 2), the
contribution to TSP from the natural sources(soil and marine
aerosol) was about 30%, the contribution from other four
sources was 25%, and the remaining 45% was from the other
various small particulate-generating factories and secondary
aerosols formed in the atmosphere.
4)
Kadowaki reported that according to the bimodal model
the percentage: contribution to TSP from the natural sources
was estimated to be 50% in spring when yellow sand phenomena
often appeared and 35 to 40 % in the other seasons.
These values of contribution were calculated with respect
to the particulate matter of smaller than lOym in size.
To know real influences of particulate matters on human
environment, the contribution from emission sources as a
function of particle size(for example, 2ym>, 2ymŁ)will be
more useful. As the result, the contribution from the natural
sources to coarse particles and the contribution from auto-
mobile exhausts and secondary aerosol formed in the atmosphere
to fine particles are to be much higher. This approach will
be very useful to propose a more reasonable emission control
for particulate matter. Furthermore, the estimation of
environmental effect of suspended particulate matter(SPM)
should be done by taking into consideration the quality of SPM
besides the quantity of SPM. Kasahara suggested that to
estimate the environmental effect the size and the chemical
compos it ion. of SPM should be taken into account as well as
the concentration-..-
Analysis of Carbon of Particulate Matter
It is expected the contribution to fine particles arises
mainly from secondary aerosols formed from gas-particle
transformation and diesel exhausts. To make this clear,
PROCEEDINGS—PAGE 108
-------
characterization of size-segregated samples and determination
of appropriate index element or index material for particles
from diesel exhausts will be required. It is considered that
the main components of the secondary aerosols are SO«27 NO3~,
Cl~, NH<, and organics. While analytical methods for inorganic
ions are almost established, for organics improved methods
making it possible to determine the quantity of organics by
using only a little amount of samples are desired, since the
conventional solvent extraction method needs a significant
amount of samples and the quantity of extract depends on the
kind of solvent used. In this connection, various analytical
methods for particulate carbon were reported * that the quantity
of organic carbon in stead of that of total organics were
measured, together with the quantity of elemental carbon which
was considered as the main component of diesel exhaust particles
and as one of the important substances relating to climate
change.
Such kind of studies have been performed also in Japan.
Ohta et al. determined the quantities of total carbon(Ct)/
elemental carbon(Cei) and organic carbon(Corg) with an NC-
analyzer and a FID-GC. Ct was obtained from direct analysis
of a quartz fiber filter of SPM, Cei was analyzed after pre-
treatment of the filter at the temperature of 300°C for 30
minutes (organic carbons were removed), and corg=Ct~Cel*
Sakamoto et al.7) determined both Corg and Cei by using
a Thermal Carbon Analyzer consisting of a NDIR for detection
of CO2 and a thermal volatilization apparatus(Corg : thermal
volatilization at 450°C, Cel : combustion at 850°C). The time
required for heating in this method is only several minutes
which is extremely shorter than the other methods. This makes
it possible to reduce the possibility that organic carbons
are changed to elemental carbons owing to carbonization.
The diurnal variatiqns of Corg and Cel of SPM determined by
using the Thermal Carbon Analyzer are shown in Fig.l.
* Cf. "Particulate Carbon - Atmospheric Life Cycle -", ed.
G.T.Wolff and R.L.Klimisch, 1982.
PROCEEDINGS—PAGE 109
-------
The contribution from automobile exhausts has been
determined from the ratio of particle emission from gasoline
powered vehicles to that from diesel engine ones,according to
the CEB method where Pb emitted from gasoline powered vehicles
used as index element. Since the ratio of leaded gasoline to
all gasoline decreases to only 3 % at the present time, another
new index material will be required. The possibility of
adoption of Cei as the index material is now being discussed,
since ca. 70 % of fine particles from diesel exhausts is
PROCEEDINGS—PAGE 110
-------
31. Secondary Formation of Particulate Matter
Gas-Particle Equilibrium for NH^NOa and NHAC1
Eq.l shows the reaction of gaseous HN03 with gaseous NH3
to form particulate NH<,N03. However, since -it has been
pointed out by Stelson et al.(1979) that there is an equilib-
rium for this reaction, studies on a relationship between
concentration of HN03 or N03 in the atmosphere and ambient
temperature have widely been performed in Japan.
8) • —
Kadowaki suggested major reactions for HN03 and NO3
formation in the atmosphere(Table 3) and the formation
pathway of the secondary aerosols was explained as shown in
Fig.2. For particulate NH4C1, an equilibrium of eq.2 proba-
95
bly exists. Oka et al. observed the same relationship
between [NH3] [HCl] and 1/T as that reported for NH«NO3 by
Stelson et al.(1979) .
HN03(g) + NH3(g) > NHfcNO3(s) (1)
HCl (g) + NH3(g) * NH..C1 (s) (2)
Kara et al. considered NO3~ and Cl~ in the range of fine
particles as NHANO3 and NH6C1, respectively, and defined the
f-value for NO3~ or Cl~ as shown in eq.3. The temperature
[fine-particles]
f-value - [gaseous] + [fine-particles]
dependence of each f-value was interpreted in terms of
equilibrium of eqs. 1 and 2 (Figs.3-5).
Measurement of [NH*"1"] and [NH3] as notified in eqs. 1
and 2 was carried out by Iwase et al. The temperature
dependence was found similar to that for [NO3 ] and [Cl ] ,.
that is; in summer the concentration of gaseous NH3 was
relatively high and-'that of particulate ammonium salts was
high in winter.
Since the product of the reaction(eq.1),NHfcN03 is
deliquescent at around 25°C and a relative humidity of ca.65%
because of its hygroscopic nature, the equilibrium may
strongly depend on the value of relative humidity in the
range of relative humidity higher than 65%. Iwamoto et al.
PROCEEDINGS—PAGE 111
-------
examined the temperature and humidity dependences of NH3 and
HNO3 in air after passing through a NH<,NO3 impregnated filter.
It was found that loss of NH*NO3 occured stoichiometrically
according to eq.l and that HN03 concentration remained
constant regardless of the wide change in the air flow-rate
and a strong humidity dependence was observed over 60% of
relative humidity. The result of the temperature and humidity
dependences was in good agreement with that derived from
thermodynamic analysis of an aqueous NH<.NO3 solution which
was considered as an non-ideal -solution by Stelson and Seinfeld
(1982)(Figs.6-7). Consequently, when field measurement data
are analyzed, not only temperature but also relative humidity
should be taken into account. Since in most of the field
studies above mentioned, the collection of samples was done
by means of dual filter method and the collection period wae
very long(a half day - one week), it is likely not to estimate
accurately the effect of the artifact arising from the varia-
tion in temperature and relative humidity during the sampling
periods.
Formation of N03~ from Reaction of NOa with Marine Salt
With respect to atmospheric particulate matter collected
with an Andersen air sampler in Kobe(Kobayashi et al. ) and
8)
in Nagoya(Kadowaki ), it was estimated from the value of
*
Cl, that NaCl particles were partly converted to NaNO3
particles. Kobayashi et al. found a positive relation between
Cl, and NOa concentration as Martens et al.(1973) did and
suggested the possibility of NO3 formation through the
following reactions(eqs. 4~6) .
3NO2 + H2O »• 2HN03 + NO • (4)
NaCl + HNO3 » NaNOj, +HC1 (5)
3NOa + H2O + 2NaCl > 2NaNO3 + 2HC1 + NO (6)
* Cl, (pmol/m3) was calculated according to the following
equations.
cltheor - 1.17xNac_obs
clloss = cltheor ~ clc-obs
Cl~ of seawater/Na >
x(Na atomic weight(23.0)/C1 atomic weight(35.5))
1.17 = (Cl of seawater/Na+xi.8)
PROCEEDINGS—PAGE 112
-------
8)
However, Kadowaki pointed out that the remarkable
increase in Clloss, especially in summer may be due to
acceleration of the reaction in eq.5, accompanied by the
evaporation of HNO3 from particulate nitrates and the increase
of HNO3 formed in photochemical and heterogeneous reactions
at higher temperature.
Conversion of Nitrogen Oxides to Nitric Acid and Nitrates
In Fig.8, a simplified mechanism for nitrates(HN03 and
N03 ) formation from NOX is shown.
The formation path has been explained in terms of fN
(conversion from NOX to N03~)(Grosjean and Friedlander(1975))
or in terms of f (N03~/NOX) (Appel et al. (1978)). As" mentioned
above, however, because of the equilibrium between particulate
NO3 and gaseous HNO3-T it is more reasonable to consider the
sum of NO3 and HNO3 than to consider each species. Sakamoto
14)
et al. defined f^' as shown in eq.9 and measured the
diurnal variations of fN' (Figs.9-10). It is noted that for
[NO3~-N]
f -
N ~
, _
f ~
[NOX-N] -r [N03~-N]
[N03"-N]
[NO-N]
X
f «- [N03~-N] + [HNOa-N] __ . .
N [NOX-N] + [N03--N] + [HNOS-N] / ' ~V '
the date in summer shown in Fig. 9 the contribution of PAN to
fjg1 is considered small since the concentration of PAN
measured at some intervals was always lower than 1 ppb.
In general, f N ' is high at daytime and becomes low at night,
and a positive relationship between fjj1 and O3 concentration
is obtained. This suggests the possibility that the conver-
sion of NOX to HN03 and/or N03~ occurs (NOX+OH,03 - » HN03)
at daytime when the photochemical potential is high.
If the values of fN* with respect to the same air mass
at different sites along the trajectory are measured, the
conversion rate from NOX to HN03 and N03 can be calculated
from ffj' and the transportation time of the air mass.
On the assumption that the each result in Fig. 9 or 10 was
obtained for the same air mass, the conversion rate was
obtained to be ca. 6 % h"1 for the summer data (Fig. 9) and ca.
PROCEEDINGS—PAGE 113
-------
3%h l for the data in May(Fig.lO). These values are
considered reasonable, taking account of the difference in the
intensity of sunlight and in O3 concentration. In addition,
the value of 6 % h"1 is similar to the value of ca. 5 % h"1
obtained from the data in September-October for urban plume
by Spicer et al.(1978).
In Japan there are little isolated urban plume of
sufficient residence time, but some studies have been carried
out to obtain the conversion rate by chasing a polluted plume
over big cities with airplane. From measurement of NO3~
concentration in air over southern Kanto areas by Suzuki
et al. (August,1980), the formation rate was obtained to be
0.5-1.0 %h"1 and the formation rate of HNO3(g) was obtained
as ca. 20%h-1 on the basis of OH radical concentration
estimated from HC composition. This formation rate is between
the values of 24 %h~*(Spicer(1979)) and 13 %h"1(Meagher et al.
(1981)).
Secondary Formation of Organic Particles
Not a lot of studies have been done with respect to
organic composition in atmospheric particulate matter, but
a few studies about secondary formation of organic particles
by photochemical reaction have been done. The ratio of
primary organic particles to the secondary ones varies with
areas and seasons, but it is estimated that the ratio of
secondary organic particles to total organic particles will
increase in summer at urban areas with high photochemical
potential.
Sakamoto et al. thought most of secondary organic
particles formed by a photochemical reaction were carbonyl
compounds, and suggested that a relative values(RC=Q)fthe
ratio of integrated intensity of carbonyl ir peak(vc=0)to
that of methylene ir peak(VcH2) can be used as a parameter
showing the contribution of secondary organic particles.
This is also applicable to organic nitrates. Grosjean and
Friedlander(1975) suggested a positive relation between VC=Q
intensity of organic particles and O3 concentration. The value
of Rp_o here is in a good correlation with O3 concentration,
which is one of parameters showing photochemical potential
(Table 4). Moreover, with regards to samples in the fine
PROCEEDINGS—PAGE 114
-------
particle regions, a correlation coefficient between RC=O and
03 concentration was obtained as 0.83(n=15)(significant), but
for the coarse particles such a correlation was not found.
This may suggest the possibility of heterogeneous nucleation
that organic carbonyl compounds formed in a photochemical
reaction condense on the fine particles with large surface
18) „ ...
areas. From organic acidic components in the atmospheric
particulate matter collected on August, 1978, several kinds of
a,u)-dicarboxylic acids (succinic acid, glutaric acid and adipic
acid) were identified,which were considered as final products
of HC and O3 or OH reactions(Table 5). '
(C=O) ' Av 1/2 (C=0)
^-~u emax(CH2)-AvV2 (CH2)
1/cl • log -^/I (C=O) • Av V2 (C=O)
1/cl- log Io/~L (CH2) - Av 1/2
_ log To/T (C=O) 'Avi/2 (C=O)
log L/I(CH2) .Avi/2 (CH2)
Formation of Particles from Cyclohexenes
Bandow et al. and Sakamoto et al. ' studied a
chemical reaction mechanism for formation of particles from
cyclohexenes in photoirradiation of cyclohexenes-NOx-air
system. Hexanedialdehyde(OHC(CH2)VCHO) in dry air system and
adipic acid(HOOC(CH2)ACOOH) in a humidified air systern_were
formed from cyclohexene. Glutaric acid(HOOC(CH2)3COOH) and
PAN were formed from 1-methylcyclohexene. This suggests not
only the scission of 1,2-double bond but also that of 1,6-
22)
single bond can easily occur in 1-methylcyclohexene(Fig.11).
These products can be explained by the reaction pathways
almost same to that proposed by Grosjean and Friedlander(1980),
that is ; at the initial reaction stage HC is consumed mainly
with OH radical addition to the double bond.
Even if equimolar NOX,S02 and cyclohexene are completely
transformed to particulate matter, the weight ratio of the
particulate matter is not equivalent but is calculated as
1 : 1.5 : 2.4(=N03~ : S04a~ : adipic acid). It is thus considered
that because of the largest value of 2.4 the contribution to
secondary particles from organic particles will be important.
It is necessary to study the photochemical particle forma-
tion ability for various kinds of HC which may be emitted
from sources.
PROCEEDINGS—PAGE 115
-------
Formation of Particles from Reduced Sulfur Compounds
In addition to a lot of studies on photooxidation of S02,
there has been growing interest in photochemical reaction of
reduced sulfur compounds which are biologically generated and
are important compounds in consideration of sulfur balance'in
the atmosphere(Cox et al.(1980),Niki et al.(1980,1982)).
Hatakeyama et al. ~ studied a photochemical reaction
of dimethyl sulfide, methane thiol, or dimethyl disulfide-NO-
air system and found H2SO<, and CH3S03H which was detected in
atmospheric pariculate matter by Panter et al.(1980). A
possible formation pathway was shown in Fig.12 where the
reaction was initiated by addition of OH to S. ' In this
reaction, the simultaneous formation of SO2 in addition to
CH3SO3H and H2SO<, is of interest,
Formation of Particles in Photoirradiation of HC-NO-SOg-
Air System
By means of a rotatable aerosol chamber which makes the
residence time of particles in air longer(Volume 4m3,S/V=4.8m"1),
26 271
Izumi et al.- ' conducted the photoirradiation experiments
of C3H«-NO-SO2-dry air system to study photochemical oxidation
of S02 and particles formation. The SO2 .oxidation rate can
not be explained only in terms of the reaction of SO2 with OH
radical and a good correlation was found between SO2 decay
rate and the amount of C3H6 consumed in O3-C3H6 reaction.26'
In a dark reaction of O3-C3H6-SO2-dry air,27^ at the initial
stage very fine particles were mainly formed and they grew
with reaction time(Fig.13). By use of steady state method
wi.th respect to an active intermediate(P*) formed from O3 and
C3H6 and from a series of experiments with different S02
concentrations, they have shown that about a half of P*(k*/k=
0.92) was consumed for oxidation of SO2 to SO3(eq.l2).
C,H6 + O, —* P* -—(11)
P* + S02 -^-* S03 + Products (12)
P* + SO2 k'> SO2 + Products (13)
PROCEEDINGS—PAGE 116
-------
Transport and Removal of Particulate Matter
Since all big cities and industrialized zones in Japan
are located along coasts, the air pollutants from emission
sources are transported by sea-land breezes etc., transformed
with chemical and physical processes and removed with various
processes. Several studies on fate of gaseous pollutants have
been done by NIES group and others, but only a few studies
have been done with respect to particulate matter.
TO \
Kadowaki et al. studied the change in the composition
of particulate matter by comparing that of an urban area with
that of mountainous and coastal areas which were about 70 km
distant from the urban area. They found that in the course of
the transport process from the urban area to mountainous area,
soil particles and marine-salt particles in the range of
coarse particles were selectively removed. They also pointed
out some transformation during the transport process since SG.1
existed as NH^HSCU in the urban area and did as (NH4) 2SOi, in
the mountainous area,
q\
Oka et al. collected atmospheric particulate matter
with an NH3 denuder and measured quantitatively H2SOA mist and
S0«.2~ concentrations of size-segregated particulate matter.
From the particle size distribution observed, they suggested
the possibility that fine particle H2SCU mist reacted with NH3
in the course of condensation to become large size particle
of SO*2-.
Tamaki and Hiraki studied the effect of rainfall on
the concentrations of gaseous NOX and particulate NO3~ in the
atmosphere. The average concentration of N03~ during rainfall
decreased to about a half of its initial concentration by
about 10 mm rainfall and to about a quarter just after rainfall.
This was explained from the fact that NO3 is contained
mainly in coarse particles and is not produced in air during
rainfall, since it is formed photochemically.
PROCEEDINGS—PAGE 117
-------
HI. Artifacts in Sampling Atmospheric Particulate Matter
with Filter Collection
Since the middle of 1970's, it has been reported in
succession that the positive interference in,, sampling
atmospheric particulate sulfate and nitrate resulted from
adsorption of S02 and HN03 on or reaction with the filter
materials. As a result of many active studies on positive
and negative artifacts in sampling sulfate and nitrate, it
was found that several (quartz-fiber and Teflon) filters gave
relatively little artifact.
The polycyclic aromatic hydrocarbon (PAH) content of particulate
matter emitted into the air by combustion of fossil fuels
is receiving increasing attention. Furthermore, due to "better
fuel economy of diesel engines than that of gasoline engines,.
the increase in dieselized fraction of automobile will increase
the emission of particulate originating in vehicular engines*
Recently, many studies on the filter collection of PAH and it^
carcinogenic nitro-derivatives, which present mostly in
respirable size and may constitute a significant inhalation
health hazard to the human population, have been reported.
Sampling Atmospheric PAH with Polyurethane Foam Plugs
It is difficult to determine the real ambient concentra-
tion of PAH with collection of a particulate matter on filters
by using conventional high-volume air sampler due to its high
vapor pressure. Yamasaki et al. determined the concentra-
tions of 3- to 6-ring PAHs both in the vapor phase(PAHxvap)
and in the particulate phase(PAHxpat) with the collection
method using a combination of glass-fiber filters(GFs) and
two layers-polyurethane foam plugs(PUFPs) as back-up.
This method permitted a higher flow rate in sampling particles
than the collection method using various polymer beads
(Broddin et al.(1980), Robertson et al.(1980), Lindgren et al.
(1980)) and cooling collection systems(Handa et al.(1980)31^).
In their result, at ambient temperature levels substantial
amounts of 3- to 5-ring PAHs were found in the vapor phase
depending upon temperature(T,K) and 6-ring PAHs were all
found in the particulate phase(Table 6). Aspects of the
formers in ambient air were considered to be explained by the
PROCEEDINGS—PAGE 118
-------
Langmuir adsorption concept (eq. 14) . thus, it was suggested
that correction of various PAH data obtained by conventional
methods might be possible using the concept.
109 CPAHxpat]/[TSP] = -VT + B —(14)
[PAHxvap] : PAHX concentration (ng/m3) in the gas phase
[PAHxpat] : PAHX concentration (ng/m3) in the particulate phase
[TSP] : TSP concentration (ng/m3)
A,B : Constant
Formation of Nitropyrene as Artifact during Collection
of Particulate Matter
It has been reported by Pitts et al. (1978) that formation
of nitro-derivatives by the possible reactions between PAHs
and nitrogen oxides, which were representative atmospheric
pollutants,, on filter for collection of particulate matterc
Carcinogenicity of 1-nitropyrene and 3-nitrof luoranthene
recently reported by Kawashi et al. (1981) induced the great
concern about the possible health effects by these compounds
in diesel exhaust particles collected,,
Since the first report by Pitts et al. appeared in 1978,
an increasing number of papers have been published on both
the increase in mutagenicity and the amount of nitro-
derivatives by the exposure of diesel exhaust particulate
matter and/or PAHs on filters to diesel exhaust gas or controll-
ed gas containing NOa, SO2, HN03/ etc. Also, identification
of 1-nitropyrene in atmospheric and diesel exhaust particulate
matter was reported by several workers. However, the study
on the collection method without artifact nitration of PAHs
during sampling will be very significant since nitro-derivatives
of PAHs are formed by the reaction of PAHs even with sever.al
ppb of HNC-3 due to the result reported by Pitts et al. (1978) .
Sakamoto et al'.'33* reported the conversion efficiency of
pyrene to nitro-derivative by exposure of controlled humidifi-
ed gas containing 50 ppm of N0a (Table 7) and indicated that
the conversion efficiency decreased significantly by the treat-
ment of quartz-fiber filter with K2CO3 (Table 8). Recently,
.the systematic study on collection and analytical methods for
nitro-derivatives of PAHs in atmospheric and diesel exhaust
particulate matter was started in Japan.
PROCEEDINGS — PAGE 119
-------
REFERENCES
1) Y.Matsuo, M.Fujiwara, and T.Higuchi, J.Environ.Pollut.Control.,
3.4_, 951 (1978) .
2) A.Mizohata and
-------
19) K. Sakamoto, S. Sasaki, S. Otsuka, I. Iwamoto, and N. Yamaki,
41th Spring Annual Meeting Abstracts of Chem. Soc. Jpn., No.
3C10, Apr. 1980.
20) H.Bandow and H. Akimoto, 23th Conference Abstracts of Jpn. Soc.
Air Pollut., No. 439, Nov. 1982.
21) K. Sakamoto, T. Sutoh, S. Otsuka, I. Iwamoto, and N. Yamaki,
21th Conference Abstracts of Jpn. Soc.Air Pollut., No. 420,
Nov. 1980; K. Sakamoto, T. Sutoh, N. Fukushima, Y. Matsuda,
S. Otsuka, I. Iwamoto, and N. Yamaki, 43th Spring Annual
Meeting Abstracts of Chem. Soc. Jpn., No. 3007, Apr. 1981.
22) K. Sakamoto, T. Sutoh, S. Otsuka, I. Iwamoto, and N. Yamaki,
22th Conference of Jpn. Soc. Air Pollut., No. 724, Oct. 1981.
*
23) S. Hatakeyama, M. Okuda, and H. Akimoto, Geophys. Res. Lett.,
1, 583(1982).
•it
24) S. Hatakeyama and H. Akimoto, Int. Sympos. Abstracts on Chem.
Kinet. related to Atmos. Chem., No. 16, June 1982, Jpn.
25) S. Hatakeyama and H. Akimoto, 23th Conference Abstracts of Jpn.
Soc. Air Pollut., No. 442, Nov. 1982.
26) K. Izumi, M. Mizuochi, T. Fukuyama, K. Murano, and M. Okuda,
45th Spring Annual Meeting Abstracts of Chem. Soc. Jpn., NO.
3S15, Apr. 1982.
27) K. Izumi, M. Mizuochi, K. Murano, and T. Fukuyama, 23th
Conference of Jpn. Soc. Air Pollut., No. 837, Nov. 1982.
28) S. Kadowaki, S. Imai, and K. Yoshimoto, ibid.. No. 353, Nov.
1982.
29) M. Tamaki and T. Hiraki, J. Chem. Soc. Jpn., 1982. 1252.
30) H. Yamasaki, K. Kuwata, and H. Miyamoto, Bunseki Kagaku, 27,
317(1978).
31) T. Handa, Y. Kato, T. Yamamura, T. Ishii, and K. Suda, Environ.
Sci. Technol., 14, 316(1980).
32)*H. Yamasaki, K. Kuwata, and H. Miyamoto, ibid., 16, 189(1982).
33) K. Sakamoto, K. Mita, N. Harayama, S. Otsuka, I. Iwamoto, and
N. Yamaki, 23th Conference Abstracts of Jpn. Soc. Air Pollut.,
No. 852, Nov. 1982.
* These literatures are English and. the others are Japanese.
PROCEEDINGS—PAGE 121
-------
Table 1. Contributions of major sources to TSP.
Component
Main components taken into concideration
Soil particles
Marine aerosols
Iron-and-steel industry
Refuse incineration
Fuel oil combustion
Gasoline engine automobile
Total primary aerosol predicted
Observed TSP
Balance not accounted for by resolution
Remaing components
Diesel engine
Secondary aerosol components
Sulfur dioxide — So!"
Nitrogen oxides — NO!
Hydrocarbons particulates
Various small aerosol generating factories
Predicted
TSP
contribution
(ug/m'J
26.1
1.5
4.3
2.4
1.9
2.8
39.0
79.0
40.0
14.2
9.3
7
?
7
Percentage
contribution
(%)
33.0
1.9
5.4
3.0
2.4
3.5
49.4
100.0
50.6
18.0
11.8
20.8
Table 2. Contributions of major sources to TSP.
Year
Leaded-/all gasoline (I)
TSP(pg/m')
To^cl Contribution^
Soil particles
Marine aerosols
Iron-and-steel industry
Refuse incineration
Fuel oil combustion
Automobile exhaustfdiesf! and)
^gasoline /
Others
1975
18
49.4
pg/ra* 1
10. 120. S
1.5 3.0
2.3| 4.6
2.0- 4.2
3.2! 6.5
4.8J 9.8
25.4:51.4
1976
13.5
57.5
ug/m» *
13.0-22.7
0.71 1.2
2.6J 4.6
2.8! 4.9
t
2.8! 4.8
5.4! 9.4
1
30.252.5
1977
10.5
43 .-1
ug/m* %
8.4J19.5
0.8; 2.0
Z.i: 4.8
i
3.9] 9.0
1.9; 4.5
3.9; 9.1
22.051.1
1978
8.0
53.7
ug/m' %
11.0(20.6
1.4J 2.6
2.6J 4.9
1.81 3.4
1
1.9; 3.6
7.3JL3.6
27.651.3
1979
5.5
39.0
ug/»* t
10.426.8
1.4; 3.6
1.6,' 4.2
t
2.0| 5.0
1.4! 3.5
4.6^1.8
17. 645.0
PROCEEDINGS—PAGE 122
-------
Table 3. Major reactions of nitric acid and nitrate aerosols
In the atmosphere.
Fotnutiod
Formation
of ntlfic ac
Ptllti of focmaiioa
id by pliofoclienucil otuJ*iioA
of mu*e Kid bjr thermal mctioa
Viporiiuion of niuie
Formation
Fofnuiioa
of niu.it
of ivilt • ic
«y droplet
MfOsol by homof8/3, 1979
(4 or 6h)
Correlation (Number of
coef f Iclent (r) sample (n) )
0.95 (7)
0.79 <7>
0.85 (7)
0.84 (10)
0.8S (9)
0.58 (9)
0.83** (below 1.8pm) (15)
0.24 (above 1.8pm) (15)
0, concentration
fOil ...
*• ^ max
C<)l1 2 or Sheave
t°'| 24h-avt .
L0>J Uh-ave11'
*• '^ 4 or Shrave
*. *J 4 or 6h-ave
a) Sampling site: Saitama Univ. Sampling method : high-volume
air sampler or Andersen high-volume air sampler.
b) 12h-averaged 0, concentration during 6:00 an—6:00 pa.
Table 5. Dlcarboxylic acids and aliphatic acids
Identified in the atmospheric aerosols.
Dicarboxylic Acid
Malonic Acid
Methylmalonic Acid
Succinic Acid
Melhysuccinic Acid
Glularic Acid
2-.3-Me(hylgIutaiic Acid
Adipic Acid
Dimelhylglutaric Acid
2-,3-Methylidipic Acid
Pimelic Acid
Suberic Acid
Azclaic Acid
Sebacic Acid
Authors
Crosjean el oL
(1978)
O
0
o
O
0
O
O
O
O
O
0
O
O
Cionn a aL
(1911)
'•'.'.
O
O
O
Schuetzle ef al.
(1975)
O
O
0
O
O
Sakamoto el aL
(1980)
O
O
O
O
PROCEEDINGS—PAGE 123
-------
Table 6. PAHs in the ambient air trapped on the GF and PUFPs.
PAH
Phenantrene+Anthracene
Fluoranthene
Pyrene
Benzo(a)fluorene+Benzo(b)fluorene
Chrysene+Benzo(a)antnracene+Triphenylene
Bezo (a ) py rene+Benzo (e ) pyrene
o-Phenyl enepyrene+Unknown
Benzo(ghi )perylene+Anthanthrene
PAHs trapped
(ua/1000 «*)
GF
1.16
0.95
0.99
0.30
2.34
2.99
2.91
3.05
PUFP-1
69.8
24.6
18.5
3.58
3.05
0.17
n.d.
n.d.
PUFP-2
34.6
0.13
0.07
n. .
n. .
n. .
n. .
n. .
Total
(ug/100 m»]
105.6
"25.7
19.6
3.78
5.39
3.16
2.91
3.0S
X on
GF
1.1
3.7
5.1
5.3
43.4
94.6
100
100
Ring
number
3
4
4
4
4
5
6
6
Sampling site : Environmental Pollution Control Center(1-3-62, Nakamichi, Higashinari-ku,
Osaka-shi, Osaka), Sanpling date : July 6-7, 1977. Temperature : (22.9-32.4T.). Volume of
air sample : 1114 n5; ** Unknown substance was surmised benzo(b)chrysene due to reference
(6). PUFP-1; First PUFP. PUFP-2 : Second PUFP, n.d. : Not detected (below; 0.05ug/1000 m1)
Table .7, Conversion of pyrene to nitropyrene.
F1lter(pH)
1st
' — -
PA
PA
— -
2nd
>TQ (55)
TO
AE fSS)
AE
TBG (5.1)
TCG (5.4)
FP («)
Tine
(h)
20
21
72
Conversion
(*) (n)
24 (4)
6 (2)
...16. (2)'
3 (2)
6 (2)
6 (2)
0
0
AMOunt of pyrene/47nt filteri
Table 8, Conversion of pyrene to nitropyrene
'on several treated filters.
Exposure
Method
Through
filter
Onv filter
Tissuquartz
filter
HCl-treated
Non-treated
K,CO, -treated
Non-treated
K, CO, -treated
pR
5.1
4,9
6.9
10.1
6.9
10.1
10.3
Conversion
(X) (n)
47 (2)
49 (2)
35 (5)
0.1 (2)
39 (4)
0.1 (2)
0-1 (2)
Amount of pyrene/47u+ filtenSmj.
PROCEEDINGS—PAGE 124
-------
60
40
20
racel
E3Corg
EH NO;
Ea so; " r
J
/T
d A " >
'\3pfi\V- •'•• -•
F^jinnff ._Ł*• — i
mUTĄ7/ '/. Y/
.— 1
/
\J
V' ^c^
w
x2-a
ia^
^
;v
n
li
?
x
v^
Q
y
3
I
?
t
2
y>
\ w
80^
E
CT»
13.
z
ex
40"
0
, — , — c\j n ^f 1/1 ^o co Q~*
7 ^ < / i i i / )
o . — i — rsi n -*r "^ ^ co
u^ !2 j
July
August
F1g.l
. Chemical composition of SPM,
Sampling date: sunnier in 1981.
Sampling site: Saltama Univ.
Sampling method: hiqh-volumt
air sampler.
Porlicle d>om*1«r ( >mi
< F(nt parliclts 1^ Loant porttcU* :
Fig. 2. Schematic diagram of suggested mechinismj-
for sulfate, niu-ate. and chloride formatioi1
in uib«n air at Nagoyt
•as
Cf
rf>
o o ,
NOT
- o^1
Fig.3. Ambient temperature
vs. f-value of Cl..
Fig.4. Ambient temperature
vs. f-value of NOl.
Kig.f). Dependence of theoretical
vapor pressure (pHr^ anŁl
Puun ) on ambient tempers-
HNUi
ture In eqs. 1 and 2.
PROCEEDINGS -PAGE 125
-------
Fig.6. Temperature dependence of [HN03] in the
purified air passed through NH,N05 loaded
TQ filter;at R.H.OJ.face velocity4.7cm/sec.
lOtng impregnated,( ) calculated, from
Stelson's equation.
" 0 20 .40 60
Relative humidity(1)
Fig.7. Relative humidity dependence of [HNOs] in the
purified air passed through NH»NO» impregnated
TQ filter;at 25"C,face velocity 4.7cm/sec.1Drag
impregnated,( j calculated values from
Stelson's equation.
PANS.
RONO
. RONO, X""
»? HO,
• * *
X \P«A
NO . ' NOi
hv \
\
V
"HOT"
OH
"H,O
_ ^^
NaO,
V»»P
I HNO.
/
/
X
Soil et;
.H,SO. H^-
NH,,NaCl /
etc. y
Fig.8. Mechanism for nitrates formation
in the atmosphere.
100
0
is
(«*)
10
t
so
(«)
to
M
10
10
0
11
Tix «>
!•*/«•!
M
It
It
I
II
Fig.9. Diurnal variations of f^, fN>
and selected pollutants at
Saitama Univ.(July 31^August
2..1979).
(ppb)
100
50
0
20
10
50
40
30
20
10
0
HNO,(ppb)
18 6 18 6 18 6
Time of day(h)
18
Fig.10. Diurnal variations of f^. IH,
and selected pollutants at
Saitama Univ.(May 27^30,1980).
PROCEEDINGS—PAGE 126
-------
•oi
ex.
— » PAN
+• CH.CO-
I T v,n,v,u- CH,CO. +
\xr—
'COOH
C'OOH
COOH
Fig.11. Possible formation pathways for glutaric
acid and FAN.
00-
CH.SCH,
OH
NO —J-.NO,
0-
I
CH.SCH,
OH
CH.SCH, + OH
I
CH.SCH, -
OH
CH.SOH + CKS
.
ICH.SO.H]
0-
CH. +-SCH, .CH, + SO,H-^r(HTj
OH JO,
|SO,| + HO,
Fig.12. Possible formation pathways for
methylsulfonlc acid, s.ulfurlc acid,
and sulfur dioxide.
io4-
i lo3-
I ,flt
10
^ . , . ^ .
• 16' >67 io' io2 10' ioz 10'
Factlel* DU»ctec Inn)
Fig. 13. f.rtlcl. >1» distribution
of photoch*«Ic«l ««ro»ol for«ij
fro- C3H6(1.0pp.)/HO(0.2pp.)/S02
(O.lpp«)/drr «ir (HjO
-------
APPENDICES
PROCEEDINGS—PAGE 129
-------
JAPAN-US JOINT CONFERENCE
ON
PHOTOCHEMICAL AIR POLLUTION AND
AIR POLLUTION-RELATED METEOROLOGY
Conference Room
National Institute for Environmental Studies
16-2 Onogawa, Yatabe-cho, Tsukuba-gun
Ibaragi Prefecture, Japan
December 1-2, 1982
AGENDA.
Wednesday, December 1, 1982
Acting Chairman: Dr. T. Okita
10:00 - io:uo
- 11:00
11:00 - 12:00
12:00 - 13:50
13:30 - 15:00
15:00 - 15:30
15:30 - 17:00
Opening Remarks
Introduction of Participants
Election of Session Chairman
Approval of Conference Program
Refreshments
Dr. J. Kondo
( Director of NIES )
Session Chairman.: Dr. H. L. Wiser
Wiser
17:00 -
The Problems of Acid Bain in Japan
Lunch
Progress in Photochemical Air Q,ualit7
Simulation Modeling
Researches on Acid Rain in Japan
Refreshments
Urten Ozone Modeling Developments in
the U. S.
A Numerical Simulation of Local Wind
and Photochemical Air'Pollution
Transport and Transformation Of Air
Pollutants tsy Land arid Sea Breezes
Mr. S. Kato
Dr. K. L. Demerjian
Mr. T. Komeiji
Dr. B. Dimitriades
Dr. P. Kimura
Dr. H. Tsuruta
PROCEEDINGS—PAGE 131
-------
Thursday, December g, iggg Session Chairman: Dr. N. Murayama
9:00 - 10:50 Evaluation of Eight Linear Regional- Dr. K. L. Demerjian
scale Sulfur Models by the Regional
Modeling Subgroup of the United States/
Canadian Work Group 2
Field Studies on Photochemical
Air Pollution in Japan Dr. S. Wakamatsu
10:30 — 11:00 Refreshments
ir.OO - 12:30 US Studies on Stratospheric Ozone Dr. H. L. Wiser
Intrusion of Stratospheric Ozone
into the Troposphere Dr. H. Muramateu
The Vertical Distributions of
CFzC/a. CFC/4 arid NŁ over Japan Dr. M. Hirota
12: 30 - H :00 Lunch
Tsukuba - - - »• Tokyo
PROCEEDINGS—PAGE 132
-------
THE PROBLEMS OF ACID RAIN IN JAPAN
Presented by S-Kato
Environment Agency
Japan
PROCEEDINGS—PAGE 133
-------
I. Actual occurrence of acid rain found to date in Japan
1. Damage caused by acid rain in Japan
In Japan, no damage by acid rain to the ecosystem; forest,
agricultural products, fish, etc., has been reported, though
it has become a problem in North Europe and North America.
Therefore, in this field, acid rain is not a social problem
yet in Japan. The problem of acid rain occurring in Japan
to date is irritation to the eyes and the skin due to mist
and drizzle occurring in the metropolitan area, etc. from
1973 to 1975 (see Table 1).
Since this was considered to be caused by irritants such as
hydrogen ions (H+) contained in mist and drizzle, it became
necessary to clarify the mechanism of pollution generation by
these materials. The Environment Agency designated this kind
of pollution "wet-air pollution" (health hazard due to contam-
inated precipitation) and carried out relevant studies for the
5 years from 1975 to 1979 fiscal year.
(Note) Differences in adverse effects due to acid rain between
Japan and USA found to date.
0 Japan Irritation caused to humans (irritation
to the eyes and the skin)
o USA Damage to the ecosystem (lakes and marshes,
forest, etc.)
2. On wet-air pollution studies
Studies on wet-air pollution are intended to clarify the genera-
tion mechanism. When this is considered, it is necessary to
clarify (1) weather conditions liable to cause mist or drizzle,
(2) the production of irritants and absorbtion into precipitation,
and (3) the production mechanism of irritants during precipitation.
PROCEEDINGS—PAGE 135
-------
To clarify the conditions causing actual damage, it is necessary
to know (4) the concentrations of irritants in precipitation and
(5) the concentrations of those irritants which cause damage.
To clarify the above points, a series of studies were made.
Irritation to the eyes and the skin was surmised to be caused
not only by the action of acid materials in precipitation but
also the mutual action of such coexisting materials as formaldehyde
(HCHO), formic acid (HCOOH), and hydrogen peroxide (l^C^), etc.
and the production mechanism of these materials was taken to be
as shown in Fig. 1.
3. Results of researches
(1) Weather conditions liable to cause wet air pollution
Weather conditions on days when reports were made were
investigated, and it was found that irritation occurred
under certain weather conditions (when a weak front with
high humidity slowly approached a stagnant high pressure
area and humidity was high.
(2) Results of rainwater analysis
Rainwater was analysed from 1975 to 1979 fiscal year.
Especially, pH showed a very low value of about 4 on
average (see Table 2).
(3) Relation between the concentrations of irritants and its
effects
In an experiment using rabbits, irritation to the eyes
could be confirmed only at very high concentrations of
irritants.
PROCEEDINGS—PAGE 136
-------
(4) Generation, diffusion and production of irritants in
the atmosphere
Among irritants and their precursors, the distributions
of the generation sources of those other than sulfur
dioxide and nitrogen oxides were not known, and sufficient
data could not be obtained on the drift and diffusion of
these materials. For this reason, the relation between
sources and effect could not be clarified. Furthermore,
also with regard to the production of irritants in the
atmosphere, the actual production mechanism and production
rates could not be determined by field studies. Especially
the transformation at high humidity in cloudy weather was
not clarified, being left as a problem to be solved in the
future.
(5) Absorbtion of irritants into precipitation and production
mechanism during precipitation
The pH of rainwater was found to relate to sulfate ions
and nitrate ions, and low pH is presumed to be caused by
the absorbtion of sulfate mist and gaseous nitric acid.
(6) Change of wet-air pollution
Adverse effects on human health from wet-air pollution
occurred often in the period from 1973 to 1975, but since
then, none were reported except one case in 1976. However,
recently in June, 1981, a case where several persons
reported adverse effects occurred in Isezaki City, Gunraa
Pref.
PROCEEDINGS—PAGE 137
-------
II. Planning future research
The Environment Agency is planning to conduct studies concerning
acid rain from fiscal 1983 for 5 years. The outline is as follows
(for the entire flow; see attached sheet).
1. Atmosphere system
To find measures to prevent influences on the ecosystem, etc.,
the actual conditions of acid rain and the balance of pollutants
must be clarified, and the production mechanism of acid rain
must be:known. For these purposes, the studies with the following
contents will be made.
(1) Analysing pollutants and rainwater
Pollutants and rainwater will be sampled at 12 places
throughout the country, to analyze pH, sulfate ions,
nitrace ions, chlorine ions and ammonium ions, to determine
the conditions of pollution.
(2) Measuring the distributions of pollutants by altitude,
and arranging and analysing meteorological data
The concentrations of pollutants (sulfur dioxide, nitrogen
oxides, ozone, sulfate ions, nitrate ions and nitric acid)
in the sky will be investigated by altitude, to determine
their distribution, and meteorological data (wind direction,
wind velocity, temperature and humidity) will be arranged
and analyzed.
(3) Measuring the settling rates of pollutants
To clarify the balance of pollutants in the atmosphere,
the settling rates of pollutants (sulfur dioxide, nitrogen
oxides, sulfate ions, nitrate ions and nitric acid) on the
ground surface will be examined.
PROCEEDINGS—PAGE 138
-------
(4) Studies concerning the process of absorbtion of pollutants
into rainwater, and so on
Studies concerning the process of absorbtion into rainwater
of various pollutants (sulfur dioxide, nitrogen oxides,
sulfate ions, nitrate ions and nitric acid) causing the
production of acid rain will be made.
2. Land water system
(1) Survey of actual conditions
To determine the actual conditions of land water pollution
due to acid rain throughout the country, the pH, electrical
conductivity, metal concentrations, etc. will be measured
in lakes, marshes, rivers, etc. in mountainous areas little
affected by artificial pollution, and the species and
numbers of fish, shells, and other animals living there
will be investigated.
(2) Studies to clarify the acidifying mechanism
Based on the results of the survey of actual conditions,
studies will be made successively in the land water areas
of lakes, marshes, rivers, etc. where acidifying continues,
to clarify the acidifying mechanism of land water.
3. Soil system
Considering the weather conditions such as precipitation, land
use, and so on, 5 typical areas will be selected throughout
the country, and studies will be made on the physicochemical
natures of rainwater and soil, and biological natures of soil,
vegetation, etc. to determine the actual conditions of soil
environment worsened by acid rain, and to clarify the influences
on the soil ecosystem and vegetation.
PROCEEDINGS—PAGE 139
-------
Table 1 Number of reported sufferers
Fiscal
year
1973
1974
1975
1976
Month and date
6.28 * 6.29
9.13
7.3 ^ 7.4
7.18
Others
6.25
Others
8.16
Number of reported
sufferers
540
7
Subtotal 547
32,546
506
129
Subtotal 33,181
143
101
Subtotal 244
1
G™n* 33,973
total '
pH of
rainwater
2^3
-
3^4
3^4
3^4
5.6
PROCEEDINGS—PAGE 140
-------
Table 2 pH of rainwater in Kanto Area
(3 mm from initial rainfall)
N. Fiscal
^\vear
Place \^
Mito
Shimodate
Utsunomiya
Tochigi
Maebashi
Tatebayashi
Urawa
Kumagaya
Ichihara
Togane
Chiyoda
Tama
Kawasaki
Yokohama
Hiratsuka
Mean
1975
3.3 ^ 5.8
(4.8)
3.4 ^ 6.6
( - )
3.2 ^ 6.3
(3.9)
3.6 -v 6.3
(4.0)
4.4 * 6.5
(5.0)
3.3 ^ 6.1
(3.9)
3.1 * 4.4
(3.7)
3.1 ^ 6.7
(3.7)
3.7 ^ 5.7
(4.3)
3.9 ^ 6.0
(4.3)
3.9 ^ 6.6
(4.4)
3.5 -v 6.6
(4.3)
3.5 ^ 6.4
(3.9)
3.3 ^ 4.4
(3.7)
3.4 ^ 5.4
(4.0)
3.1 * 6.7
1976
4.0 ^ 5.2
(5.0)
3.8 ^ 5.4
(4.3)
3.3 -v 4.0
(3.9)
3.6 ^ 5.2
(4.0)
3.9 ^ 4.0
(3.9)
3.6 ^ 4.2
(3.9)
3.6 ^ 4.7
(4.0)
4.7 ^ 6.5
(5.0)
4.4 ^ 6.8
(4.9)
4.3 ^ 6.4
(5.0)
4.1 ^ 5.9
(4.8)
4.7 ^ 6.8
(5.2)
3.6 ^ 5.3
(4.1)
3.3 ^ 4.0
(3.6)
3.7 ^ 4.8
(4.0)
3.3 ^ 6.8
1977
4.6 ^ 6.8
(5.0)
4.0 ^ 7.4
(4.7)
4.1 ^ 6.3
(4.6)
3.9 ^ 6.2
(4.6)
4.6 ^ 6.1
(5.0)
5.2 ^ 5.8
(5.5)
3.7 ^ 6.0
(4.3)
3.8 ^ 6.5
(4.7)
3.7 -v 5.7
(4.3)
4.0 ^ 5.8
(4.6)
4.0 ^ 5.6
(4.4)
3.5 ^ 6.9
(4.0)
4.0 ^ 4.8
(4.3)
3.0 ^ 3.6
(3.3)
3.8 ^ 4.2
(4.0)
3.0 ^ 7.4
1978
4.2 ^ 6.0
(5.4)
4.2 ^ 5.9
(4.8)
3.3 -v 6.2
(4.0)
3.7 -v 6.3
(4.2)
3.6 ^ 4.6
(3.9)
3.7 ^ 4.9
(4.3)
3.5 ^ 6.1
(4.0)
3.9 -v 6.6
(4.3)
3.6 ^ 6.4
(4.2)
3.3 ^ 6.1
(5.2)
3.5 ^ 7.0
(4.1)
3.4 ^ 4.4
(3.8)
3.5 ~ 4.7
(3.9)
3.4 ^ 4.5
(3.8)
3.1 ^ 5.6
(3.6)
3.1 ^ 7.0
1979
4.3 ^ 6.7
(5.0)
4.2 ^ 6.7
(5.1)
3.9 ^ 6.5
(4.4)
4.3 ^ 7.0
(4.9)
4.3 ~ 6.3
(4.7)
4.5 'v 6.2
(5.2)
4.1 ^ 6.3
(4.7)
4.6 -v 6.6
(5.1)
3.8 ^ 5.0
(4.0)
4.5 ^ 5.5
(4.8)
4.0 ^ 6.5
(4.5)
4.2 'v 6.1
(4.7)
3.3 ^ 5.3
(3.6)
4.0 ^ 6.5
(4.5)
4.5 ~ 5.8
(4.9)
3.3 ^ 7.0
Mean
4.8 ^ 5.4
3.9 ^ 5.1
3.9 * 5.1
4.0 ^ 6.3
3.9 *> 5.0
3.9 ^ 5.2
3.7 ^ 4.7
3.7 ^ 5.1
4.0 -v 4.9
4.3 ^ 5.8
4.1 ^ 4.8
3.8 -v 5.2
3.6 ^ 4.3
3.3 ^ 4.5
3.6 ^ 4.9
3.3 ^ 6.3
n
54
52
80
75
33
82
71
53
99
74
61
72
70
60
62
998
(Note) 1) The values in parenthese are mean values.
2) The period of survey was from the latter half of June to the first
half of July.
PROCEEDINGS—PAGE 141
-------
Primary
materials
SO2, NOx
HO*., RCHO
Hydrocarbons ,
NH3
Sea salt
particles, etc
Transfor-
mation
Secondary
materials
H2S04, Sol
HN03, N03
ECU, NH4
RCHO, RCOOH
03, H202,
etc.
D
u
c
t
issoluti
ptake an
ondensa-
ion
Drift
Diffusion
Dissoluti
and uptak
Transformation
in
°n' precipitation
»- Pt/Mirl
and mist
'. 3,. Drizzle
on and rain
.e
Effects]
Transformation
in rainfall
Fig. 1 Assumed production mechanism of wet-air pollution
PROCEEDINGS--PAGE 142
-------
/Y
S
•Utsunomi
f ) 'Utsur^omiya
^—\ MaebashiC . ,^
} \ To^h^gi
^ ,-'^^-^V'- ^ShimodJ,te
Y
• KumagVya
(N._
• UrWa
Kawasaki
Yokohama
Fig. 2 Surveyed places
PROCEEDINGS—PAGE 143
-------
(ATTACHED SHEET) Acid rain research flow
Survey of
actual conditions
Examination of
production system
Examination of
influences
Forecast and
evalution
Preventive
Measures
Atmosphere
system
Land
water
system
Soil
system
Whole
Measuring the water
quality of precipi-
tation throughout
the country (pH,
electrical conduc-
tivity, etc.)
Measuring the qual-
ity of land water
in rivers, lakes
marshes, etc. in
mountainous areas,
etc. throughout the
country (pH, elec-
trical conductivity,
metal ions, etc.)
Clarifying the drift,
diffusion and setting
rates or pollutants,
and the processes of
absorbtion into rain-
water, etc.
Studies to clarify
the acidifying me-
chanism of land
water in model
lakes, marshes,
etc.
Studies to clarify the changes in soil
properties due to acid rain in fixed places
(pH, Ex-Ca, active aluminium, etc.)
Studies on the eco-
system of land water
of rivers, lakes,
marshes, etc. in
mountainous areas
throughout the
country (numbers
and species of fish,
benthos)
Studies on the in-
fluences on soil
ecosystem and
vegetation
Review of existing literature (domestic and abroad)
n
o
f»
en
rt
H«
3
00
B
P.
01
3
00
(0
3
M>
i-"
1
O
(D
(O
01
O
l-l
0>
h"
rt
O
M
H-
00
H
(0
i
rt
§
01
CO
C
fl>
(0
rt
fl>
3
01
00
01
M
rt
00
3
n>
^
01
rt
H-
O
3
CO
O
l-l
O
(D
M
-------
PROGRESS IN PHOTOCHEMICAL AIR QUALITY
SIMULATION MODELING
presented by K.L. Demerjian
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 145
-------
PROGRESS IN PHOTOCHEMICAL AIR QUALITY SIMULATION MODELING
J.H. Shreffler, K.L. Schere, and K.L. Demerjian1
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711 U.S.A.
INTRODUCTION
Over the past 15 years we have witnessed the development of
atmospheric dispersion models which include increasingly complex
approximations of photochemistry. Unlike existing Gaussian or
analytic schemes, the models simulate atmospheric chemical reac-
tions, some of which act at very rapid rates, and therefore involve
time steps much shorter than are generally regarded as necessary
for transport and dispersion of inert species. The computational
demands rise rapidly with the number of chemical species considered,
and it is not unusual to see present-day models for urban and
regional scales having computer simulation speeds comparable to the
real-world events. Computer restrictions have tended to engender
decisions to diminish the spatial resolution of the models in
order to increase the number of species and reactions treated.
We have recognized for a number of years that within the class
of existing or conceived photochemical air quality simulation
models (PAQSM's) there are only several basic approaches. The
United States Environmental Protection Agency (U.S. EPA) in the
mid-1970's reviewed the various urban scale models which were in
existence and chose three, embodying distinct approaches, for
further refinement, development, and evaluation. In the case of
the EPA, the ultimate goal was to provide regulatory tools for
control of photochemical pollutants, 03 in particular. The models
authors are on assignment from the National Oceanic and
Atmospheric Administration, U.S. Department of Commerce.
PROCEEDINGS—PAGE 147
-------
were to be. used to predict the problems created by new sources or
the benefits of controlling existing sources. Shortly thereafter,
work also began on a regional-scale (1000 km) photochemical model
to be used initially in the northeastern U.S. Interest in larger
scales derived from a growing awareness that elevated 03 levels
may be found over large regions and at considerable distances from
urban centers (Wolff et al-, 1977). The goal of the regional
modeling effort was again to allow defensible regulatory decisions,
but the complexity of the problem and its immense resource require-
ments precluded study of several alternate approaches as was done
with the urban models.
The purpose of this paper is to describe the PAQSM's emerging
from research and development projects of the U.S. EPA. The pre-
sentation will not attempt a detailed discussion of chemical mech-
anisms or numerical schemes. Rather, emphasis will be placed
on the models' basic structures, data requirements, computer re-
quirements, and problems encountered in applying them. Also, we
will describe the two major field programs which have been con-
ducted by EPA to support testing and evaluation of the models.
COMPONENTS OF MODEL SIMULATIONS
PAQSM's may differ,in complexity and therefore data and re-
source requirements through deliberate choices of the builders.
For example, restriction to a simple box-model makes spatial re-
solution of emissions and winds unnecessary. Furthermore, data
requirements will vary greatly depending on the goal of the simula-
tions* If model development and evaluation are planned, then the
data needs are extensive. On the other hand, if the model is
already viewed as a reliable operational tool, data needs are
greatly reduced as climatological scenarios may be used for
worst-case and average events- In the latter case, predicted
pollutant levels are only checked for reasonableness against pre-
vious records and not against specific observations. Photochemical
air quality simulations-share similar attributes, and requirements,
regardless of the particular model used. Generally, all PAQSM's
are logical frameworks which synthesize; information from meteor-
ology, sources, and air quality and predict the photochemical
pollutant concentrations consistent with expressions for governing
physical laws. The simulation is therefore an interplay of four
equally important components.
PHYSICAL LAWS: Physical laws are expressed as mathematical equa-
tions governing atmospheric motions (transport and diffusion) and
the chemical kinetics. The equations enforce mass conservation
and promote appropriate chemical transformations in a manner
consistent with theoretical considerations and observational data.
PROCEEDINGS—PAGE 148
-------
Kinetic mechanisms generally have been developed by comparing
numerical predictions to concentrations resulting from carefully
controlled smog chamber experiments. It is important to note that
equations of both 'the meteorological and chemical aspects of the
problem are approximations. For example, mixing rates are usually
difficult to prescribe, and hydrocarbon (HC) species must be lumped
in terms of reactivity classes.
SOURCE EMISSIONS: Pollutant emissions are an important component
in the application of a PAQSM and must be compiled on spatial and
temporal scales comparable to the resolution inherent in the model
structure. Photochemistry involves a strong diurnal cycle which
dictates a temporal resolution on the order of 1-h in the emis-
sions inventory. Inventories usually are divided into the source
types of area, point, and line. Area sources include many diffuse,
population-related sources such as those associated with space
heating or residential automotive traffic. Point sources are
large, identifiable sources such as power plants, and line sources
refer to major automotive thoroughfares. A typical inventory for
a large city may include several hundred point sources. Unless
the model has the unusual capacity to treat line sources, they may
be incorporated into the area source inventory. For PAQSM's the
emphasis is usually on developing emissions for HC, NOX, and CO,
although mechanisms generally include other species such as S02
and SO^. The HC emission is broken into reactivity classes con-
sistent with the kinetic mechanism (e.g., olefins, paraffins,
aromatics, formaldehyde, and other aldehydes). Also, the NOX
must be appropriately split between NO and N02-
METEOROLOGICAL FACTORS: Meteorological factors account for the
transport and diffusion of pollutants and influence reaction rates
in the chemical mechanism. These data are introduced into the
model from observations and are not predicted. Important param-
eters include the wind field, inversion height, atmospheric sta-
bility, solar radiation, water vapor, and temperature.
AIR QUALITY MEASUREMENTS: Ambient concentration measurements are
needed to set initial conditions, boundary conditions, and as
evaluation data. Photochemical simulations usually begin with an
initial set of observations used to assign concentrations of impor-
tant 03 precursors, HC and NOX, within the modeled domain. As
an alternate strategy, the model could generate its own initial
state by allowing sufficient time for emissions to establish suit-
able concentration levels. However, this strategy is usually
impracticable when considering other aspects of the problem such as
ventilation rates, especially on the urban-scale. Boundary condi-
tions refer to the concentrations assigned at the extremities of the
modeled region, both in the horizontal and vertical. Concentrations
at the upwind boundary of a grid-model domain are of obvious impor-
PROCEEDINGS—PAGE 149
-------
tance. Moreover, the concentrations aloft may also play a role
as they are entrained during the diurnal growth of the mixed layer.
Ambient data must have a spatial and temporal resolution at least
comparable to the model output if evaluation of the model is to be
attempted-
DRBAN MODELa
The U.S. EPA currently has interest in three urban models
embodying distinctly different approaches and levels of complexity.
Those models are:
Photochemical Box Model (PBM) - a single cell Eulerian model
constructed by EPA.
Lagrangian Photochemical Model (LPM) - a multi-level parcel
model originally developed by Environmental Research and
Technology, Inc.
Urban Airshed Model (UAM) - a multi-level, Eulerian grid model
originally developed by Systems Applications, Inc.
The versions of the models at EPA have been structured to easily
use urban data compiled during the Regional Air Pollution Study
(RAPS) and have been subject to continuing modifications. The
RAPS will be discussed in a later section. In this section, the
basic frameworks and requirements of the models will be surveyed.
Photochemical Box Model
The Photochemical Box Model (PBM) is a single cell Eulerian
air quality model whose purpose is to simulate the transport and
chemical transformation of air pollutants in smog prone urban
atmospheres. The model's domain is set in a variable-volume,
well-mixed reacting cell where the physical and chemical processes
responsible for the generation of 63 by its HC and NOX precursors
are mathematically created - These processes include the transport
and dispersion of pollutant species through the cell, the injection
of primary precursor species by emission sources, and the chemical
transformation of the reactive species into intermediate and second-
ary products. They are schematically illustrated in Figure 1. In
a typical application of the model, the horizontal length scale of
the single cell is. about 20 km and the vertical scale is time-
varying, equivalent to the depth of the mixed layer. The model
domain is centered on the city such that the area encompasses most
of the major emissions sources. Source emissions are assumed to
be distributed uniformly across :the surface face of the cell.
PROCEEDINGS—PAGE 150
-------
Fig. 1. Schematic of the domain and processes in the Photochemical
Box Model.
While the data requirements of the PBM are not particularly
rigorous, three data preprocessors must be executed prior to each
model simulation. The first accesses the 1-h average air quality
and meteorological data base forming the initial concentrations of
participating species, determining the 1-h average wind vectors
that guide transport through the model domain, and choosing the
hourly updated concentration of species at the upwind boundary of
the box- The first preprocessor also forms similar spatial aver-
ages of concentrations of monitored pollutant species for compari-
son with model predictions.
A function of the second preprocessor is to determine 10 min
average values of total solar radiation from pyranometer measure-
ments. From these values the time-varying photolytic rate con-
stants can be determined through a series of empirical and theo-
retical relationships. The temporal resolution of these data is
greater than for most other data because of the rapid chemical
reactions which are associated with these rate constants. This
preprocessor also calculates the diurnal growth pattern of the
mixed layer, or the depth of the model domain. To perform this
calculation, the morning minimum and afternoon maximum mixed layer
heights are provided by the user- These heights were determined
in the RAPS application by studying the vertical structure of the
temperature and moisture from radiosondes. Releases occurred
PROCEEDINGS—PAGE 151
-------
throughout the day at 6-h intervals beginning 1 h before sunrise.
The model fits a piece-wise linear function between the specified
minimum and maximum depths that simulates the actual mixed layer
growth as observed by acoustic sounder, lidar, and radiosonde
techniques• Ten-minute averages of mixed layer height are formed
for the PBM; photolytic rate constants for the corresponding time
period are integrated through this depth.
The third preprocessor is responsible for forming hourly
emissions source terms for CO, NO, N02, and non-methane HC. Even
though the emissions may have spatial resolution, all source emis-
sions, both area and point, within the model domain are summed to
provide a total emission rate for each primary pollutant species
at every hour of simulation.
The PBM may be executed when the data files from the three
preprocessors have been created. The only other relevant informa-
tion needed at this time is the concentration of 63 at the top
boundary of the modeling domain, that pollutant having been trans-
ported into the area by winds aloft during the night. This concen-
tration is determined from the 03 measurements at the far upwind
surface monitoring sites after the nocturnal temperature inversion
has eroded and the air aloft mixes down through the atmosphere to
the surface. Concentrations of other pollutants at the top of the
modeling region are assumed to be negligible. The simulation is
typically started at 0500 CST and continues through 1700 CST.
The PBM contains a chemical kinetic mechanism with 36 reac-
tions and 27 reactive species (Demerjian and Schere, 1979). The
set of equations describing the rates of change of the concentra-
tions of these species is numerically solved at time steps on the
order of 10 min- From these solutions the model determines the
1—h average predicted concentration of each modeled species.
Lagrangian Photochemical Model
The Lagrangian Photochemical Model (LPM) was developed by
Environmental Research and Technology, Inc. of Westlake Village,
California and adapted under contract with EPA for use with the
RAPS data base. (The LPM is essentially identical to the general-
use model named ELSTAR.) The LPM envisions a portion of the atmos-
phere as an identifiable parcel which can be tracked from early
morning to the late afternoon. As the parcel moves over the
various emissions sources, pollutants are assimilated, vertically
mixed, and subjected to photochemical reactions in the presence
of solar radiation (see Figure 2). The LPM is attractive relative
to grid models in that it is fairly simple to execute and uses a
moderate amount of computer time. On the other hand, the LPM
PROCEEDINGS—PAGE 152
-------
SOLAR RADIATION AND
TEMPERATURE IS GIVEN
ASA FUNCTION OF TIME
TIME-DEPENDENT ATMOSPHERIC
MIXING AND CHEMICAL REACTION
IS COMPUTED FOR AIR PARCEL UP
TO THE MIXING HEIGHT h
SPACE/TIME TRACK
THROUGH THE SOURCE
GRID IS DERIVED
FROM WIND DATA
POLLUTANT INFLUXES AT ANY
ELEVATION ARE IMPOSED BY THE
EMISSION SOURCE FUNCTIONS
Fig. 2. Schematic of the modeling concepts of the Lagrangian
Photochemical Model.
calculates concentrations only within a parcel and not over a
complete spatial field-
The LPM is executed using a series of program modules. They
are METHOD, EMMOD, and KEMOD, sequentially performing calculations
on meteorology and air quality, emissions, and photochemistry. The
input and running procedures described by Lurmann et al• (1979) have
been generally followed, although some modifications were deemed
necessary as more experience with the model was acquired (Lurmann,
1980; 1981).
The first step in setting up a simulation is to determine the
starting point of a parcel so that it will arrive at a specified
point at an assigned time. A backward trajectory can be generated
by the METHOD. The parcel trajectory is usually determined by
1/R^ weighting of winds from the closest three stations. However,
experience with wind data suggests that even closely situated
stations can show large, unexplained differences in wind vectors
from time to time- Reliance on a single anomalous station, if the
parcel has a close approach, can create an erratic trajectory. To
eliminate the possibility of such vagaries, it may be prudent to
compute a single resultant wind vector from the wind station network
and assign it to all stations for a given hour. The resultant gives
a general movement of the air mass over the region.
PROCEEDINGS PAGE 158
-------
Because of the difficulty in solving the chemistry set with
minimal solar radiation, the start time of a parcel (which is on
the hour) must be at least 10 min past local sunrise- Once the
start position is set, the METHOD is run in a forward-trajectory
mode until 1800 CST or the parcel leaves the region.
Mixing heights are computed from radiosondes released approxi-
mately 1 h before sunrise and at 6-h intervals thereafter. The
temperature and wind profiles from the balloons are processed
automatically by the model to give vertical eddy diffusivities
throughout the day. The trajectory information from METHOD is
stored and is available to EMMOD to enable that module to obtain
the emissions that the parcel will encounter. METHOD also stores
the observed pollutant concentrations along the trajectory using a
1/R.2 weighting scheme on the closest three stations. These
observations are available for setting initial conditions in the
parcel and later comparisons with the model predictions• In most
cases, the parcel starts in a relatively clean rural environment,
and levels of HC, NO, and NO2 are assumed to decrease with height
according to an assigned "formula. On the other hand, ozone is
depleted near the ground at night; thus, the initial 03 increases
from the surface to a value at 400 m which is equal to the observed
1000-1200 CST surface concentration upwind of the city (Shreffler
and Evans, 1982).
The EMMOD creates a record of the emissions entering the par-
cel as it traverses its trajectory. These emissions are used by
the KEMOD to simulate the photochemical reactions which will take
place as a function of time, yielding concentrations of 39 chemical
species at 30-min intervals. A-total of 65 reactions are modeled.
Urban Airshed Model
The .Urban Airshed Hodel (UAM) is a three-dimensional (3-D)
grid-type, or Eulerian, PAQSM developed by Systems Applications,
Inc. (SAI) of San Rafael, California. The structure of the model
consists of a latticework array of cells (see Figure 3), the total
volume of which represents an urban-scale domain and in which the
physical and chemical.processes responsible for photochemical
smog are mathematically simulated. These processes include the
advection of.pollutant species through the modeling domain, the
species entrainment from aloft by a growing mixed layer, the diffu-
sion of material from cell-to-cell, the injection of primary source
emissions into- the modeled volume, and the chemical transformations
of reactive.species into intermediate and secondary products.
The horizontal dimensions of each cell are constant but the heights
of the cells vary throughout a model simulation as the depth of
-the mixed layer in the UAM changes accordingly.
PROCEEDINGS—PAGE 154
-------
Fig. 3. Schematic of the grid domain used in the Urban Airshed
Model.
In typical applications, the area modeled is about 60 x 60
km, and each individual cell is 4 km on a side in the horizontal.
Vertically, there are four layers of cells in total; the bottom
two layers simulate the mixed layer and the top two represent the
region immediately above the mixed layer. The 3-D grid model is a
sophisticated type of PAQSM and provides both spatially and tempo-
rally resolved concentration predictions. Thus, the UAM attempts
to estimate the 1-h average observed concentration of a pollutant
species at each monitoring site within the model domain.
The package of computer programs constituting the DAM actually
contains 12 data preprocessing programs as well as the simulation
model. The data requirements for applying the model ar« rather
intensive. The preprocessors (PP's) access surface-based, hourly
air quality and meteorological data base, the upper air pibal and
radiosonde data, and the source emissions Inventory for the neces-
sary parameters, and process the parameters as required by the
simulation model. A brief description of the PP's follows in order
to convey a sense of the UAM data requirements.
The chemistry PP sets up the rate constants and other param-
eters related to the chemical kinetic mechanism within the model.
This mechanism, named Carbon Bond II, is a state-of-the-art set of
chemical reactions describing the NOX-HC-03 species interactions
PROCEEDINGS—PAGE 155
-------
within a photochemically active atmosphere. The mechanism was
developed at SAI by Whitten et al. (1980). The terrain PP de-
scribes the spatially varying types of surfaces within the modeling
domain, reflecting the transition from urban to suburban and final-
ly to rural land uses. The spatial resolution of these land forms
should be on the same scale as the grid cell size. Both the chem-
istry and terrain PP's need only be executed once for a particular
model application. All of the remaining PP's, however, must be
executed prior to each model simulation within the application.
The diffusion-break and region-top PP's describe the varia-
tions in the mixing heights and modeling region depths, respective-
ly, over the domain during the course of the simulation. Hourly
values of the mixing height must be supplied at representative
locations on the grid- It is the responsibility of the user to
specify these depths in a meaningful manner from upper air soundings
or equivalent data* The meteorological PP describes the temporal
variation of vertical temperature gradient, stability class, atmos-
pheric pressure, water vapor concentration, and the N02 photo-
lysis rate constant. Values for these parameters do not vary
spatially in the model- The top-concentration PP specifies the
concentrations of principal species at the top of the modeled
region throughout the simulation. In most cases these concentra-
tions will be close to the clean air background values for these
species, although substantial concentrations of 03 can result
from advection over the region. Its value is determined from the
03 measurements at the far upwind surface monitoring sites after
the nocturnal temperature inversion has eroded and the air aloft
mixes to the surface.
The air quality, temperature and wind speed PP's all require
data from a surface monitoring network. Hourly averages of observed
species concentrations are objectively interpolated across the model
grid by the air quality PP to produce a field of initial concentra-
tions for the UAM. Typically this initial field is applied near
sunrise. As the density of monitoring locations increases, the
reliability of the objective interpolation does likewise. The
temperature-PP produces gridded fields of surface temperature at
each hour of model simulation. These are required for both the
wind and point source PP's. Finally, the wind PP assimilates all
available surface and upper air wind speed and direction measure-
ments and produces a gridded field of u and v wind components at
each of the model's four vertical levels. This objective wind
field analysis routine was developed by Anderson of SAI (Killus et
al-i 1977). Vertical wind velocities are calculated internally by
the UAM from mass continuity considerations. Anderson's interpola-
tion routine smooths the data in such a way so as to eliminate
unrealistic vertical motions. It simulates the mesoscale urban
circulation patterns through the use of the surface temperature
PROCEEDINGS—PAGE 156
-------
patterns. The boundary concentration PP determines the modeled
species concentrations along each of the domain's borders for
each hour of model simulation. A vertical concentration gradient
is also described.
The last two PP's needed before the model is executed deal
with source emissions. The area sources, including highway and
line sources, and the point sources are treated separately. Hourly
emission rates of all primary pollutants for each grid are calculated.
Organic HC emissions must be distributed into particular structural
classes.
Each PP generates a data file required by the UAM. Model
simulations begin at 0500 CST and continue through 1700 CST. The
model numerically calculates the rates of change of species concen-
trations at time steps on the order of several minutes, and from
these determines the 1-h average predicted concentrations.
Status of the Urban Models
The three models discussed in this section are viewed by EPA as
being in final form. That is, substantial breakthroughs in chemical
mechanisms or the compilation of a clearly superior data base will-
be needed before revisions would be considered. Although the model
preprocessors are geared to easy use of the RAPS data, it would not
be difficult to adapt the models to other locations and other forms
of data. In the next section the RAPS data set will be described
along with some results of model applications. Before proceeding,
we review some problems which arose during the testing of the models
and look at estimated resource requirements.
Because of its relative simplicity, the PBM is attractive for
use where resources are limited and results do not need to be highly
detailed. With this model, it should be noted that emissions
control strategies are difficult to target on particular sources
because of the lack of spatial resolution. Also, the highest 03
concentrations in the domain, upon which decisions may be more
critically centered, are lost in the averaging assumptions. Final-
ly, the simplified wind field (a single vector for each hour) and
large size to the Eulerian domain make the model best suited for
low-wind or stagnation episodes. These, of course, are usually
conditions most favorable to high 03 concentrations.
Compared to the UAM, the LPM may seem less resource intensive.
However, the data requirements are nearly identical. The LPM may be
used along a single trajectory, perhaps that trajectory encountering
the observed maximum concentration, and therefore use much less
computer time. If multiple trajectories are executed to display
PROCEEDINGS—PAGE 157
-------
concentrations over the region, such as the UAM does in a single
simulation, then computer resources and preparation time will grow
accordingly.
A number of problems have become apparent in the development
and testing of the models. For example, the LPM seems quite sensi-
tive to initial conditions since the parcel retains all pollutants.
Thus, care must be taken to put realistic vertical distributions of
precursors in the parcel at the beginning of the simulation day.
The LPM has five vertical levels in which to distribute the pollut-
ants. As originally conceived the model did not include any lateral
diffusion, and this produced unrealistically high 03 concentrations
since all emissions were retained in a fixed volume. The model now
allows the parcel to expand laterally in a manner commensurate
with empirically determined diffusion. Whereas this feature con-
siders dilution of a parcel when beyond the high emissions area,
it introduces a problem in setting side boundary condition concen-
trations representative of the air that is entrained into the
parcel. Our approach has been to set the initial parcel size
about equal to the downtown area of highest emissions (5 x 5 km for
RAPS) and treat the entrained air as having background values of HC
and NOX. The parcel is viewed as a segment of the urban plume.
However, large power plant sources are liable to present conditions
seriously in conflict with that assumption from time to time. This
example evinces a fundamental truth about all of the PAQSM's; there
is an element of art in their application which is unlikely to be
eliminated without jeopardizing the models' flexibility in treating
different situations.
The. UAM as originally constructed suffered from substantial
numerical diffusion. We believe this potential problem should be
thoroughly investigated in any grid model of this sort. The tend-
ency of the DAM to underestimate ozone maxima in our tests (see
results in next section) was suspected of being due to the spurious
diffusion. However, the underestimation was only slightly improved
by a new and accurate numerical scheme, which, incidentally, in-
creased computer time by 15 percent (Schere, 1982).
The choice of which particular model to use in a specific
application involves not only the accuracy of the model but also the
resources required to operate it. The models discussed here have
resource requirements correlated with their level of complexity. In
terms of man-months needed to set up a single day simulation and
computer time expended (minutes of CPU on a Univac 1100/82) the
approximate requirements are:
PBM —, 0.15 man-month 1 minute CPU
LPM —-— 0.20 man-month 10 minutes CPU
UAM — 0.50 man-month 110 minutes CPU
PROCEEDINGS—PAGE 158
-------
The initial man-month investment needed to become adequately
familiar with the model perhaps exceeds the simulation set-up time
by at least a factor of 10.
The three models discussed here have all shown themselves to be
acceptable tools for analysis of urban 03 air quality. The specific
configuration of an application along with the quantity and quality
of related data and resources available to the user must all be
considered in the final selection of a model. For an indication
of average 63 air quality in an urban area under stagnation condi-
tions or as a screening method for a more complex model, the PBM is
appropriate. The choice of a trajectory model, such as the LPM,
or a grid model, like the UAM, might well be decided by resource
requirements or by the number of proposed simulations. In any
event, the user of any of these models must have a strong scientific
background and exercise extreme care in implementing the air quality
simulations.
THE RAPS AND URBAN MODEL EVALUATION
Beside the physical laws expressed in the PAQSM, essential
components of air quality simulations include emissions, meteorol-
ogical and ambient concentration data. In this section we describe
an extensive field effort for gathering data on an urban scale and
some of the modeling results based on that data.
The Regional Air Pollution Study (RAPS) was carried out in the
St. Louis area during 1974-1977- St. Louis is a city of negligible
topographic relief in the central U.S.; about 2-3 million people
reside in the metropolitan area. Although augmented by a variety of
special studies, the principal data base was gathered by a 25
station network comprising the Regional Air Monitoring System
(RAMS). Measurements included both meteorological parameters and
pollutant concentrations. The purpose of RAPS was to develop a data
base for testing numerical air quality simulation models and
especially models involving the complex photochemical reactions
leading to 03 formation. The RAPS expenditures totaled about $25
million over a five-year period, and the data, base remains the most
comprehensive available.
RAMS measurements, recorded as 1-min averages, included wind
speed and direction, temperature, vertical temperature difference,
solar radiation and concentrations of 03, CO, HC, NOX and S02-
The sampling manifold used by all gaseous monitors had an intake
at about 4 m from the ground. Figure 4 shows the layout of the
RAMS network. A more detailed description of the RAPS program-and
the RAMS instrumentation is given by Schiermeier (1978).
PROCEEDINGS--PAGE 159
-------
Fig. 4. The St. Louis area with locations of the RAPS stations
101-125.
The 1-min average values of all parameters were objectively
screened. The screening excluded values which were null, part of a
calibration set, taken during a period of excessive drift, outside
instrument limits, taken under abnormal instrument status, or
indicating persistence. The remaining data were used to form 1-h
averages, and the 1-h average archive was generally the one used in
model evaluation.
Overviews of the RAPS 03 and NOX data are given by Shreffler
and Evans (1982) and Shreffler (1982). Figure 5 displays a 4-rao
record of ozone taken in mid-1975. The time series are for the
inflow concentration (that from upwind rural areas) and the maximum
1-h concentration recorded at any station in the network. The
locations of the maxima, especially the strong peaks, usually show
PROCEEDINGS PACK
-------
I
m"
o
N
O
4/30
Fig. 5.
5/20
6/9
6/29
7/19
8/8
8/28
Time series of the daily maximum 1-hr ozone concentration
(solid) and the daily rural inflow ozone concentration
(dashed) for May - August 1975 as measured in the RAPS.
clear downwind relations to high emissions areas- The differences
between concentrations in the two series may be ascribed to net
enhancement of 63 levels due to the urban emissions.
In preparation for model evaluation, the entire data base was
examined to establish suitable case-study days. In all, 20 days
were selected for simulations based on high observed 63 maxima and
adequate availability of data. The maximum observed 03 on these
days ranged from 160 to 260 ppb •
The three urban models (PBM, LPM, and UAM) were adapted to
access the RAPS data archive, which included a detailed emissions
inventory with resolution to 1 h and 1 km (Littman, 1979)• Al-
though precise input requirements differ among the models, effort
was made to supply information in a consistent form for all.
Thus, the simulations were done in parallel, using the same basic
data sets, and the results give a comparison among models as well
as an indication of accuracy against observed concentrations.
As examples of model results, Figures 6 and 7 show observed
and predicted 03 values for the PBM and LPM on a single day, 1
October, 1976. Stagnation conditions prevailed on that day. For
the PBM, the time series show concentrations within a fixed box
20 km x 20 km over the city center. Maximum and minimum observed
PROCEEDINGS—PAGE 161
-------
PBM SIMULATION-761001
a'
Q.
ro
O
oo
10.0 12.5 19.0
TIME, HOURS (CST)
17.5
20.0
Fig. 6. PBM predicted ozone (solid) compared to measured domain
average (circles) and maximum and minimum within the
domain (dashed). This result is for the 1 October 1976
RAPS simulation.
concentrations over all the stations (13) in the box are given by
the dashed lines. The average observed concentration is given by
circles, and the model prediction is given by the solid line- For
the LPM, the concentrations refer to a moving parcel with initial
horizontal dimensions 5 km x 5 km. Predicted values are given for
two vertical levels in the parcel, the surface (L-l) and about 200 m
(L-3). This particular parcel arrives at the station (102) ob-
serving the maximum 1-h 03 value at the time of that observation
(1400 CST).
The ability of the models to reproduce observed 03 maxima
over the 20 test days is summarized by statistics given in Table 1.
Complete results of the urban model evaluation are now available in
report form (Schere and Shreffler, 1982).
PROCEEDINGS—PAGE 162
-------
.4
Q.
Q.
o
M
O
.1
0.0
76275. RAMS 102 AT 1400CST. START 0700
I | I I I I
OBS
PRED L-1
PRED L-3
i i i i
i i i
j ii i i [ill i
\
&JO 7Jt> 10.0 12.5 15.0
HOUR, CST
17.5
20.0
Fig. 7. LPM predicted ozone at the surface (L-1) and 200 m (L-3)
compared to the observed ozone. This result is for the
1 October, 1976 RAPS simulation.
Table 1. Statistics on residual concentrations (observed minus
predicted) of maximum ozone (ppb) from 20 days of RAPS
data- The observed maxima ranged 160-260 ppb-
PBM
LPM
UAM
AC
s.d.(AC)
TAcT
-12
39
29
5
58
50
62
35
62
PROCEEDINGS—PAGE 163
-------
A REGIONAL SCALE (1000 KM) MODEL
The U.S. EPA is presently developing a model that can guide
the formulation of regional emissions control strategies by esti-
mating the effect of sources on concentrations in remote regions,
determining the pollution burden that cities impose on distant
neighbors, and eventually analyzing the effect of emissions on
acid rain, visibility and fine particles. The utility and credi-
bility of the model will be determined primarily by the extent to
which it accounts for all the governing physical and chemical
processes* Accordingly, the model is formulated, in principle, to
treat all of the chemical and physical processes that are known,
or presently thought, to affect the concentrations of air pollut-
ants over several day/1000 kilometer scale domains. Among these
processes are (not necessarily in order of importance):
1. Horizontal transport;
2. Photochemistry, including the very slow reactions;
3. Nighttime chemistry of the products and precursors of photoc-
hemical reactions;
4. Nighttime wind shear, stability stratification, and turbulence
"episodes" associated with the nocturnal jet;
5. Cumulus cloud effects - venting pollutants from the mixed
layer, perturbing photochemical reaction rates in their shadows,
providing sites for liquid phase reactions, influencing changes
in the mixed layer depth, perturbing horizontal flow;
6. Mesoscale vertical motion induced by terrain and horizontal
divergence of the large scale flow;
7. Mesoscale eddy effects on urban plume trajectories and growth
rates;
8. Terrain effects on horizontal flows, removal, diffusion;
9. Subgrid scale chemistry processes resulting from emissions
from sources smaller than the model's grid can resolve;
10. Natural sources of HC, NOX and stratospheric 03;
11- Wet and dry removal processes, e.g., washout and deposition.
Of the eleven processes listed above, only the first and last
have been treated in any detail in the regional scale models of air
pollution developed to date. In fact, a review of these models
(see, for example, reviews by Drake and Bass in Henderson et al.
PROCEEDINGS—PAGE 164
-------
1980) reveals that virtually all of the Eulerian type models are in
essence simply expanded urban scale models. They account for the
physical processes that are active during daylight hours and within
10 km or so of a source, but they neglect both the processes that
are important beyond this distance and those that are active at
night.
The U.S. EPA has taken the approach of developing a truly re-
gional model, allowing the processes described above to influence
the structure of the model, rather than trying to force the pro-
cesses into an existing urban structure. The original goal of this
work was to develop a specific model of regional scale photochemical
air pollution. However, as the work progressed and new developments
and ideas continually emerged, the need was seen for a general
modeling framework within which the various physical and chemical
processes that play important roles could be treated in modular
form- This would permit ongoing incorporation into the model of
state-of-the-art techniques without the need to overhaul the model
each time. The structure and modular form of the Regional Oxidant
Modeling System (ROMS) are unique for studying regional scale
pollution. Complete documentation of the ROMS is being prepared
(Lamb, 1982a, 1982b).
When this model development work was initiated some four years
ago, an attempt was made to derive from the observational evidence
available at that time an estimate of the minimum vertical and
horizontal resolutions necessary to describe regional scale air
pollution phenomena. The aim was to arrive at the best compromise
between the restrictions imposed upon the model by computer time
and memory limitations and the need to describe as accurately as
possible all of the governing processes cited above. Careful
review was made of the NO, 03 and meteorological data reported
in Siple (1977) by the participants of the 1975 Northeast Oxidant
Transport Study. The characteristics of the 03 distribution de-
scribed in those reports would require at the very least a threer-
level model—one level assigned to the surface layer, another
level to the remainder of the daytime mixed layer, and an additional
layer atop the mixed layer. The top level would be used in conjunc-
tion with the mixed layer to account for downward fluxes of stratos-
pheric 03 as well as upward fluxes of 03 and its precursors into
the subsidence inversion layer above. Material that entered this
top layer could be transported by winds aloft to areas outside the
modeling region; it could enter precipitating clouds and be rained
out of the atmosphere; or it could undergo chemical transformation.
Representing the subsidence inversion, where cumulus clouds often
form under stagnant high pressure conditions, the top level of the
model would be instrumental in simulating the chemical sink effect
of heterogeneous (within cloud droplets) reactions among 03, its
precursors, and other natural and pollutant species. Including
PROCEEDINGS—PAGE 165
-------
cloud effects in the model would be especially important in future
simulation of S02 and sulfates.
Having three layers in a model is insufficient in itself to
simulate all relevant phenomena. For example, three layers of
constant thickness are incompatible with the spatial and temporal
variability that the radiation inversion and mixed layer thickness
are known to have. What is needed in the model is three "dynamic"
layers that are free to expand and contract locally in response to
changes in the phenomena they are intended to treat. The model
discussed here possesses this property. Figure 8 illustrates the
vertical structure of this model and the physical phenomena that
each layer is intended to simulate. The surfaces that comprise the
interfaces of adjacent layers in our model are variable in both
space and time in order that each layer can keep track of the
changes that occur in the particular set of phenomena that layer is
designed to describe -(summarized in Figure 8). A consequence of
this structure is that the volumes of the grid cells vary in both
space and time. By contrast, in conventional models the grid
network and cell volumes remain fixed and surfaces such as the mixed
layer top move through the grid system- In the following several
paragraphs, we elaborate on some of the phenomena cited in Figure 8
that our model will take into account.
During the day the highest layer shown in Figure 8a represents
the synoptic scale subsidence inversion, which may or may not
contain cumulus clouds. Stratospheric 63 is transported downward
through this layer and anthropogenic 63 and its precursors can be
carried into it by cumulus clouds or penetrative convection. The
base of this layer is normally 1 to 2 km above ground level- Below
it pollutants are kept well mixed vertically by turbulent convec-
tion. If the winds -are strong or the surface heat flux is weak,
wind speed and direction may vary appreciably within the first
several hundred meters above ground. There is usually also a marked
difference in the wind speed and direction between the inversion
layer and the mixed layer below. Over large lakes and along sea
coasts there is frequently a second inversion layer below that
generated by synoptic scale subsidence. This lower inversion is
produced by sea or lake breeze regimes, and it restricts the verti-
cal mixing of pollutants emitted over the water and within several
kilometers inland from the water's edge.
Air drawn into young cumulus clouds originates primarily in
the lower portion of the mixed layer. Fresh emissions of NOX and
HC can be transported by the vertical currents that feed these
clouds from ground—level to altitudes well above the top of the
mixed layer in one steady, upward motion. In the process little
or no mixing with aged pollutants in the mixed layer occurs. At
night, cumulus clouds usually evaporate, and when they do they
PROCEEDINGS—PAGE 166
-------
r Function!
Llyt' 3
2 U(«**'O ti*"tpo*t b*
ent f«ftfl l*«itpo««k ««| ph*M (h«NW|l
l.y~ 0
Fig. 8. Schematic illustration of the dynamic layer structure of
the regional model and phenomena each layer is designed
to treat: (a) daytime, (b) nighttime.
PROCEEDINGS—PAGE 167
-------
leave behind products of liquid phase reactions that can be trans-
ported hundreds of kilometers before sunrise.
Dramatic changes occur in the mixed layer at night. With the
onset of surface cooling following sunset, a stable layer of air
forms near the ground that quenches the vertical momentum fluxes
that give rise to frictional drag on the horizontal flow. With
retardation forces eliminated, the wind just above the stable layer
accelerates giving rise to the phenomenon known as the nocturnal
jet. Wind speeds in the core of the jet, which usually lies between
300 and 500 meters above ground, may be 10-15 m/s while at the
same time the air is nearly calm at the surface. Emissions from
tall stacks and from sources within the urban heat island enter the
jet region at night. There they react with aged pollutants from the
previous day and are transported considerable distances by the
strong flow. The remnant of the previous day's mixed layer above
the jet is isolated from the influence of fresh emissions and it
moves at a slower speed than air below.
Sporadic episodes of turbulence in the shear layer beneath the
nocturnal jet are a mechanism by which 03 and constituents of urban
plumes are brought to ground-level at night. There, deposition on
surfaces and reactions with emissions of small, low-level sources
occur. This sporadic mixing process is perhaps the only mechanism
by which the reservoir of aged pollutants aloft can be depleted at
night.
One point that we wish to emphasize here is that one-layer
regional scale air pollution models are incapable of simulating
the effects on pollutants like 03 of the vertical segregation of
aged and fresh emissions that occurs at night. Being cut-off from
contact with the ground and fresh NOX emissions, 03 above the
nighttime radiation inversion is free to travel great distances
before it is mixed vertically by convection the following day.
The effect of this nighttime segregation of pollutants is to extend
greatly the effective residence times of species like 03 in the
lower troposphere. Consequently, a multi-layered model seems to be
essential to simulate accurately the long range transport of photo-
chemical air pollutants.
As now planned, the horizontal resolution of the model is about
18 km. The resolution should be as high as possible to mitigate
the effects of subgrid scale concentration fluctuations. A scheme
to treat subgrid chemistry is implemented in Layer 0, adjacent to
the ground. Layer 0 is treated diagnostically in the governing
equations, and also handles surface depletion. In modeling atmo-
spheric processes over 1000 km scale regions, effect of the earth's
curvature must be taken into account- Model equations are trans-
formed into a curvilinear frame in which latitude, longitude and
PROCEEDINGS—PAGE 168
-------
elevation are coordinates and the basis vectors point north, east
and vertically upward at every point on the earth. We have chosen
this frame because it is a natural one from which transformations
to any rectilinear system are easily performed. Also, is the
frame in which worldwide meteorological data are reported.
There are three basic problems that must be overcome to make
operational a model as large and comprehensive as the one we are
developing.
First, due to the large number of processes that we plan to
treat, our model is rather complicated. In order to alleviate the
problems that this might cause in operating the model and in making
future refinements, we have structured it so that its central core
consists solely of a set of algorithms for solving the coupled set
of generalized finite difference equations that describe processes in
each of its layers. The modeling functions of describing the mixed
layer dynamics; topographic effects "on winds; chemistry; cloud
fluxes, etc. will be handled by a set of special processors that are
external to the central model and which feed the model key variables
through a computer file. Within this framework the techniques used
to describe the various physical processes can be altered without
overhauling the model itself. An additional advantage is that
execution times are greatly reduced when several runs of a given
scenario are to be performed in which only one or a few parameter
values are altered.
A second problem is limitations of computer storage capacity.
To simulate air quality over the northeastern United States with the
horizontal resolution we desire, our model has roughly 10^ grid
points and treats 25 (eventually more) chemical species. Thus, the
concentration variables alone require 250K words of storage and
this is just under the working limit of 260K words of memory on
EPA's Univac computer. To accomodate a model of the anticipated
size we have developed special techniques for handling the modeling
domain in piecewise fashion.
Finally, the empirical data needed to parameterize some of the
physical phenomena cited earlier are not presently available. To
remedy this, EPA initiated project NEROS (Northeast Regional Oxidant
Study) to collect during the summers of 1979 and 1980 the meteoro-
logical and chemical data required to formulate the model. A second
goal of NEROS was to gather the data required to perform comprehen-
sive test runs and evaluation exercises of the model (see Clark and
Clarke, 1982; Clarke et al. 1982).
Aircraft sampling during NEROS was designed to gather evalua-
tion data for the regional model. Using a trajectory model and
radiosonde data, flight tracks were directed to obtain Lagrangian
PROCEEDINGS—PAGE 169
-------
TIME(EST)
1151-1359
1709-1919
2308-0104
0704-0928
1309-1S18
Fig. 9. Aircraft flight tracks, ozone concentrations (ppb), and
air trajectories (6-hr segment given by (D)) for 3-4
August 1979.
sampling of an air mass. Figure 9 shows computed trajectories and
corresponding aircraft flight tracks with 03 concentration iso-
pleths on one day. Figure 10 gives the 63 concentrations in a
vertical cross-section along one of the flight tracks (E-F). The
significant 03 concentrations at the level of the nocturnal jet
are evident in these data taken near midnight.
The model is currently set up with a 60 x 42 array of grid
cells. Figure 11 shows the region of the northeast U.S. being
modeled as well as N02 isopleths resulting from a short test run.
Although the model is functioning, substantial development work
must yet be accomplished. Preprocessors must be thoroughly checked
for accuracy, a refined emissions inventory (including biogenics)
must be finished, and the NEROS data base must be analyzed. We
PROCEEDINGS -PAGE 170
-------
3800
2500 —
2000
, 1500 — \ _
•:. - >-
Fig. 10. Cross-section of ozone concentrations (ppb) for the track
E-F in Fig. 9. Aircraft path is given by solid line,
and the dashed line shows an elevated inversion.
Fig. 11. The grid of the regional model and NC>2 isopleths from a
test run. The isopleths are for 0800 local time and
reflect sources about urban areas.
PROCEEDINGS- PAGE 171
-------
believe these tasks can be completed in mid-1983. After that we
would need about another year of model running and evaluation
before the model would be operational and could conceivably be
transferred to other user groups (perhaps mid-1984).
CONCLUDING REMARKS
Over the past decade, focus of research, development, and
evaluation efforts concerning photochemical models has been ini-
tially on the urban 03 problem and more recently on the regional 03
problem. Because of the earlier emphasis on the urban scale, there
now exist in operational form several PAQSM's which are designed to
simulate urban photochemistry and which have been applied in the
U.S., Europe, and Australia; We have summarized the attributes of
three such models which have emerged from programs of the U.S. EPA
and been evaluated with a comprehensive urban data base. The models
have been shown to be effective in predicting urban 03 concentra-
tions. The choice of which model to use will depend, in part, on
the spatial resolution required and the computer resources available.
The impetus for development of a regional model resulted from
a growing awareness that constituents of photochemical smog travel
long distances, and control of locally generated 03 may not be
sufficient to alleviate air quality problems. On the regional
scale, the multi-day nature of long-range transport necessitates
consideration of factors which may be neglected for a one-day
episode on .the urban scale. These processes include slow photochem-
ical reactions and segregation of surface and upper layers at night.
Testing and evaluation of the regional model using data of the NEROS
program will begin in 1983.
In confining our review to progress with EPA models, we do not
mean to exclude other approaches which are being considered but
believe that whatever choices are made the number of basic frame-
works must remain quite limited due to the large resource require-
ments of PAQSM's. Thus, refinement of a single approach is pre-
ferred over a proliferation of models with slightly different char-
acteristics. We have emphasized that successful application of a
PAQSM involves the use of an extensive, and usually expensive,
data base including information on meteorology, air quality, and
source emissions. We believe that commitment to gathering such
data should stand equally with efforts at- developing the mathe-
matical constructs and computer programs which constitute the
PAQSM.
PROCEEDINGS^—PAGE 172
-------
REFERENCES
Clark, T.L., and J.F. Clarke, 1982: Boundary layer transport of
NOX and 03 from Baltimore, Maryland-A case study. Proceedings
of Third Joint Conference on Applications of Air Pollution
Meteorology, San Antonio, Texas, American Meteorological Society.
Clarke, J.F., J-K.S. Ching, R.M. Brown, H. Westburg, and J.H. White,
1982: Regional transport of ozone. Proceedings of Third Joint
Conference on Applications of Air Pollution Meteorology, San
Antonio, Texas, American Meteorological Society.
Demerjian, K.L. and K.L. Schere, 1979: Application of a photochemi-
cal box model for 03 air quality in Houston, Texas. In Pro-
ceedings of Ozone/Oxidants: Interactions with the Total Environ-
ment II, Houston, Texas, October 1979, Air Pollution Control
Association, pp. 329-352.
Henderson, R.G., R.P. Fitter and J. Wisniewski, 1980: Research
Guidelines for Regional Modeling of Fine Particulates, Acid
Deposition and Visibility, Report of a Workshop held at Port
Deposit, MD October 29-November 1, 1979.
Killus, J.P., J.P. Meyer, D.R. Durran, G.E. Anderson, T.N. Jerskey
and G.Z. Whitten, 1977: Continued research in mesoscale air
pollution simulation modeling: Volume V—Refinements in numeri-
cal analysis, transport, chemistry, and pollutant removal.
Report No- ES77-142, Systems Applications, Inc., San Rafael, CA
94903.
Lamb, R.G., 1982a: A Regional Scale (1000 km) Model of Photochem-
ical Air Pollution, Part I: Model Formulation. EPA Report (in
press), U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina.
Lamb, R.G., 1982b: A Regional Scale (1000 km) Model of Photochem-
ical Air Pollution, Part II: Procedures for Model Operations,
Validation and Refinement, EPA Draft Report (June 1982), U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina.
Littman, F.E., 1979: Regional Air Pollution Study-Emission inven-
tory summarization. Report No. EPA-600/4-79-004, U.S. Environ-
mental Protection Agency, Research Triangle Park, NC 27711-
Lurmann, F., D. Godden, A.C. Lloyd, and R.A. Nordsieck, 1979: A
Lagrangian Photochemical Air Quality Simulation Model. Vol.
I-Model Formulation, Vol. II-Users Manual. EPA-600/8-79-015a,b
(available NTIS).
PROCEEDINGS—PAGE 173
-------
Lurmann, F., 1980: Modification and Analysis of the Lagrangian
Photochemical Air Quality Simulation Model for St. Louis.
Environmental Research and Technology, Inc. Document No. P-A095.
Westlake Village, CA. 25pp.
Lurmann, F., 1981: Incorporation of Lateral Diffusion in the
Lagrangian Photochemical Air Quality Simulation Model. Environ-
mental Research and Technology, Inc. Document No. P-A748.
Westlake Village, CA. 32pp.
Schere, K.L., 1982: An evaluation of several numerical advection
schemes. Submitted to Atmospheric Environment.
Schere, K.L. and J.H. Shreffler, 1982: Final Evaluation of Urban-
Scale Photochemical Air Quality Simulation Models. EPA Report
(in press), 249 pp.
Schiermeier, F.A., 1978: Air monitoring milestones: RAPS field
measurements are in. Environmental Science and Technology, 12,
664-651.
Shreffler, J.H., 1982: Observations and modeling of NOX in an
urban area. Proceedings of the U.S.-Dutch Symposium on NOX.
Maastricht, The Netherlands, May 24-28, 1982.
Shreffler, J.H. and R.B. Evans, 1982: The surface ozone record
from the Regional Air Pollution Study, 1975-1976. Atmospheric
Environment, 16, 1311-1321.
Siple, G.W., 1977: Air Quality Data for the Northeast Oxidant
Transport Study, EPA-600/4-77-020, Environmental Protection
Agency, Research Triangle Park, North Carolina.
Whitten, G.Z., J.P. Killus, and H. Hugo, 1980: Modeling of Simu-
lated Photochemical Smog with Kinetic Mechanisms - Vol. 1.
report No- EPA-600/3-80-028a, U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711.
Wolff, G.T., P.J. Lioy, G.D. Wight, R.E. Meyers and R.T. Cederwall
1977: An investigation of long—range transport of ozone across
the midwestern and eastern United States. Atmospheric Environ-
ment. 11, 797-802.
PROCEED INGS— PAGE 174
-------
RESEARCHES ON ACID RAIN IN JAPAN
Presented by T- Kome iji
Tokyo Metropolitan Research Institute for Environment Protection
Japan
PROCEEDINGS—PAGE 175
-------
1. INTRODUCTION
Increases of acidity in precipitation and its influences have posed
serious environmental problems in various countries in the world,
and there is now an urgent requirement to clarify the actual condi-
tions and to take preventive measures. In Japan, many people had
reported pains in the eyes, sore skin and other irritation; in 1973
in Shizuoka and Yamanashi Prefectures, and in 1974 and 1975 in the
Kanto Area. These occurrences forced many people to recognize that
pollution from precipitation and its influences were serious in
Japan, and subsequently, local private organisations and the Environ-
1) 2)
ment Agency started research
The problem of acid rain in Japan characteristically showed itself
as a direct influence on the human body, unlike the increase in
acidity in lakes and marshes which influenced fish and decreased
forest productivity as observed in Northern Europe and North
Eastern America. For this reason, the major subject in the researches
into acid rain in Japan has been to clarify the actual pollution level
of the initial precipitation which caused irritation to the human
body.
Since this irritation showed itself, many studies have been made, and
a large amount of information has been obtained. This report introduces
the present situation of the problem of acid rain in Japan, mainly in
reference to the reports of those studies.
2. OCCURRENCE OF IRRITATION CAUSED TO THE HUMAN BODY BY ACID RAIN IN
JAPAN
The occurrence of irritation to humans from acid rain is shown in
Tables 1 and 2^' 3^. As can be seen from Table 1, 31,815 persons
reported irritation on July 3, 1974 mainly in the south of Tochigi
Prefecture. Symptoms were irritation of the eyes and sore areas on
the skin. On July 4, the affected area spread to the south of Kanto.
PROCEEDINGS—PAGE 177
-------
On July 18, there were similar reports mainly in the south of
Tochigi Prefecture and also in Saitama Prefecture. In 1975, on June
25 reports came from the widest area in the year, and many people
were affected.
To investigate the causes, precipitation was sampled and the com-
ponents measured in the Kanto Area. The sampling method and
measurement methods were standardized in the Environment Agency's
2)
wet air pollution survey in 1975 .
This wet air pollution survey was carried out for 5 years from
1975 to 1979, for 10-day periods from the latter part of June to
the early part of July when reports of irritation seemed most
*
frequent. The precipitation sampler used in the survey is shown
in Fig. 1. It is designed to allow sampling, especially of the
initial precipitation, separately since the irritations were said
to have occurred especially at the beginning of precipitation.
Places surveyed are given in Fig. 2, and the items and methods of
measurement are .shown in Table 3.
The resultant pattern of irritation caused by acid rain in 1974 and
1975 obtained by the wet air pollution survey by the Environment
Agency and other surveys are shown in Tables 1 and 2. The measured
results of precipitation components are shown in Tables 4 and 5.
As can be seen from Table 4, the pH of the precipitation on July
3, 1974 was as low as 3.1 in Ome and Utsunpmiya, and sulfate ion
concentrations were respectively 30 and 37 vg/ml. In the case of
the precipitation on July 4, the measured pH values were 3.5 in
Chiyoda, 3.0 in Chofu, 3.7 in Kawasaki and 3.6 in Tatebayashi,
respectively, being low. The sulfate ion concentration measured
simultaneously was as high as 23 pg/mA both in Chiyoda and Chofu.
As shown in Table 5, the pH values of precipitation on June 25,
1975 in the respective places were low, being 3.3 in Ome, 3.1 in
Kumagaya and 3.3 in Tatebayashi. Sulfate ion concentration was
PROCEEDINGS—PAGE 178
-------
very high, being 34.2 yg/m& in Kumagaya and nitrate ion concentra-
tion, too, was as high as 26.4 in the same area. As mentioned, the
pH values of precipitation sampled in places where the irritation's
caused by acid rain were reported, on the days when it occurred
were low, and sulfate ion and nitrate ion concentrations were very
high.
For the precipitation of June 25 1975, sulfate ions, nitrate ions
and chloride ions were taken up as components relating to the drop
in the pH of precipitation, and the relation between the-ratios
4)
of the three components and pH ranges is shown in Figs; 3 and 4
From Fig. 3, it can be estimated that sulfate ion and nitrate ion
concentrations contribute to the drop in pH since these concentrations
in the anion concentration of precipitation with pH4 or less are high
while chloride ion concentration is low. However, as for the cause
of irritation, we cannot conclude that irritation is caused by acid
substances in the precipitation only, and it is surmised that
irritants such as formic acid and formaldehyde are also having an
effect4*• 5).
3. PRESENT SITUATION OF RESEARCH INTO PRECIPITATION IN JAPAN
Studies on precipitation in Japan have been made systematically since the
reports of the irritations from acid rain in 1974 and 1975, but
research into the causes of acid precipitation have only just begun.
Studies made in Japan are introduced below in reference to the category
of the research.
3.1 Chemical components of precipitation
The places in Japan where chemical components of precipitation
quoted in this report were measured are shown in Fig. 5.
3.1.1 pH
The pH of precipitation was continuously measured for several
years in Yokkaichi and Kumamoto in the early 1960s '
PROCEEDINGS—PAGE 179
-------
Figs. 6 and 7 show the results of those measurements. According
to Fig. 6t the, pH of precipitation in Yokkaichi was about 6 in
1961, and suddenly dropped thereafter, reaching about.4 in 1966.
In the case of Kumamoto City in Fig. 7, the pH, which had remained
at about 7 in the period from 1963 to 1966, tended to drop from
1967, to about 4.5 in 1972.
In addition to the above, event precipitation was sampled over a
long period of time to enable measurement of the components of
precipitation. The Tokyo Metoropolitan Research Institute for
Environmental.Protection,(TMRIEP) has sampled the initial 1 mm
precipitation, initial 5 mm precipitation and event precipitation
o\
since 1973, for measurement . The results of Ph measurement are
shown in Table 6. The changes in the annual mean value of pH in
4 places in Tokyo are shown in Figs. 8, 9 and 10. According to.
Fig* 8 (1) which shows the yearly changes in the annual mean pH
to initial (0 to 1 mm) precipitation, the pH can be said to have
remained almost the .same in the 4 places, showing no large yearly
variation. By place, the pH values (4.8 to 5.2) in Chiyoda were
found to be obviously higher than the values of 4.2 to 4.5 in the
other three places.
According to Fig. 8 (2) which shows the minimum pH values of (0 to
1 mm) precipitation, the minimum value did not show any large
yearly variation, either.
As for the difference between places, as in case of the mean
values, the pH in Chiyoda tended to be higher than that in the
other places.
According to Fig. 8 (3) which shows the maximum pH values of
(0 to 1 mm) precipitation, unlike the former two cases, the
maximum value changed greatly between years, and large changes
were observed in Chiyoda and Chofu.
PROCEEDINGS—PAGE 180
-------
As shown in Figs. 8 (1), (2) and (3), the pH of (0 to 1 mm)
precipitation became low in the outer suburbs and the hilly
areas compared with Chiyoda which is in the center of Tokyo.
According to Fig. 9 (1) which shows the yearly changes in the
annual mean pH value of (0 to 5 mm) precipitation, the pH tended
to drop in the period from 1974 to 1978, but rose a little in
1979 and 1980. It can be said that throughout the period, the
pH remained at almost the same level. By place, as for (0 to 1
mm) precipitation, the values in Chiyoda were high, and those in
Chofu and Ome were low.
According to Fig. 9 (2) which shows the yearly changes in the
minimum value, the pH dropped most in Chofu and Ome in 1974 and
1975, and tended to rise a little thereafter.
According to Fig. 10 (1) which shows the yearly changes in the
annual mean pH of event precipitations at one time, observed for
(0 to 1 mm) and (0 to 5 mm) precipitations, the annual mean pH
of event precipitation tended to rise in 1979 and 1980. By
place, the pH values in Chiyoda in 1978 and 1979 were lower than
those in the other places, showing that the differences in the
pH of event precipitation between the respective places were
different from that for the above mentioned pH of initial precipita-
tion.
The TMRIEP took measurements in control places, in addition to
8)
the survey made in Tokyo . The results are shown in Fig. 11.
According to Fig. 11 (1), in the case of Ogasawara by the sea,
the monthly mean pH values of (0 to 1 mm) precipitation were
distributed around neutral (pH 7), being from 5.6 to 9.4.
The pH values in Yagisawa, a mountainous area in Fig. 11 (1)
ranged from 4.0 to 4.9, being on the same level as those in the
outer suburbs and the hilly areas of Tokyo. This shows that the
PROCEEDINGS—PAGE 181
-------
acid precipitation covered a very wide area. Also the pH values
for (0 to 5 mm) precipitation in Fig. 11 (2) show the same trend
as for (0 to 1 inm) precipitation.
Among the precipitation surveys made in the Kanto Area, a year-
9)
long survey was made in Urawa . Fig. 12 shows the results of
measurement. According to Fig. 12, the annual mean pH values of
(0 to 1 mm) precipitation in Urawa were lower than the pH values
in places in Tokyo, being especially low in 1975. As for the
yearly change, the pH of precipitation in Urawa tended to rise.
Measurement was made also in Koenji, Tokyo . The results are
shown in Table 7. According to Table 7 which shows the yearly
change in the pH of precipitation, the pH values remained almost
the same at about 4.5. The mean pH value of event precipitation
for the 6 years from 1973 to 1978 in Koenji was 4.52, which is the
same as the mean pH value 4.52 of event precipitation in the period
from 1978 to 1980 in Chiyoda. Though there may have been some
differences in measuring conditions, etc., the mean pH value of
precipitation in Tokyo Ward was about 4.5.
The Meteorological Agency measured the components of precipitation
in Ryorii Iwate Prefecture "as part of an WHO background survey
The results are shown in Fig. 13. According to Fig. 13,' the
monthly mean pH values of precipitation in Ryori were almost 4.5
or higher, and the annual mean pH value is estimated to be 5.0 or
higher.
The pH values of precipitation measured in other cities in Japan
12)
are shown in Table 8, together with components of precipitation '
13), 14), 15), 16)^ According to Table 8, all the mean pH values
for the measurements were 4.9 or lower. From this, it can be
estimated that acid precipitation is increasing also in local
cities.
PROCEEDINGS^-PAGE 182
-------
3.1.2 Electric conductivity
Measurement of electric conductivity has been more frequently
made since the reports of irritations from acid rain in 1973.
Q\
The results of measurements done by the TMRIEP are shown in
Table 9. Yearly changes in electric conductivity are shown in
Figs. 14, 15 and 16.
According to Fig. 14 (1) which shows the yearly changes in the
annual mean electric conductivity of (0 to 1 mm) precipitation,
the values show a slight downward trend in respective places.
By place, the values are decreasing in the order Chiyoda/Chofu,
Ome and Okutama.
According to Fig. 14 (2) which shows the yearly changes in the
minimum electric conductivity of (0 to 1 mm) precipitation, the
values tended to increase in Chiyoda, but have remained almost
the same in Ome and Okutama. As for the differences by place,
clear differences in the descending order of Chiyoda, Chofu, Ome
and Okutama are observed. The clear differences in the minimum
value by place show the background values of pollution of precipi-
tation.
According to Fig. 14 (3) which shows the changes in the maximum
electric conductivity of (0 to 1 mm) precipitation, it changed
greatly between years as with the pH values, and both the yearly
changes and the differences between places showed no clear
trends.
According to Fig. 15 (1) which shows the yearly changes in the
annual mean electric conductivity of (0 to 5 mm) precipitation,
the values in Chiyoda and Ome tended to decrease gradually.
Differences between places in the descending order of Chiyoda,
Chofu, Ome and Okutama was indicated very clearly. From this,
PROCEEDINGS—PAGE 183
-------
a close relationship between the electric conductivity ot precipi-
tation and human activities can be estimated.
According to Fig. 15 (2) which shows the changes in the minimum
electric conductivity of (0 to 5 mm) precipitation, the values
tended to decrease until 1977, but from then remained almost the
same.
Figs. 16 (1), (2) and (3) show that the mean and minimum electric
conductivities of event precipitation remained the same, and that
differences in the descending order of Chiyoda, Ome and Okutama
clearly existed between the places.
The yearly changes in the electric conductivity in Urawa are
9)
shown in Fig. 17 as for pH . As shown in Fig. 17, the electric
conductivity of initial (0 to 3 mm) precipitation in Drawa tended
to decrease in the period from 1975 to 1980, but has increased
in the last year. With increase in precipitation from 1 mm to
2 mm, then to 3 mm, the electric conductivity showed a clear
decrease. The difference between the electrical conductivity
for initial 1 mm precipitation and that for initial 2 mm precipi-
tation was observed to be especially large.
The TMRIEP also took measurements in control places, and the
. Q\
results are shown in Fig. 18 '. According to Fig. 18, the
electric conductivities in Ogasawara as a control place by the
sea were large, and the mean values throughout the period were
90.2 for (0 to 1 mm) precipitation and 60.6 for (0 to 5 mm)
precipitation, showing the same level of values as in Chiyoda.
In Yagisawa, a mountainous area, the mean electric conductivity
of (0 to 1 mm) precipitation was 30.6, and that of (0 to 5 mm)
precipitation was 18.9, being a little lower than the values in
Okutama.
PROCEEDINGS—PAGE 184
-------
3.1.3 Sulfate ions, nitrate ions, chloride ions and ammonium ions
Survey results on the concentrations of sulfate ions, nitrate
ions, chloride ions and ammonium ions which were measured most
frequently in the surveys on the components of precipitation
are introduced below.
The results of measurements made by the TMRIEP are shown in
Figs. 19, 20, 21, 22 and 23 and Table 10. According to Fig.
19, the highest ion concentration among the components of
initial (0 to 1 mm) precipitation in Chiyoda was sulfate, being
followed by chloride, nitrate and ammonium in this order.
According to Fig. 20 which shows the yearly changes in the
chemical components of (0 to 5 mm) precipitation, the con-
centrations of sulfate ions and chloride ions showed similar
changes, decreasing until 1978, and increasing a little in
1979 and 1980. The concentration of nitrate ions showed a slight
upward trend from about 1976. The concentration of ammonium ions
remained almost the same. Sulfate ions showed the highest con-
centration, being followed by chloride, nitrate and ammonium in
this order (in pg/mŁ). Taking the mean values from 1974 to 1980,
the respective component ion concentration can be arranged in the
order sulfate 8.6 (pg/mJZ.) , chloride 4.9, nitrate 4.1 and ammonium
1.3.
According to Fig. 21, the concentrations of components of event
precipitation in Chiyoda showed the same order as in Fig. 19.
However, with respect to yearly changes, sulfate ion and chloride
ion concentrations changed greatly, and nitrate ion and ammonium
ion concentration changed little.
The concentrations of components by place in 1980 will be
compared. For (0 to 5 ram) precipitation, as shown in Fig. 22,
sulfate ion and chloride ion concentration showed a similar
PROCEEDINGS—PAGE 185
-------
downward trend, and this trend increased according to movement
from urban center to inland area in the order Chiyoda, Ome and
Qkutama. The rate of decrease in ammonium ion concentration
was smaller than for the former two, but the downward trend was
the same. Nitrate ion concentration was the highest in Ome,
unlike the three other components. For the event precipitation
shown in Fig. 23, the differences in the concentrations of com-
ponents among the three places were observed to be small,
excluding nitrate ion concentrations.
As can be seen from Figs. 19 to 23, the change in pattern of
sulfate ion and chloride ion concentration are similar in the
same places and between different places. Both are estimated to
be similar with regard to transport and diffusion in the air and
the mechanism of absorption in precipitation. Ammonium ion
concentration showed the smallest yearly change and the difference
between places. Nitrate ion concentration showed the highest in
the intermediate position between urban center and mountainous
area, and this is surmised to be caused by the difference in
transport, diffusion and process of reaction from sulfate ions.
Takeuchi measured the chemical components of precipitation in
Kichijoji, Tokyo and reported the yearly and monthly changes in
sulfate ion and chloride ion concentrations . These measured
values are shown in Tables 11 and 12. According to Table 11,
sulfate ion concentration was high in 1970 and chloride ion con-
centration was high in 1972. But clear yearly changes could not
be seen. In reference to the monthly mean values for 7 years in
Table 12, both sulfate ion and chloride ion concentrations tended
to drop in August, September and October. The mean value of
sulfate ion concentration over 7 years was reported to be 4.8
(ug/mJl), and this coincided well with the mean value of 4.82 for
sulfate ion concentration for the 3 years from 1978 to 1980
reported by the TMRIEP. The mean value of chloride ion concentra-
tion for 7 years in Kichijoji was 0.9, and this is smaller than
PROCEEDINGS—PAGE 186
-------
the 1.27 for event precipitation in 1980 in Ome shown in Table
10. It was reported that chlorine compounds in the air increased
in Tokyo during this period '.
Saitama Prefectural Pollution Center measured the components of
initial precipitation in Urawa. The results are shown in Figs.
9)
24, 25, 26 and 27 . These graphs show that sulfate ion and
chloride ion concentrations were lower than those in Chiyoda
(Fig. 19), but that nitrate ion concentration was about the same,
and that ammonium ion concentration was higher. With regard to
yearly changes, sulfate ion and nitrate ion concentrations tended
to decrease until 1979, but increased again 1980. Chloride ion
and ammonium ion concentrations remained almost the same. As for
the rates of decrease of component concentrations of 1 mm, 2 mm
and 3 mm precipitation in Urawa, very large rates of decrease
were observed between 1 mm precipitation and 2 mm precipitation.
Kanagawa Prefectural Pollution Center measured the components of
19)
precipitation and reported the results . The results are shown
in Tables 13 and 14. The ion concentrations of the three com-
ponents, sulfate, nitrate and chloride were lower than those of
Chiyoda shown in Table 10.
The concentrations of sulfate ions and nitrate ions in precipita-
tion measured in various places in Japan shown in Table 8 were
lower than the concentrations of event precipitation in 1980 in
Chiyoda and Ome, Tokyo shown in Table 10.
A survey of inland regions made by Tamaki, et al. in Hyogo
• 15,
8)
20)
Prefecture was reported . The results are shown in Table 15.
The TMRIEP took measurements in control places, and the results
are shown in Table 15. In Ogasawara a control place by the sea,
as shown in Table 16, chloride ion concentration was 9.9 (yg/mŁ),
about double that of Chiyoda, sulfate ion concentration about
PROCEEDINGS—PAGE 187
-------
1/3, nitrate ion concentration about 1/5 and ammonium ion con-
centration about 1/15. In Moshi, as shown in Table 15, sulfate
ion concentration was about 1/3 of that of Chiyoda, nitrate ion
concentration about 1/2, chloride ion concentration about 1/5
and ammonium ion concentration about 1/3. Similarly, in Yagisawa,
compared with Chiyoda, sulfate ion concentration was about 1/11,
nitrate ion concentration about 1/7, chlorine ion concentration
about 1/10 and ammonium ion concentration about 1/11.
3.1.4 Other substances
Organic acids, aldehydes, etc. can all be considered as possible
irritants in addition to inorganic acid materials. They were
also measured, and the results were reported. Measured values
obtained so far are shown in Table 17. A case where formaldehyde
2)
showed a high value of 2.7 (yg/mZ) in Chiyoda was reported .
Acrolein concentration was reported to be 0.23 in Kanagawa Pre-
5) 23)
fecture . The mean value in Hiratsuka was reported to be 0.045
Formic acid concentrations were reported to be 0.11 to 0.9 in
21)
Kobe . Hydrogen peroxide concentrations measured by Yoshizumi
22)
were found to be 0.0001 to 1.06 ' (in yg/mi).
The concentrations of formaldehyde and formic acid shown here are
high enough to induce irritation in the eyes, according to Kurokawa,
et al.
3.3 Mutual relations between components of precipitation
The composition ratios in equivalent of sulfate ions, nitrate ions
and chlorine ions in precipitation obtained in the wet air pollution
24)
survey by the Environment Agency are shown in Figs. 28 and 29
Triangle diagrams indicating cases of irritation are shown in Figs.
3 and 4.
Fig. 4 shows high sulfate ion rates, and Fig. 3 shows high nitrate
PROCEEDINGS—PAGE 188
-------
ion ratios. Fig. 3 shows cases of reported irritation in the Kanto
Area on June 25, 1975, with low chloride ion ratios and much precipi-
tation of 4 or lower pH. Fig. 28 shows a case where precipitation of
4 or lower pH was little observed, and in this case, the chloride
ion ratios tended to be high.
In the composition of anions in precipitations of 4 or lower pH, as
can be seen in Figs. 3 and 29, the chloride ion concentration con-
sidered to have originated from the sea was low, and the concentrations
25)
of sulfate ions ' and nitrate ions reported to have artificially
originated were high. As indicated here, the pH of precipitation in
the Kanto Area was found to drop when sulfate ions and nitrate ions
were contained in high ratios in precipitation.
Correlation coefficients obtained between hydrogen ion concentration
and
19.
0Ł\
and other ion concentrations in Urawa are shown in Tables 18 and
Table 18 shows that the correlation of hydrogen ion concentration
with other ion concentrations in (0 - 1 mm) precipitation in Urawa
was relatively good. Table 19 shows correlation coefficients obtained
similarly for precipitation of 4 or lower pH, and it can be seen that
the correlation in 1972 in Table 19 is far better than that in Table
18.
3.4 Relation between the increase in precipitation and concentrations of
components of precipitation
In general, it is recognized that according to the increase in
precipitation from the beginning of precipitation, concentrations
of components in the precipitation drop.
Continuous samples of fractions of a single precipitation and con-
27)
centrations of components measured in Chiyoda are shown in Fig.
30. As shown in Fig. 30, the pH of precipitation in Chiyoda became
PROCEEDINGS—PAGE 189
-------
high once initially with the increase in precipitation and showed a
gradual decrease thereafter. Electrical conductivity decreased
relatively constantly with increase in precipitation. Chloride
ion concentration decreased very much in the beginning, and almost
dropped lower than the detection limit for a precipitation higher
than 20 mm. The rate of decrease in sulfate ion concentration is
surmised to be about the same as that of chloride ion concentration,
but at a precipitation higher than 20 mm, the concentration remained
higher than for chloride ions. The component with the smallest rate
of decrease was nitrate, and the ion concentration tended to drop
gradually with increase in precipitation.
8)
From the results of measurement made by the TMRIEP annual mean
values of, pH and electric conductivity per 1 mm of precipitation
were obtained for one year of precipitation by amount of precipita-
tion, to determine the relations between the increase in precipitation
with pH values and electric conductivities in the respective places.
They are shown in Figs. 31 and 32. According to Fig. 31, in Chiyoda
in the urban center, pH was rather high in the (0 to 1 mm) precipita-
tion of 1978 and the beginning of 1979, and dropped with increased
precipitation. In event precipitation, it showed a rising trend
again. As mentioned above, in the beginning of precipitation in
Chiyoda in the urban center, as can be seen also in Fig. 30, pH was
rather high and tended to drop with increased precipitations, show-
ing a relationship opposite to the decrease of general components of
precipitation with increased precipitation. This phenomenon is
assumed to have been caused by the fact that the precipitation in
the urban center contained large amounts of buffer substances such
28)
as dust acting to neutralize the pH toward the alkali side . With
regard to this matter, in the cases of Ome and Okutama shown in
Fig, 31, pH showed a rise consistently with increased precipitation.
The electric conductivity shown in Fig. 32 decreased consistently
in all places.
PROCEEDINGS—PAGE 190
-------
With regard to the rates of decrease of electric conductivity
against precipitation in respective places, the values for Chiyoda
and Chofu were larger than Ome which was larger than Okutama. The
rate of decrease was observed to drop according to the increase in
the distance from the source of artificial ion generation.
In order to show that, in the urban center, there are many buffer
substances which are believed to suppress the drop of pH of precipi-
tation for initial precipitation, the relationship between hydrogen
ion concentration and electric conductivity are shown in Figs. 33,
34 and 35.
Fig. 33 shows that electric conductivity was very large for hydrogen
ion concentration in (0 to 1 mm) precipitation in Chiyoda, in the
city center and Chofu, a suburb town, and that electric conductivity
decreased for hydrogen ion concentration in Ome and Okutama with
increased distance from the city center.
As shown in Fig. 34 for (0 to 5 mm) precipitation and Fig. 35 for
event precipitation, with increased precipitation, electric conducti-
vity showed a trend approaching a linear relationship between the
hydrogen ion concentration of hydrochloric acid and electric con-
ductivity in all places.
In Fig. 35, it can be seen that many electrolytes in precipitation
existed as acids in Okutama, the farthest place from the city center.
Furthermore in all places irrespective of, the amount of precipita-
tion, with the increase in hydrogen ion concentration, the relation
between hydrogen ion concentration and electric conductivity
approached that of hydrochloric acid.
3.5 Amounts of chemical components falling due to precipitation
The quantities of chemical components contained in the precipita-
g\
tion are shown in Table 20.
PROCEEDINGS—PAGE 191
-------
As shown in Table 20, the amounts of chemical components in the
precipitation were 0.91 to 1.3 (g/m2 year) of sulfate ions, 0.45
to 0.67 of nitrate ions, 0.62 to 0.85 of chloride ions and 0.12
to 0.18 of ammonium ions. The amount of sulfate ions was the
largest, followed by chlorine ions, nitrate ions, ammonium ions
and hydrogen ions in this order. Also for (0 to 5 mm) precipita-
tion and overall precipitation, the order of ion component content
was the same as for (0 to 1 mm) precipitation.
Though the sampling times were different, mean values were calculated
to calculate the ratios of the amounts in (0 to 1 mm) and (0 to 5
mm) precipitation to those in event precipitation, as shown in Table
21. According to Table 21, irrespective of precipitation, the com-
ponent showing the largest amount was sulfate ions, accounting for
about 502 of the whole, followed by chloride ions, nitrate ions,
ammonium ions and hydrogen ions in this order. As for the ratios of
the quantities of chemical components in (0 to 1 mm) and (0 to 5 mm)
precipitation to the quantities in event precipitation, nitrate ions
showed the highest value in (0 to 1 mm) precipitation, followed by
chloride ions, sulfate ions or ammonium ions, and hydrogen ions in
this order, and nitrate ions showed the highest value in (0 to 5 mm)
precipitation, followed by ammonium ions, sulfate ions, chloride
ions and hydrogen ions.
This shows that in Chiyoda, the quantity of sulfate ions contained
in precipitation was the largest of the chemical components of
precipitation, and that nitrate ions were removed at the highest
rate at the beginning of precipitation.
The TMRIEP measured sybstances falling when dry (during non-precipita-
29)
tion) and falling when wet (during precipitation), and the results
are shown in Tables 22, 23 and 24.
According to the tables, anions such as sulfate ions fell mostly
during precipitation, and this trend was more clearly observed in
PROCEEDINGS—PAGE 192
-------
Ome and Okutama with increased distance from the city center,
rather than in Chiyoda, in the city center. Furthermore, as can
be seen in Table 24, in Chiyoda and Ome, the ratios of anions
as a percentage of the weight falling were very high. It was
clarified that anions such as sulfate ions, which are closely
relating to hydrogen ions as mentioned, fell in very large quan-
tities to the ground during precipitation.
3.6 Frequency distribution of component concentrations of precipita-
tion
The frequency distribution of measured electric conductivities
of precipitation completed by Kanagawa Prefectural Pollution
23)
Center is shown in Fig. 36. The frequency distributions-
obtained from the results of precipitation measurement made by
the TMRIEP and plotted on log-linear paper are shown in Figs.
37 and 38.
In Fig. 37 (1), two distributions separated at pH5 were observed.
In Fig. 37 (2), electric conductivities showed a logarithmic-
normal distribution.
In Chiyoda, hydrogen ions show a distribution with high frequencies
in concentrations lower than those of the logarithmic-normal
distribution type. Electric conductivities showed a logarithmic-
normal distribution. From these results, the distribution of
component concentrations in precipitation is surmised to be close
to a logarithmic-normal distribution.
3.7 pH distribution by raindrop size
To clarify the causes of irritations due to acid rain in Japan,
individual raindrops must be sampled, to measure components such
as the pH of the raindrop, to determine the distribution, aside
from bulk sampling as done hitherto for precipitation sampling.
PROCEEDINGS—PAGE 193
-------
A method of examination was contrived. In this method, a raindrop
is received on parafilm, and the pH of the raindrop is measured
in reference to the color due to the dissolution of the pH indicator
uniformly dispersed on the parafilm into the raindrop
The discoloration ranges of the indicator used are shown in Fig.
39, and pH distributions of raindrops are shown in Fig. 40. As
shown in Fig. 40, the sizes of raindrops were less than 1.0 mm.
As for the pH distributions by particle size, low pH values (1.2
to 1.8) appeared at a frequency of 65% in the particle size range
of 0.2 mm or less. With the increase in size of the raindrops, low
pH values appeared less, and high pH values appeared more. In the
largest particle size range from 0.8 to 1.0 mm, pH values lower
than 3.9 did not appear, and most values were 4 or higher.
From this example, it was clarified that when the mean pH of bulk
rainwater was about 3.8, the pH values by raindrop sizes were
distributed in the range 1.2 or lower and the in the range 4.8 or
higher. There was observed a trend for the width of the distribution
of the pH raindrops to be wide in a place where artificial pollution
30)
was high, and narrow in a place with low pollution .
4. PROBLEMS IN THE FUTURE
As introduced here, research into acid rain in Japan has just started,
and there remain many problems yet to.be solved.
Problems to be solved are:
1) Clarifying the causes of the increase in acid precipitation
Relations with artificial sources
Transport dispersion and reaction in air, and the mechanism of
absorption into precipitation
PROCEEDINGS—PAGE 194
-------
2) Surveys of actual conditions
Sampling, measuring methods, survey items
Field survey techniques such as selection of survey areas
Distribution of areas covered by acid precipitation (local and
national)
3) Examination of models
Material balance
4) Influences
Atmosphere
Hydrosphere
Biosphere including ground surface, soil, etc.
With regard to the above, mutually related studies must be
promoted.
5. CONCLUSION
Studies concerning acid rain in Japan have been mostly done in-
dividually, and only a few systematic studies have been made.
The latter includes the Wet-Air Pollution Survey by the Environment
Agency and a Joint Survey by Tokyo and six prefectual government in
the Kanto Area.
Studies to be promoted in future must more specifically and systemati-
cally aim at solutions of problems with clear objectives.
Efforts must be made in this direction.
PROCEEDINGS—PAGE 195
-------
REFERENCES
1) "Acid Rain (Wet-Air Pollution) in the Kanto Area", Council for Preven-
tion of Pollution in Tokyo and Three Prefectures, Air Pollution Section
of Pollution Control Promotion Headquarters of Governments in Kanto
Area, March, 1975
2) "1975 Report on the Wet Air Pollution Survey", Wet Air Pollution
Examination Committee, Air Quality Bureau of Environment Agency,
March, 1976
3) "Acid Rain (Wet-Air Pollution) in the Kanto Area in 1975", Council for
the Prevention of Pollution in Tokyo and Three Prefectures, Air
Pollution Section of Pollution Control Promotion Headquarters of
Governments in the Kanto Area, October, 1976.
4) Toshikazu Ohtaki: "Wet-Air Pollution - Acid Rain -, Pollution and
Control Measures", 13, 732-750 (1977)
5) Michiko Kurokawa, et al.: Formaldehyde concentration in rainfall and
irritation to the eyes", Air Pollution Study, 10, 86 (1975)
6) Katsumi Yoshida: "Acid rain and the Morning Glory", Air Pollution
News, No. 66 (1971)
7) Teiji Nishi: "Air Pollution in reference to pH of rainwater", Air
Pollution News, No. 64 (1971)
8) "Results of investigations concerning acid rain" (1), (2), (3) and
(4), Tokyo Metropolitan Research Institute for Environmental
Protection Institute, (1975 - 1982)
9) Kazuko Mizukami: "A survey on the components of rainwater (4th re-
port)", Annual Report of Saitama Prefectural Pollution Center,
60-66 (1981)
PROCEEDINGS—PAGE 196
-------
10) Katsuko Saruhashi and Teruko Kanazawa: "pH of precipitation, Weather",
25, 784-786 (1978)
11) Report of Ocean Pollution Observation by the Meteorological Agency,
Marine Department of Meteorological Agency, (1976-1980)
12) Yoko Kitamura, et al.: "A study on the distribution of air pollutants
in the environment - Properties of rainwater in Miyagi Prefecture",
Report of Miyagi Prefectural Pollution Control Technical Center, No.
8 (1979)
13) Hikaru Satsumabayashi and Kazutoshi Sasaki: "Chemical components of
precipitation in Nagano City", Report of study by Nagano Prefectural
Sanitation and Pollution Research Institute (1979)
14) Kanji Masamichi, et al.: "A study on components of rainwater (3rd
report)", Annual Report of Fukui Prefectural Pollution Center, 7,
(1977)
15) Tsunao Suetsugu, et al.: "A survey of actual conditions of acid
rain"
16) National Council on Studies of Pollution: "A study on the distribu-
tion of air pollutions in the environment", (1977)
17) Nobuyuki, Nakai and Ushio Takeuchi: "Chemistry of rain and air
pollution, Chemistry", 29, 418-426
18) Tetsuhito Komeiji, Isao Koyama, Mie Kyoda, Tatsukichi Ishiguro,
Morio Kadoi and Yoshiyuki Oinuma: "Examination on the use of the
corrosion of metallic materials as an index of air pollution",
Pollution Study Report of Tokyo Metropolitan Research Institute
for Environmental Protection Institute, 63-89 (1976)
19) Kanagawa Prefectural Pollution Center: "A survey on wet-air pollu-
tion", Research Report on Air Pollution by Kanagawa Prefectural
Government, 19th report, 110-120 (1977)
PROCEEDINGS—PAGE -197
-------
20) Motonori Tamaki and Takatoshi Hiraki: "Ion composition in rainwater
in the hinterland of an urban area", Environment Technology-, 9,
865-871 (1980)
21) Hiroshi Sakurada, Kozo Shintani and Hiroshi Watanabe: "Concentra-
tions of formaldehyde and formic acid in air and rainwater", A
collection of summaries of lectures at the 23rd meeting of the Air
Pollution Society, 265 (1982)
22) Kunio Yoshizumi: Private letter
23) Kanagawa Prefectural Pollution Center: "A survey on wet-air pollu-
tion", Research Report on Air Pollution by Kanagawa Prefectural
Government, 18th report, 46-66 (1976)
24) 1976 Research Report on Wet-Air Pollution, Committee for the
Examination of Wet-Air Pollution, Air Quality Bureau of Environment
Agency (1977)
25) Nobuyuki Nakai, Naoko takahashi and Ushio Takeuchi: "Sources of
sulfate ions in precipitation and air pollution", Geochemistry
(Special issue for environmental problems), 118-124 (1975)
26) Kazuko Mizukami arid Yasuo Kaneko: "A survey on the components of
rainwater (3rd report)", Annual Report of Saitama Prefectural
Pollution Center, 60-65 (1978)
27) Tetsuhito Kameiji, Tadashi Sawada, Toshio Ohira, Kazuyoshi Hirosawa
and Morio Kadoi: "A survey on the components of rainwater", Annual
Report of Tokyo Metropolitan Research Institute for Environmental
Protection, 6, 104-112 (1975)
28) Tetsuhiko Komeiji, Saburo Fukuoka, Yoshitsugu Nakano, Kunihiko
Asakino and Toshio Ohira: "A study on the pollution of rainwater
and its mechanism", Annual Report of Tokyo Metropolitan Research
Institute for Environmental Protection, 7, 27-37 (1976)
PROCEEDINGS—PAGE 198
-------
29) Tetsuhito Komeiji, Isao Koyama, Nobuko Watanabe and Tatsukichi
Ishiguro: "Pollution characteristics of falling matter in precipi-
tation, etc. by place", Annual Report of Tokyo Metropolitan
Research Institute for Environmental Protection, 81-88 (1982)
30) Tetsuhito Komeiji, Isao Koyama, Tatsukichi Ishiguro and Morio
Kadoi: A collection of summaries of lectures at the 23rd meeting
of the Air Pollution Society, 258 (1982)
PROCEED INGS—PAGE 199
-------
ta
w
a
Table 1 Reported Influences of acid tain on the human body
(1974)
Month/
date
7/3
7/4
7/5
7/6
7/13
7/14
7/17
7/18
7/20
Prefecture
Tochlgi
Ibaragi
Saltama
Gunma
Total
Ibaragi
Tokyo
Kanagawa
Chlba
Total
Ibaragi
Chiba
Total
Chiba
Saltama
Saltama
Tochlgi
Tochlgi
Saltama
Total
Saltama
Grand Total
Number of
reporters
28,762
1,793
1,120
140
321
203
187
20
731
3
9
12
4
3
1
71
225
281
374
506
33,144
Time of
occurrence
14:00 1> 18:00
15 : 30 v 9:00
19:30-021:30
llsOO
10:30-015:15
10:00-016:00
10:00-016:00
21:00
Synptoma
Irritation of the eyea,
soreness to the akin,
offensive odor
Irritation of the eyea,
painful arms
Byes smarting, hoarseness
of the throat
Pain in the eyea
Irritation in the eyea
11
tt
smarting on the arms
Irritation in the eyea
Irritation in the eyea
Irritation in the eyes,
inflamed part above the
eyea
Irritation in the eyea
ii
ii
»
ii
it
eyea bloodshot
Table 2 Reported Influences of acid rain on the human body
Month/
5/3
5/19
6/24
6/25
6/26
7/10
Prefecture
Saltama
Ibaragi
Saltfltna
Tochlgi
Saitama
Tokyo
Kanagaua
Total
Gunma
Sal tama
Grand total
Number of
1
72
9
90
43
9
1
143
18
1
244
Time of
occurrence
JilO* 7:30
7:00-014:10
13:05^19:50
8:00-017:00
11:201-12:30
16:00 •o 23:00
15:00
17:00
Symptoms
Pain in the
eyea, watering
eyea.
Pain in the
eyea, watering
eyea .
Pain in the
eye"
Rain in tne
eyea
Rain in the
eyes
Pain In the
eyea, watering
eyea.
Skin smarting
Pain In the
eyes
Pain in the
(1975)
Remark*
1 in Kavaguchl
72 in Koga
7 in Kumagaya,
2 in Chichibu
90 in Kanuma
30 in Fukaya,
12 in Enaminura,
1 in Aaaka
4 in Horitna,
1 in Chuo, 1
in.Chofu, 1 in
koganei, 1 in
Tanaahi, 1 in
Tachikawa
1 in Totauka
18 in Olzuml
1 in Omiya
-------
Table 3 Measurement Items and methods for preclpltaton
3)
"0
S)
o
n
M
M
O
HH
O
OT
13
>
O
W
Item
Method
pll
Electrical conductivity
Sulfate Ions (S042")
Nitrate Ions (N03-)
Chlorine Ions (Cl~)
Ammonium Ions (N'll^+)
Formaldehyde (HCHO)
Glass electrode method
Conductivity meter (25°C)
Barium chloride turbldlmetry
(glycerol-alcoliol method)
Sodium sallcyfate method
Mercury (II) thlocyanate method
Indophenol method
m-amlnophenol method
Table A Results of analyzed components of rainwater In July, 197A
Tokyo
Tochigl
Tokyo
Kana-
gawa
Chlba
Sal-
tama
Place sampled
TMR1EP
Owe City
Prefectural Office
Ota-ku
Itabashl-ku
TMRIKP
Shlnagawa-ku
Onie City
Oiofu City
Musasliino City
Mltaka City (Mure)
Pollution Center
(Head Office)
Pollution Center
(llelhln Branch)
Pollution Center
(Shonan Branch)
Yokohama City
(Sanitation Research)
Yokosuka City
(Sanitation Research)
Kawasaki City
(Pollution Research)
Kawasaki City
(Monitor Center)
Kawasaki City
(Saiwai-ku)
Kawasaki City
(Tama Ikuta)
Funabashi City
Chlba City
Pollution Research
Institute)
Honjo Public Health
Center
Date of
sampling
2 •>- 3
3
3
A
A
A
A * 5
A -v, 5
A
A
A
A * 5
A
A -v, 5
A -v 5
A -v. 5
A ->, 5
A -v, 5
A -v 5
A -v, 5
A i 5
A -<. 5
PH
A23
3.1
3.1
3.5
A. 5
A. 2
3.0
A. 9
3.6
3 -x. A
A. 2
4.3
3 -v- «
A.O
3.7
A. 2
A. 2
3.8
39.5
A.OC
6.7
A. 5
SO^-Z
(ppm)
-
30.0
37
16.7
22.6
26.6
3.0
22.6
7.9
16.7
-
6.5
A. 5
-
-
27.0
-
-
-
-
-
1A.O
-
NOj-
(ppm)
-
8.7B
7.AB
3.6*
8. IB
7.9»
2.AB
8. 5B
-
-
-
-
0.5*
-
-
-
-
-
-
-
-
-
-
ct-
(ppm)
-
10.0
-
5.0
5.0
-
-
2. A
7.0
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Remarks
-------
Place sampled
Gunma
Ibara-
gi
Tokyo
Kana-
gawa
Chtba
Salt ami
Gunma
Ibara-
gi
Tochlgl
Haebashl City
Kiryu City
Ota City
laezakl City
Tatebayaahl City
Annaka City
Koba City
TMRIEP
Ota-ku
Ome City
Cho(u City
Pollution Center
(Head Office)
Pollution Center
(Shonan Branch)
Yokohama City
(Sanitation Research)
Yokosuka City
(Sanitation Research)
Klsarazu City
lion Jo Public Health
Center
Kumagaya City
Takasakl City
Maebaahl City
Ota City
Tatebayashl City
Annaka City
lahloka City
Mi to City
Prefectural Office
Date of
sampling
4
4
4
4
4
4
4 -v. 5
IB* 19
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
pH
7.0*
7.0*
5.6*
5.6*
3.6
4.2*'
3.9*
6.07
6. 59
3,61
4.5
3.8
3.6
3.7
4.3**
6.2
3*4
3.2
3.9
7.0*
3.6*
4.2*
3.4**
4.0
4.2
3
80^-2
(ppn)
_
-
-
-
-
_
-
42.2
30.6
7.3
24.0
25.0
11
L5
77.5
-
-
-
-
- '
-
-
-
-
-
N03-
(ppm)
_
-
-
-
-
_
-
9.61
3.4*
13. 0»
-
11.0*
14*
20*
-
14.1*
-
-
-
-
-
-
-
-
-
-
Cf
(ppm)
_
-
-
-
-
_
-
11.4
4.8
6.1
18.9
4.1
-
6.3
-
-
-
-
-
-
-
-
-
-
-
-
Remarks
Table S Concentrations of rainwater components for the
Initial 1 mm rainfall on June 25. 1975
Tokyo
gawa
Chiba
Saita-
ma
Tochi-
B!
Ibara-
gi
Gunna
"^ Item
Plsce-^^^
Chlyoda-ku
Tama New Town
Ota-ku
Chofu City
One City
Itabashl-ku
Yokohama City
(Asahi-ku)
Kawasaki City
Yokohama City
(Sanitation
Research)
lliratsuka City
Klsarazu City
Ichlhara City
Yachlyo City
Ichlkawa City
Sahara City
Togane City
Urawa City
Kumagaya City
Utsunoralya
City
Tochigl City
Hi to City
Koga City
Tatebayashl
City
pll
4.2
1.5
6.6
3.4
3.3
5.9
3.7
3.7
3.8
3.6
4.1
4.6
3.9
3.7
4.0
4.0
3.5
3.1
3.5
3.6
3.3
3.4
3.3
SO^-
(Mg/ml)
49.0
36.1
44.8
31.8
13.2
12.5
27.3
30.3
15.0
18.0
7.2
6.8
12.4
15.5
3.5
ND
40.0
34.2
30.1
20.0
17.0
25.0
21.0
N03-
(ug/nO
21.3
20.3
16.5
19.7
13.2
3.7
17.4
17.6
11.0
17.0
5.6
6.7
13.0
14.8
3.7
4.5
17.2
26.4
5.8
4.9
19.7
30.2
12.0
«-
(US/ml)
10.2
13.2
5.8
10.6
3.2
1.9
3.3
6.4
3.5
6.9
5.5
1.3
3.8
3.7
1.6
2.5
B.3
4.0
3.2
0.9
1.4
1.8
ND
NH4+
(MB/»*>
1.3
2.6
1.8
1.8
4.3
2.6
8.2
8.2
-
3.6
2.1
1.4
1.5
4.2
0.3
0.2
7.5
4.7
3.8
2.2
1.6
1.9
2.6
IICHO
(MgM)
0.2
1.2
1.0
1.1
0.3
-
i.i
1.1
0.4
0.7
0.2
0.3
0.3
0.4
0.2
0.2
0.9
1.1
0.6
0.5
0.9
0.9
ND
Time of
sampling
11)15
13 *
lOiOO
13>00
20100
22:00
13:10
12 US
-
16(30
lliOO
10 tOO
22 tOO
12 tOO
14 tOO
10:30
16 tOO
13*16
12*21
12*15
9*10
9*10
13 * 16
Renarka
Places where irritations were reported;
Kaahlma, Tochlgl Pref. 8:00*17:00 90 persons
Fukaya, etc., Saltama Pref. 11:00*13:00 43 persons
Chofu. etc., Tokyo 16:00*23:00 9 persona
Totsuka, Kanagawa Pref. 15:00* 1 person
-------
Table 6 (1) pH of (0 Co 1 mm) precipitation in Tokyo
Chiyoda
Year
1975
1976
1977
1978
1979
1980
1 Chofu | Ome :
Mean Number of Min.-Max. • Mean • Number of Min.-Max. Mean Number of
samples i i samples samples
4.8
4.8
4.8
4.8
5.2
4.9
63
70
67
75
79
76
i 3
': 3
' 3
: 3
4
3
.8-7
.6-7
.9-6
.6-4
.0-7
.6-8
.8 ; 4.5 89
.1 4.3 j 123
.9 4.5 | 61
.8 4.2 i 79
.2 : - i
.2 ! - '
: 3.4-10 4.4 99
: 3.3-6.9 4.5 112
3.6-6.9 4.3 106
! 4.4 99
j - 4.4 122
Min.-Max.
3.5-7.7
3.6-6.3
3.6-6.5
3.4-6.1 !
3.9-7.6
3.4-6.8
Okutama
Mean Number of Min.-Max.
samples
-
-
4.3
4.2
4.5
4.5
-
-
90
95
120
113
3
3
3
3
- -
-
.3-7.2
.5-7.5
.5-7.5
.3-7.8
Mean: Arithmetic mean
Table 6 (2) pH of (0 to 5 mm) precipitation in Tokyo
Tear
1973
1474
1975
1976
1977
1978
1979
1980
Mean
4.6
4.9
4.5
4.6
4.5
4.4
4.5
4.7
Chiyoda
Number of
samples
8
63
63
70
67
75
79
76
Min.-Max.
3.9-6.3
3.5-7.2
3.6-7.9
3.6-7.1
3.6-6.9
3.3-9.2
3.6-7.2
3.6-8.2
Mean
4.4
4.5
4.4
4.4
4.4
4.2
-
-
Chofu
Dumber of
samples
14
96
89
123
61
79
-
-
Min.-Max.
3.9-5.0
3.0-7.4
3.4-10
3.3-6.9
3.6-7.1
3.4-7.2
-
-
Mean
4.5
4.8
4.4
4.5
4.3
4.2
4.5
4.5
Ome
Number of
samples
16
87
99
112
106
69
99
122
Min.-Max.
3.7-5.6
2.5-8.0
3.1-7.7
3.6-6.3
3.4-6.5
3.3-6.2
3.3-7.6
3.4-6.8
Mean
-
-
-
-
4.4
4.3
4.5
4.5
Okutama
Number of
samples
-
-
-
-
40
95
120
113
Min.-Max.
-
-
-
-
3.3-7.2
3.5-7.4
3.5-7.7
3.3-7.8
Mean: Arithmetic mean
1) In 1973 and 1974, 0-4.6 mm was regarded as 0-5.0 mm.
PROCEEDINGS—PAGE 203
-------
Table 6 (3) pH of event precipitation in Tokyo
1978
1979
1980
Mean
4.3
4.6
4.8
Number of
samples
75
79
76
Kin. -Max.
3.3-9.2
3.6-7.2
3.6-8.2
Mean
-
4.7
4.7
Number of
samples
-
99
122
Min.-Max.
-
3.3-7.6
3.4-6.8
Mean
4.5
4.8
4.7
Number of j Min.-Max.
samples ;
95 j 3.5-7.4
120 ! 3.5-7.7
113 ! 3.3-7.8
1) Calculated from the measured values for every 1 mm precipitation.
extracted from the measured values for every 1 mm precipitation.
Min.-Max. values were also
Table 7 pH of precipitation in Koenji, Tokyo (Meteorological Research Institute)
Period
Jul
Jan
Jan
Jan
Jan
Jan
to
to
to
to
to
to
Dec,
Dec,
Dec,
Dec.
Dec.
Jul.
1973
1974
1975
1976
1977
1978
Number of samples
" i
86 '
68 .
83 :
69 I
32 ;
pH
Mean '
4
4
4
4
4
4
.57
.65
.42
.47
.57
.51
1
!
*
i
i
i
i
i
3
3
3
3
3
3
Range
.60-
.48-
.34-
.71-
.64-
.90-
6
6
6
6
6
5
.80
.60
.30
.34
.86
.80
Jul, 1973 to
Jul, 1978
364
4.52
Table 8 Analytical results of rainwater components In local citie
(The values indicate the mean values of all measured values)
^"\^^ Itea
Prefecturfe-^.
Miyagi
Nagano
Fukul
Gifu
Saga
Place
Sendal City
(General Sanita-
tion Center)
Nagano City
(Sanitation and
Pollution Research
Institute)
Fukui City (Pol-
lution Center)
Gifii City (Pol-
lution Research
Institute)
Karatsu City
(Pollution
Center)
Period
Jul to Nov.
1979
Aug. 1975 to
Jttl, 1976
May, 1977 to
Feb. 1978
Apr to Oct.
1976
May to Nov.
1975
Rainwater sampl-
ing method
Rainwater sampler
which allows
sampling every 1
urn rainfall
Sampling by a
dust jar for
every rainfall
Rainwater sampler
which allows
sampling every 1
ma rainfall
ditto
Rainwater sampler
which allows
sampling every
0.5 n rainfall
PH
4.75
4.62
4.39
4.52
4.90
S042~
(vg/nt)
4.73
2.9
4.2
5.5
2.70
N03~
(ug/ml)
1.3
1.16
1.68
2.4
0.52
ct-
(ug/ml)
3.71
0.76
3.94
1.9
0.85
NH4+
(ug/nJO
0.32
0.54
0.83
PROCEEDINGS--PAGE 204
-------
Table 9 Electric conductivities of precipitation in Tokyo
Table 9 (1) Electric conductivities of (0 to 1 mm) precipitation
(at 25°C)
tear
1975
1976
1977
1978
1979
1980
Chiyoda
Mean
120
82
83
84
92
72
Number of
samples
72
74
67
75
79
76
Man. -Max.
10-510
8.8-200
16-250
25-280
21-670
20-190
Chofu
Mean
98
84
89
94
-
-
Number of
samples
90
125
62
78
-
-
Min.-Max.
11-580
4.3-830
9.8-200
14-280
-
-
Ome
Mean
73
55
62
65
44
57
Number of
samples
99
73
106
68
99
122
Min.-Max.
2.3-450
4.1-170
4.9-250
5.4-280
3.4-310
3.6-270
Oku tana
Mean
-
-
36.6
34.7
35
34
Number of
samples
-
-
91
95
120
113
Min.-Max.
-
-
2.2-200
2.4-130
2.5-310
2.3-280
Table 9 (2) Electric conductivities of (0 to 5 mm) precipitation
Tear
1973
1974
1975
1976
1977
1978
1979
1980
Chiyoda
Mean
69
57
69
48
49
43
49
39
Number of
samples
8
63
72
74
67
75
79
76
Min.-Max.
5.5-125
7.4-192
12-510
9.2-200
5.3-250
5.3-280
5.1-670
5.1-190
Chofu
Mean
37
40
62
40
54
51
Number of
samples
90
12.5
62
78
Min.-Max.
12-62
70-660
3.0-580
2.6-830
5.0-200
3.7-280
One
Mean
41
51
44
35
40
37
30
38
Number of
samples
16
87
99
73
106
68
97
122
Min.-Max.
12-130
3.6-330
2.3-490
3.8-170
1.3-250
3.3-280
2.8-310
2.4-300
Okutama
Mean
25
22
23
23
Number of
samples
91
95
120
113
Min.-Max.
1.7-200
1.3-130
1.6-310
1.4-280
PROCEEDINGS—PAGE 205
-------
Table 9 (3) Electric conductivities of event precipitation
Year 1
1978 '
1979 |
1980 !
Mean
26
32
28
Number of
samples
75
79
76
Hin.-Max.
5.3-280
4.8-670
5.1-190
Mean
-
22
21
Number of
samples
-
99
122
Min.-Max.
-
2.8-310
2.4-300
Mean
13
12
15
Number of
samples
95
120
130
Min.-Max.
1.4-130
1.6-310
1.4-280
Table 10 Chemical Components of precipitation in Tokyo
Chiyoda
>>8recipita-
tion —
Year ^v^
1974
1975
1976
1977
1978
1979
1980
Mean
S042-
(0-lam)
18.7
(71)
18.0
(73)
13.6
(52)
17.1
(196)
(O-Sarn)
9.25
(300)
10.1
(270)
8.15
(270)
7.81
(207)
6.75
(282)
8.34
(267)
9.18
(316)
8.57
(1912)
Total
3.12
(972)
4.01
(1247.5)
6.64
(1452.7)
4.82
(3672.2)
N03-
(0-lnrn)
9.36
..(71)..
6.73
(73)
6.72
(52)
7.68
(196)
(0-5«n)
4.18
(295)
4.09
(270)
3.36
(270)
4.33
(212)
3.95
(282)
4.84
(266)
4.29
(316)
4.14
(1911)
Total
1.79
(972)
2.06
(1249.1)
1.62
(1452.7)
1.81
(3673.9)
Cl~
(0-laa) (0-5nm)
12.2
(71)
8.96
(73)
9.30
(52)
10.2
(196)
5.26
(280)
6.14
(268)
4.62
(272)
4.80
(212)
3.32
(281)
6.10
(265)
4.22
(316)
4.90
(1894)
Total
1.78
(972)
4.58
(1247.1)
2.04
(1452.7)
2.83
(3671.9)
One
1980
6.04
(411)
*-25
(1026.9)
5.03
(411)
2.96
(1026.9)
Okutaaa
1980
4.28
(424)
3.82
(1123.5)
2.27
(424)
1.32
(1123.5)
2.06
(411)
1.27
(1026.9)
NH4+
(0-lnm)
1.68
(71)
2.42
(73)
1.98
(52)
2.04
(196)
(0-Sim.)
1.38
(229)
1.20
(270)
1.34
(210)
0.87
(281)
1.47
(273)
1.39
(316)
1.26
(1579)
Total
0.55
(972)
0.54
C1250.1)
0.79
C1452.7)
0.64
(3674.9)
1.11
(411)
0.56
(1026.9)
0.83
(424)
0.47
(1123.5)
0.31
(424)
0.21
(1123.5)
Concentration in Vg/mt. The values In the parentheses indicate the amount* of precipitation.
PROCEEDINGS—PAGE 206
-------
Table 11 Yearly changes in the concentrations of sulfate ions and chloride
ions in precipitation (1967-1973. Kichijoji, Tokyo)
Year
SOA2~ (wg/mŁ)
ci- ( " )
1967
A.O
0.5
1968
5.9
0.1
1969
4.8
0.5
1970
7.1
1.2
1971
4.4
0.8
1972 | 1973
4.1
1.6
3.8
0.1
Table 12 Monthly mean values of the concentrations of sulfate ions and
chloride ions in precipitation (1967-1973, Kichijoji, Tokyo)
Month i Jan. | Feb. ! Mar. ' Apr.
S042- (pg/mi) ; 4.0 j 6.0 j 4.6 ! 4.7
Ci~ ( " ) 0.9 | 2.1 } 1.1 | 1.2
May.
4.9
1.0
Jun.
4.8
1.2
Jul.
5.4
0.9
Aug.
3.3
0.4
Sept.
4.0
0.4
Oct.
3.3
0.3
Nov.
4.7
1.2
Dec.
4.3
1.9
Mean value
over 7 vears
4.8
0.9
Table 13 Mean values and variations (Jan to Dec, 1975)
(Initial 1 am precipitation)
Component
pH
Electric con-
ductivity
(umho/cm)
(ppm)
N03
(ppm)
ct-
(ppm)
HCHO
(ppm)
Place of
measurement
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Arithmetic
mean value
X
4.48
5.18
4.73
104
94.3
9.0
12.3
12.1
6.4
6.1
8.9
9.7
9.5
11.1
0.24
0.25
0.37
Geometrical
mean value
xct
4.40
5.10
4.66
85.0
78.8
5.3
8.9
9.7
4.4
3.4
5.6
7.2
6.3
7.7
0.13
0.15
0.38
Standard
deviation
a
0.90
0.96
0.80
62.9
55.4
13
10.0
8.2
5.7
5.7
8.6
6.8
8.8
9.7
0.27
0.26
0.25
Standard devia-
tion percent
A
20
19
17
60
59
144
81
68
89
95
97
70
93
87
113
102
68
Number of
samples
n
62
85
81
59
85
62
52
59
61
53
53
61
51
59
62
54
42
PROCEEDINGS--PAGE 207
-------
Table 14 Ranges
Component
PH
Electric con-
ductivity
(umho/cn)
S04*-
(ppm)
N03-
(pp»)
cz-
(ppm)
HCHO
P»)
Place of
measurement
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Yokohama
Kawasaki
Hiratsuka
Min. and max. values
Max. value
7.4
7.3
6.7
367.0
295.0
- '
91
48
42
23
24
45
27
46
50
1.1
1.1
1.1
Min . value
3.3
3.5
3.4
15.1
6.4
-
0.3
0.9
2.0
0.0
0.1
0.3
0.6
0.8
0.1
0.00
0.01
0.00
Ranee
R
4.1
3.8
3.3
351.9
288.6
-
90.7
47.1
40.0
23.0
23.9
44.7
26.4
45.2
49.9
1.10
1.09
| 1.10
Table 15 Average concentrations of main ions in the rain water
(Mita City 1976)
S042 (wg/mt)
NH4 (wg/Bl)
N03 (wg/»O
N02 (yg/Bl)
a (ug/al)
Hg (ug/ai)
PB
E.C. *)
(V3/0)
Summer
(Ho. 1 6)
1.73 (6)
0.40 (6)
1.34 (6)
<0.001 (6)
0.54 (6)
0.18 (6)
4.43 (6)
22.78 (6)
Autumn
(Ho. 7 14)
0.78 (2)
0.13 (8)
0.42 (8)
<0.001 (8)
0.54 (8)
0.22 (7)
5.07 (8)
11.10 (8)
Total
1.49 (8)
0.25 (14)
0.82 (14)
<0.001 (14)
0.54 (14)
0.20 (13)
4.80 (14)
16.10 (14)
The numbers of samples are in parentheses.
*) Electric conductivity
PROCEEDINGS—PAGE 208
-------
Table 16 Concentrations of precipitation components in control places
1978
Jun.
Jul.
Aug.
Sept.
Oct.
Mean
SO
Yagisawa
0.57
(36)
0.69
(19)
1.02
(24)
0.63
(35)
1.05
(16)
0.75
(130)
,2-
Ogasawara
2.3
(16)
-
2.8
(12)
-
3.7
(13)
2.9
(41)
N0:
Yagisawa
0.49
(36)
0.82
(19)
0.86
(24)
0.26
(35)
0.94
(10)
0.58
(124)
r
Ogasawara
1.4
(16)
-
-
0.3
<")
| 0.91
| (29)
a
Yagisawa
0.26
(36)
0.12
(19)
0.63
(24)
0.64
(35)
0.85
(16)
0.48
(130)
Ogasawara
6.3
(16)
-
8.4
(12)
15.7
(13)
9.9
(41)
NH4
Yagisaga
0.12
(31)
0.092
(19)
0.18
(24)
0.086
(35)
0.097
(16)
0.11
(125)
+
Ogasawara
0.08
(16)
0.09
(12)
-
0.07
(13)
0.08
(41)
In Yagisawa, (0 to 6 mm) precipitation was taken as one sample.
In Ogasawara, (0 to 5 mm) precipitation was taken as one sample.
The values in parentheses indicate amounts of precipitation in mn.
Note: Mean of Chiyoda (73-80)
(0 to 5 mm)
Total precipitation
(78-80)
8.57
4.82
4.14
1.81
4.90
2.83
NH4+
1.26
0.64
Table 17 Examples of measured components of precipitation
(ug/ml)
Flace
Tacebayashi
Drawa
Kumagaya
Chiyoda
Kanagava
Kobe
Chiyoda
Hiratsuka
Period
June, 1975 to
July, 1979
ditto
• ditto
ditto
November, 1981
to July, 1982
June to
September 1982
July to
December, 1974
Formaldehyde
<0.05 ~ 0.76
<0.1 t. 2.2
<0.1 *• 2.0
<0.1 ^ 2.6
1.9
0.017 •v.
0.21
0.53
Acrolein
0.23
0.045
Formic acid
0.11 -v. 0.9
Hydrogen
peroxide
0 -v. 1.06
Measurer
Survey by Environment
Agency2)
n
n
it
Kurokawa, et al.5)
Sakurada, et al.2D
Yoshizunrf.22)
Kanagava Prefectural
Pollution Center23)
PROCEEDINGS—PAGE 209
-------
Table 18 Correlation coefficients between hydrogen ion concentration and respective
materials (initial 1 mm precipitation)
^
1975
1976
1977
Electric
conductivity
0.93
0.87
0.51
S0.2-
0.59
0.78
0.27
1*03-
0.78
0.82
0.26
! ci-
1 0.25
j 0.35
0.12
^
0.59
0.75
0.04
N
41
48
44
Table 19 Correlation coefficients between hydrogen ion concentration and respective
materials (rainfall of pH 4 or lower)
fear ^^~
1975
1976
1977
Electric
conductivity
0.84
0.91
0.66
s°'"
0.45
0.79
0.54
N03"
0.73
0.89
0.43
ci-
0.33
; o.4i
0.36
-+
0.48
0.79
0.22
N
48
51
27
Table 20 Amounts of chemical components falling In precipitation (Amount in g/a? year,
precipitation in ma)
Chemical
coniponccc
(0 to 1 mm) SO*2-
Precipita-
tion 1103-
Cl~
SH4+
H*
Precipita-
tion
(0 to 5 —0 S042~
Preelplta-
tion 1103-
CJT
BH4+
H+
Precipita-
tion
Complete S0i2-
Precipita- 1103-
clon Ct-
HH4+
#•
Precipita-
tion
1974 1975
1.33
(48)
0.665
(60)
0 . 645
(51)
0.120
(32)
0.00113
(13)
71
2,78 2.77
1.254 1.104
1.58 1.66
0.372
0.00369 0.00815
300 270
j
1976
1.31
(59)
0.478
(52)
0.636
(50)
0.177
(54)
0.00116
(17)
73
2.22
0.913
1.25
0.326
0.00668
272
Qiiyoda
1977
0.913
(55)
0.450
(49)
0.623
(61)
0.133
(40)
0.00119
(16)
67
1.66
0.918
1.02
0.284
0.00718
212
1978
0.00119
(11)
75
1.90
(62)
1.11
(64)
0.94
(55)
0.245
(46)
0.0126
(26)
282
3.03
1.74
1.700
0.537
0.0487
972
1979
0.00050
(5.8)
79
2.28
(46)
1.32
(52)
1.67
(29)
0.401
(77)
0.00787
(25)
273
5.01
2.S7
5.73 -
0.524
0.0314
1250
• Ome
1980 • 1980
j
i
;
,
0.00096
(15)
76
2.90 2.48
(30) (57)
1.36 Z.07
(47) (68)
1.33 0.847
(46) (65)
0.439 0.456
(39) (79)
0.00675 0.0139
(29) (63)
316 411
9.65 4.36
2.35 3.0*
1 2.91' l.JU
1.14 0.579
6.0230 O.OMO
1453 1027
Oku-
tama
1980
1.81
(42)
o.yt>4
(65)
0.353
(67)
0.133
(56)
o.ui^tr
(52)
424
4.29
i.*y
0.236
U.U^46~
1124
Note: The values in parentheses of (0 to 1 BB) precipitation are ratios (Z) to (0 to 5 an) precipitation.
The values la parentheses of (0 to 5 m) precipitation are ratios (Z) to complete precipitation.
PROCEEDINGS—PAGE 210
-------
o
Table 21 Amounts of chemical components falling
Chiyoda, Tokyo (1974 - 1980)
in
Precipita-
tion
(0 to 1 mm)
precipita-
tion
(0 to 5 mm)
precipita-
tion
precipita-
Chemlcal
component
S042-
N03-
«~
NII4+
H+
Total
501,2-
NOj-
ci-
NH4+
11+
Total
S0$2-
N03-
ct,-
NI14+
H+
Total
Mean amount
1.19
0.532
0.702
0.144
0.000992
2.57
2.40
1.15
1.22
0.349
0.00752
5.13
6.32
2.26
3.55
0.771
0.0327
12.9
X to the total
chemical com-
ponents
46
21
27
5.6
0.04
47
22
24
6.8
0.15
49
18
28
6.0
0.25
X to the corre-
sponding amount
of overall preci-
pitation
19
24
20
19
3.0
38
51
34
45
23
Table 22 Amounts of respective components falling by place
(wet and dry)
(In mg/n>2/month)
Chlyoda
Fine Rainy
SO/r 664 . 3
NOj- 187.1
ct-
NH4+
K+
127.1
44.0
12.2 '
Na+ I 38.9
901.6
238.6
231.7
117.4
7.9
66.8
Ca2+ j 120.2 ; 78.1
Mg2+
FeZ-t-
Mn2+
Pb2+
Zn2+
Ni2+
12.5
2.7
0.98
0.27
5.2
0.059
14.2
4.1
1.2
_..„._
8.2
0.050
Ome
Fine | Rainy
357.6 I 735.3
165.4 | 255.6
57.2 1 139.4
19.0 119.1
17.2 i 7.7
22. .2 39.2
50.8 452.9
5.9 7.4
1.5 3.0
0.28 0.58
0.033 0.76
1.7 4.0
0.023 0.22
Okut
Fine
55.1
70.8
28.4
21.0
14.2
15.5
22.1
2.7
1.2
0.12
0.058
0.76
0.026
a ma
Rainy
204.2
136.4
56.0
67.6
13.1
31.0
232.6
4.6
2.3
0.36
0.15
1.6
0.18
td
to
-------
W
W
O
01
Table 23 Ratios of the respective componenta by place
(In X)
80^2-
N03~
ci-
NH4+
K+
Na+
CaZ+
M82+
Fe2+
Mn2+
Pb*+
Zn2+
HI
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Fine
Rainy
Chiyoda
42.4
57.6
44.0
56.0
35.4
64.6
27.3
72.7
60.7
39.3
36.8
63.2
60.6
39.4
46.8
53.2
39.7
60.3
45.0
55.0
9.1
90.9
38.8
61.2
54.1
45.9
One
32.7
67.3
39.3
60.7
29.1
70.9
13.8
86.2
69.1
30.9
36.2
63.8
10.1
89.9
44.4
55.6
33.3
66.7
32.6
67.4
4.2
95.8
29.8
70.2
9.5
90.5
Okutana
21.2
78.8
34.2
65.8
33.6
66.4
24.2
75.8
52.0
48.0
33.3
66.7
8.7
91.3
37.0
63.0
34.3
65.7
40.0
60.0
27.9
72.1
32.2
67.8
12.6
87.4
Table 24 Concentrations of the respective components by
place
(In X)
S0<,2-
N03-
c -
NH$-
K+
Na+
Ca2+
Mg2+
Fe*+
MnZ+
Pb2+
Zn2+
Ni2+
Chiyoda
Fine
54.6
15.4
10.5
3.7
1.0
3.2
9.9
1.0
0.2
0.1
0
0.4
0
Rolny
53.9
14.3
13.8
7.0
- 0.5
4.0
4.7
0.8
0.2
0.1
0.2
0.5
0
Ome
Fine
51.2
23.7
8.2
2.7
2.5
3.2
7.3
0.8
0.2
0
0
0.2
0
Rainy
41.7
14.5
7.9
6.8
0.4
2.2
25.7
0.4
0.2
0
0
0.2
0
Okutama
Fine
23.7
30.4
12.2
9.3
6.1
6.7
9.5
1.2
0.5
0.1
0
0.3
0
Rainy
27.2
18.2
7.5
9.0
1.8
4.1
31.0
0.6
0.3
0.1
0
0.2
0
-------
Hard glaaa
U'yreg) ,
Cover
-'
-_.
JO
:
-.
a
5
01
a
•-
-500
-I"
Water nampler
Float
Volume of ahndrd
purl Ic.n, SO rr
Water aa>pllng
buttle
Fig. 1 Precipitation aanpler
Nakanojo
Okuf»raj
. •
Chlyodi? 'Y«chlyo
; •. » • • JMilknwn
gan«
(Tokyo)
Chlyoda: Urban center
i
Cliofu: Suburb* City
One: Quaal-Btountalnoua area
Okutaiu! Houncalnoua area
Fig. 2-(2) Surveyed placee
-------
0 10 20 30 *0 50 60 70JJOJO 100
ci-(t)
0 10 20 30 40 50 60 70 80 90 100
—- ct-<*>
Ft.. 3 Co-po.ltion r.tlo. In .,Ulv.Un. Fig. 4 Co.*o.ltlon r.tlo. of .nlon. In
of .nlon, in r.inv.cer (Joae r.lnw.ter (June 30. 197
25. 1975)
-------
pll
6.0
4.0
pll
7
6
5
4
3
61 63 65 67
Fig. 6 Changes In the pll of rainwater
in Yokkalchl City (Mean values
of 18 placea in the city, by
Yoshlda)
-a
w
CO
o
M
t\3
o-i
63 64 65 66 67 6S 69 70 71 72
Fig. 7 Changes in the pll of rainwater
in Kumamoto City (Mean values
of 5 to 7 placea In the city,
Surveyed by Nishi, Sanitation
Bureau of Kumamoto Municipal
Office)
pH
7.0
6.0
5.0
4.0
i Hhiyoda x Ome
kChofu ° Okutama
75 76 77 78 79 80
(1) Changes in the annual mean pll of
(0 to 1 mm) precipitation
pH
5
75 76 77 78 79 80
(2) Changed in the minimum pll value of
(0 to 1 mm) precipitation
Fig. 8 Changes in the pH of Initial
(0 to 1 mm) precipitation
(1) Mean pH values of (0 to 5 mra)
precipitation
73 74 75 76 77 78 79
80
4.0
3.0
(2) Minimum pll valuea of (0 to 5 mm)
precipitation
8
75 76 77 78 79 80
(3) Changes in the maximum pH values of 7
(0 to 1 mm) precipitation
73 74 75 76 77 78 79 80
(3) Maximum pll values of (0 to 5 ran)
precipitation
73 74 75
76
77 78 79 80
Fig. 9 Changes In the pH of (0 to 5 mm)
precipitation
-------
•H oi e
o
"E.
Goo
• x o
e
o
«U •*
O u
r- C C
pH
9
8
7
6
5
4
0 Yagisava 9
• 8
• 7
•
o 5
„ ° o °
00 04
I 1 1 1 I "
.
• • « *
.
00 o 0 o ° ° ° ° 0
•775 6 7'786 7 8 9 10 '767 8 9 10 '77 5 6 7 '78 6 7 8 9 10
(1) Mean pH values of (0 to 1 m) precipita-
tion in control places
(2) Hean pH values of (0 to 5 mn) precipitation in
control places
Fig. 11 pB values of precipitation in control places
PROCEEDINGS—PAGE 216
-------
•*)
r*
00
23
n n
iii
1976 1977 1978 1979 1980
1234 5 67 89 10 1112 12 3* 5 6789 10 1112 123* 567 89 1011 12 123*56789 10 1112 123*5678
Fig. 13 pH of rainwater In a background area (Ryori, Sanriku-cho, Iwate
Pref., Surveyed by Meteorological Agency)1!)
PROCEEDINGS—PAGE 217
-------
I
M
M
O
O
w
N
w
00
120
a
100
80
60
<°
20
75 76 77 78 79 80
(1) Changes In the mean electric conductivity
of (0 to 1 mm) precipitation
30
20
10
75 76 77 78 79 80
(2) Change! In the minimum electric conductivity
of (0 to 1 mm) precipitation
500
§ 400
300
200
100
»580
670
75 76 77 78 79 80
(3) Changes In the maximum electric conductivity
of (0 to 1 mm) precipitation
100
80
60
40
20
0 - 5 mm Mean
73 74 73 76 77 78 79 80
0 - 5 mm Min.
1
20
10
600
500
•=•
JJ 400
"5
a
u 300
w
200
100
73 74 75 76 77 78 79 80
73 74 75 76 77 78 79 80
Fig. 14 Changes In the electric conductivity of (0 to 1 nun)
precipitation
Fig. IS Changes in the electric conductivity of (0 to
5 mm) precipitation
-------
O
tfl
N>
tŁ>
600
EC
30
20
10
20
^^^ 15
* 10
5
500
400
300
200
* • — •• 100
* x
.
- A
• A
/ \
' / \
/ r~k
/r
/T
78 79 80 78 79 80 78 79 80
(1) Mean electric con- (2) Minimum electric (3) Maximum electric con
ductlvltiea of event conductivities of ductlvlties of event
precipitation event precipita- precipitation
tion
100
80
60
Initial 1 mm precipitation
,' ,\ \ _.-« Initial 2 mm precipitation
/ '^ *••-'" ,S Initial 3 mm precipitation
T .
75 76 77 78 79 80
Fig. 17 Annual changes in electric conductivity (Urawa)
O
w
M
a
Fig. 16 Changes In the electric conductivity of event
precipitation
-------
120
o o o
o «o ~o
•H
(U3/nn)33
40
20
120
100
• 1
• ^80
I 60
° o o 40
• • 0 o
0 ° 20
• Ogasawara
O Yagisawa
.
9 ° •
o o * 00° °o
'767 8 9 10*775 6 7 '786 7 8 9 10
Mean electric conductivity of (0 to 1 m)
precipitation in control places
Mean electric conductivities of (0 to 5 am)
precipitation in control places
Fig. 18 Electric conductivities of precipitation in control places
1 f ^
• •
n a
So *
II s
» • > •
% i 2 §
*+-,. '*«
H- • o-
So
g.§ tf
iS S
Z
o
PROCEEDINGS—PAGE 220
-------
O
M
M
a
o
n
10
'78 79 80 Year
Fig. 21 Mean values of
components of event
precipitation
10
o
16
4
2
4 ct-
p N03-
a
o
3 Place
10
8
6
Ł 4
* «-
Fig. 23 Mean values of com-
ponents of event
precipitation (1980)
o
•rl
6
Fig. 22 Mean values of components of
(0 to 5 mm) precipitation
(1980)
t-o
-------
12
10
Ł 8
/A X
'
75 76 77 78 79 80
Fig. 24 Yearly change* In sulfate
ion concentration (Urawa)
u
"• (2)
— (3)
75 76 77 78 79 80
Fig. 26 Yearly changes In chloride
ion concentration (Urawa)
10
8
?
o.
~ 6
fc
85 4
""'—-
75 76 77 78 79 80
Fig. 25 Yearly changes In nitrate
Ion concentration (Urawa)
, _., --(2)
75 76 77 78 79 0
Fig. 27 Yearly changes in ammonium
Ion concentration (Urawa)
0 A100 «u2-
10
0 10 20 30 40 50 60 70
9° 1°°
Cl~ (I)
-------
o
n
ra
0 A 10° so<2-
100
n 10 20 30 40 50 60 70 80 ?Q_ 100
-------
pH
7
5 -2.
'78-
• Chiyoda
* Om*
Okutama
120
100
"g 80
"a
^ 60
o
u
•H 15 -3 Preeipita-
Tear
Fig. 31 Relation between precipitation and pH
40
20
•Chiyoda
AChofu
xOme
OOkutama
• Precipita
'tion
Tear
Fig. 32 Relation between precipitation and
electric conductivity
80
70
60
Z X
20
10
• Chiyoda
AChofu
XOme
OOkutama
(HCt)
0.01 0.02 0.03 0.04 0.05 0.05 0.07 H+
Ug/»l
Fig. 34 Relation between hydrogen ion concentration
and electric conductivity in (0 to 5 mm)
precipitation
PROCEEDINGS—PAGE 224
-------
100
90
80
70
o 60
a
50
40
30
20
10
x
x
f» Chiyoda
'* Chofu
I * Ome
L° Okutama
CfiCi)
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
H* (ug/ml)
Fig. 33 Relation between hydrogen ion con-
centration and electric conductivity
in (0 to 1 mm) precipitation
70
60
"s 50
u
•a
,3 40
u
u
30
20
10
o Chiyoda
x Ome
O Okutama
(HC1)
0.01 0.02 0.03 0.04 0.05 0.06 0.07
H+ (wg/mt)
Fig. 35 Relation between hydrogen ion con-
centration and electric conductivity
in event precipitation
3
—
r
i
i
r
40 60 80 100120140160*190
; (proho/cm
o
rsi
Fig. 36 Frequency distributions
hydrogen ion
(2)
PROCEEDINGS—PAGE 225
-------
201
18
16
14
12
10
8
6
4
2
0
302:
(H+)
27
24
I
21
18
15
12
6
T~l ^
1
r
u
0.1 1 1 10 100 1000 10
pH7 pH6 pH5 pH4
(EC)
Ti
m
100 1000
(uy/cm)
Fig. 37 Frequency distributions of hydrogen ion
concentrations and electric conductivities
(1979, (0 to 1 ma) precipitation)
1000
pH3 1000
PH4 100
PH5 10
pH6
pH7
.01
10
20
30 40 50 60 70 80
90
99
Fig. 38 (1) Frequency distribution of hydrogen ion concentrations in Chiyoda
(1979, (0 to 1 mm) precipitation)
99.S
FKOCKKDINGS PAGE 226
-------
1000
EC
100
10 -
•*•
i :
4-
.
-f 4 "*
L 10
I '
j
1
1
!
I
1
1
|
1 |
1
1 <
, 1
|
1
1
\ |
1 ;
' t*
\ -t*
1 _(*+**" i
-W- '
-*•"*"* '
i '
•f j
r 1
i i
1 i
1 1
1
1 1
1
I |
1 1
]
20 30 4
I
1 1
1 I
,111
t 1 "*
1 ' 1
I 1 i i
I > 1 1
i 1 1 i
1
: !
+ '
1 i
1 1 i 41
: i ^ i
II1 + !
! i : y
II; i
i ; ! 44
i *tf -M- i
1 '. ttMM |
I t^jif* 1
j..ffH*rt*+^ !
.rf* ' i I ' 1
•-•^ | i
i i i
1 i
iii1 '
1 ! 1
' 1
I t i i 1
,, i
i i ; i i
i : i i
i
' } !
i i i i ,
i '
iii
i i
i i
i i
1 ' i '
t < i i • i
0 50 60 70 80 90 99 9
z
9.9
Fig. 38 (2) Frequency distribution of electric conductivities in Chiyoda
(1979, (0 to 1 mm) precipitation)
PROCEEDINGS PACK 227
-------
July 4, 1981; Tokyo Metropolitan Pollution Research Institute;
Llglit Rain; Rain started at 10:30, and sampling started at 10:35;
nean pll 3.80
(Particle alze
distribution)
(pH distribution)
Particle size - 0.2>
i 0.2 0.4 0.6 0.8(mm)
0.2 i i i i
0.4 0.6 0.8 1.0
0.2 - 0.4
61.2
14.2
19.5
1.2 1.82.73.3 3.9 4.2 4.8
i i i i i i i
1.8 2.7 3.3 3.9 4.2 4.8 pll
46.7
14.6
21.8
16-8 5.2 2.8 3-6
1.2 1.82.7 3.3 3.94.2 4.8
i i i i i i i
1.8 2.7 3.3 3.9 4.2 4.8 pH
0.4 - 0.6m
(1)
36.4
.28.2
1.2 1.8 2.7 3.3 3.9 4.2 4.8
i i i i i i
1.8 2.7 3.3 3.9 4.2 4 8
0.8 ~ 1.0m
0.6 - 0.8m
,28.1
34.1
1.2 1.8 2.7 3.3 3.9 4.3 4.9
f * i i t t t
1.8 2.7 3.3 3.9 4.3 4>8
54.8
43.1
2.2
172 178 2.7 3.3 3.9 4.2 4.8
t f i i i i t
1.8 2.7 3.3 3.9 4.2 4.8
Fig. 40 Raindrop size distribution and pH distributions by
raindrop sizes.
PROCEEDINGS--PAGE 228
-------
C Ref. 3
892 Journal of the Meteorological Society of Japan Vol. 59, No. 6
A Numerical Model of Acidification of Cloud Water
By Sachio Ohta and Toshiichi Okita
Department of Sanitary Engineering. Hokkaido University, Sapporo 060
and Chiaki Kato
Sumitomo Denko Ltd., Konohana-ku, Osaka
(Manuscript received 22 April 1981, in revised form 21 September 1981)
PROCEEDINGS—PAGE 229
-------
Abstract
The numerical model calculation was conducted cf pH of cloud water on the assumption
that the cloud droplets formed on H2SO4 and other sulfate and nitrate nuclei dissolve acidic
and alkaline gases to achieve gas-liquid equilibrium in a very short time. With the initial
concentration of gases and particulate components as found in the real atmosphere the pH
of the clcud water is calculated to be three. The calculation inetrprets pH, and the concen-
trations of SO42~, NO3~ and other species in cloud water observed on a mountain fairly
well.
1. Introduction
The acidification of the precipitation was
reported from Europe, U.S.A. and Canada. In
Japan three episodes occurred in which people,
particularly students who were cycling, walking
and working in outside air were suffered from
eye- and skin-irritation as a result of the attack
of contaminated drizzle droplets. The dates of
the episodes, the places where the episodes oc-
curred (c.f. Fig. 1) and the number of those who
complained the irritation are as shown in Table 1.
Since the pH of the drizzle droplets was as
low as about three, we assumed that hydrogen
ion in the droplets might be responsible for the
irritation as well as formaldehyde and other
irritants. For the study of the cause of the
irritating precipitation Japanese Environment
Agency organized a study group and since 1975
during rainy season of June and July three
dimensional study of atmospheric gaseous and
particulate species and the species in precipitation
have been made in the Kanto area (Fig. 1). As
members of the group we made measurements
of acidic and alkaline gaseous and particulate
species and of pH and the concentration of
various species in cloud water on the top of Mt.
Tsukuba (altitude 870m).
Then in order to know the effect of the gaseous
and particulate species on the pH of the cloud
water, an equilibrium model was formulated to
describe the pH of the cloud water in terms of
the concentration of gaseous and particulate
species in the atmosphere. It was also found
that the dissolution of NQj into water is un-
important for the formation of NOs~ in the
water.
2. Determination of concentration of chemical
species in cloud water
In the rainy season the top of Mt. Tsukuba
was frequently covered with a cap cloud. In
June and July of 1975 through 1978 sampling
of the cloud water was made. The concentrations
of the species in the cloud water are reproduced
from the reference of Okita and Ohta (1979) as
shown in Fig. 2 in which the data are arranged
in the order of diminishing H+ concentration.
It was found that high H+ concentration was
accompanied by high NO3~ concentration and
high SC>42~ concentration exceeding NH4 + con-
centration on normality basis in range A. Cl~
concentration was independent of H"*" concentra-
tion. It was also found that in case of high H +
concentration (range A) the ionic balance was
approximately established between H + , NH4 + ,
SO42-, NO3~ and Cl~ ions. Therefore it seems
that the incorporation of acidic particulate and
PROCEEDINGS—PAGE 231
-------
December 1981
S. Ohta, T. Okita and C. Kato
Fig. 1 Map of Kanto area, x: Industrial area.
Table 1 Dales, places and number of people who complained eye- and skin-irritalion due
(o contaminated drizzle.
Date
Place
Number of people
June 28 and 29, 1973 j Several cities and towns on the coast
of Suruga Bay
Uenohara
Area covering Ashikaga. Sano, Ohira,
Koga and Kumagaya
Tokyo and Yokohama area
Kanuma
Kumagaya
July 3 tnd 4, 1974
June :5, 1975
441
32,730
144
gaseous species such as H:SO4 and HNOj into
the cloud droplets would give rise to acidic drop-
lets. In the ranges of less H* concentration (B
and C ranges) SO42~ concentration was usually
the same as NH4* concentration or slightly lower,
which presupposes that most of the S(V~ was
incorporated into the droplets in the form of
(NH«);SO4. In these ranges NOj~ concentration
was usually much lower than in range A.
3. Determination of equilibrium vapor
concentration of NOj over its dilute aqueous
solution
Since gaseous HNOj. HC1. NHj etc. may dis-
solve into cloud droplets to have influence on the
pH of the cloud water, the solubilities of these
gases into water are important data for the
formulation of the equilibrium model. The
solubilities of SO., HC1 and CO2 were sum-
marized by Orel
-------
Journal of the Meteorological Society of Japan
Vol. 59, No. 6
10
I10'3
110-*
o
10-2
§10'3
§
.210-*
2
"c
§10-*
10'6
Fig.
2 Concentrations of species in cloud
water arranged in the order of diminish-
ing H+ concentration.
A: pH<3.3, B: 3.3ŁpH<4.0,
C: 4.0^pH<5.0, D: 5.0
• H+, ANH4+,
xci-.
Fig. 3 Experimental setup for measurement
of NO2 solubility.
B: Tedlar bag, Ia: Bubbler containing dis-
tilled water, Ib: Bubbler containing Saltz-
man solution, F: Rotameter, P: Pump.
bag was passed through the bubblers /„ and 7b
with flow rate of 0.33 L min"1. 40 mL of distilled
and deionized water and Saltzman's reagent were
put into /a and /t, respectively. Whether the air-
water equilibrium was attained in /« was checked
by successive measurements of the concentration
of NOs sampled in Ib. It was found that within
20 minutes the concentration of NOj became
invariant. The high and low concentrations of
N0a~ in /„ were respectively determined by
NOj~ ion electrode and naphthylethylene diamine
(NEDA) colorimetric method. High NOj~ con-
centration was determined by NO»~ ion electrode.
In the case of low concentration of NO*~ it was
reduced to NO2~ by cupper-cadmium column
and then determined by NEDA method. pH
was measured on a pH meter.
Orel et al. (1977) expressed the air-water
equilibrium of NO2 as follows,
ZNOj+HzO ^=1 HNO2+H++NOj-
„ [HNOa]IH+][NOr]
/CNO, •• -
where PNOt is the pressure (atm) of NOj. HNOj
is dissociated as follows,
• 5.1xlO-j—
10~8atm and pH = 5.0 into this equation, NOS~
PROCEEDING6—PAGE 299
-------
S. Ohta, T. Okita and C. Kato
10-9
10°
101 102
N02 Concentration ppm
Fig. 4 Relation between -^NOJ and
concentration.
and NaNO3 may be neglected.
It is assumed that the air parcel containing
above mentioned gaseous and aerosol species
ascends along a side of a mountain and that the
cloud of water content of W (gm~3) is formed
at a level which is somewhat lower than the top
of the mountain. The initial supersaturation in
the cloud is difficult to measure, but Scott et al.
(1979) estimated that it is greater than 0.2%.
Therefore in the present calculation the initial
supersaturation is assumed to be 0.2%. It is
further assumed that no entrainment of gaseous
and aerosol components occurs after the forma-
tion of the cloud.
The radius r of solution droplets containing
m (g) of solute which is in equilibrium with the
air of relative humidity H is given by (Mason,
1971)
JL f ( 2ff'Mv \1
x|l+- ""'""
NO*
concentration becomes to be an order of 10~16
mol L"1 which is very small in comparison with
the NO3~ concentration in cloud water as shown
in Fig. 2.
4. A numerical model of acidification of cloud
water
In order to interprete the pH and the con-
centration of gaseous and aerosol species in the
atmosphere the following numerical simulation is
conducted.
It may primarily be assumed that in the atmos-
phere the gaseous species SO-,, CO* NO2, HNOS,
NH» and HC1, and aerosol components H2SO4,
(NHOaSO^ NasSO*. Nf^NOj and NaNO3 are
the major species affecting the pH and content
of ionic species of the cloud droplets.
As described in the previous section our experi-
ment shows that the solubility of NO2 in water
is so low that its effect may be neglected. Na2SO4
and NaNO3 may be produced by reacting H2SO4
and HNO3 with NaCl in sea-salt particles, so that
most of them would be in the micron-size range.
However, since the size distribution of SO42~
measured on Mt. Tsukuba showed that most of
the SO42~ was submicron in size (Okita, 1978),
therefore in inland area the presence of Na*SO4
(1)
where Mm and Mw are respectively molecular
weight of solute and water, p^ and pi! respec-
tively the density of water and solution, a sur-
face tension of solution, i van't HofFs factor,
R' gas constant and T is absolute temperature.
Fig. 5 is the relation between m and r derived
from equation (1) for the solution of (NH4)2SO4,
HaSO4 and NH4NO3. The measurements of the
size distributions of SO42~ and NO-T conducted
on the top of Mt. Tsukuba (Okita, 1978) and at
Nagoya (Kadowaki, 1976 and 1977) indicated
that more than 95% by weight of SO42~ and
NO3- is in the range above 0.1 //m in particle
radius. Since the relative humidity during the
measurements was usually below 80% most
aerosol mass might be consisted of particles with
mass of H2SO4, (NH4)2SO4 or NH4NO3 larger
than 10-15g.
The equilibrium radius r of solution droplet
when relative humidity exceeds 100% as calcu-
lated using equation (1) is shown in Fig. 6. This
figure indicates that the particles containing more
than 10-»5 g of H2S04, (NH4)2SO4 and NH4NO3
can be activated as condensation nuclei with the
supersaturation exceeding 0.2%.
In view of the above conclusion it may be
assumed that almost all the aerosol mass of
H;>SO4, (NH4)..SO4 and NH4NO3 may be incorpo-
rated into cloud droplets as condensation nuclei.
PROCEEDINGS—PAGE 234
-------
Journal of the Meteorological Society of Japan
Vol. 59, No. «
Droplet Radius r (cm)
Fig. 5 Mass of solute as a function of equi-
librium radius of solution droplet and
relative humidity H.
On forming a cloud droplet around the nucleus
it seems that water soluble gases are also incorpo-
rated into aqueous phase in a very short time
to establish gas-liquid phase equilibrium of these
gases.
The equations for the chemical equilibrium
of SOz-NHa-COa-HCl-HNOrHaO system are
10'
10'
Oroplet Radius r (cm)
Fig. 6 The equilibrium relative humidity (or
supersaturation) as a function of droplet
radius for solution droplets containing
indicated mass by m.
given in Table 2. The values of KM and
were measured by Davis and de Bruin (1964).
Kka was measured by Hales et al. (1979). Other
Henry's constants and dissociation constants are
those described in Table III of Orel et al.'s paper
(1977). 7-+, j- _ and 7-2- are the activity co-
efficients which are given by the following equa-
tions according to the Debye-Hiickel theory,
iog,o r« - logio 7-,- - - A vT*Y(i + vT)
+Q.2AW, (z-1,2,-), (2)
where A= 0.509 and / is ionic strength which
is given by
Table 2 Chemical equilibria in SO2-NH3-CO2-HC1-HNOS-H2O system.
Reaction
NHj(g) + H2O
CO2-H20;±H
HNO3(g) + H2O ;; HNO>-H2O
Equilibrium constant expression
Value'of equilibrium
I constant at 20°C
/C«,=[H*][S01'-]rtr.-/CHS05-]r-
KIC = [H *][HCO,->+r-/TCO, • H20]
X« = [HCl-H2O]/PHci
«,i=[H+]CCl-]r+r-/CHCl-H20]
K»»=CHNO3-HZO]//'HNO,
«,»=[H*][NO,-]7+r-/[HNO3-H2O]
1.008x10-"
1.24
1.27x10-'
6.24x10-'
92.9
1.774 x 10-»
3.4xlO-«
4.45 x10-'
4.68x10-"
19.0
1.3xlO«
2.1xlos
15.4
PROCEEDINGS—PAOI
-------
December 1981
S. Ohta, T. Okita and C. Kato
(3)
in which [/] and z{ are the concentration and
valency of i ions respectively.
In the cloud water the following electro-
neutrality is also established, that is
[H+]+[NH4+]=[OH-] + [HSOj-]
+2[SO»'-] + [HCOr]+2[COj2-]
+ [Cl-] + [NOr]+2[SO4*-] (4)
Substituting the expressions for the equilibrium
constants in Table 2 the equation (4) can be
written as
t
Since only a small amount of SO4*~ is pro-
duced by the oxdation of SO2 in cloud droplets
within several minutes (for example, Larson et al.
(1978)) it may be assumed that all the SO42~
in the droplets comes from aerosol SO42~. If
the pressure Pk of the gases k under equilibrium
condition are known, equation (5) may be solved
with respect to [H+] for specified values of ;•+,
j- and jz— According to the calculation, Pt
in the cloud are expressed by
P°NHixlO«. 2000m,j . 1000/w.
_ 24.04W
PNH,=
Pso,-
fso.xlO8
24.04W
r+r- 24.04 w
(7)
P°co. x 108
24.04^
108 '
24.04^
24.04 W
108 '
/"HNO.xlO* lOOOmn
24MW
KihKhh
(8)
(9)
(10)
24.04 W
where Pfc° is the initial pressure of the gas *, W
liquid water content (gin-*), ma=msb/(a+b),
and m, and m, are the concentrations (g m~s)
of aerosol SO42- and NO3~ components in the
atmosphere respectively. The letters a and b are
respectively the concentrations (molm"3) of
H2SO4 and (NH4)2SO4, and M, and Mn are re-
spectively molecular weight of SO42~ and NO3~.
The SO42~ concentration in cloud water is
given by
[SO42-] - 10QQm,/M,W. (11)
Substituting equations (6)-(ll) into equation
(5), then using equations (2) and (3), [H+] may
be obtained. The concentrations of the various
ions may then be calculated from equations
shown in Table 2. More details of the principles
of the model calculation has been described
elsewhere (Okita 1974, Okita and Ohta 1980).
The model calculation was conducted for the
four cases shown in Table 3, where R=a/(a+b)
Table 3 Initial concentrations of gases and aerosol components.
Case
P°so,
fNH,
P°HC1
P*co,
P°HNO3
m>
m*
R
W
(atm)
(atm)
(atm)
(atm)
(atm)
(ug m-1)
(/«g m-1)
{=a/(a + b)}
(g m-»)
I
2x10-'
2x10-'
10-'
3.3xlO-4
variable
20
0.6
0.5
0.2
n
2x10-'
2x10-'
10-'
3.3xlO-«
5xlO-»
20
0.6
variable
0.2
in
2xlO-«
5x10-'
io-»
3.3xlO-«
5 x 10-»
20
0.6
variable
0.2
IV
2xlO-»
2xlO-»
io-»
3.3xlO-«
variable
10
0.6
0.5
0.2
PROCEEDINGS—PAGE 236
-------
which shows the contribution of HaSC^ on total
sulfate.
5. Results of calculation
The results of the model calculations are shown
in Figs. 7-10.
Case I: Fig. 7 indicates that with increasing
P°HNOj both [NO3~] and [H+] increase. The
computed pH are 2.73 and 2.54 with P0HNo3 of
5xlO~9 and 10~8atm respectively. On the other
hand, there is little dependence of [SO42~],
[NH4+] and [Cl~] upon P«HNoj.
Case II: Fig. 8 indicates the increase of [H+]
and decrease of [NH4+] with increasing R. There
ID'2
Journal of the Meteorological Society of Japan
10-2
Vol. 59, No. 6
o
iio'3
I
'c
O
O
10'
•X -
P°
"HNOj
10
atm)
Fig. 7 Calculated concentrations of species
in cloud water in case I of Table 3.
XC1~.
10*1-1—
o
c
°10
o
o
'3
10'
x—x—x—x
-I • • •
-X—X X .
8
| 10'3
o
"c
01
c
O
u
T
x— x— x
— x — * - x — x— x— x -
0
0.5
R
1.0
Fig. 8 Calculated concentrations of species
in cloud water in case II of Table 3.
Symbols in the figure are the same as
in Fig. 7.
0.5
R
1.0
Fig. 9 Calculated concentrations of species
in cloud water in case HI of Table 3.
Symbols in the figure are the same as
in Fig. 7.
10-2
c
o
"o 10'3
I
3
iff
1-4
-X -
10
HN03
(xiO~9 atm)
Fig. 10 Calculated concentrations of species
in cloud water in case IV of Table 3.
Symbols in the figure are the same as
in Fig. 7.
is little variation of [SO42-], [NO3~] and [Cl"]
with R. If all the sulfate is sulfric acid, pH is
calculated to be 2.54.
Case III: In comparison with Fig. 8, Fig. 9
indicates that with the increase of P0NHo3 [NHi+]
increases and [H+] decreases whereas little change
occurs of [SO42-], [NO3-] and [Cl~]. The figure
also indicates that the pH values are 2.90 and
2.64 with R of 0.5 and 1.0 respectively.
Case IV: In comparison with Fig. 7, Fig. 10
shows that when the initial concentration of
SO42~ is reduced by half [SO42-], [NrV] and
[H + ] decrease whereas [NO3-] and [Cl~] are
not subject to change. The values of pH are
PROCEEDINGS—PAGE 237
-------
December 1981
S. Ohta, T. Okita and C. Kato
respectively 2.87 and 2.62 with P^HKO, of 5x
1C-9 and ID-8 atm.
6. Comparison of calculation with measured
concentrations
The minimum pH of the cloud water collected
at Mt. Tsukuba was 2.80 (Fig. 2). The measure-
ment at Mt. Tsukuba (Okita, 1968) indicated
that the mean cloud water content was about
0.2 g m~3. The estimaiton of the cloud water
content from the amount of water collected on
wire net during the 1975-1978 observation period
gave about the same content as the previous
measurement.
On the top of Mt. Tsukuba the highest HNO3
concentration was 1.5xlO~9atm, but it was as
high as 6.5xlO-aatm at height of 700m near
Mt. Tsukuba as measured on board a helicopter
(Okita and Ohta, 1979). On the top of Mt.
Tsukuba NH3 concentration was as high as
(2-8)xlO~9atm probably due to its generation
on the mountain, but the measurements using
helicopter and tethered balloons indicated that
it was about 2xlO~9atm at height of 500-
1000 m in the free atmosphere (Okita and Ohta,
1500
0.5 .
(SO!")-(NHt)
1.0
Fig. 11 The vertical distributions of the ratio
of difference of equivalent concentrations
of «V~ and NH4+ to that of SO42~
measured in July 5-7, 1978 using a heli-
copter over the Kanto plain.
• July 5, O July 6, X July 7.
Table 4 Comparison of calculated and measured concentrations of species in cloud water.
Initial values in air
Calculated values
Measured values in cloud water
SO* 2xlO-» atm
NHj 2xlO-» »
HNOa 5xlO-»
HC1 10-»
COt 3.3xlO-«
W 0.2 g m-»
SO,*- 20 /tg m-J
R 0.5
NOr 0.6 tig m-1
in cloud air
SOt 1.9999x10-' atm
NHa 4.91x10-'* »
HNO, 5.46x10-" »
HC1 1.37xlO-'« "
COt 3.30xlO-« *
in cloud water
pH 2.73
H* 1.87xlO-'molL-'
OH- 6.21xlO-« »
HSO,- 1.94x10-'
SO,*- 8.62x10-"
SO,*- 1.04xlO-» "
NO," 1.09x10-'
NH,+ 1.51xlO-J
CI- 2.08x10-*
HCOr 3.08x10-'
CO,*- 1.02xlO-'« //
I 4.42x10-'
r*,r- 0.9305
n+,rz- 0.7497
pH 2.80
H+ 1.59xIO-> mol L->
SO,*- 0.89x10-' *
NO,- 1. 10x10-'
NH,» 1.11x10-' >,
Cl- 1.83xlO-«
PROCEEDINGS—PAGE 238
-------
Journal of the Meteorological Society of Japan
Vol. 59, No. 6
1979). The measurements on the top of Mt.
Tsukuba indicated that the maximum SOa and
HC1 concentrations were about 2xlO~9 and 10~9
atm respectively. The maximum SO42~ concentra-
tion was 21 jug m~3 on the top of Mt. Tsukuba.
No measurement has been made of the con-
centration of HjSOi particles, but at Mt. Tsukuba
SO42~ concentration was, on the average, about
30% higher than that of NH4+ on normality
basis (Okita and Ohta, 1979). Further, it may
be presumed that some H2SO4 collected on filter
was reacted with NH3 gas to produce (NHOzSC^
or NH4HSO4. The concentrations of SO42~ and
NHi* components of aerosols in the atmospheric
boundary layer was measured over the Kanto
plain 50 km west of Mt. Tsukuba in July 5-7,
1978 by using a helicopter. The vertical distri-
butions of the ratio of difference of equivalent
concentrations of SO42~ and NH4+ to that of
SO42~ are shown in Fig. 11. Above the height
of 500 m, this ratio becomes 0.80, but at Mt.
Tsukuba the ratio seems to decrease slightly on
account of the generation of gaseous NH3 from
the surface of the mountain. Therefore, there
is a possibility that more than a half of the
SO42~ would be H2SO4. The concentration of
NOs~ was usually much lower than that of
so42-.
The above measurements indicated that there
is a possibility of the formation of cloud of
liquid water content of 0.2 g m~3 in the air of
P*so,=2x10-'atm, P°HNO,=5x IQ-'atm, P°NH3
=2x10-'atm, P°co, = 3.3x IQ-'atm, P°Hci = 10-'
atm, ms=20/*gm-3, R=0.5, mn = 0.6 pgm-3. The
calculated gas-liquid equilibrium concentrations
are as shown in Table 4, which correspond to
the case of P°HNC>3 = 5 x 10"9 atm in Fig. 7.
Table 4 indicates that HNO3 and HC1 gases
are almost completely dissolved into cloud drop-
lets and more than 99.7% of NH3 is also
absorbed into the droplets. On the other hand,
SOi and CO2 are scarcely absorbed and thus the
concentrations of HSO3-, SO32~, HCO3~ and
CO32- are very low. The pH value is 2.73 which
is slightly lower than the minimum pH of 2.80
observed at Mt. Tsukuba. In addition, the
calculated concentrations of SO42~, NH4 + , NO3~
and CI~ are close to their measured concentra-
tions in the cloud water of pH 2.80.
Therefore it may be concluded that low pH
of cloud water observed on Mt. Tsukuba may
be interpreted by the incorporation of acidic
aerosol such as H2SO4 and acidic gases such as
HNO3 and HC1 into the cloud droplets even in
the presence of neutralizing gases such as NH3.
7. Discussion and conclusion
Our model calculation indicates that the pH
of cloud water of less than three may be inter-
preted by the formation of cloud droplets upon
H2SO4 nuclei and dissolution of HNO3 into the
droplets if both species are present in the atmos-
phere with concentration of about ten ppb or
fig m~3 even in the presence of NH3.
The measurements of gaseous and paniculate
species and of the species in cloud water on the
summit of Mt. Tsukuba and in the free atmos-
phere near Mt. Tsukuba confirmed the validity
of our model calculation.
However, our equilibrium model may only be
applied to the cloud which elapses within 20
minutes after its formation. When the cloud
stays longer the chemical reaction within the
cloud droplets such as those between HSO3~ and
O3 and between NO2~ and O3 would produce
additional H2SO4 and HNO3. Further closed
model may not be applied to long-lived cloud.
In addition our model may be applied only to
the A range in Fig. 2 and for the interpretation
of higher pH value other mechanism must be
incorporated.
Moreover, more detailed knowledge of the
vertical and horizontal distributions of H2SO4
panicles and HNO3 gas and the mechanism of
their formation-in the atmosphere are necessary
in order to more clearly understand the mecha-
nism of the formation of acid precipitation.
Acknowledgements
The authors express their thanks to the
Japanese Environment Agency for their financial
support of their work.
References
Davis, W., Jr., and H. J. de Bruin, 1964: New ac-
tivity coefficients of 0-100 percent aqueous nitric
acid. J. Inorg. Nucl. Client., 26, 1069-1083.
Hales, J. M., and D. R. Drewes, 1979: Solubility
of ammonia in water at low concentrations. At-
mospheric Environment, 13, 1133-1148.
Kadowaki, S., 1976: Size distribution of atmospheric
total aerosols, sulfate, ammonium and nitrate
particles in the Nagoya area. Atmospheric En-
vironment, 10, 39-43
, 1977: Size distribution and chemical
composition of atmospheric particulate nitrate in
the Nagoya area. Atmospheric Environment, 11,
671-675.
Kato, S., N. Ono, and H. Tohata, 1978: A basic
PROCEEDINGS—PAGE 239
-------
December 1981
S. Ohta, T. Okita and C. Kato
study of treatment of NOx in wet system (Ab-
sorption of NOx into and subsequent reaction
with water). Abstracts of 19th Annual Meeting
of the Japan Air Pollution Soc., 469 (in Japa-
nese).
Larson, T. V, N. R. Horike, and H. Harrison, 1978:
Oxidation of sulfer dioxide by oxygen and ozone
in aqueous solution: a kinetic study with signifi-
cance to atmospheric rate processes. Atmospheric
Environment, 12, 1597-1611.
Mason, B. J., 1971: The physics of clouds, Oxford
Univ. Press, p. 26.
Okita, T., 1968: Concentration of sulfate and other
inorganic materials in fog- and cloud-water and
in aerosol. 1. Meteor. Soc. Japan. 46, 120-127.
, 1974: Mechanism of formation of acid
rain (II). Taikiosen-kenkyu, 9, 176 (in Japanese).
1978: Recent investigation of transfor-
mation and deposition of atmospheric constitu-
ents. Tenki, IS, 110-119 (in Japanese).
Okita, T., and S. Ohta, 1979: Measurement of ni-
trogeneous and other compounds in the atmos-
phere and in cloud-water: A study of the mecha-
nism of formation of acid precipitation. Nitro-
geneous Air Pollutants, Chemical and Biological
Implications. Grosjean, D. ed., Ann Aabor Sci-
ence, 289-305.
, 1980: Acid Rain. Analysis of Mechanism
of Air Pollution. Suzuki, T. ed., Sangyo Toshyo,
203-231 (in Japanese).
Orel, A.E., and J.H. Seinfeld, 1977: Nitrate for-
mation in atmospheric aerosols. Envir. Sci. Tech-
nol., 11, 1000-1007.
Scott, B.C., and N. S. Laulainen, 1979: On the
concentration of sulfate in precipitation. /. Appl.
Meteor., 18, 138-147.
ffl
ft -
PH3
4*-, NO,-
pH,
PROCEEDINGS—PAGE 240
-------
URBAN OZONE MODELING DEVELOPMENTS IN THE U.S.
presented by B. Dimitriades
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 241
-------
This discussion will deal (a) with the results from recent efforts in the
U.S. to develop and evaluate urban 63 models, (b) with the main imperfections
in the 03 models currently in existence in the U.S., and (c) with some sugges-
tions for additional research needed in this area that could be done independently
or cooperatively with Japan.
There are three types of urban 03 models now in use or contemplated in
the U.S.:
(A) Mechanistic or EKMA-Type Models
These models predict 03 concentrations from VOC and NOX concentrations in
ambient air. The precursor concentrations are either measured directly
or are estimated from emission rates through simple dilution calculations.
The relationships between 63 and precursor concentrations are used in the
form of 63 isopleth curves derived from the chemical mechanism. Dispersion
is not treated by the model but advection effects are treated in the
"trajectory" form of the EKMA model. For application, we need to know
primarily the max. 63 concentration (or 90-percentile or etc.) and the
VOC/NOX ratio during 6-9 am. For more accuracy, we need to input also
information on background pollutant concentrations, diurnal mixing height
variation, and post-9-am hourly emission rates. The model is used in a
relative sense, that is, to predict Oo changes from precursor changes.
Ideally, therefore, the model should be evaluated by comparison against
observations of 03 air quality changes. There are several USEPA documents
describing the EKMA model, which have been sent to Japan.
(B) Air Quality Simulation Models (AQSM)
Unlike EKMA, the AQSMs treat emission dispersion in detail, which enhances
model validity but it also adds considerable complexity. The AQSMs treat
chemistry with detail which currently is comparable to that of the EKMA
chemistry. Unlike EKMA, however, the computer capacity demands associated
with the greater spatial detail of the AQSMs put a limit to the detail
with which AQSMs can treat chemistry. The AQSM models are designed to
predict absolute air qualities, and their evaluation, therefore, is done
by comparison against absolute air quality observations.
(C) Photochemical Box Models (PBM)
These are models which with respect to conceptual validity and complexity
lie between EKMA and AQSM. They treat chemistry with detail comparable
to that of EKMA, and advection as with the "trajectory" form of EKMA;
they do not treat horizontal or vertical dispersion. Unlike EKMA, the
PBM does not use 03-isopleths to compute 63 from VOC and MOX inputs; it
relies instead on built-in mathematics. Also, unlike EKMA, it has been
designed to predict absolute air quality; its evaluation, therefore, is
based on comparison against absolute air quality observations.
PROCEEDINGS—PAGE 243
-------
Evaluation status and availability of these models is as follows. Three
AOSMs and one PBM have been evaluated against field data taken in the St. Louis
area during 1974-1977 (RAPS data). This evaluation and results are discussed
in detail in the EPA Report "Final Evaluation of Urban Scale Photochemical
AQSMs" (sent to Japan), and in a summary form in a report by Schere and Shreffler
(attached here). Briefly, the three AQSMs were a Langrangian model developed
by ERT (ELSTAR) , and two Eulerian, grid-type models developed by Lawrence
Liver-more Laboratory (LIRAQ) and by Systems Applications, Inc. (SAI), respec-
tively. The fourth model was a PBM using a chemical mechanism developed by
Demerjian. First results from the LIRAQ evaluation showed such a poor model
performance that they discouraged further evaluation of the model. Results
from the other model evaluations are illustrated in Figure 1 and Tables 1 and
2. Of these models, the ELSTAR is now available for use; the SAI and PBM will
take a few more months before they are completely documented and ready for use
by others.
Evaluation of the EKMA model was discussed in detail in a previous presen-
tation in this Conference (Photochemical Pollution Panel). Briefly, evaluation
efforts in the last two years included a comparison of the EKMA mechanism with
other mechanisms currently in existence in the U.S., and evaluations against
field data. The mechanism intercomparison effort revealed substantial disagree-
ments suggesting that most or all of the intercompared chemical mechanisms
have inaccuracies, a problem that, obviously, affects not EKMA only, but all
models that include an C^-chemistry component. The field validations were all
of limited conclusiveness for a number of reasons, discussed in detail in an
EKMA Workshop conducted by USEPA in December, 1981 (Workshop Proceedings have
been sent to the Japanese Environment Agency). The EKMA model and guidelines
for its use are currently available for use.
^perfections in the current models exist in all their components, that
is, in the emissions, emission dispersion, and chemistry modules. The chemical
composition of the VOC emissions, as measured currently, has uncertainties,
one problem being that the currently available emission composition data were
taken for the most part from laboratory testing of emissions and do not
necessarily reflect actual composition in the ambient air. The horizontal
adyection model component also is subject to errors because of "artificial
diffusion" caused by the finite difference numerical advection algorithm.
Lastly, the chemistry module is subject to uncertainties, as revealed from the
EKMA mechanism intercomparison study.
While some research is being conducted in the U.S. in response to the
above problems, we believe that additional efforts done cooperatively by Japan
and the U.S. will be of considerable value. Specifically, as we have already
discussed in the Photochemical Pollution panel meeting, we propose a cooperative
mechanism intercomparison study that will include this time the Carter/Akimoto
(and other) mechanism(s) not included in the U.S. studies. We would propose,
further, testing of existing urban 03 models against air quality data for
Japanese urban areas (such as Tokyo). In a more general sense, we would
invite our Japanese colleagues to participate in efforts to develop standard
sets of chemical kinetic data, of smog chamber data, and of field data, against
which existing and new urban 03 models could be evaluated.
PROCEEDINGS—PAGE 244
-------
PBM SIMULATION-750522
» » I
-{-[• i i i i i i i i r i i i i
/\,^
\ Model Prediction^
' I i i r i I i i i i I i "j. i i I t j I i I i i i t
7-5 10.0 12.5 15.0
TIME, HOURS (CST)
17.5 20.0
Figure 1. PBM simulation results for 03 - Day 142 of 1975.
PROCEEDINGS—PAGE 245
-------
TABLE 1. A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA FROM
THE ELSTAR FOR 20 DAYS SELECTED FROM 1975-1976.
Day
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
Hour
12
14
12
15
13
14
11
15
13
14
14
15
16
15
15
12
13
13
11
14
Station
101
112
125
124
118
113
118
121
121
121
122
122
115
114
120
108
117
109
120
102
Observed
4-meters
(ppm)
0.20
0.20
0.16
0.21
0.18
0.18
0.21
0.17
0.19
0.23
0.26
0.20
0.22
0.22
0.16
0.17
0.17
0.23
0.18
0.24
LPM Predicted
Level-1 Level-3
(ppm) (ppm)
0.07
0.15
0.25
0.18
0.21
0.11
0.14
0.16
0.18
0.19
0.29
0.18
0.21
0.15
0.11
0.10
0.08
0.20
0.30
0.26
0.13
0.31
0.24
0.18
0.22
0.13
0.15
0.16
0.17
0.19
0.30
0.18
0.23
0.15
0.12
0.11
0.09
0.22
0.29
0.30
PROCEEDINGS—PAGE 246
-------
TABLE 2. A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA FOR THE SAI
Julian
date4
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
Hour
(CST)
12
14
11
10
13
14
11
15
13
14
12
14
16
15
15
12
13
13
11
14
Staff on
101
112
12lb
119&
118
113
118
121
121
121
121&
114&
115
114
120
108
117
109
120
102
Observed
at 4-meters
(ppm)
0.195
0.202
0.142
0.171
0.184
0.185
0.209
0.166
0.193
0.233
0.179
0.172
0.221
0.223
0.155
0.170
0.166
0.225
0.176
0.244
UAM
Specific
(ppm)
0.116
0.156
0.083
0.154
0.132
0.141
0.128
0.118
0.095
0.133
0.146
0.142
0.121
0.160
0.143
0.128
0.050
0.077
0.119
0.220
Predicted
Independent
(pgm)
0.238
0.243
0.166
0.209
0.234
0.252
0.195
0.149
0.214
0.205
0.178
0.312
0.190
0.174
0.169
0.155
0.073
0.124
0".203
0.246
fTwenty days selected from 1975 and 1976.
Overall maximum at 122, 12-3, 124, -or 125; outside SAI domain.
PROCEEDINGS—PAGE 247
-------
URBAN-SCALE PHOTOCHEMICAL MODEL EVALUATIONS FOR OZONE
USING RAPS DATA
Kenneth L. Schere and Jack H. Shreffler
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Environmental Protection Agency
Research Triangle Park. North Carolina 27711
1. INTRODUCTION
The Regional Air Pollution Study (RAPS) was conducted in the
St. Louis area over the period 1974-1977 ISchiermeier, 1978).
RAPS was designed to provide a comprehensive data set for the
testing and evaluation of numerical air quality simulation models
on an urban scale* This paper describes some of the evaluation
tests that have been performed by EPA on three such models* One,
a relatively siciple box-type model, was constructed at EPA's
Meteorology Laboratory. The second, a Lagrangian model* was
developed by Environmental Research and Technology, Inc. of Santa
Barbara, California, and the third, a 3-D grid model, was
developed by Systems Applications, Inc. of San Rafael, California.
This paper summarizes the results of simulations for 20 days
where maximum 1-hr o/one concentrations were observed at 0.16 to
O.Z6 ppa in St. Louis* Generally these days exhibited stagnation
conditions with little cloud cover and represented the situations
conducive to production of photochemical smoy from local emissions
CTable 1). The simulations were designed to evaluate the ability
of the 3 models to reproduce the observed ozone maxima.
2. PHOTOCHEMICAL BOX MODEL
The Photochemical Box Model CPBM) is a single cell Eulerian
air quality simulation model whose purpose is to simulate the
transport and chemical transformation of air pollutants in smog
prone urban atmospheres. Ihe model's domain is set in a variable
volume, well nixed reacting cell where the physical and chemical
processes responsible for the generation of ozone by its
hyarocarbon and nitrogen oxides precursors are mathematically
created*
In the application of the model to the St. Louis RAPS data
PROCEEDINGS—PAGE 248
-------
TABLE 1. A SUMMARY OF WIND SPEED (WS), WIND DIRECTION (WD), TEMPERATURE,
SOLAR RADIATION (ALL WAVELENGTHS). AND MAXIMUM AFTERNOON MIXING
HEIGHT (MH) FOR THE 20 DAYS EXAMINED.*
Date
Day
(Julian)
WS
(m/s)
WD
(deg)
Temp
(°C)
Solar
(ly/min)
Max MH
(m)
5/22/75
6/27/75
7/01/75
7/02/75
7/03/75
7/26/75
7/28/75
8/09/75
8/18/75
8/19/75
9/08/75
6/07/76
6/08/76
7/13/76
7/29/76
7/30/76
8/12/76
8/13/76
8/24/76
10/01/76
142
178
182
183
184
207
209
221
230
231
251
159
160
195
211
212
225
226
237
275
1.1
0.4
1.4
1.4
1.8
1.0
2.0
0.4
1.6
1.3
1.8
1.0
1.3
2.3
0.3
1.7
2.3
1.1
1.3
0.6
224
245
70
15
324
139
18
88
167
168
181
129
284
145
251
205
253
273
110
222
29
29
29
30
30
26
X
26
27
28
25
25
27
28
25
X
29
30
28
22
1.12
0.96
0.99
0.92
0.85
0.98
0.98
0.98
0.96
0.95
0.89
1.06
1.01
1.02
0.53
0.82
0.70
0.86
0.82
0.78
1504
1822
2606
2488
1875
1477
1909
1195
1488
1052
1797
1972
1772
1853
1706
1X4
730
1427
2124
527
*Meteoro1ogical variables (except MH) are network averages over
the period 0700-1359 CST.
PROCEEDINGS—PAGE 249
-------
base the horizontal Length scale of the single cell was 20 km and
the vertical scale uas time-varying, proportional to the aepth of
the oixea layer* The model domain was centered on downtown St.
Louis such that the 20x20 km area encompassed most of the major
emissions sources on either side of the Mississippi River. Source
emissions were assumed to be distributed uniformly across the
surface face of the cell. Twelve of the surface monitoring
stations as well as one upper air sounding location were also
located vithin the cell's boundaries.
The P&K contains a chemical kinetic mechanism with 36
reactions and 27 reactive species. The set of equations
describing the rates of change of the concentrations of these
species is numerically solved at time steps on the order of 10
minutes. Model simulation is started at 0500 CST and continues
through 1700 CST. From these solutions the model then determines
the houi—average predicteo concentrations.of all modeled species.
The simulated concentrations from the PBM represent spatial
averages over the volume of the box.
Table 2 presents a summary of the predicted and PEN
domain-average observed ozone maxima and an elementary statistical
analysis of the results tor all 20 modeled days. The 'specific'
model predictions are ttoose that correspond to the same hour as
the observed maximum and the 'inoependent' predictions represent
the peak at any hour of the simulation. The specific and
independent predictions may not coincide, indicating a phase lag
between observed and predicted ozone peaks. This lag often
appears -hen the maximum observed ozone within the model domain
occurs before noon. Statistics on the residual concentration have
been computed for both the specific and independent predictions.
Both the average signed resi ana I and absolute residual, and
standard deviation (s.o.) are presented for the analyses.
Ttie average residual is negative in both the specific and
independent cases, indicating an overprediction of ozone, although
the specific value is half the independent value. The value of
the average absolute residuals are both different than the signed
residuals. This implies that there were underpreoictions as well
as overpredictions among the individual days. The magnitude of
the s.d. is slightly greater than the average residual. The
discrepancy on Day 251 accounts for a good portion of the s.d.
The average value of the model's overprediction for ozone over all
20 days is 23*.
3. LAGRANGIAN PHOTOCHEMICAL MODEL
The Layrangian Photochemical Model (LPM) envisions a portion
of the atmosphere as an identifiable parcel. As the parcel moves
over the region, pollutants are assimilated, vertically mixed, and
subjected to photochemical reactions in the presence of solar
PROCEEDINGS—PAGE 250
-------
TABLE 2. A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA FOR THE PBM
Julian
date3
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
Hour
(CST)
12
14
13
13
13
14
10
13
13
10
15
13
14
15
15
12
12
15
11
14
Observed
at 4-meters
(ppm)
0.137
0.159
0.094
0.115
0.114
0.140
0.108
0.132
0.072
0.084
0.084
0.125
0.162
0.151
0.094
0.127
0.111
0.144
0.115
0.183
PBM
Specific
(ppm)
0.125
0.162
0.128
0.070
0.182
0.144
0.087
0.125
0.096
0.059
0.205
0.203
0.166
0.132
0.084
0.114
0.088
0.171
0.132
0.216
Predicted
Independent
(ppm)
0.148
0.173
0.140
0.097
0.182
0.145
0.122
0.125
0.119
0.101
0.205
0.210
0.170
0.137
0.088
0.121
0.092
0.171
0.156
0.223
aTwenty days selected from 1975 and 1976.
For the data displayed above:
AC = Obs - Specific
1C" = -0.012
s.d.(AC) = 0.039
= 0.029
AC = Obs - Independent
1C = -0.024
s.d.(AC) = 0.036
lC = 0.031
PROCEEDINGS—PAGE 251
-------
radiation. The parcel expands along its trajectory at a rate
proportional to a Gaussian signa-y derived from the stability
class. The LPM is attractive relative to arid models in that it
is fairly simple to execute and uses a moderate amount of computer
tine .
The model is executed using a series of modules, sequentially
performing calculations on meteorology and air quality* emissions
and photochemistry. The input and running procedures described by
Lurmann et al. (1979) have been followed generally, although some
modifications were deemed necessary as more experience with the
RAPS data base was acquired.
The first step in setting up a simulation is to determine the
starting point of a parcel so that it will arrive at a specified
point at an assigned time. A backward trajectory is generated by
inverse-distance squared weighting oi winds from the closest 3
monitoring stations. However, experience with wind data suggests
that even closely situated stations can show large unexplained
differences in wind vectors from time to time. Reliance on a
singlet anomalous station, if the parcel has a close approach, can
create an erratic trajectory. To eliminate the possibility of
such vagaries, it was decided to compute a single resultant wind
vector from the network and assign it to all stations for the
hour. The start time of a parcel (which is on the hour) must be
at least 10 rain past local sunrise. Once the start position is
set, the model is run in a for*ard-trajectory node until IbOO CST
or the parcel leaves the region.
In nost cases, the parcel starts in a relatively clean rural
environment, and levels of hydrocarbons, nitric oxide and nitrogen
dioxide are assumed to decrease with height according to an
assigned formula. Cn the other hand, ozone is depleted near the
ground at night; thus, the initial ozone profile increases with
he ight.
Table 3 presents a summary of the predicted and observed ozone
maxima and a statistical evaluation of the results for the 20 days
of the stuay. The prediction refers to the model prediction at
the tirce and position of the observed ozone maximum. The
resiouals are calculated for both the first (L-1) and the third
(L-3) vertical mode I level predictions. The average residual for
L-1 inaicates slight underprediet ion , while the average residual
for L-3 is essentially zero. For both levels, the s.d.'s are
identical, and the average absolute deviations are nearly so.
A. URBAN AIRSHED KODEL
The Urban Airshed Mooel (UAM) is a 3-D grid-type, or Eulerian,
photochemical air quality simulation model (PAGSM). The structure
of the model consists of a latticework array of cells, the^ tot-l
volume of which represents an urban-scale domain and in which the
PROCEEDINGS—PAGE 252
-------
TABLE 3- A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA
FROM THE LPM FOR 20 DAYS SELECTED FROM 1975-1976.
Day
Hour
Station
Observed
4-meters
(ppm)
LPM Predicted
Level-1 Level -3
(ppm) (ppm)
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
12
14
12
15
13
14
11
15
13
14
14
15
16
15
15
12
13
13
11
14
101
112
125
124
118
113
118
121
121
121
122
122
115
114
120
108
117
109
120
102
0.20
0.20
0.16
0.21
0.18
0.18
0.21
0.17
0.19
0.23
0.26
0.20
0.22
0.22
0.16
0.17
0.17
0.23
0.18
0.24
0.07
0.15
0.25
0.18
0.21
0.11
0.14
0.16
0.18
0.19
0.29
0.18
0.21
0.15
0.11
0.10
0.08
0.20
0.30
0.26
0.13
0.31
0.24
0.18
0.22
0.13
0.15
0.16
0.17
0.19
0.30
0.18
0.23
0.15
0.12
0.13
0.09
0.22
0.29
0.30
For the data displayed above:
AC = Obs minus L-l Pred AC = Obs minus L-3 Pred
AC" = 0.023 ATT = 0.005
s.d.(AC) = 0.058 s.d.UC) = 0.058
TACT = 0.052 TACT = 0.050
PROCEEDINGS—PAGE 253
-------
physical and chemical processes responsible for photocheaical smog
are mathematically simulated. The horizontal dimensions of each
cell are constant but the heights of the celts vary throughout a
• ode I simulation as the depth of the mixed layer in the (JAM
changes accordingly.
In the application of the model to the St. Louis data base,
the area modeled was 60 k.n wide and Ł0 km long. Each individual
cell was 4 km on a side in the horizontal. Vertically* there were
4 layers oi cells in total; the bottom 2 layers simulated the
mixed layer and the top 2 represented the region immediately above
the mixed layer. The domain of the UAH was centered just west of
downtown St. Louis and included the entire metropolitan area.
The package of programs constituting the UAM actually contains
12 data preprocessing programs as well as the simulation model.
The uata requirements for applying the model are rather intensive.
The preprocessors access the RAPS surface-based hour-average air
quality and meteorological data base, the upper air pibal and
radiosonde data, and the source emissions inventory for the
necessary parameters, and process the parameters as required by
the simulation model. Simulations begin at 0500 CST and continue
through 1700 CST. The model numerically calculates the rates of
change of species concentrations at time steps on the order of
several minutes, and then Determines from these the hour-average
predicted concentrations.
Table 4 presents a summary of the predicted and observed ozone
maxima and a statistical analysis of the UAH results for all 20
modeled days. The 'specific* mooel predictions are those that
correspond to the same location and time as the observed maximum
ana the "independent* predictions represent the maximum
hour-average ozone generated by the model at any lowest layer grid
cell at any time during the simulation. The specific and
independent model predictions, as seen in Table A, typically do
not coincide. This inaicates that the maximum value of ozone
produced by the model did not correspond either in space or time,
or both, to the maximum value observed in the atmosphere.
Statistics on the residual concentration, as shown below, have
been computed for both the specific and independent predictions.
Moreover, the average signed residual and s.d. are presented for
the analysis.
Values of the average residual for • the specific and
independent cases, C.062 and -C.006 ppm respectively, differ
noticeably. The independent predictions show more promise toward
maximum ozone simulation. The average absolute residual is much
larger than the average signed residual in the independent case,
indicating both over- and underpredictions among the 20 days
tested. In the specific case however, the average signed and
absolute residuals are identical, signifying that the model has
unaerpredicted consistently over all 20 days. The s.d.'s of the
specific and independent resiouals are 0.035 and 0.053 ppn
respectively. The model's average performance for maximum ozone
PROCEEDINGS—PAGE 254
-------
TABLE 4. A SUMMARY OF PREDICTED AND OBSERVED 03 MAXIMA FOR THE UAM
Julian
date*
142 (1975)
178
182
183
184
207
209
221
230
231
251
159 (1976)
160
195
211
212
225
226
237
275
Hour
(CST)
12
14
11
10
13
14
11
15
13
14
12
14
16
15
15
12
13
13
11
14
Station
101
112
12lb
119&
118
113
118
121
121
121
12lb
114b
115
114
120
108
117
109
120
102
Observed
at 4-meters
(ppm)
0.195
0.202
0.142
0.171
0.184
0.185
0.209
0.166
0.193
0.233
0.179
0.172
0.221
0.223
0.155
0.170
0.166
0.225
0.176
0.244
UAM Predicted
Specific
(ppm)
0.116
0.156
0.083
0.154
0.132
0.141
0.128
0.118
0.095
0.133
0.146
0.142
0.121
0.160
0.143
0.128
0.050
0.077
0.119
0.220
Independent
(ppm)
0.238
0.243
0.166
0.209
0.234
0.252
0.195
0.149
0.214
0.205
0.178
0.312
0.190
0.174
0.169
0.155
0.073
0.124
0.203
0.246
aTwenty days selected from 1975 and 1976.
bOverall maximum at 122,123,124, or 125; outside UAM domain.
For the data displayed above:
AC = Obs - Specific AC «
IF = 0.062 1C" «
s.ti.UC) = 0.035 s.d.UC) =
IC = 0.062 |AC| »
Obs - Independent
-0.006
0.053
0.041
PROCEEDINGS—PAGE 255
-------
over all simulations shows a 32.AZ underprediction in the specific
case and a 4.4X overprediction in the independent.
5. CONCLUSIONS
The evaluation of the P&K, LP« and UAM for ozone is now
completed by the Meteorology Division of EPA. A full report
documenting the study is currently being prepared (Schere and
Shreffler, 1982).
Evioence shows the PBM to be a useful tool in assessing
area-wide urban air quality for photochemical ly reactive
pollutants, especially in stagnation conditions. In order for
this model to be a better aid to the regulatory communityt the
relationship between the average ozone, concentrat ion predicted
within the model domain ano the maximum ozone level observed at a
single station must be stuoied.
The LPM also shows promise to be an effective aid in
understanding urban ozone production. The model is relatively
easy to use, inexpensive to execute, and seems immune to various
execution errors which tend to arise unexpectedly in complex
computations of this sort.
From the work performed on the UAH it is clear that the
potential use of a grid-type PAQSM such as this one is great,
although the complexity of the model often makes the solution of
problems that arise wore difficut. The model evaluation effort
discussed here has shown the UAK to simulate ozone maxima in an
independent sense in St. Louis quite well, although generally
underpredicting in a specific sense.
REFERENCES
Lurwann, F., 0. Godden, A.C. Lloyd, and R.A. Nordsieck, 1979:
A Layrangian Photochemical Air Quality Simulation Model.
Vol. I-Sodel Formulation, Vol. Jl-User*s Manual.
EPA-60G/8-79-015a,ij (available from NTIS as PB 300470
and PB 300471).
Schere, K.L. and J.H. Shreffler, 1982: Final Evaluation
of Urban-Scale Photochemical Air Quality Simulation
Models. EPA Report, U.S. Environmental Protection
Agency, Research Triangle Park, NC (in press).
Schiermeier, F.A., 1976: Air monitoring, milestones:
RAPS field measurements are in. Environ. Sci. Technol
12. 644-651.
PROCEEDINGS—PAGE 256
-------
A NUMERICAL SIMULATION OF LOCAL WIND AND
PHOTOCHEMICAL AIR POLLUTION
presented by F. Kimura
Meteorological Research Institute
Japan MA
PROCEEDINGS—PAGE 257
-------
A Numerical Simuration of Local Wind and
Photochemical Air pollution
Fuj i a K i mura
Meteorological Research Institute. Tsukuba, Japan
Presented at 7th Japan-US Joint Meeting on Air Pollution Related
Meteorology and Photochemical air pollution (Nov.29-Dec.2*1982«
in Tokyo and Tsukuba)
l.Intraduct i on
A number of different numerical models on the photochemical smog
have been developed during the last ten years in the U.S.A. and also
in Japan. Some models among them have fairly succeeded to
simulate ozone or other pollutants concentration.
fit the same time, however, many difficult problems to be
solved have been pointed out. One of the most important
common problem of these models is a lack of metearo103ica1
data such as ii?p = r uinds vertical diffusivityi mixing depth etc. It
is well known that the accuracy of these data has a very strong
effect on the results. This problem is more important in case of
*ir pollution in complex local wind system including land and sea
breeze and mountain and valley wind. However, it is very
high cost to make field observation of three dimensional wind sxstem.
In this study, numerical simulation on the photochemical air
Pollution under an idealized local wind system is carried out by a
Ivio step numerical model in order to make clear the
fundamental characteristics of the effect of the local wind upon the
Photochemical air pollution.
Fi-a.l is a flow-chart of the model. The first step is a three-
PROCEEDINGS—PAGE 259
-------
dimensional time-dependent local wind model, which calculates wind
velocity and vertical diffusion coefficient under the given conci-
sion such as large scale wind, topography and ground surface condition.
The second step is a photochemical air pollution model which is also
three dimensional model by using the result of the first step model as
input data. The second step model is further constructed with two
sub-models* a diffusion and advection model and a photochemical
react i on model.
INPUT synoptic wind
I topography etc.
LOCAL WIND
MODEL
OUT PUT
I
(u.v.vtt
V K2 \
u.v.w
Kz.6,q
INPUT--source data
p^
lOTGCHEMICAL SMOG MODEL
DIFFUSION
ADVECTION
MODEL
__J PHOTO C HEM JCAL1
[4 [REACTION
x IMODEL
OUT PUT
NO. NO2
03
concentration
Fig. I A Mow-chart of
s tep model
) fc1 > tjj n
We assumed an idealized photochemical reaction in place of the
actual reaction which is very complex. Some models of the photochemical
reaction have been developed well, but not compleatly established
yet. Moreover, they need detail infomation on the emission
data espetially on non-CHA hydrocarbons, which is difficult to
obtain. In this study, we do not intend to simulate
the reaction process in detail, but outline of the
reaction process. The idealized photochemical reaction model is simple'
PROCEEDINGS—PAGE 260
-------
but the feature of \i, espetialy ozone concentration behaviour) is
very close to the artual reaction. The idealized reaction model is
derived from more complex Photochemical reaction model by a kind
of mathematical apploximation.
2. Numerical model
1) Local wind model
The governing equations and numerical integration scheme of
the model are almost the same as the local wind model developed
by Kikuchi et.al(1981). The equations are Boussinesq hydrostatic
equations which are written in the terrain-following coordinate
system.
Equations of mortion:
Equation of -thermodynamics:
Continuity equation:
w
Equation of mixing ratio:
- -' -
where z* is the terrain-following vertical coordinate defined as
Z* = Zr^1 h =Zr -Z*
in which ZT and z^ are the level of the top and the around
surface of the model atmosphere respectively. w* is the vertical
zf-velocity, and other symbols are of usual.
PROCEEDINGS—PAGE 261
-------
The considered area is 450 km square and coverd by 30xo0
arid points at the interval of 15 km in the horizontal plane. The
vertical thichness of the model atmosphere is 6000m and divided into
15 layers with larger interval in upper layers.
The ground surface temperature 60 is predicted by the force
restore method(Bhumral ker, 1975) . which agrees very well to the
multiple soil layers model (Deardorf f , 1978) . 8e Is written as
foil ows:
Where. S is total solar radiation* L is net Ions wave radiations
H is sensible heat flux. IE is latent heat flux. 6, is the soil
temperature to be constant in the certain depth, and C, and C2 are
the soil characteristic parameters.
The vertical diffusion coefficient, which is very important
to vertical transport of heat and pollutants, is calculated
from the turbulent closure model (level 2} by He! lor and Yamada.
2.) Diffusion and advection model
The equation is written in the terrain fallcwins coordinate
system which is the same as the local wind model.
,
a* d? JZ" h
where, Cjf.Q^and R^ are concentration, source emission and formation
rate by the reaction of the i-th pollutant.- respectively. Boundary condi-
tion at the around surface is given by the following equation which
means surface dry deposition.
where v^Cz,). is the deposition velocity definded at the level of l
-------
calculated by the followind equation under the assumption of a constant
flux lexer of pollutans.
where u* is friction velocitx and k is Karman constant.
3) Idearized photochemical reaction model
It is well known that formation of ozone by photochemical
reaction depends on in i t i al . cancentra't i on of NOx and non-CH4
hydrocarbon! composition of the hydrocarbon and also light intensity.
In the case of hydrocarbon excess^ the maximum ozone concentration
does not stronaly depend on the initial hydrocarbon. Since observed
iata in a city or near~a city usual. 1 x. shows high hydrocarbon concen-~
lration> we assumed that hydrocarbon concentration is always
sufficient hish and it can be eliminate form independent variables
af the reaction model.
From this paint af view, we further simplify photochemical
reaction models and constract an idearized reaction model, fit the first.
W evaluate the budset of NO.N02.03 and 0 by the reactions. The
lost important reactions are the follows:
NO, t \\y -A^ Ntf -t 0 CO
tftO^tM -^^ 0? 1- M [«J
NO + 03 -J**- NO, t 0* (3)
The other reactions are classified followins 8 STOUPS:
Sum of the oxidation reactions of NO except reaction [3] are
represented by Rtl NO), where,
R, = ks CMO.; + h« CRO.j t h^ CRO,') + 2 k^CN03j + •- ^
Intal rate of decay reactons of N02J03,NO,0 are represented bx R,[NO}
l^03|,Ri[NO) , and Rf L 0) , respec t i ve 1 x and total formation rates of NO,N02
PROCEEDINGS—PAGE 263
-------
and 03 are represented bx F\ .Fj and Fj respect ivel x.
F2 =
F»H
Conservation equations of NOsNOZ^OS and 0 are written as fallows.
- CNOJ = -k3CNOJ C031 tk/ CNOJ -R/ CNOJ - R^CNOJ t F/_
7
/0
Al thosh the value of each term stronslx depends an the
initial concentration and the elapsed time, the order of them are
shown under the equations in a cas= of txpical initial concentra t i cr<
and reaction time.
Terms of reactionCl} - [3} which are Iar3er than others, can
be eliminated and ue obtain the follouina equations which include onlx
slow reaction.
To /0
= R/CNOJ -
/o
where.
PROCEEDINGS—PAGE 264
-------
Eiuati ons Cn) and (.23; express the most important process of the photo
chemical reaction. We neslect smaller terms of eqs. L20)«-(.il) and
get the followins equations after some substitution.
CNOO - ^ - /%-CPOKNOxJ
C N 0 J = (NOxJ - S/2 + /s_^ - CPO) (NOO
where
S= CPOJ t CNOx] -
This set of equations must be sood approximation of the reaction»
but it is not closed, because R, ,R^ and Rj are complex functions
of unknown concentrations. Then, we introduce new assumption:
RJ.RT. and R; are assumed to be parameters instead of functions
of unkown cancenira t i ons and to be approximated as follows:
These equations must be the simplest form
which do not spoil the fundamental charac terer i s t i cs of the
reactions. The order of the constant r( - r^ can be estimated
b> more deta i 1 react i on model , i.e., (r, , nj_ , r, ) =
(1^ 10 , 10 ), but optimum values of them must be decided by
comparison with experiments.
Fig. 2 shows the relationship of the maximum ozone concent-
ration and the intial NOx concentration under hydrocarbon excess
PROCEEDINGS—PAGE 265
-------
0.4
~ 0.2
c
a.
(A) C2H4
/
/
(B)C3H5
x
2 0.4
0.2
(C) 1-C4H8
^\
o'V*1
a/'''
•(D)1-CSH10
'°
/
O/
014 9162536 1 4 9162536
(NOx)g (10"2ppm)
J3.2 Mdximim ozone cancentration
and initial NCx in HC excess
case 3iven bx a chamber simulation
after Sakamaki, et al.(1981)
Black circles arc calculated value
bx the ida Iized reaction model
S. 0.4
z
ttl
u
z
O
u
0.2
0.0
k.» 0.175 min
E
E
c.
a
rt"
O
x
(3
2 3
TIME (HR)
0 0.1 0.2 0.3
[N0x)0 (ppm)
Fig.3 The same as Fi9.2j but
the maximum ozone fomation rater
i 3. u.One of the solution of the
idealized photochemical
rede t i o mode I
PROCEEDINGS—PAGE 266
-------
condition given bx a chamber experiment bx Sakamaki et al.(1981).
Black circles are calculated value bx the idealized reaction model.
when the pararameters r. *r-.and r- are assumed to be 3.6.0.1 and 0.08
Calculated ozone concentration is almost proportoinal to
square root of initial NOx concentration. This agrees with
the experimental result. The experiment
also shows week dependencx of the maximum ozone concentration
on a species of hxdrocarbon.
Fig.3 shows the maximum formation rate of ozone. Althogh
the experimental result depends on, a species of hxdrocarbon.
calculated values are seemed to agree well to the
typical value in the low NOx level.
The experimental maximum formation rate of ozone is almost
proportional to K| value, and the calculated one shows the same manner
is.the experimental result. A solution of the ideaized reaction mode]
is shown in Fig.4. Time variations of concentrations have
similar feature as those of the chamber experiments.
3L Photochemical air pollution in the two dimensional land and sed
freeze
ftt the first, we discuss the photochemical air pollution in a
fcro dimensional land and sea breeze;which must be the simplest
local wind. Simulated area is 225km in horisontal scale and divided
Jflto 30 grid points with interval of 7.5 km. We assume the source
irea which exists from the coast to 30 km inland. Emission rate is
assumed to be 200 rrf (7.Łbn)~Z for each of NO and MR ( non-reactive
roHutant) from 0700 until 2000 and zero in the other period.
fe the ground surface, the deposition
felocitx of ozone is assumed to be 0.3 cm/sec. And kj value
PROCEEDINGS—PAGE 267
-------
is assumed as follows:
k| /k3 = 0.013 S (ppm)
where S (cal/cm min) is total solar radiation.
At the first, zero synoptic wind is assumed over all area.
Fig.5 shows distributions of wind velocity* ozone, NOx and
NR at 0600 on the second day, i.e., 30 hours after numerical
integration started. Week land breeze exists broadly in the
lowest 500m layer and weeker counter flow appears above
the land breeze layer.
Ozone concentration is low mear the around? but high
in the layer of 500m to 1500m above the source
area. This ozone is due to the primary pollutant emitted on the
day before. NOx and NR concentrations in the figure are also due
to the day before, and difference of them means removal by the
reaction, because the emission rate of them are same each other.
fit 1200 (Fig.6), sea breese appeares near the coast and mixing layer
develops on the land, which enhances vertical diffusion. 03
is formed by the reactions and mixing with ozone in the upper layer?
so that very hish concentration appeares en the source area, but
NOx concentration is not high because of vertical diffusion and
decay.
At 1500 (Fig.7) sea breeze penetrates over 50 km inland with
strong horizontal velocity (over 5 m/sec) and a sea breeze front
is formed with notable upward motion. As a result? concentrations
of ozone and non-reactive pollutant decrease in the low level
on the source area? bu't inclease on the downwind. Concentration
in the upper layer on the source area is still high and it is slowly
moving toward the sea.
At 1800 (Fig,8). sea breeze penetrates further inland, but
PROCEEDINGS—PAGE 268
-------
NR (ppb)
NOx(PPb)
oU
2
1
UJ
O3(PPb]
0.20 M
Q-_-_
~75SOURCE 15° 225
HORIZONTAL DISTANCE(Km)
1 ' 3- ':> Ui I': I r i l)u I i on of w i nd vn I or: I I y.
o/ono- l\l()x rind NR (tmn-rpdC L i Ve'
I'ci I I 1.1 I ,ui I ) -i i ven by I.In;, mode I
;"•
NOx(ppb)
I-
X
a
LU
O3(ppb)
0.2 0 ui,-,
J_LL ^ii ^
7G
souncE
1 2
225
l-IOniZONTAL DISTANCE(Km)
-------
-s
S3
O
o
~
IS
C:
"0
o
X
O
UJ
0 *
d=---—--------75-gouncE 15° 22!
HORIZONTAL D I S TANCE(Km)
SOUPCE ISU 225
HORIZONTAL DISTANCE(Km)
I i •:. f! I hr 'i.imr ,r; I i 'i. l:>5 l)ii I a I IHWO.
-------
100-
150
(J
o
UJ
a
2
Fig. 9 Diurnal variation ti f wind velocity
& n d c o 111:e 111. r & t i o n B A 1 111 e 1 e v c 1 o f
'25in> X Is the distance fron tlie coast
a
a.
a.
•z.
O
ac
H
Z
Ul
o
o
o
100
100
100
24
Flti.lW 'I he siime as f:lg.9p but v«r-lur> distance
from I.he const
-------
the mixing laxer changes graduallx to stable laxer. It makes hish
NOx concentration and further low 03 concentration, or sometime zsro
ozone, in the low level on the source area. In the upper layer-
however, hish 03 concentration is kept until the next morning.
Simiar phenomenon is often observed actualx.
Fig.9 shows diurnal variation of wind velocitx and concentration
at the lowest level (25m) in the source area. Ozone concentration
has a peak at about 1100> and NOx and NR have bi-modal variation
with two peaks in the morning and eveing. These diurnal variation
are commonlx observed in urban area. At about 0800,land breeze changes
to sea breeze, which becomes stronger at about noon when ozone
concentration begins to decrease. The maximum velocitx of sea
breeze is about 5 m/s, and sea breeze continue until midnight.
Fig.10 shows the same as Fig.9» but at the other points. X in
the fisure is the distance from the coast. Peak time of ozone
concentration shifts afterward with increasing distance from
the source area* and at far points, we can fined another small
peaks at about noon. The small peak at noon is due to the
effect of the previous dax.
The figure shows also that the diunal variation of wind velocitx
also shifts afterward with incleasing X. This is one of the
common characteristics of land and sea breezes.
The effect of week large scale wind is investigated. Large
scale wind modifies both of 1 and and sea breeze and distribution
of concentrations. Fig.ll-a shows the lowest level wind
velocitx at point x=ll km in three cases: large scale wind Up =
0J-0.3 and -1.0 m/sec. Minus of U$ means direction of the
larse scale wind is the same as land breeze. Modification
of the lowest level wind is small, but 03 concentration is
PROCEEDINGS—PAGE 272
-------
stronalx modified (Fig.lib). Week sea-ward larse scale wind
enhances ozone concentration and shifts the peak time afterward in
the source area.
10
n
•v
E
o
o
_I
111
a
z
-5
— 0 m/s
0.3 m/s
1 m/s
12
18
24
TIME
Fis.lla The effect of week I arse scale
wind on the wind speed at X=ll km.
150
24
Tl M E
Fig lib The effect of *eek lar-ae scale
wind on ozone conentrst.on.
PROCEEDINGS—PAGE 273
-------
4. Application to Tokyo metropolitan area
Manx researchers have been pointed out that the photochemical
air pollution of Tokyo metropolitan area is strongly affected by
the local winds. It is also suggested that horizontal distribution:
of the maximum ozone concentration in a polluted day can be
classified into the following three types.
Type Is High ozone concentration appeares broadly in Kanto
pi a i ns.
Type II:High ozone concentration appeares only in the southan
part of the plains.
Type IIIiHigh ozone concentration appeares onlx in the northan
part of the plains.
For examle, Figs.12 and 13 show the maximum ozone concentrations in
a typical day of type I (17,Jun,1979) and Type II (5,Jul,1978),
respect ively.
All of three types appear under such the almost same meteoro-
logical condision as mild synoptic wind on a clear day. But the
local wind systems are appricably different each other. Fig.1/1
and 15 show observed surface wind on the same days as Fiss. 12 and
13, respectively. On the days of type I> wind is usually calm
all over Kanto plains in the early morning, and southerly sea
breezes begin to blow from Sagami bay and Tokyo bay in the
before noon. The maximum velocity of the sea breezes reaches
about 5 m/sec. On the northan part of the plains:, wind direction
shifts to south-easterly in the before noon. This pattern often
continues until night. On the days of type II, northerly
wind prevails over the plains in the early morning* but it is nat
strong. After that? easterly sea breeze from the eastan ocean
PROCEEDINGS—PAGE 274
-------
prevails over the plains. The southerly sea breeze from Saaami bax
begins to blow later.than that of type I. It often begin in the
afternoon, and usualy covers only the south-west part of the Plains.
On the days of type III, wind system is quite similar to the case
of type I, unless the southerly component is stronger.
Numerical simulation is carried out in the area shown bx
Fi3. 16, which is much larger than the Kanto plains. Because
it has been suggested that the local wind system in the plains
is affected by the mountain area in central Japan. The
figure also shows an area source (shadowded area) and elevated
point sources (black circles). The emission rate is given
by roush estimation with.diurnal variation. Figs. 17 and 18 show
calculated wind velocities in the height of 25 m for the cases
that the large scale wind is zero (case A) and NNUI 3 m/sec (
case B), rewpectively. Cases A and B are corresponding to the
Types I and II respectively. Characteristics of the wind systems
in the cases A and 8 agree fairly with the those of the type I
and II, respectively, which are mentioned above.
Figs 19 (a),(b),(c) and (d) are calculated ozone concentration
in the case A for 0900, 1200, 1500 and 2130, respectively. Ozone
concentration begins to inclease at about 0900 on the source area,
and it reaches the maximum value at about 1200. The polluted air
mass is transported toward north by the southerly sea breeze,
but the ozone concentration slowly decreases in the afternoon.
Figs 20 (a)-(d) are the same as Figs 19 (a)-(d), but in the
case B. On the Sagami bay, ozone concentration begins to inclease
at about 0900, because NOx emitted from the source has been trans-
Ported to south by the northerly wind in the early morning.
PROCEEDINGS—PAGE 275
-------
Although hish ozone concentration covers the southern part of Kanto
plainsi it does not so to the northan part until nisht. This agrees
with the fundamental characteristics of the txpe II.
PROCEEDINGS—PAGE 276
-------
n
••
•v
C
...
- :
Fig. 12 I lit? max I mum ozone t:nnt:eii Ira II mi
in a typical day of type I ( 17, Jun, IV
I lie samp as Fig. 12. but
C:i.Jul . 1970)
J]
-------
UK91S
,
< r, ^ 'r-_tx .v
,
/" : ^'"Acc
ig.U Observed uind velocitx on 17 Jun, 1979 CtxPe I)
PROCEEDINGS—PAGE 278
-------
v> i <
i\l x--x Xr5\1'
U'.------- '-7 V
'~'"r' t '. ',„ o x' ^_ -^^
' 2
Fig. 15 Observed uind velocitx on 5 Jul.1978 (txpe II).
PROCEEDINGS—PAGE 279
-------
iii i
•; i O
140
V\KANTO PLAIN
VooOOuncE
OOO AIIU A
35'
THE PAC I P 1C OCEAN
Fls.16 Tlio numerical Simula t lun tii>i>llp(l nri-
Is an urea source and black circles nrp
sciurce (h =
Sli.iclnuided ti
v.ilfd niiinl
PROCEEDINGS--PAGE 280
-------
=2SM
WIND VEL3C1TT, LTal2.,.r.-?5H Us CALM HINDVE,OCiT
J.T=15~.r =2SM U= CALM WIND VELOC1TT. LT = 2l~.r=25M U= CALM WIND VELOCITY
Fis.17 Calculated wind velocity In the case fi. which is f • 2~*/
-------
= 25M
r r
MNE3H HIND VELOCITT L7=lL.r=2SH U= NNE3M WIND VELOCITY
\
LT:15».r=25M U= NNE3M HIND VELOC 1 T T . L T = 2 [ „,?• :
-_ x
U = NNE3M HIND VELOCITT
Fig.18 Calculated wind velocitx in the case 8. which is
cor respondins to txpe II. (see Fis.15)
PROCEEDINGS--PAGE 282
-------
-
•
.
-:
--
D
•v
n
....,• - .. !M . >
ft I • , ' ,' -'" -.>.'• V / (l
-; i »u >> / .'
-------
V
LT=15.,,,Z- =25M U= CALM
03
PPfl
Fig.19 (c) Calculated ozone concentration
at 1500 in the case 0
LTr2l ...,Z- =25M U= CALM 03
Fig.19 (d) Calculated ozone concentration
at 2100 in the case f\
-------
•--
...
n
;v
c
M
•
•
\/
LT=9.,.. Z" =2511 U= HNC3M 03
prn
Fit).20 ( a ) Cii I eul * ted ozone cmicen tra U un
,1 I W9H0 in the case H
"WW
N
I. 1 = |,?....r =?5H Ur HHE3H 03
1*1.20 (I)) (xii I r:u I o (r?(J o^nnp r one c-n t r a t i on
.1 I \'SVM In I lie c-ir.c I)
-------
H
n
a
• •
:
...
a
...
'
-
\/
MV ^\-> ,.] i n .u
,;' ^-VvO 'A\
•
LT=I5.,..Z- s25H U= NNE3H 133
TIMl.
frlg.20 (c) Calctilntcd ozone cmicen tra 11 un
.1 t l'.)HW in tlie c«isi.' [j
>
l.T=2U..Z* =
U- MHE3H 03
(d) Calculated tiztine concentration
in
C,IH«» (J
-------
TRANSPORT AND TENS FORMATION OF AIR POLLUTANTS
BY LAND AND SEA BREEZES
Presented by H- Tsuruta
Yokohama Research Institute for Enviromental Protection
Japan
PROCEEDINGS—PAGE 287
-------
TRANSPORT AND TRNSFORMATION OF AIR POLLUTANTS
BY LAND AND SEA BREEZES
1. INTRODUCTION
Most of urban or industrial areas in Japan are located along
the coastal region, and mountain areas are located close to the rural
or urban area. Air pollutants, emitted from those source areas, are
expected to be transported by meso-scale wind circulation, such as
land and sea breeze, mountain and valley wind, and larger-scale one
coupled between the two systems,
In the Kanto region, it has been reported that the polluted air
mass with with high oxidant concentration above lOOppb formed in the
Tokyo Metropolitan area has been transported as far as Tochigi 01
Nagano area, 100 km north from Tokyo, by southerly winds ( Fig. 1 ).
In the coastal region in Japan, high oxidant concentration has
been also measured during the sea breeze especially in the summer season,
as shown in Fig. 2 and Fig. 3. Lyons and Olsson have studied the
transport mechanism of air pollutants during the lake breeze along the
2)
Lake Michigan. Lyons and Cole have recently reported that the same
phenomena as mentioned above in Japan have been observed in the coastal
area of Lake Michigan. The transport mechanism of air pollutants by
meso-scale circulation, however, was hardly known due to the lack of
continuous data over the sea. Photochemical reaction must be considered
in the course of transport of the polluted air mass.
We have challenged to reveal its mechanism using the research vessel.
Results of preliminary surveys conducted in 1976 and 1977 were reported
PROCEEDINGS—PAGE 289
-------
at this fifth Conference. In those experiments,it has been reyealed
that the primary pollutants emitted mainly from Keihin industrial area
are transported to Sagami Bay by offshore flow from midnight to early
morning, and high oxidant concentrations are observed over Sagami Bay
in the daytime when it becomes calm after offshore wind. In this reportc
a brief outline of the results obtained from two experiments conducted
in 1980 and 1981.
2. EXPERIMENTAL
A two-day field program around Sagami Bay area was planned in the
summer of 1980 and 1981, by the observational group in the special research
project, " Study of meso-scale atmospheric pollution in Kan to District"..
Ocean Research Institute of Tokyo University observed vertical wind
structure over Sagami Bay at the fixed station, the research vessel
" Tansei-Maru " , belonging to Ocean Research Institute, as shown in Fig. 4.
National Institute for Environmental Studies launched pilot balloons at
six locations, Yokohama National University also observed vertical wind
profile at another six locations. Yokohama Research Institute for Environ-
mental Protection measured concentrations of gases and aerosols in
ambient air at the fixed station in Sagami Bay, using the research vessel.
National Institute for Environmental Studies and Yokohama Research Institute
for Environmental Protection measured with light airplane, three dimensional
distribution for meteorological parameters ( including real wind field ),
air pollutants, respectively. Ozone sonde was used to measure the diurnal
variation on vertical distribution of ozone concentration at one station
only in the summer of 1981-
PROCEEDINGS—PAGE 290
-------
These field studies were conducted, 12-13 August 198Q, 9-11 August
1981, respectively. The results on the overall field and -theoretical
studies, and on the three-dimensional wind structure and its diurnal
variation on meso-scale circulation will be reported elsewhere, and the
results on measurements of air pollutants will be shown hereafter.
3. RESULT AND DISCUSSION
3-1 Diurnal Variation of Gaseous Pollutants over Sagami Bay
Measurements for gaseous pollutants were made in ambient air over
Sagami Bay, which were N0x> CH^, NMHC, C2~C5 light hydrocarbons, CO, HCHO,
03, HN03 and #2°2 ^ H2°2 and HN°3 were measured in cooperation with
Dr. Yoshizumi, Tokyo Metropolitan Research Institute for Environmental
Protection ). The data obtained in the summer of 1980 and 1981 are shown
in Fig. 5 and Fig. 6, respectively.
Meso-scale wind circulation was developed during both period for
observation, as shown Fig. 7. It was clearly found that the offshore
wind continued to blow during the morning, the onshore wind during the
afternoon, as shown in Fig. 5 and Fig. 6. In the coastal area of Sagami
Bay, warnings against photochemical air pollution were issued as a result
of high oxidant concentrations ( more than 0.12 ppm hourly average value)
only on two days of 9 Aug. and 10 Aug., in 1981. The following discussion
will be made mainly on the data obtained in the summer of 1981.
The concentrations of primary pollutants, such as NO , CH,, NMHC and
X *T
CO, were higher during the land breeze than during the sea breeze. On the
other hand, levels of the secondary pollutants such as 0^, were higher
during the sea breeze than during the land breeze.
PROCEEDINGS—PAGE 291
-------
The 0, concentrations, after reaching to the maximum in the early
afternoon, were gradually decreased, but were higher than the background
level of 0_, about 20-30 ppb observed in the marine atmosphere. These
phenomena mean that, as the air with the primary pollutants advects
further offshore, 0^ is formed in the air with photochemical reaction,
and that in. the afternoon, the air with its burden of 0, advects landward
because there is almost no scavenging process over the sea. Just after
the offshore wind begins to blow, the 0_ values are drastically decreased
below the background level due to the reaction with NO emitted at the
urban area. During a few hours after sunrise, the photochemical
equilibrium is established between NO, NO- and 0_ as shown in Fig* 5,
if the 0, concentration is not relatively high levels. In the late.
morning when the solar radiation energy is increasing rapidly, this
equiliblium can be no longer maintained as the reaction of hydrocarbons
with NO becomes more active, and the 0,, formation rate increases rapidly
as shown in Fig. 5.
The concentration of HNO~ was higher during the onshore flow than
during the offshore flow, as presented in Fig. 6, shows that UNO. is
formed with the reaction ( 1 ) ,
N0 + OH
which is one of the net loss reaction for NO .
x
HO was measured with the equipment on board, which was made by
4) 5)
Yoshizumi after improving the Kok's measurement system. The H_0
concentrations were presented in Fig. 6. The ambient H_0. values were
very higher during the sea breeze than during the land breeze, as well as
the HNO, values. As the diurnal variation of the H-0_ values were similar
PROCEEDINGS—PAGE 292
-------
to that of 03 and the ratio of 03 to H^ is about ten, H202 is produced
from the following photochemical reactions:
HCHO + hv > H + CHO ( 2 )
CO + OH > H + CO ( 3 )
2 + H02 * H2°2 * °? C 5 3
Kok has already measured the ambient HŁ02 concentrations in the California
South Coast Air Basin in the daytime during the summer seasont and has
discussed the correlation of the H^ values with the 07 values with
great ambiguirv =
The HCHO concentration, which is not only a primary jpollutant but also
a secondary produced pollutant, was slightly higher during the offshore flow
than during the onshore flow-
C2~C5 hydrocarbons were analyzed in real time with FID-GC on board
during the 2nd survey in 1981. As the ratio of the concentration of
C3H6 and ^2H4 t0 that °f C2H2 were Ver7 lower in the daytime than in the
nighttime, it is assumed that most of C-H, and C.H, were consumed with
36 24
the reaction of OH or 0,, near the source area or in the course of transport.
3-2 DIURNAL VARIATION OF AEROSOLS OVER SAGAMI BAY
Aerosols in ambient air over Sagami Bay were collected and analyzed
for S04 ~, N0~, Cl~, NH* and metals ( Na+, Mg2+, K+, Ca2+, Fe, Mn, Zn,
Pb, Ni, V ) and PAH ( BaP, BghiP ). Fig. 8 and Fig. 9 show the diurnal
variation of the concentrations of these species in aerosols. Diurnal
pattern of total mass concentrations ( TSP ) were so similar to that of
such a primary pollutants in gases as CH, or CO.
2-
But the SO, concentrations in aerosols were much higher during the
PROCEEDINGS—PAGE 293
-------
sea breeze than during the land breeze as well as that of HNO,. It is
2-
re as on able that SO, in aerosols was formed mainly with photochemical
reaction in the day of high oxidant levels. N0~ and Cl , on the other
hand, showed the pattern of primary pollutant, but the ratio of these
concentrations during sea breeze to those during land breeze was very
lower than in the case of trace gases such as CO or CH,. From these
phenomena, loss mechanism for N0_ and Cl must be considered:
2NO~ + H2S04 > 2HN03(g) + S0^~ ( 6 )
2C1~ + H2SOA > 2HC1 (g) + S0^~ ( 7 )
Sea salt particles are expected to be found in the samples during
2+ +
the onshore flow. As shown in Fig. 9, that the ratio of Mg to Na
was almost nearly the value of 0.13 in the sea water during the onshore
flow, is the demonstration of generation of sea salt particles. Therefore,
chlorine loss also occurs with the reaction between NaCl and H.SO, ( 7 )
or HNO.(g) during the onshore flow.
Fe was used to be the index for soil particles in the first
approximation. Enrichment factors for Mn and Zn were calculated to be
about 2 and 100, respectively, over Sagami Bay. It is clear that the
origin of Mn and Zn are soil and anthropogenic sources respectively.
4. DIURNAL VARIATION OF SURFACE OXIDANT CONCENTRATIONS IN MONITORING
STATIONS AND TRANSPORT OF POLLUTED AIR MASS
Fig. 10 shows the surface oxidant concentrations near the shore line
and the inland areas, on 9-10 August, 1981. After sunrise, oxidant
concentration increased gradually. Just after the front of sea breeze
passed by, the levels of primary pollutants such as SO , CH, and CO
increased stepwise, and oxidant value reached to the maximum 1-3 hours
PROCEEDINGS—PAGE 294
-------
later. With trajectory analysis as shown in Fig. 11, the polluted air
mass with its burden of primary pollutants, starting from Keihin industrial
area in the late morning to the south by northerly wind, were transported
over Sagami Bay at about noontime, and was advected to the shoreline
with its burden of photochemical oxidants such as 0_, HNO , H 0 and
2-
SO^ aerosols obsserved on the research vessel, by onshore flow.
In the evening, the oxidant concentration near the surface rapidly
decreased with the reaction of NO, and the cooled air sloping down the
mountain areas moved slowly from the northwest, reached to the shoreline
at midnight. The oxidant concentration resulted in the drastic decrease
over Sagami Bay just after the offshore flow with its burden of fresh NO^,
Fig. 12 shows the schematic pattern of the diurnal variation of NO and 0-
near the surface of the inland and the sea, on the day when the land and
sea breeze is developed and on the day when the southerly wind is prevailing,
respectively.
5. DIURNAL VARIATION OF VERTICAL DISTRIBUTION OF OXIDANT CONCENTRATIONS
Measurements on vertical distribution of oxidant concentrations
were made at the station Y in Fig. 11, from noon on 9 August to noon on
11 August, in 1981, using the ozone zonde. Fig. 13 shows the vertical
profile on wind, temperature and oxidant concentrations at station Y.
During the nighttime, oxidant concentrations at the height of
500-800 m remained to be the value above 100 ppv observed after the front
of sea breeze passed through, while the oxidant value was recorded in
minimum below the 300m height.
PROCEEDINGS—PAGE 295
-------
Vertical distribution of oxidant was uniform in the mixing layer,
in the morning after breaking up of temperature inversion and when the
offshore wind was blowing. This phenomena is due to the vertical mixing
of aged pollutants above the temperature inversion layer with the fresh
air pollutants near the surface, with the thermal convection.
6. THREE DIMENSIONAL PROFILE OF THE 0., AND N0,r CONCENTRATIONS IN THE
J itJ~""~l~" t-j-.
FOUR STAGE OF MESO-SCALE CIRCULATION
During two experiments,measurements for 0, and N0_ were made with
airplane to obtain the horizontal distribution at the constant height of
300m and 1,000m, and the vertical distribution over the three points
of bay,shoreline and inland. The flight was done in the four stage of
the day, which means the offshore flow stage, alternative stage from
offshore flow to onshore flow, the onshore flow stage and alternative
stage from the onshore flow to the offshore flow, respectively.
Fig. 14 and Fig. 15 show examples of vertical cross section of
8)
0_ and NO concentrations and the wind structure analyzed by Fujibe ,
J X
in the onshore flow stage and in the alternative stage from the onshore
flow to the offshore flow.
Fig. 16 shows the schematic pattern on the vertical cross section
for the 0_ and NO concentrations in the four stage of the raeso-scale
circulation, perpendicular to the shoreline.
In the first stage, the primary pollutants are advected from inland
to the sea,and the level of NO are higher in the land than over the sea,
and than above the mixing layer. On the other hand, the levels of 0~
are lower in the land than over the sea and than that above the mixing
PROCEEDINGS—PAGE 296
-------
layer which is 20-30 ppb of 0_. The mixing layer is of a few hundred meters
height.
In the 2nd stage, the photochemical process is predominant. The 0_
concentration is increased rapidly both over the land and the sea, and
.mixing layer is developed to the 1,000m height or so over the land by
strong thermal convection
In the 3rd stage, as the onshore flow advects the secondary pollutants
to the inland horizontary and aloft especially near the front of sea
breeze, the concentration of 0., in the lower layer over the sea is
decreased. If the return flow is present, 0_ alofted is advected to
the sea gradually. The NO levels are lower than in the first stage
,0*
as NO- is consumed with photochemical reaction.
J&
In the 4th stage, as the large scale offshore wind begins to blow
the higher concentration of 0» alofted is advected to the upper part of
the sea, and the layer of the higher 0. level is formed both over the
land and the sea. But the 0 levels near the surface in the land is
decreased by the reaction with NO.
Then, in the first stage on next day, it is characteristic that
the aged pollutant layer with the high 03 level is present above the
mixing layer. After breaking up of the inversion, these aged pollutants
influence the photochemistry and the 03 level near the surface with
vertical mixing.
PROCEEDINGS—PAGE 297
-------
7. SUMMARY
In the summer of 1980 and 1981, a field study for diurnal variation
of air pollutants over Sagami Bay and three dimensional stucture of air
pollutants on meso-scale circulation was conducted using the research
vessel and the aircraft.
The high concentrations of 0_, H 0 ,HNO_ and SOf were observed
during the sea breeze in the afternoon. As the diurnal variation of
their concentrations are very similar, all of them are mainly produced
with photochemical reactions in the daytime of summer,
0,, which is formed in the urban and industrial areas, are transported
three-dimensionally with meso-scale circuration induced by land and
sea breezes, and the aged pollutants with its burden of 0, were present
above the mixing layer in the next morning after the day when land and
sea breezes were generated.
PROCEEDINGS—PAGE 298
-------
REFERENCES
1) Lyons, W.A. and L.E. Olsson; Mesoscale air pollution transport
in the Chicago Lake breeze, J. Air Poll. Cotrol Assoc«s 22, 876-881
(1972),
2) Lyons, W.A. and H.S. Cole: Photochemical oxidant transport: Mesoscale.
lake breeze and synoptic-scale aspects, J.. Appl. Meteor., 15, 733-743
C1976),
3} Tsuruta, H., H. Maeda and M. Ohta: Observation on transport anc*
background concentrations of atmospheric pollutants over Sagami Bay
and the Izu Islands Sea, Papers presented at the fifth US-Japau
Conference on Photochemical Air Pollution. 1980,
4) Yoshizumi, K.: in preparation.
5) Kok, G.L., T.P. Holler, M.B. Lopez, H.A. Nachtrieb and M. Yuan:
Chemiluminescent method for determination of hydrogen peroxide
in the ambient atmosphere, Environ. Sci. Technol., 12, 1072-1076(1978)
6) Kok, G.L., K.R. Darnall, A.M. Winer, J.N. Pitts.Jr. and B.W. Gay;
Ambient air measurements of hydrogen peroxide in the California
South Coast Air Basin, Environ. Sci. Technol., 12, 1077-1080 (1978).
7) Kadowaki, S.: Behavior of sea salt particles in urban air, Chemical
Society of Japan, 141-146 (1980).
8) Fujibe, F.: in preparation.
PROCEEDINGS—PAGE 299
-------
H(NMVH)
^1000
1000-500
500-200
200- 50
50- 10
.; i
1 Map of NO^ emission rate in the Southern part of Kanto areo
100
—
.
C:
••
O
50
HIRATSUKA
AUG 1973
\
00 06 12 18 24
Time of Hour
Fig. 2 Diurnal variation of oxidant concentration
when the land and sea breeze are generated C
when northerly wind is prevailing ( ) and
when southerly wind is prevailing ( )_ .
PROCEEDINGS—PAGE 300
-------
Fig. 3 Areas where high oxidant concentrations have been
measured during sea breeze, in Japan.
A — Sendai Bay,
C — Sagami Bay,
E — Ohsaka Bay,
G — Seto Inland Sea,
I — Suruga Bay,
K — Fukui coastal area.
B — Tokyo Bay,
D — Ise Bay,
F — Kii Channel,
H — Lake Biwa,
J — Toyama Bay,
PROCEEDINGS—PAGE 301
-------
Fig. 4 Map of experimental area.
g) Fixed station, " Tansei-Maru "
• station for pilot balloon
-
z
v
Monitoring station for air pollutants
Point for 0 measurement
Station for 0- sonde
Point for vertical flight
Course for horizontal flight
Area abive 600m height.
PROCKEDINGS—PAGE 302
-------
1.4
1.0
0.6
f i
140
rT 100
60
20
_ 80
a
S 40
80
3
5 40
80
J3
I *°
2.8
2.4
2.C
"i 1.6
CL
-5 1.2
0.8
0.4
10
S
5 90
i 60
> 30
NO.
NO
CH4
NMHC
HCHO
BghiP
TSR
BaP
00 06 12 18 00 06 12 18 00
12 Aug. 1980 13 Aug. 14 Aug.
Fig. 5 Diurnal variation of air pollutants,
(. 12 - 13 August, 1980 )
PROCEEDINGS—PAGE 303
-------
12
00 12 00 12 00
09 Aug. 10 Aug.
12
Fig. 6 Diurnal variation of air pollutants
( 09 - 11 August, 1981 )
PROCEEDINGS—PAGE 304
-------
- 4 J
1 * d
I
V .
\ ] i
'"» /
(
(
*— 4 "%
ST
06 h
V J 12 Aug.
X C
7
/^T~'-X *x_
C ,y (
(
<
— 4 r,
j
*V 06 h
13 Aug.
. (
'
1
, ,^L -
(, '
/
(
(
s
/ < — 4 ">,\
/
1 5 h
'
^ /'
/
r, /
./
(
\
^
< — *%
/ r*v
V\ 15 h
O u •
' :• : ,^
i ' ' ;,/
1 •-.- 4
V7
/ /'
/ J^ ^
' <
i
S
. — 4 "X.
06 h
09 Aug
C
,'• ^
-I >
x
^ •"-^" \" |~ i S^
> 7
' (
^
; : J
*-*7
V^^^LA ! <
r \^
t ^ • (
/
'
C
V,
//' v.
'
(
H
^— .
06 h
10 /
,
(1) 12 - 13 August 1980 _ 1() August ;
"0
>
« Fig. 7 Surface wind vectors
-------
12 00 12 00 12 00 12 00
9 AUG.1981. 10 AUG. 11 AUG.
Fig. 8 Diurnal variation of aerosols over Sagami Bay.
(9-11 August 1981 )
PROCEEDINGS—PAGE 306
-------
O)
12 00 12 00
9 Aug. 10 Aug.
12 00
11 Aug.
Fig. 9 Diurnal variation of aerosols over Sagami Bay.
(9-11 August 1981 )
PROCEEDINGS PAGE 307
-------
00 06 12 18 00 06 12 18 00 00 06 12 18 00 06 12 18 00
09 Aug. 10 Aug. 09 Aug. 10 Aug.
(l) so.
(2) 0.
Fig. 10 Diurnal variation of SO. and 0 concentrations around the coastal area.
(9-10 August 1981 )
PROCEEDINGS—PAGE 308
-------
(1) The hour where the onshore flow passed by
and the trajectory of the air reached to shoreline at 14 h.
Z) 124 / 142 ,110?
(2) Maximum 0 concentration
and the hour where maximum concentration was recorded.
Fig. 11 Analysis of monitoring data on 9 August 1981.
PROCEEDINGS- PAGE 309
-------
0
Inland
200
100
2
a.
a
0
200
NO:
- 100
NO;
Inland
Sea
r
00 06 12 18 00
Time
Fig. 12 Schematic diurnal variation of 0, and NO concentrations
J Ji
over the land and the sea.
Solid line ( ) corresponds with the day when the land and
sea breeze occurs and dot line ( ) with the day when the
southerly wind blows all day.
PROCEEDINGS—PAGE 310
-------
1 500 -
1,000 -
(1)5
500 -
12 14 16 18 20 22 00 02 04 06 08 10 12 U 16
TIME I hour. JSTI
)• 9 AUG 1981 )• 10 AUG 1981 H
10 AUG 1981
10.
no
TIME {hour JST I
( 3 )
9 AUG 1981
Fig. 13 Diurnal variation of vertical profile
of (1) wind, (2) temperature and (3) Oxidant
at station Y on 9-10 August , 1981.
PROCEEDINGS--PAGE 311
-------
1 5h
1630
1730
2. 0
*
>
x
X
\
t
f 5 *
til
1 f #
> X <•
.-]
.--I
« 1
2. 0 - >
1.5
1.0
0.5
NOv
Sea <
_
5
^> 10
• — ^_ ^K>
> I n nH
•*-- * x x x *•
1 . 5 • - Z "T * 1 ! t C
i.O — *****C1
H * " V V * "
•V 4 4 f \-
o- s - - T y : ; ; ; :
D.0Jtiilli5;
34°40' 35°00' f 3^
C .3 -, ^ .. ...
Oca
^_ ^^
' ' ^
' -r '
H
M
*20t
x- ^ ^
^* .*- -*-
k
» i }
J t J--j
> ^ y
||
/" X S
anIlM^liPi!l!iI&
35°40'
!6h
Fig. 14 Vertival cross section of 0_ and NO concentrations and
3 x
the wind structure in the 3rd stage.
( 9 August 1981 )
-------
10cm s
0026— 0113
Sea
Land
OOh
5ms
2.
1 .
1.
0.
0.
0
s
0
5
0
c^^n^!^
- — +-+—+—
* — f « 1
~*~-^mtmz
~~ * T V V
*• N »• v
—
*-
^
~"
X
N
V
•*
•^
—-
~*
*
"*^
<
-r
A
•*
^
*-
~*
4
>
>
r
*"
-~
.
<
•/
y
fi
-r- •*"
_,. X
•*—
< •* —
«___"*~~-
«^_"*~~~»
•»-~r*^--
^ X
r 4
4 *
^ ^
T
X ^ *
Jt Jf "
•*• T ^
X v1 x
V^ v
X^^-
t K V
j : :
< ^ ^
4 k "^
34°40'
35°00'
Sea
35'
35°401
Land
Fig. 15 Vertical cross section of 0_ and NO concentrations and
J X
the wind struture in the 4th stage.
( 10 August 1981 )
PROCEEDINGS--PAGE 313
-------
stage
NOx
Sea
Sea < I > Land
Fig. 16 Schematic pattern of the vertical cross section for
the 0. and NO concentrations in the four stage of
J X
the meso-scale circulation.
PROCEEDINGS—PAGE 314
-------
EVALUATION OF EIGHT LINEAR REGIONAL-SCALE SULFUR
MODELS BY THE REGIONAL MODELING SUBGROUP OF THE
UNITED STATES/CANADIAN WORK GROUP 2
presented by K.L. Demerjian
Environmental Sciences Research Laboratory
USEPA
PROCEEDINGS—PAGE 315
-------
EVALUATION OF EIGHT LINEAR REGIONAL-SCALE SULFUR MODELS BY THE
REGIONAL MODELING SUBGROUP OF THE UNITED STATES/CANADIAN WORK GROUP 2
Francis A. Schiermeier Prasanta K. Misra
United States Subgroup Co-Chairman Canadian Subgroup Co-Chairman
U.S. Environmental Protection Agency Ontario Ministry of the Environment
Research Triangle Park, NC 27711 Toronto, Ontario M5S 1Z8
The Atmospheric Sciences and Analysis Work Gi~oup 2, formed under the
United States/Canadian Memorandum of Intent on Transboundary Air Pollution,
was charged with describing the transport of air pollutants from their
sources to final deposition, especially in ecologically sensitive areas.
Eight linear regional-scale models developed by Canadian and United States
scientists were applied by the Regional Modeling Subgroup of Work Group 2
using standardized 1978 emissions and precipitation input data sets.
Model results were evaluated with currently-available January, July, and
annual 1978 observational data sets.
Concentrations and depositions of sulfur compounds as well as source-
receptor relationships (transfer matrices) were calculated by the eight
long-range transport models using simplified formulations. These were
state-of-the-art linear models in which scavenging and chemical transfor-
mation processes were treated linearly as a first approximation.
For the 1978 data set, most of the models appeared to perform rela-
tively better in predicting the deposition of sulfur in precipitation than
in predicting sulfate concentrations in ambient air. Based on available
1978 wet deposition measurements, the models were able to reproduce the
correct order of magnitude of the large time and space-scale features of
measured wet sulfur deposition patterns. In the construction of unit
transfer matrices, the models examined by the Regional Modeling Subgroup
predicted generally similar relative impacts on receptor regions in terms
of ranked order of importance, although variations existed among models in
the absolute magnitudes of the transfer matrix elements.
PROCEEDINGS—PAGE 317
-------
Introduction
The Atmospheric Sciences and Analysis Work Group 2 was one of five work
groups established under the Memorandum of Intent (MOI) on Transboundary Air
Pollution, signed by the governments of Canada and the United States on
August 5, 1980. The objectives of the work groups were to synthesize
available knowledge about the causes and effects of transboundary air
pollution, with initial emphasis on acid deposition, for use by the govern-
ments of the two countries in negotiating a bilateral air quality agreement.
Under the auspices of Work Group 2, the Regional Modeling Subgroup conducted
the Phase III evaluation of eight linear regional-scale sulfur transport
models using standardized input and validation data sets. The Modeling
Subgroup consisted of some members of Work Group 2 (restricted to employees
of the Canadian federal or provincial governments and the United States
federal government) and participating modelers. The participating modelers
included non-government scientists directly involved in the operation of the
Phase III selected models and those interested in advancing the science of
model evaluation and intercomparison.
It was not possible (due to practical modeling considerations as well as our
incomplete understanding of the phenomena) to provide models for all of the
pollutants of interest or to incorporate all the processes mathematically
into operational regional models. For example, deficiencies in both emission
inventories and in our comprehension of transformation and deposition
processes precluded the development of quantitative models for acid nitrate
deposition. Nor was it possible to incorporate the detailed atmospheric
chemical reactions between SO , NO , volatile organic compounds (VOC),
oxidants, and their acidic reaction products. The models used by Work
Group 2, therefore, contained simplifying assumptions which were based upon
our current understanding of the phenomena of long-range pollutant transport.
The results of the Phase III evaluation of the eight long-range transport
models are described in the Regional Modeling Subgroup Final Reportl
and are summarized in the Atmospheric Sciences and Analysis Work Group 2
Final Report.2 The information contained in the Subgroup Report is repre-
sentative of the current state of knowledge in modeling long-range transport
of air pollution, given the limited time and resources available to conduct
the evaluation and to prepare the report.
Model Profiles
Eight regional-scale sulfur transport models (Table I) developed by Canadian
and United States scientists were selected by Work Group 2 based on cri-
teria established during earlier phases of the MOI. These models were
applied by the Regional Modeling Subgroup using standardized input and
validation data sets to fulfill the MOI terms of reference. The techniques
of simulating the transport, diffusion, transformation, and deposition of
pollutants among the models were quite varied since: (1) the best modeling
techniques were not clearly discernable; (2) the natures of all relevant
physical and chemical processes were not well understood; (3) each modeler
made various assumptions to simplify the complex processes; and (4) each
model was developed independently.
PROCEEDINGS—PAGE 318
-------
The models are "linear" in the sense that chemical transformations and
scavenging are expressed as first-order processes where the rate constants
are assumed to be independent of emissions and co-pollutant concentrations.
This assumption of linearity that is incorporated into all eight models
used in the Phase III effort may be questioned. However, nonlinear para-
meterizations for models have either not been available or have not been
feasible for incorporation into regional-scale transport models of the
types used here at the present state of the art. The question that is
subject to individual scientific judgment is whether the linear sulfur
models represent a reasonable first approximation in the absence of opera-
tional nonlinear models. It is not clear that nonlinear effects would
invalidate the general monthly and annual results of these linear models.
Each of the models simulated the transport, diffusion, transformation, and
deposition of sulfur compounds. Monthly-averaged concentrations and
depositions, as well as monthly source-receptor relationships (transfer
matrices), were generated for central and eastern North America. The AES,
ASTRAP. CAPITA, MOE, and RCDM models also included varying portions of
western North America in the model domain. The ASTRAP, CAPITA, ENAMAP,
MEP, MOE, and RCDM models were source-oriented, which facilitated genera-
tion of concentration and deposition fields, while the AES and UMACID
models computed the concentrations and depositions at user-specified
receptor points.
Most of the models utilized gridded SCK point and area source emissions as
input and, with the exception of the ASTRAP and MEP models, treated the S02
emissions within a given grid cell as one virtual point source emitting
pollutants at one level or within one layer. The ASTRAP and MEP models
distributed the emissions vertically as a function of stability and stack
characteristics. The emission grid resolution varied from model to model
within the range of 70 to 190 km. In addition to S02> the AES, ASTRAP,
CAPITA, ENAMAP, MOE, and UMACID models were capable of including primary
sulfate emissions. During Phase III, the MOE model assumed a constant
SO 2~/SO ratio of 0.02, while the CAPITA model assumed that 1% of the
total sulfur emitted was sulfate. The UMACID model varied the percentages
of sulfur emitted in the form of sulfate from 1% in the winter to 5% in the
summer. The AES, ASTRAP,' and ENAMAP models required sulfate emissions
input; if none were available (as was the case in Phase III), sulfate
emissions were not considered.
Six of the models required objectively-analyzed meteorological observations,
while the other two (MOE and RCDM) used the statistical characteristics of
long-term climatological data. For example, the MOE model made assumptions
regarding long-term wind statistics and the RCDM model used monthly and
seasonal resultant wind vectors at each source emissions area to define the
pollutant transport process. The MEP model used 6-hourly surface pressure
data to generate wind fields, while the remaining five models used analyses
of actual wind data (surface and/or upper air) observed at 3- to 24-hourly
intervals. Precipitation data input requirements ranged from 3-hourly
analyses to the average durations of wet and dry periods over seasonal or
annual time periods. The analyses of the wind and precipitation data were
performed by preprocessors and were not modules within the models.
PROCEEDINGS—PAGE 319
-------
The mixing heights parameterized in the AES, ENAMAP, RCDM, and UMACID
models were treated as seasonally dependent constants. In these models, no
diurnal fluctuations were considered; the heights reflected the estimated
average mixing height over the simulation period. Only the AES model used
monthly gridded values of the mixing height based on seasonally stratified
climatological data. The CAPITA model used seasonally dependent mixing
height values but also considered a nocturnal lowering of the mixing height
to 300 m, thus simulating the release of nocturnal emissions above the
surface layer. The ASTRAP and MEP models considered a diurnally fluctuating
mixing height as well as the occurrence of a nocturnal inversion. A user-
specified, fixed mixing height of 1000 m was used in the MOE model for the
Phase III application.
The MOE and RCDM models used analytical functions to determine the dis-
tribution of mass in space and time after emission, while the other models
treated the emissions as discrete puffs. The mass of sulfur in each puff
was determined from the time increment (3 to 12 hours) and the emission
rate. The AES model allowed input of pollutants to boxes which were trans-
ported across emission areas at 3-hourly time steps, while the other five
models actually simulated the transport of individual pollutant puffs.
Horizontal dispersion was simulated in different ways. The AES and ASTRAP
models calculated long-term dispersion directly through the distribution of
simulated trajectories. The MEP model assumed a Gaussian distribution about
the simulated plume centerline. The CAPITA model used constant daytime and
nighttime horizontal dispersion rates, plus vertical shear overnight. The
RCDM model assumed a constant horizontal dispersion rate, while the ENAMAP,
MOE, and UMACID models used time-dependent dispersion parameters.
Vertical dispersion in all the models except ASTRAP occurred instantaneously
within the specified mixed layer, resulting in a homogeneous distribution of
pollutants in the vertical. Additionally, the MEP model allowed pollutant
input above the mixed layer. The vertical dispersion in ASTRAP, which
considered nine sublayers within the boundary layer, was dependent upon a
diurnally varying stability profile.
The rate of transformation of SO to sulfate was defined to be a constant
1% per hour in the AES, ENAMAP, HOE, and RCDM models. The transformation
rate varied seasonally in the remaining models. In addition to the seasonal
variations, the ASTRAP, MEP, and UMACID models also considered diurnal
variations in the transformation rates.
Dry deposition of sulfur compounds was simulated using a constant deposition
rate for S0_ and another for sulfate in the AES, MOE, and RCDM models. The
CAPITA and ENAMAP models used seasonally dependent dry deposition rates,
while the remaining models used diurnally and seasonally dependent rates.
The RCDM model also used the total dry time (total time in a period minus
the total precipitating time) to calculate dry deposition.
All but the CAPITA and MOE models simulated wet sulfur deposition based on
the actual precipitation rate over a specific period times a scavenging
coefficient. The CAPITA model computed wet deposition using the 6-hourly
probability of the occurrence of various intensities of precipitation across
each 127-km grid square. The MOE model used an average precipitation rate
PROCEEDINGS—PAGE 320
-------
based on climatological data times a scavenging coefficient, and assumed
that the probability of precipitation scavenging was related to the
durations of wet and dry periods along the pollutant trajectory. Wet
deposition in the RCDM model was simulated by the product of the scavenging
coefficient, the total precipitation for the simulation period, and the
ratio of total time to wet periods. The MEP model also considered the pH
of the precipitation based on seasonal pH observations and the ambient
temperature as factors for determining the amount of wet sulfur deposition.
The AES, MEP, and MOE models systematically added a constant background
sulfur contribution to the annual wet sulfur deposition.
These descriptions of the model input data requirements, model parameteri-
zations, and model output address specific Phase III applications of Work
Group 2. In a series of applications of different scenarios, the nature
of the available input data, the time and cost constraints, and the
desired model output often dictate the form of model parameter!zations.
Thus the features used in the Phase III model parameterizations may differ
from those used in previous phases and are subject to change for subsequent
applications.
It should be noted that these eight long-range sulfur transport models
represent the various types of currently available models capable of
simulating sulfur transport, diffusion, transformation, and deposition.
The variations in techniques and parameterizations of the models reflect a
breadth of scientific opinions and judgment amongst the modelers. As the
pertinent physical and chemical processes and relationships become better
understood, these models will likely be modified to reflect the resulting
gains in knowledge.
Phase III Input Data Bases
The standardized 1978 input data sets for the Phase III modeling effort
included sulfur emissions and precipitation data. In the Phase III
Canadian SO emissions inventory representative of 1978, point sources
were locatea by latitude/longitude while area source emissions were
gridded on a 127-km spacing. The area emissions were calculated for 1976,
but were not expected to differ greatly from 1978. All large point sources,
specifically major power plants and smelters, were updated to their 1978
emission levels while smaller point sources were reported for 1976, the
most recent year for which data were available.
The United States SO emissions inventory for 1978 was given for each
eastern state as total emissions for the utility and non-utility sectors.
The utilities inventory was based on 1978 fuel consumption, plant flue gas
desulfurization equipment, and other fuel characteristics. Although the
non-utility emissions estimates were derived from 1980 data, their use was
recommended by Work Group 3B in the absence of better information.
A further disaggregation of the emissions for the eastern United States
was performed by the Regional Modeling Subgroup. Emissions from the
200 largest utility emitters for 1978 were gridded based on the source
latitude/longitude. The remaining utility emissions were distributed as
state wide percentage changes in the 1979 emissions of plants not included
in the top emitters for 1978. The non-utility emissions were disaggre-
gated within each state by scaling the non-utility source totals to
PROCEEDINGS—PAGE 321
-------
the emissions inventory prepared for the Multistate Atmospheric Power
Production Pollution Study (MAP3S). In the MAP3S inventory, point
sources were located by latitude/longitude and area source emissions
were presented as county totals located at the area centroid of each
county.
The western United States emissions distributions were provided by Work
Group 3B. The emissions from large point sources, except utilities
emitting more than 25,000 (short) tons per year, were taken from the
1978 National Emissions Data System (NEDS) files and were identified by
latitude/longitude. The large power plant emissions were estimated
separately by an EPA contractor. All other emissions were extracted
from NEDS and provided by Air Quality Control Region. Because of this
piecemeal approach, standardized western United States emissions were
not available in model compatible form in time for incorporation into
the Phase III model runs. Thus, different emissions estimates for the
western states—if any—were used by the participating modelers.
To facilitate analysis of some modeling results, the 40 emissions source
regions in the United States and Canada (Figure 1) were grouped into
11 aggregate regions (Table II). The uncertainties in the emissions
data were not available at the time of model evaluation. It is not
unlikely that these uncertainties are functions of source regions, which
would affect the model results significantly since a constant uncertainty
factor would only introduce a bias. Furthermore, since emissions were
only available as annual totals, model predictions cannot be expected to
accurately simulate monthly or seasonal emissions variations. To properly
analyze the behavior of model predictions, one would need to know the
magnitudes and spatial variability of all the emissions uncertainties.
The amount of sulfur deposition predicted by the models was a function
of the estimates of precipitation used as input. Therefore, it was
important to make independent, rigorous estimates of actual precipitation
at the targeted sensitive receptors as well as at the sites used to
evaluate model performance. In addition, since the meteorological
representativeness of the period of simulation was of special interest,
the precipitation amounts for 1978 were compared with 30-year normals to
determine how precipitation amounts for this year compared with statistical
averages.
At the thirteen model evaluation sites in eastern North America, the
precipitation data obtained from rain gauge values were corrected for
biases in rain gauge type and for wind effects. At the nine targeted
sensitive receptors used for construction of transfer matrices, areal
estimates of the monthly and annual precipitation amounts were determined
by averaging the daily total observed precipitation from all recording
sites in an 80 by 80 km grid cell centered on the receptor. These
values were similarly adjusted to reflect unbiased precipitation amounts.
The observed precipitation during 1978 at the ten Canadian evaluation
sites was on the average 50% above the 30-year normal in January, 20% below
the normal in July, and about normal for the year. At the targeted
sensitive receptors (Figure 2), five of the areas had January 1978 totals
that exceeded the 30-year average by more than one standard deviation,
showing that the month was exceedingly wet. For the month of July 1978
PROCEEDINGS—PAGE 322
-------
and for the year 1978, only two sites had values that fell outside (below)
one standard deviation from the average values.
Phase III Network Observations
The evaluation of model estimates with observations was a critical aspect
of assessing model performance. The Regional Modeling Subgroup agreed in
Phase III to compare results from the selected models with network data in
eastern North America for January, July, and annual 1978. Different net-
works were surveyed and evaluated based on availability of data, sampling
period, geographical coverage, data capture and the degree of local
effects. The Electric Power Research Institute's Sulfate Regional Experi-
ment (EPRI SURE) and the MAP3S sampling networks in the northeastern
United States and the Canadian Network for Sampling Precipitation (CANSAP)
network in Canada were selected for the model evaluation.
In order to proceed with the model evaluation within the Phase III report
deadline, the Regional Modeling Subgroup decided on criteria for data
representativeness and capture, and .on methods for estimating uncertainties
in the data. The criteria, methodology, and resulting evaluation data for
each network are described in detail in the Regional Modeling Subgroup
Final Report.^
The only network meeting the selection criteria for air quality data was
the EPRI SURE, which measured SO- and sulfate concentrations in the
eastern United States on hourly and 24-hourly intervals, respectively.
During the period under consideration, the nine Class I sites operated for
the entire year while the 45 Class II sites only operated during four
intensive periods (January 10 to February 10, April, July, and October 1978),
It was agreed not to use the SO data because these measurements were
subject to influence by local sources, and because 50% of the SO- measure-
ments were below the detection limit of 3 ppb.
Sulfate measurements for both Class I and Class II EPRI SURE sites were
used to generate evaluation data sets for January and July of 1978 while
only Class I sites were considered for the annual average estimates. To
generate a monthly average concentration, a site was required to have a
minimum data capture of 65% for the month except for the Class II sites in
January, where data capture was relaxed to 55% (since these sites only
operated a maximum of 22 days during that month). At least nine months of
data were required for valid annual estimates. No regionally representa-
tive air quality measurements were available in Canada, except for one
site in the EPRI SURE network.
The CANSAP network measured concentrations of sulfur in precipitation and
precipitation amounts, permitting estimates of sulfur deposition for the
sampling period (usually monthly). CANSAP had 20 sites operating within
the region east of Manitoba. Ten of these were rejected because of poor
handling, insufficient data, location in urban or industrial areas, or
influence by local sources. Three of the remaining sites were considered
questionable and flagged as such, but were included in the model evalua-
tion exercise. For the ten acceptable sites, it was required that for the
monthly sample, the monitor must have operated a minimum of 20 and a
maximum of 35 days per sampling month and have a minimum collection effi-
ciency of 25%. Collection efficiency is defined as the ratio, expressed
PROCEEDINGS—PAGE 323
-------
in percent, of the precipitation amount recorded by the sampler to the
actual amount estimated from the catch in a co-located rain gauge, where
the catch was adjusted based on collector type, site exposure, and wind
effects.
CANSAP wet deposition samples at coastal sites were corrected for con-
tributions from sea salt, while corrections for under-catch and evaporation
were made at all ten sites. Resulting values for wet deposition were
expressed as kilograms of sulfur per hectare. Annual amounts were obtained
by simply scaling the collection period (9 to 11 months) amounts to the
annual precipitation at the site.
The MAP3S network sampled on a daily basis. Samples with a catch of less
than 50% of the nearby rain gauge volume were not used for deposition
calculations. When the daily sampler catch exceeded 50% of the rain gauge
catch, the volume of the rain gauge was used for deposition calculations.
Total deposition for a specific month was calculated only when the total
sampler catch for the month was a minimum of 90% of the rain gauge volume.
Only four sites reported wet deposition for July 1978, but one of these was
rejected because of an unresolvable difference of precipitation between the
reported sample catch and the official rain gauge report at the same site.
January and annual 1978 wet deposition estimates were only available for
one MAP3S site and hence, were not included in the data set for model
evaluation.
The data from the CANSAP, MAP3S, and EPRI SURE networks are believed
insufficient for conclusive model evaluation. The periods chosen, two
sample months and the entire year, were too short, while the spatial dis-
tribution of the data was also poor. Air quality data were only available
in the eastern United States while the precipitation chemistry data were
mainly for Canada. The uncertainty, especially in the CANSAP data, was
high, being about a factor of two in many cases based on under-catch
alone. In addition, a bias in the annual estimates is likely when only
nine months are required for the annual sample and when the missing months
fall completely within one season. Based on these considerations, the
model evaluations to be described below can only serve as a demonstration
of the statistical methodology to be used and as a preliminary indication
of model performance.
Transport Winds and Precipitation Fields
An attempt was made to compare the methods which were used by each model to
analyze the transport winds and precipitation fields, and to compare the
dispersion which was either explicitly expressed or was inherent in the
scatter of sequential trajectories. The purpose was to offer insight into
the manipulation of the standardized meteorological data by the different
models in the hope of explaining some of the differences in results between
the models.
Each modeler was requested to calculate trajectories at time steps of
12 hours up to a total of 96 hours for January and July 1978. The origins
of the trajectories were chosen to be Sudbury, Ontario and St. Louis,
Missouri. The former site was selected because it represented a signi-
ficant point source in a region with relatively sparse wind data, while the
latter represented an area of relatively good wind data availability in
all directions. The resulting mean displacements for the various models
PROCEEDINGS—PAGE 324
-------
varied considerably, both in distance and direction, depending upon the
type of data used to generate the wind flow. Such discrepancies in the
calculation of mean trajectories were a definite factor in producing dif-
ferences in model concentrations and depositions.
Some of the models displayed westward displacements after long eastward
travel times; this was an attribute of the analyses for this trajectory
exercise and did not represent what was used in the model evaluation
effort. Pollutant puffs transported to the northern or eastern boundaries
were dropped since computation of trajectories outside the modeling domains
were terminated, thereby biasing the trajectory statistics toward the lower
wind speeds used in this exercise.
The dispersion over the extended period of analysis caused by shifting wind
patterns was computed explicitly by two of the models, but was implicit in
the sequential trajectory calculations of the remaining models. To compare
the effective dispersion in January and July 1978 from Sudbury and
St. Louis, calculations were performed of the square root of the sum of the
variances in the x (east) and y (north) directions. As with the trajectorie
the values of the standard deviations from model to model varied widely,
depending on the methodology that was employed to calculate the transport
field.
An analogous test was made for the second meteorological input common to
the models, namely precipitation. The precipitation gridding process is
quite dissimilar between models, which can lead to significantly different
results. For example, some models gridded hourly precipitation data to
compatible time steps and grid configurations. Since these gridding
processes and configurations were dissimilar, the precipitation amounts
derived for grid cells encompassing wet sulfur deposition receptors usually
differed between models. Other models relied on a statistical treatment of
the stochastic properties of precipitation to estimate the precipitation
occurrence and amounts along the trajectories. This analysis involved
specification of the probabilities of changing from wet to dry periods and
vice versa, and the frequency distribution of precipitation amounts.
Therefore, similar precipitation data sets obtained and gridded independ-
ently can lead to discrepancies in the modeling results.
Model Evaluation
The physical and chemical processes associated with the long-range trans-
port and deposition of pollutants are extremely complex and, at the present
time, our understanding of them is limited. Because of this limitation,
the framework of long-range transport models will have to be modified to
accommodate new information as it becomes available. This suggests that
our confidence in model simulation is directly related to the amount of
evaluation that has been performed.
A set of criteria for model evaluation was formulated by the Regional Model-
ing Subgroup, based on the recommendations of the Workshop on Dispersion
Model Performance sponsored by the American Meteorological Society. The
criteria were based on the differences between observed and model-predicted
monthly and annual values for ambient sulfate concentrations and for
wet sulfur deposition. The sulfate concentrations used for model evalua-
tion were obtained from the EPRI SURE monitoring network as shown in
PROCEEDINGS—PAGE 325
-------
Figure 3; the number of values available were 29, 47, and 9 for January,
July, and annual 1978, respectively. All but one of the sites were located
in the United States. The networks of CANSAP (ten sites in Canada) and
MAP3S (three sites in the United States) were used as the source of wet
sulfur deposition values for the model evaluation exercise (Figure 4); the
number of values available were 7, 13, and 8 for January, July, and annual
1978, respectively. The statistical tests and evaluation results are
described in detail in the Regional Modeling Subgroup Report.1
Although significant limitations exist in the input emissions and preci-
pitation data as well as in the measurement data used for evaluation, some
conclusions can be drawn regarding the overall performance of the models.
Collectively, the models appeared to perform better for wet sulfur deposi-
tion than for ambient sulfate concentration prediction. This is somewhat
surprising because wet sulfur deposition is episodic in nature, whereas the
model results were aggregated as non-episodic or longer-term. Ambient S0_
was not considered for the evaluation exercise due to unavoidable contami-
nation of data from nearby sources and errors in measurements at very low
concentrations.
The evaluation data set for annual 1978 contained only eight points for wet
sulfur deposition and nine points for sulfate concentrations. This did not
constitute a sufficient data set from which to draw statistical conclusions
on the relative performance of models. Therefore, although some models
appear individually to perform better than others, no firm conclusions
should be drawn without more extensive evaluations. Clearly, while a start
has been made toward developing evaluation statistics, further testing must
be done to provide reliable quantitative information about model performance.
Transfer Matrices
The MOI terms of reference required Work Group 2 to recommend tools for
preliminary assessment activities. This included the evaluation of observa-
tions and the estimation of emissions reductions that would be needed in
source regions in order to achieve proposed reductions in air pollutant
concentrations and deposition rates necessary to protect sensitive areas.
The principal tools available for this assessment are air quality simula-
tion models (including the long-range transport models used in this exer-
cise) , local and mesocale models, and transfer matrices. Transfer matrices
should be viewed primarily as a convenient form for either representing the
results of a model or for applying those results to the study of emissions
reduction scenarios.
The effect that emissions of pollutants from any source (or group of sources
in a region) will have in producing ambient concentrations and deposition
of pollutants at some other receptor location is known as the source-
receptor relationship. That relationship is determined by the directions
that pollutants are carried by the winds and also by the dispersal, chemical
transformation, and removal of pollutants along the way. It should be
understood, however, that nonlinear processes and interactions among pollut-
ants are important to the actual, absolute source-receptor relationships.
To the extent that such nonlinear processes are important, one cannot
define a simple relationship among.the regions because the effect that
emissions from one source have on any receptor depends somewhat on various
pollutants emitted from other sources.
PROCEEDINGS—PAGE 326
-------
Transfer matrices are based on the principle of linear superposition of
pollutant emissions from various sources. This means one assumes that
concentrations and/or depositions at a receptor location are the sum of
partial contributions, where each contribution is proportional to emissions
from a source or group of sources in an upwind region. Thus, a coefficient
of proportionality exists connecting each source region with each receptor.
The array of these coefficients connecting all sources with all receptor
locations of interest constitutes a normalized or unit transfer matrix.
The eight long-range transport models were used to calculate transfer
matrices for ambient sulfate and for wet and dry sulfur deposition for the
periods of January, July, and annual 1978. Nine targeted sensitive
receptors were originally selected by the Canada/United States Research
Consultation Group as receptors sited in regions of northeastern North
America which are sensitive to acidic deposition. Figure 2 shows the
locations of these receptors. The 40 by 9 matrices of Phase III pertained
to groups of all SO. sources over 40 source regions (Figure 1) and to the
nine targeted sensitive receptor areas. Smaller 11 by 9 matrices were
also produced by aggregating the 40 source regions into 11 larger source
regions (Table II). Tables of normalized transfer coefficients (matrices)
for wet and dry sulfur deposition and for ambient sulfate concentration
calculated by the eight long-range transport models are presented as
Appendices to the Regional Modeling Subgroup Final Report.1
The Phase III matrices showed substantial variation in the absolute size
of transfer coefficients from one model to another which is to be expected
since the annual matrices are based on various meteorological periods and
different analysis methods using the 1978 input data. Hence it is difficult
to ascertain correct values for the absolute size of source-receptor rela-
tionships.
If emissions reduction studies are not based on absolute values of source-
receptor relationships, the demands on accuracy from matrix elements are
less critical. For example, if prescribed emission reductions were estab-
lished for large regions, then matrix elements might be used for only a
qualitative indication of the relative importance of different source
regions to given receptor areas. For purposes of qualitative assessment,
perhaps only the relative position of rank orders among source regions
would be of concern. Rank order simply means that the source regions are
placed in order according to the value of the matrix elements pertaining
to a given receptor. Thus, if the value for source A exceeds that for
source B, it is ranked above B irrespective of the absolute size of the
difference. Preliminary analysis indicated that the eight models produced
matrices with reasonably consistent rank orderings.
Transfer matrices represent the results of model calculations and generally
suffer from any limitations in the model and input data used to calculate
them. Some additional limitations are inherent in the matrices as a result
of definitions and approximations made in their calculation, which include
the lumping of emissions sources in each source region and averaging
pollutant values over space and time in the receptor area. These specific
matrix limitations are not at issue, but need to be accounted for when
attempting to use the matrices in practical applications.
PROCEEDINGS—PAGE 327
-------
More controversial questions about matrices relate to interpretation of
the significance of matrix elements and to their validity as assessment
tools. Normalized matrix elements explicitly identify quantitative rela-
tionships between source regions and receptor regions and tend to imply a
linear relationship between emissions and air quality or deposition.
While those relationships are also imbedded in the parent model if it is
one of the linear types used in Phase III, the source-receptor relationship
is not so obvious in aggregate model results. Predictions of absolute
values of concentrations and depositions can be compared with observed
data, but neither the linearity nor the partial contributions from each
source region can be directly verified by comparison with currently
available data. In the absence of such direct validation, there is dis-
agreement about the degree of confidence that can be placed in the linearity
of prediction and in the inferred source-receptor relationships based upon
given level of agreement between model calculations and current absolute
observations.
Conclusions
While differing opinions persist within the modeling community as to the
proper method and the statistics to be used for evaluation and intercom-
parison of model results, the Regional Modeling Subgroup designated
specific evaluation criteria for performing this task for the eight
selected models used in the Phase III MOI application.
It is generally accepted that one should expect model predictions to
deviate from measurements because a practical model cannot incorporate
even our current understanding of the relevant chemical and physical
processes involved in long-range transport and deposition of pollutants,
and because available observational data are insufficient to estimate the
ensemble average which the model is designed to predict. Due to these
deficiencies, it is not possible to quantify the uncertainties in model
predictions based on the differences between model predictions and obser-
vations (residuals).
One of the original tasks for the Regional Modeling Subgroup was to recom-
mend the best model(s) for future use in emission reduction scenarios.
Towards this end, the input parameters of the models were displayed and
model evaluation criteria were standardized to ascertain model accuracies.
However, due to uncertainties in the emissions inventory and precipitation
data used by the models as well as in the measurement data used for
evaluation, a ranking of models in terms of their absolute performance
could not be made during the Phase III effort.
Furthermore, the assumption of linearity by the models provides for rate
constants that are independent of emissions and co-pollutant concentra-
tions. The question is raised as to whether the linear models of sulfur
transport represent reasonable first approximations in the absence of
operational nonlinear models. Complete analysis of the impact of non-
linearity was beyond the scope of the Phase III effort, although addi-
tional sensitivity analyses using the models and evaluations against
measurement data could be expected to shed more light on this matter. For
the present, however, it thus becomes 'subject to individual scientific
judgment whether or not nonlinear effects would invalidate the general
monthly and annual results of these linear models.
PROCEEDINGS—PAGE 328
-------
As a group, the simple linear models appear to be able to reproduce the
right order of magnitude of the large time and space-scale features of the
measured wet sulfur deposition patterns. The Phase III evaluations of the
models with observed data showed that collectively, most models performed
relatively better in predicting the deposition of sulfur in precipitation
than in predicting ambient sulfate concentrations.
Transfer matrices for sulfur species produced by the eight long-range
transport models revealed variations among the absolute magnitudes of the
transfer matrix elements for a single model and for the same element among
different models. It has not been possible to date to choose a best model
among the eight nor to produce with confidence a best estimate single
transfer matrix for each variable based upon a valid statistical analysis
of all model results. However, all eight models predicted generally
similar relative impacts on the receptors in terms of ranked order of
importance based on the meteorology of one given year.
The diversity of parameterizations in these eight models reflects the fact
that they were independently developed. The variations in techniques and
parameterizations reflect a breadth of scientific opinions and judgment
amongst modelers. It is not at all surprising that models show differences
in detail when applied to a single data set. On the contrary, the extent
of agreement among the model outputs and between these outputs with
observed wet sulfur deposition data is encouraging.
Acknowledgment
This paper is a synopsis of work performed by the following members of the
Regional Modeling Subgroup of Work Group 2:
Ball, Richard H. Olson, Marvin P. Shannon, Jack D.
Clark, Terry L. Patterson, David E. Weisman, Boris
Ley, Barbara Peck, Eugene L. Venkatram, Akula
Niemann, Brand L. Samson, Perry J. Voldner, Eva C.
REFERENCES
1. F. A. Schiermeier and P. K. Misra, "Regional Modeling Subgroup Final
Report," United States/Canada Memorandum of Intent, Report No. 2F-M,
198 pp (1982).
2. H. L. Ferguson and L. Machta, "Atmospheric Sciences and Analysis Work
Group 2 Final Report," United States/Canada Memorandum of Intent,
Report No. 2F, (1982).
PROCEEDINGS—PAGE 329
-------
TABLE I. TRANSPORT MODELS ASSEMBLED BY REGIONAL MODELING SUBGROUP
Model Name
Acronym
1. Atmospheric Environment Service Long-Range Transport Model AES
2. Advanced Statistical Trajectory Regional ASTRAP
Air Pollution Model
3. Center for Air Pollution Impact and Trends Analysis CAPITA
Monte Carlo Model
4. Eastern North American Model of Air Pollution ENAMAP-1
5. Transport of Regional Anthropogenic Nitrogen and Sulfur MEP
Model of Meteorological and Environmental Planning, Ltd.
6. Ontario Ministry of Environment Long-Range Transport Model MOE
7. University of Illinois RCDM-3
Regional Climatological Dispersion Model
8. University of Michigan Atmospheric UMACID
Contributions to Interregional Deposition Model
TABLE II. PHASE III SOURCE REGION GROUPINGS
Region
Number
Region
Source Areas
Included
1
2
3
4
5
6
7
8
9
10
11
Maritime Provinces
Quebec
Ontario
Western Provinces
Northeastern States
Eastern Midwest States
East Coast States
Southern and Gulf Coast States
Central States
Western Midwest States
Western States
20, 21, 22
17, 18, 19
12, 13, 14, 15, 16
10, 11, 23, 24
59, 68, 69, 70
50, 52, 58
63, 66
60, 64, 65, 67
51, 53, 54, 56, 57
55, 61, 62
71, 72, 73, 74
PROCEEDINGS—PAGE 330
-------
Figure 1. North American emissions source regions,
,'•Ł,0
Figure 2. North American targeted sensitive receptor areas
PROCKKDJNGS
331
-------
JANUARY. JULY, AND
ANNUAL
JULY AND ANNUAL
• JANUARY AND JULY
A JULY
T JANUARY
Figure 3. EPRI SURE sites used in Phase III model evaluation.
LEGEND
JANUARY, JULY, AND
ANNUAL
© JULY AND ANNUAL
O JULY
Figure U. CANSAP and MAP3S sites used in Phase III model evaluation.
PROCEEDINGS—PAGE 332
-------
FIELD STUDIES ON PHOTOCHEMICAL AIR POLLUTION
IN JAPAN
Shinji Wakamatsu
Itsushi Uno
Makoto Suzuki
Yasushi Ogawa
National Institute for
Environmental Studies
P.O. Yatabe, Ibaragi 305
JAPAN
presented by S. Wakamatsu
National Institute for Environmental Studies
Japan EA
PROCEEDINGS—PAGE 333
-------
1. Introduction
Photochemical oxidants are major air pollutants in Japan. The Japan
Environment Agency has been measuring oxidant concentration at more than
900 stations throughout Japan. Especially high concentrations are
observed around the Tokyo Metropolitan Area and the Osaka area in the
summer season, but high oxidant concentrations, more than 60 ppb,
exceeding the environmental standard are also observed in the spring
season at many monitoring stations in Japan. In this report, the
monitoring data are analyzed statistically to clarify the seasonal
variations of oxidant concentration in Japan. The aerial and temporal
scales of photochemical air pollution are also analyzed using the data
from about 300 monitoring stations and aircraft data covering the Tokyo
Metropolitan Area.
2 The seasonal variations of oxidant concentration
The general phenomenon of high photochemical smog when a
photochemical oxidant warning is issued usually occurs from May to
September. A photochemical oxidant warning is issured by the governor of
each prefecture when the hourly averaged oxidant concentration is 120
ppb or higher and when this state is likely to continue from a
meteorological viewpoint. If the warning is issued in two prefectures on
the same day, it is considered as two warning days. The total number of
warning days in 1981 was 59. Of these 41 days were issued from the
Tokyo Bay area and 16 days were issued from the Osaka Bay area. In
addition, relatively high concentration (exceeding 60 ppb) are observed
at many monitoring stations in Japan.
To clarify the seasonal variations of oxidants in Japan, daily
maximum oxidant concentration between 1976 and 1980 were averaged
monthly. These results are shown in Figure 1. The seasonal variation was
classified into three distinct types. The first type shows a peak in
the spring season between April and May. This type is mainly observed
in northern part of Japan. The second type shows double peaks
corresponding to the spring and autumn seasons and this type is mainly
observed in southern Japan. In both cases, maximum oxidant
concentrations are approximately 60 ppb. In the third type, the maximum
peak of oxidants is mainly observed in summer season having a value of
approximately 70 ppb. This third type is widely observed near the big
city areas, such as Tokyo and Osaka. The former two types are probably
due to stratospheric ozone intrusion*1.3 The difference in seasons
between the southern and northern parts of Japan may correspond to the
difference of meteorological conditions. The third type is due to
anthropogenic sources. These results shows us that the effects of
stratospheric ozone subsidence are significant in the spring season, but
negligible in the summer. Next, the aerial and temporal scales of
photochemical air pollution from the anthropogenic sources will be
analyzed focusing on the Tokyo Metropolitan Area in the summer season.
PROCEEDINGS—PAGE 335
-------
3. Temporal scale of photochemical smog
The Kanto Plain, where the Tokyo Metropolitan Area is stuated, is
one of the biggest plains in Japan. The topography of this area shown in
Figure 2. The western and northern sides are walled by mountains, 1,000
to 2,000 meter high. The eastern side is open to the Pacific Ocean, and
in the south there are two bays, Tokyo Bay and Sagami Bay. The Keihin
Industrial Complex is situated at the western side of Tokyo Bay and the
Keiyo Industrial Complex is at the eastern side of this bay. On the
southwest side of Sagami Bay, the Izu peninsula projects into the
Pacific Ocean.
Generally speaking, when the geostrophic wind velocity is below the
level of approximately 6 m/s, five different types of local wind
circulations predominate in the summer season and the dynamics of
photochemical oxidants are closely connected to them. The sea breeze
from Tokyo Bay enter Tokyo as an SE wind. The large-scale sea breeze
from Sagami Bay, designated as an S wind, arrives in Tokyo later than
the Tokyo Bay breeze. The other large- scale sea breeze is from the
Kashima Sea and this is designated as an E wind. The another two local
wind circulations are mountain and valley winds. The mountain and valley
wind from the western mountain area is seen in the early evening as an W
wind and later, large-scale mountain and valley wind from the northern
mountain area predomitates as an N, continuing till the early morning.
Depending upon the general wind direction and wind speed, these five
local wind circulation systems create an extremely complicated wind
pattern over the Kanto Plain.
Between 15 July 1981 and 20 July 1981, photochemical smog warnings
were issued successsively in the Kanto Plain under such local wind
circulating systems. Reports on health injury suspected to be due to
photochemical air pollution came from the southern Kanto district
betbween 16 July 1981 and 18 July 1981.
On 16! July 1981 and 17 July 1981, three-dimensional observation
using two instrumented aircraft were conducted. Upper wind profiles
were also measured every hour at 23 stations, and temperature profiles
were obtained using a radiosonde at 3 stations. These measurements and
data from about 300 ground level monitoring stations provided
sufficient information to clarify the mechanism of this episode on 16,
17 July 1981. The arrangement of the pilot-balloon stations, the daily
maximum concentration observed in each prefecture and meteorological
data observed at the Tokyo Meteorological Obsrvatory are shown in Figure
3. On 16, 17 July 1981, a land breeze predomitated till noon and the
wind speed was low, so that high oxidant concentration was observed in
the southern part of the Kanto district due to solar radiation and high
temperature. From 18-20 July 1981, however, a southerly wind
prevailed and relatively high oxidant concentration was observed in the
northern prefectures of the Kanto district.
Figure 4a shows the ground level oxidant concentratin pattern at
1500JST which exceeded 100 ppb and the aircraft measurement data at an
altitude of 350 in . The southerly Sagami Bay breeze was stopped by the
easterly wind from the Kashima Sea and high concentrations were observed
in the Sagami Bay sea breeze area. Aircraft data show that the extremely
high concentration, exceeding 200 ppb, were observed near the coast side
of Sagami Bay between H50-1455JST. ' Figure 4b shows the air
PROCEEDINGS—PAGE 336
-------
trajectory at the altitude of 350 meters calculated from the pilot-
balloon data. According to this trajectory analysis the extremely high
concentration air mass at 1500JST was transported from the Tokyo Bay
industrial area. This air mass was transported inland, and after
2100JST traveled down to the south due to the land breeze. This pattern
was also observed on the next day.
Figure 5a shows the air trajectory at an altitude of 150 meter
calculated using the 23 point pilot-balloon data on 17 July 1981. The
12 start points of the trajectory were set on an ellipse at 1300JST on
the shoreline of Sagami Bay, then the backward and forward trajectories
were calculated at one-hour intervals. Trajectory shows the same
circulation as on 16 July 1981. Figure 5b shows the Lagrangian
variations of NO, NO^ and Oxidant concentrations along this trajectory
line. Concentrations were obtained by interpolating the ground level
monitoring station data. In the early morning, NO showed a peak, NOz
increased corresponding to the NO decrease, and the oxidant
concentration peaked in the afternoon. These tendencies coincidence
with the smog chamber experiments.(2> The maximum oxidant concentration
was observed at 1500JST at a point about 20 km inland from the shoreline
of Sagami Bay, and according to the trajectory analysis this high
oxidant concentration was mainly caused by emissions from the Tokyo Bay
coastal industrial complexes.
To understand the two-dimensional structure along the trajectory
line, the Figure 6 was drawn. Trajectory I and II are the averaged
trajectory lines between 15 July 1981 and 17 July 1981. It is clear
that the high oxidant zones correspond to the sea breeze zone in daytime
and the high HOz zones correspond to the land breeze zone at nighttime.
NO almost reacted with 03, but on 15 July 1981, the maximum temperature
and total solar radiation were not high so that the oxidant
concentration was low compared with the other two days, and NO remained
at nighttime.
From this analysis, it is clear that the time scale of the
photochemical smog in the Tokyo Metropolitan Area is longer than one
day, and under the condition of weak atmospheric pressure gradient,
polluted air masses circulate for more than three days following the
local wind circulation pattern.
4. Aerial scale of photochemical smog
The aerial scale of the photochemical smog covering the Kanto
district can be classified according to two types of meteorological
condition. The first type is local wind circulation predominant type,
and the second is general wind predominant type. These two types are
determined using the geostrophic wind calculated from the 700 mb
constant pressure gradients. When the geostrophic wind is below 6 m/s,
the local wind circulation systems predominate in the Kanto Plain. In
this section the dynamics of the three-dimensional distribution under
the condition of weak pressure gradient and long distance transportation
phenomena in the condition of general wind predominant type are discussed.
Figure 7 shows the vertical cross section measurements using four
PROCEEDINGS—PAGE 337
-------
(3)
instrumented aircraft simultaneously. Figure 7a shows that the primary
pollutants were trapped below the inversion layer and aged pollutants
were remained above 500 m. having the maximum 0 concentration of 100
ppb at 800 m. high ,30-40 Km distance from the shore line. According to
the sea breeze penetration and development of thermal convective layer
NO was transported to the inland and upward as seen in Figure 7b-7c. In
the afternoon maximum 0 was observed in Figure 7c,and in the midnight
140 ppb 0 remained at 500 m. high, 60-100 Km distance from the shore
coast in Figure 7d. In the next morning previous bays 0 was observed at
the upper layer ( 800-1000 m. high ) having the concentration of 140 ppb
and this high 0 area was corresponded to the land breeze zone shown in
Figure 7e. Generally speaking under the condition of the weak pressure
gradient the local wind circulation predominate and high concentration
of secondary pollutants are observed at the upper, outer part of the low
level stable zone in the morning and these aged pollutants are involved
in accordance with the increase of the thermal mixed layer in the
daytime and accelate the formation of the photochenmical oxidants. From
five years of aircraft observations conducted by NIES between 1978 and
1981W , the vertical scale of the phenomenon is approximately within 2-
3 km above ground level.
Figures 8 and 9 shows the case of the general wind predominant
type. The horizontal distance between Mt.Tsukuba and Makabe is 10 km
but the time variation in Figure 8b is quite different at nighttime.
This means that the high concentrations of secondary pollutants formed
in the metropolitan area were transported traveling above the nighttime
radiation inversion area.
When the southerly seasonal wind predominates, pollutants formed in
the Kanto district are transported to the Nagano area'.6' Figure 8a shows
the movement of the oxidant peak. The peak at Omiya was observed at
1400JST, 15 June 1979 and it traveled 10 hours to Nagano. In this case,
the aerial scale of the photochemical smog was more than 200 Km.
When the north-easterly wind predominated, pollutants were
transport to Sagami Bay. These phenomena are shown in Figure 9. In this
case, pollutants from the Keihin and Keiyo industrial complexes were
transported to Izu Peninsla. Pollutants stagnated near the east coast of
this peninsla and high ozone levels, above 150 ppb, was observed. Using
these data OH radical concentration was estimated to be approximately
0.2-0.4 ppti"
5. Summary and conclusions
Monitoring station data was analyzed statistically to clarify the
seasonal variations of oxidant concentration in Japan and the temporal
and aerial scales of photochemical air pollution were also analyzed
using ground level monitoring station data and aircraft data covering
the Tokyo Metropolitan Area. Results are summarized as follows:
(1) The effects of stratospheric ozone subsidence are significant
in the spring season, but the effects are negligible in the summer
season.
PROCEEDINGS—PAGE 338
-------
(2) The temporal scale of the photochemical smog formation
mechanism is longer than one day, and under the condition of a weak
atmospheric pressure gradient, polluted air masses circulate more than
three days in the Kanto district following the local wind circulation
systems.
(3) At nighttime and early morning, high concentration of secondary
pollutants are usually observed in the upper, outer part of the low-
level stable zone and these, aged pollutants are involved in accordance
with the increase of the thermal mixed layer in daytime.
(4) The scale of the photochemical smog phenomenon in the Tokyo
Metropolitan Area is large. When the general wind prevails polluted air
masses are transported greater than 200 Km distance from the source
area and distributed up to an altitude of 2-3 Km.
These characteristic features of photochemical oxidants in Japan
will give us a firm basis for constructing a photochemical smog
formation model.
PROCEEDINGS—PAGE 339
-------
References
(1) Murao.N., Okita.T. and Ohta.S. (1982): Contribution of stratospheric
ozone on the ground-level oxidant concentration, (in Japanese)
TENKI, Vol.29, No.5, pp. 537-545
(2) Akimoto.H., Bando.H., Sakamaki,F., Inoue,6., Hoshino.M. and Okuda.M.
(1979): Photooxidation of the prophylene-nitrogen oxides-air
system studied by long-path fourier transform infrared
spectrometry. (in Japanese) Reseach Report from the National
Institute of Environmental Studies, No.9, pp. 9-27
(3) Uno,I., Wakamatsu.S., Suzuki,M. and Ogawa.Y. (1982): Distribution of
photochemical pollutants and their three-dimensional behavior
covering the Tokyo Metropolitan Area. Japan-US Joint
Conference on Photochemical Air Pollution and Air Pollution-
related Meteorology, at Tsukuba, 1-2 December 1982
(4) Wakamatsu.S., Ogawa.Y., Murano.K., Goi.K. and Aburamoto,Y. (1982):
Aircraft survey of the secondary photochemical pollutants
covering the Tokyo Metropolitan Area. To appear in Atmos.
Environ.
(5) Wakamatsu.S., Uno,I., Suzuki,M. and Ogawa.Y. (1982): The Lagrangian
observation of polluted air masses using aircraft. Japan-US
Joint Conference on Photochemical Air Pollution and Air
Pollution-related Meteorology, at Tsukuba, 1-2 December 1982
(6) Kurita.H., Sasaki,S., Hatano.S., Wakamatsu.S., Uno.I. and Ueda.H.
(1982): Movement of Oxidant in Inland Area (1) -Relation
between Oxidant Concentration in Ueda Basin and Photochemical
Oxidant in Kanto District- . (in Japanese) Annual Meeting of
Japan Society of Air Pollution, Proceedings, at Miyazaki, 9-
11 November 1982
(7) Suzuki,M., Wakamatsu.S., Uno.I. and Ogawa.Y. (1982): Evaluation of
OH radical concentration using aircraft data. Japan-US Joint
Conference on Photochemical Air Pollution and Air Pollution-
related Meteorology, at Tsukuba, 1-2 December 1982
PROCEEDINGS—PAGE 340
-------
T
80
60
40
2C
°J
3 80
& 60
fto
5 zo
i °
ao
60
to
20
0
ype II
YAMAGUCHI
13
FM AMJ J ASONO
TOKUSHIMA
U
J FMAMJ J ASONO
CSAKA MlSAKI
16
j FMAMJ j ASONO
J FMAMJ J ASON
; :
:
;
-
OITA
15
1 FMAMJ J ASONO
NAGASAKI
^__^-\ YUKiUR*
17
j FMAMJ J ASONO
-Tokyo Metropolitan
•"/< Area
LI : 7
Figure 1. Seasonal variation of oxidants in Japan. Daily maximum
oxidant concentration between 1976 and 1980 were averaged monthly.
Numeral 1n the map shows the location of the selected monitoring
station.
I'KOCKKDINGS- PACK :M 1
-------
Figure 2. Topography of the Tokyo Metropolitan Area viewed from the
south and local wind pattern. Numeral shows following ;
1. Sea and land breezes from Tokyo bay
2. Sea and land breezes from Sagami bay and seasonal wind in the summer
3. Easterly wind from Kashima sea
4. Mountain and valley wind from the western mountain area
5. Mountain and valley wind from the northern' mountain area
PROCEEDINGS- PAGE 342
-------
N37
N36°
N35'
1;
120 Pfn.n
50 kr
PACIFIC Oc*an
EI39"
EUO"
EUI*
15 16 17 18 19 20
July 1981 (Day)
July 1981
^\Hour
Date"^^
15
16
17
18
19
20
Wind direction and speed
(m/s)
0900JS7
N 2.4
NW 1.3
N 1.5
N 2.5
W 0.8
SŁ 2.6
1200JS7
S 3.4
NNW 2.6
ENE 1.8
SE 3.1
SSE 2.7
S 4.3
1500JST
SSE 3.5
SSE 2.4
SSE 4.4
SSE 5.0
SSE 5.6
S 6.6
M.T
CO
31.5
34.1
34.3
33.6
32.2
32.4
T.S.R
(MJ/ai2'
15.4
19.7
20.5
18.5
17.0
13.7
M.T: Maximum temperature ( C)
T.S.R: Total solar radiation (MJ/'ra2)
Figure 3. (a) The arrangement of the pilot-balloon station.
(b) The daily maximum oxidant concentration observed
in each prefecture in Kanto district
(c) Meteorological data observed at the Tokyo
Meteorological Observatory
I'KOCKKDINCS I'ACK .t M
-------
Trajectory
16 July 1981
Altitude 350m
Start 1500JST
Figure 4. (a) The ground level oxidant concentration pattern at
1500 JST exceeding 100 ppb and the aircraft data at 350 m. on 16
July 1981. (b) The air trajectory at the altitude of 350 m.
calculated using 23 pilot-balloon data.
Trajectory
17 Julyl981
Altitude 150m
Start 1300JST
ISO
8 12 16 20
TIMECJST)
17 July 1961
Figure 5. (a) The air trajectory at the altitude of 150 m. calculated
using 23 pilot-balloon data on 17 July 1981 (b) The Lagrangian
variation of NO, NOz . and Oxidant concentration along the trajectory
line shown in Figure 5(a)
PROCEEDINGS- PAGE 344
-------
Onidants (cob)
Trajectory I
-
0 12 0 12 0 12 0 12UST) 0
15July1981 16Julyl981 17Julyl981 18 July 1981
•
-
•
-
-
:
10-
70
u iC
C
•»
NO? (ppb)
Trajectory II
NO (pob)
Oxidants (ppb)
Wind Direction
L.B.
L.B
6 12 18 0 6 12 18 0
(Hour . JST }
• 15July1981 1 l6July1981 1-
0
rcl
i
&
*
20
•
6 12 18 0
17July1981 1
July 1)81
^^\*our
3«t«^^^
15
16
17
18
19
20
Wind direction *«
-------
31 July »»» RUM 11 OS3S-OH3 JST
1»T» RUM I? IQIO-XXtJST
'-"• :-- \ _v-»<«——
/ Nfi \ •• .11
1 Orw
t lull i C<
1 C»<«<1UI 19 «•»•
I] UrtoM 14 TlNlw** 15
II
Figure 7(a)-(e). The Oj and HOi distribution on the flight
urse H-G shown In Figure 7(f) on 31 July 1979 and 1 August
979 .using four «1rcr«ft $1«ultan1ously. Bacieally each aircraft
flew at the altitude of 350, 650. 900 and 1200 a. respectively
these data was extrapolated. Horizontal and vertical wind
shown 1n Figure 7(c)-(e) are calculated using 19 points pilot
balloon data. Dashed line in the Figure (b)-(d) shows a potential
teaperature. Flight tine table are shown .s follows;
f
(a) Run 11 0430-0630 JST 31 July 1979
(b) Run 12 0945-1130 JST 31 July 1979
(c) Run 13 1450-1630 JST 31 July 1979
(d) Run 20: 0000-0145 JST 01 August 1
(e) Run 21: 0445-0620 JST 01 AuQuSt 1
PROCEEDINGS—PAGE 346
-------
2 i. 6 8 K) 12 14 16 '.8 20 22 24
Hour(JST)
15 June 1979
Wind Direction at Mt.Tsukuba
— Sw-SE —Cilm.i—SSE- SSW-
2 '50
a
c
o
^ 100
"c.
4t
|
u 50
•I
"c
<1
•D
(5 o
--NE-.N-.
j-ENC; -wsw
s-SW
! -SW i
Makabf —>
( 70m. SU
6 12 18
I August 1979 -
6 12 18 0
Hour (JST)i
2 August 1979——
Figure 8. (a) The movement of oxidant concentration peak to the
inland area ,(b) The diurnal variation of oxidants at Mt.Tsukuba
and Makabe , (c) Location of the selected monitoring station.
I'KOOKKDINCS -PACK :M'
-------
/..D
70 Km
RUN 5
RUN]
NO,
«AUG1980
:•
RUN J
NO
SAUGH
tOOO-ll»
ZSppt
*"" t/x^X
f AUC uao. ijoo-uzory >»
SQppb
LI. JSOm
RUN*
. • -
RUN 7
NO,
( AUC 19«0
noo-mojs'
e f
* 9;(il~n(r.L Horizontal distribution of 03 and N0,at the
altitude of 350-400 n. on the Sagami Bay. Arrows in the figure
tows a wind direction. Flight time table are shown as follows;
(a) Run 03: 0400-0600 JST 06 August 1980
(b) Run 04: 0710-09CO JST 06 August 1S80
Run 05: 1000-1130 JST 06 August 1980
(d) Run 06: 1300-1420 JST 06 August 1980
(e) Run 07: 1600-1710 JST 06 August 1980
(f) Run 08: 2010-2120 JST 06 August 1980
PROCEEDINGS--PAGE 348
-------
C Ref.
A Lagrangian Observation of Polluted
Air Mass Using Aircraft
Shinji Wakamatsu
Itsushi Uno
Makoto Suzuki
Yasushi Ogawa
National Institute for Environmental Studies
P.O. Yatabe, Ibaraki 305
JAPAN
PROCEEDINGS—PAGE 349
-------
1. Introduction
To understand the photochemical smog formation mechanism in
the atmosphere it is necessary to understand the photochemical
reaction process. The photochemical reaction process is mainly
investigated using smog chamber experiments and these results are
useful to understand the fundamental characterstics of the photo-
chemical reaction processes. Nevertheless, these experimental
results are not sufficient for the total understanding of the
photochemical smog phenomena in the environment. In the case of
smog chamber studies (1) pollutants are distributed uniformly (2)
light intensity is uniform (3) the effects of atmospheric
turbulence is ignored (4) wall effects are not negligible. For
these reasons, it is necessary to validate smog chamber results
using field observation data.
For this purpose Lagrangian observation is most effective.
Calvert et al. (1976) made a Lagrangian observation using two
helicopter-tracing tetroons in the Los Angeles area (LARPP).
Decker et al. (1977) traced a photochemical reaction process
using a balloon near the St. Louis area. ( DaVinci II Project )
These are the direct methods of Lagrangian analysis. In this
paper we explain an indirect Lagrangian observational system
conducted by NIES (National Institute for Environmental Studies)
covering the Tokyo metropolitan area during 1980 and 1981.
Wakamatsu et al. (1976),(1981) showed the importance of local
climatology (sea-land breeze,mountain-vally wind, and urban-rural
interaction) in understanding the photochemical smog in the Tokyo
metropolitan area. The scale of the phenomenon is approximately
100 km by 100 km with the pollutants distributed three-
dimensional ly and the time scale of the photochemical smog forma-
tion mechanism longer than one day. The aerial scale of the
polluted air mass is usually 20-50 km almost corresponding to the
aerial distribution of the industrial complexes. In this case, we
have to know the averaged three-dimensional wind field and
averaged wind trajectory which carries the polluted air mass.
For this purpose, we made a Lagrangian observational system
using two instrumented aircraft, 23 pilot-balloons, 4 radio
sondes, 100 ground level monitoring stations and two main computer
systems.
In this paper, outlines of this system and a summary of the
observation are discussed.
2. Instrumentation and Measurement System
The aircraft measurements are performed with a twin-engine
Cessna (404-TITAN) and Aero Commander (685). The air sampling
tube was set in the nose cone of the aircraft. On the roof of the
PROCEEDINGS--PAGE 351
-------
aircraft, a UV radiometer was mounted and under the floor,
sensors for air temperature and humidity are mounted . These two
aircraft had almost identical measuring instruments and flew
about two hours alternately along the air trajectory calculated
from pilot-balloon data to obtain the photochemical reaction
processes in the traced air mass. Aircraft specifications are
shown in Table 1. Instrumentation used on the aircraft are shown
in Table 2.
(1) Gas measurement
The ozone NO, and NOx monitors were specially designed for
this aircraft study. These instruments have high response and
controlled to work under the high environmental temperature.
Electric power was supplied from the aircraft DC generator and
was converted to 400 Hz AC using a rotary inverter. These
instruments works on 400 Hz 100V AC and 28 V DC current on the
aircraft and easily to change 60-50 Hz 100 V AC for ground
calibration. The S02 monitor used was commercially available
equipment.
Pressure test and temperature test were done for all before
the observation. They were calibrated in a pressure-controlled
chamber and calibration curves were drown for each instrument.
Room temperature and manifold temperature were monitored for
the correction of temperature effects on the instruments and air
pressure was monitored for the correction of pressure effects.
Air samples for the NMHC analyzer and gas chromatographic analy-
sis of hydrocarbons were collected in a 1 liter glass vessel.
This glass vessel has two Teflon valves. Sampled air was
collected and pressurized to about 1.4 atm by a Teflon bellows
pump and this vessel was connected to the NMHC analyzer after the
flight to measure NMHC concentration, and then hydrocarbon
species were measured using GC analyzer. About 12-24 samples are
obtained in one flight.
(2) Aerosol measurement
Sulfate and nitrate concentrations were measured by ion
chromatography of aerosol samples collected on Teflon filters.
Sample flow rates were measured with a mass flow meter and
recorded on magnetic tape at three-second interval.
(3) Meteorological element, altitude and position measurement
Air temperature, relative humidity, atmospheric pressure and
UV radiation were monitored as meteorological elements. Position
data was obtained using Loran-C and altitude was calculated from
the environmental pressure data.
(4) Data processing
PROCEEDINGS—PAGE 352
-------
The measured data were averaged over 3-sec intervals and
recorded on 1/2 inch magnetic data tape in real time. Data
processing system on the aircraft are shown in Table 3(system 0).
Three micro computers (Z80A) were used for this airborne system,
CPUO acquired position data from the LORAN-C, CPU! acquired air
pollution data and meteorological data and CPU2 monitored status
information of instruments and electric power supply system.
These data processing systems are easily operated by one operator.
(5) Calibration
Before the observation, temperature and pressure tests were
conducted and calibration curves were obtained for each
instrument and these data were used to the data correction.
Before and after each flight, calibration was done using the
gas phase titration technique for ozone and NO,NOx analyzer. For
the 502, analyzer zero and span values were checked before and
after the flight.
(6) Sampling air flow and electric power supply
Sample air was led through the three Teflon pipes ( 15 mm
ID.TFE Teflon ) which projected approximately 0.5 m from the nose
cone of the aircraft . The first pipe led into the manifold and
delivered the sampling air to the measuring instruments. The
second pipe led into the high-volume air sampler to the aerosol
sampling on Teflon filter. The third pipe was used for air
collection into glass vessels and sampled air was analyzed after
the flight. The length of these pipes was approximately 5 m from
nose cone. Sampling air was supplied using aircraft dynamic
pressure for suction. These sampling air flow patterns are shown
in figure 1.
Electric power to the instruments was supplied from air-
craft's DC generator. DC current was converted to AC using three
rotary inverters. To obtain high efficency, a 400 Hz,100-120 Volt
inverter was partly used. The flow of electric power supply is
shown in figure 2.
3. Observation system and analysis
Outlines of our Lagrangian observation system are shown in
Table 4. Using this system, we are able to know tthe chemical
reaction processes in the atmosphere.
An example of data are shown in figure 3. On August 6., 1980
North Easterly general wind prevailed and pollutants from the
Keihin and Keiyo industrial complexes are transported to the
Sagarni Bay Area. These data are very useful to estimate the
photochemical reaction processes because of the absence of
PROCEEDINGS-- PAGE 353
-------
additional emission source from the sea area. Using these data
Suzuki et al. (1982) estimated OH radical concentration to be
approximately 0.2-0.4 ppt
Acknowledgements
The authors wish to express their thanks to the staff of
Kimoto Electric Co. Ltd.. for valuable cooperation during the
aircraft survey. Thanks are also given to Instrumentation &
Science Co. Ltd. (ISC) for the measurement of meteorological data
and to Institute of JUSE(Union of Japanese Scientist and
Engineerers) for the help of data processing.
References
Calvert.J.G. (1976): Test of the theory of ozone generation in
Los Angeles atmosphere. Environ. Sci. Technol., 10, 248-256.
Calvert.J.G. (1976b): Hydrocarbon involvement in photochemical
smog formation in Los Angeles atmosphere. Environ. Sci.
Technol.,10, 256-262.
Decker,C.E. (1977): Ambient monitoring aloft of ozone and
precursors near and .downwind of St. Louis. EPA-450/3-009.
Suzuki,M., Wakamatsu.S.,, Uno.I. and Ogawa.Y. (1982): Evaluation
of the OH-radical Concentration in the Polluted Atmosphere.
Japan-US Conference on Photochemical Air Pollution and Air-
Pol lution-related Meteorology, December 1-2, 1982, Tsukuba,
JAPAN,
Wakamatsu,S and Qkita.T. (1976): Vertical and horizontal
distribution of ozone covering Kanagawa Prefecture, Japan.
Memoirs of the Faculty of Eng., Hokkaido Univ., XIV, 15-24.
Wakamatsu.S., Goi.K., Aburamoto.Y., Hatano.H. and Okuda.M.
(1981): Relationship between the area! distribution of
photochemical pollutants and local wind flow covering Kanto
district.(in Japanese) J. Japan Soc. Air Pollt., 16, 146-
I 3 / •
Wakamatsu.S., Ogawa,Y., Murano.K., Okuda.M., Tsuruta.H., Goi.K.
and Aburamoto.Y. (1981b): Aircraft survey of photochemical
smog in Tokyo metropolitan area.(in Japanese) J. Japan Soc.
Air. Pollut., 16, 199-214.
PROCEEDINGS—PAGE 354
-------
AIRCRAFT SPECIFICATIONS
Instrument payload
Cessna (404-TITAN)
900 kg
Aero Commander (685)
550 kg
-1
in
c
X
c
to
Available instrument power
4.2 kVA at 28 VDC
2.7 kVA at 100 VAC
9.8 kVA at 28 VDC
6.3 kVA at 100 VAC
Sampling speed
Navigation system
85 m/s
LORAN-C, VOR, DME
85 m/s
LORAN-C, VOR
2
P>
r»-
o'
o>
sr
(A
O
n
M
O
•0
>
O
Table 1 Aircraft specifications
p
•o
p
9
m
<
-I
o
9
3
m
3
i
9
0)
> y
IS B
9 °r
O (»
•< w
-------
National Institute for Environmental Studies
Japan Environment Agency
Telephone : 0298-51-6111 Cibl« : KOGAIKENTSUKUBA
P. 0. Yatabe Tiukub* Ibiraki 300-21 J.p.n
ParameCer
Ozone
N 0
N 0 X
S 0 2
Condensation
Nuclei
Aerosol Size
Distribution
Ambient
Temperature
Ambient
Humidity
Ultraviolet
Radiation
Pressure
(Altitude)
Position
Pitching and
Rolling
Analysis
Technique
Chemi lumines cence
Chemi luminescence
Chemiluminescence
Fluorescence
Light- attenuation
Light-scattering
Platinum
Resistanse
Electronic
Capacity
Photocell
Bellows
Barometer
Loran-C
Gyro compass
Manufacture
and Model
KIMOTO
MCSAM-F
KIMOTO
MCSAM-F
KIMOTO
MCSAM-F
Monitor Labs
88SO
E/one
Rich 100
Royco
226
Deggussa
Measurement
Ranges
1-2000 ppb
( dynamic )
1-2000 ppb
( dynamic )
1-2000 ppb
( dynamic )
2-500 ppb
100k CN/cc
1.0-4. Sum dia.
-50.0-50:0'C
National 20-90 I RH
Weather Service
Eppley
UV Radiometer
Tokyo Koku
Keiki ATP-20-1
FURUNO
LC-30
Tokyo Koku
Keiki 230
0-5 mW/cm2
760-380 mmKg
(0-5400 gpm)
P + 15 deg
R f 90 deg
Time Resoonse
< 3 s
( 90 7. )
< 3 s
( 90 7. )
< 3 s
(90 7. )
5 s
( 90 7. )
3 s
( 90 7. )
3 min
( periodic )
1 s
( 90 7. )
3 s
( 90 7. )
2 s
( 90 % )
2 s
( 90 I )
2 s
( periodic )
Approximate
Resolution
1 ppb
1 ppb
1 ppb
2 ppb
1000 CN/cc
0.1 deg
3 7.
0.002 mW/cm2
1 tnmHg
(10 gpm)
0.03 min
(50 m>
P 0.1 deg
R 0.5 deg
Parameter
Sulfate
Nitrate
non-Methan
Hydrocarbon
Sampling Manufacture
Technique and Model
HighvtJlume Sampling KIMOTO 191
on Teflon Filter
Compressed Sampling
in Glass Vessel
Analysis
Technique
Ion - chroma t ogr aph
Gas - chroma t ogr aph
Manufacture
and Model
Dionex 10
Simazu GC-
Approximate
Resolution
Table 2 Instrumentation used on the aircraft
PROCEEDINGS—PAGE 356
-------
Wr
National Institute for Environmental Studies
Japan Environment Agency
: 0298-51-6111 Cable : KOCAIKENTSUKUBA
P. 0. Yit.b« Tsukub. Ib.r.ki 300-21 J.p.i,
AIRBORNE SYSTEM (SYSTEM 0)
GROUND SYSTEM (SYSTEM 1)
LORAN-C POSITION DATA
CONCENTRATION SIGNAL
OZONE
N 0 X
S 0 2
CONDESS. NUCLEI
METEOROLOGICAL DATA SIGNAL
AMBIENT TEMPERATURE
AMBIENT HUMIDITY
ABSOLUTE PRESSURE
ROOM TEMPERATURE SIGNAL
MANIFOLD TEMPERATURE SIGNAL
TEFLON FILTER SAMPLER
FLOW RATE SIGNAL
K C SAMPLING POSITION SIGNAL
GYRO COMPASS DATA SIGNAL
AEROSOL SIZE DATA
INSTRUMENTS CONDITION
DATA SIGNAL
REGULATED POUER VOLTAGE
TEMPERATURE CONTROL
POWER SUPPLY CONDITION
DATA SIGNAL
VOLTAGE
CURRENT
FREQUENCY
AEROSOL SAMPLE ON TEFLOM FILTER
H C SAMPLE IN GLASS VESSEL
=4 c P u - o ] 1
\
CRT TERMINAL
(TRACK PLOTTING)
=-j A/D CONVERTER
|,
C P U - 1 I 1
'
CRT TERMINAL
(REAL TIME MONITOR) j— — —
MAGNETIC TAPE DECK \Z 1
i_ _ _ r _
1 r
i i
i i
1 i
, i
i 1
^ DIGITAL PRINTER , |
») A/D CONVERTER ' 1
1- |
C P U - 2 1
I '
(REAL TIME MONITOR)
1 1
' I
1 i
I
_ __ — — __ —J
LINE PRINTER |
X-Y PLOTTER |
CRT TERMINAL |
I
CPU "°
I
MAGNETIC DISK |
_ _ ^>T
ION CHROMATOGRAPH
non-CHi IIC ANALYZER
CAS CHROMATOCRAPH
GROUND SYSTEM (SYSTEM 2)
OFF LINE (AFTER LANDING)
Table 3 Data processing system
PROCEEDINGS PAGE 357
-------
Aircraft Data
|System 0|
( 5 min Af
^
AC685
C-404
ter Landing)
Sonde Data
r i
Monit
t i
Meteorological Data
Pilot Balloon Data ( 23 points)
( 4 points)
Monitoring Station Data
(100 points)
Fax. Data (JMA)
Upper Data Observatory
[system 1 | System 2
Out put Ground Level Weather Map
Upper Weather Map
Vertical Temperature Profile
System 1
Vertical and Horizontal Air Pollution Map
Flight Course and Altitude
(30 min After Landing)
System 2
Vertical and Horizontal Wind Distribution
Vertical Temperature Profile (lower than 2000m)
Upper Air Flow Pattern
Ground Level Air Flow Pattern
Ground Level Air Pollution Pattern
(40 min. After Landing)
I
I Decision Making j
Flight Course for Next Flight
Table 4 Outlines of a Lagrangian observation system
PROCEEDINGS PAGE 358
-------
SJtfr
National Institute for Environmental Studies
Japan Environment Agency
-- NOSE CONE --
teflon id: 15tnm 1:5000mm
SAMPLE
MR
teflon id:15mm 1:5000mm
teflon id: 15mm 1: 5000mm
K\A—' (
GLASS
IXt VESSEL
TEFLON
VALVE
TEFLON
VALVE
-- UNDER THE FLOOR --
STATIC
PRESSURE
PORT
GLASS
MANIFOLD
O
o
UV RADIOMETER
--ON THE ROOF —
r
/
f
id: 4mm
1: 750mm
id: 2mm
1: 750mm
teflon
id: 2mm
l:750mm
teflon
id: 4mm
l:750mm
vinil
id : 8nnn
1: 1500mm
vinil
id : otnm
1: 1500mm
O7HMP1
MONITOR
N 0
MONITOR
N 0 X
MONITOR
S 0 2
MONITOR
CONDENS
NUCLEI
MONITOR
AEROSOL
SIZE
MONITOR
MANIFOLD
TEMPERATURE
MONITOR
Figure 1 Sampling air flow pattern on the aircraft
PROCEEDINGS—PAGE 359
-------
«*ftff
National Institute for Environmental Studies
Japan Environment Agency
Telephone : 0298.51-6111 Cable : KOGAIKENTSUKUBA
P. 0. YaUbe Tsukubi Ibarild 300-21 Japan
ROTARY
INVERTER
400 Hz
100 VAC
r
i
\i v
POWER
CONTROLLER
ENGINE DRIVEN
POWER SUPPLY
28 VDC
BATTERY
28 VDC
ROTARY
INVERTER
400 Hz
100 VAC
~I
/ M
POWER
CONTROLLER
50/GO Hz
100 VAC
f V
POWER
CONTROLLER
NAVIGATION
INSTRUMENTS
(LORAN-C, etc.)
1 50/60 Hz
1 100 VAC
GROUND
1
I POWER SUPPLY i
(MEASURING INSTRUMENTS)
Figure 2 Electric power supply on tfie aircraft
PROCEEDINGS—PAGE 360
-------
-
"
Figure 3. Horizontal distribution of 03 and N02at the
altitude of 350-400 m. on the Sagami Bay. Arrows in the figure
shows a wind direction.
-------
C Ref. 3
Distribution of Photochemical Pollutants and their
Three-Dimensional Behavior covering
the Tokyo Metropolitan Area
Itsushi Uno, Shinji Wakamatsu, Makoto Suzuki
and Yasushi Ogawa
National Institute for Environmental Studies
^
P.O. Yatabe, Tsukuba, Ibaraki 305,
JAPAN
PROCEEDINGS—PAGE 363
-------
ABSTRACT
The spatial distribution and transport process of photochemical
pollutants covering the Tokyo Metropolitan Area in Japan were
investigated from 31 July to 2 August 1979 using instrumented aircraft.
In the experiment, the vertical profiles of pollutants were observed
using four instrumented aircraft. This paper mainly considers the
transport process of the polluted air mass using three-dimensional
trajectory analysis. In this trajectory analysis, we determined the
hourly wind field by objective analysis techniques from pilot-balloon
observation data.
In the Tokyo Metropolitan Area, the sea-land breeze circulation is
an important factor in the photochemical oxidant formation inland, when
the geostrophic wind is weak. The nighttime radiation inversion observed
in the early morning prevents the mixing of primary pollutant emitted
from the big coastal industrial zones around Tokyo Bay. These
pollutants were then advected to the Sagami Bay area by the land breeze
and the Bay area acts as storage tank for the pollutants. These
pollutants were then converted to secondary pollutants resulting in a
high ozone air mass inland with the penetration of the sea breeze. The
sea breeze layer is thermally stable and inhibts vertical mixing of NOx.
On the other hand, at the front of sea breeze zone, a highly turbulent
area transports the NOx to 1000-1500 m above mean sea level.
Polluted ozone air masses, whose concentration exceeded 100 ppb,
were observed at 500-1000 m on both 31 July and 1 August 1979. It was
observed that the maximum ozone concentration on the second day exceeded
that on the first day. These ozone air masses contained aged pollutant
and they were entrained into the mixing layer in accordance with the
elevation of the mixing layer. This accelerated the formation rate of
secondary pollutants. These early morning, high ozone concentrations and
the storage capacity of the Sagami Bay area are important factors in the
time scale of air pollution phenomena in this region.
PROCEEDINGS--PAGE 365
-------
1. Introduction
Since 1970, complaints of eye irritaton, sore throat and the
sensation of suffocation due to the photochemical smog have been
reported every summer in Japan. Although the number of the complaints
have been decreasing, their number reached more than 1,400 in 1980.
Almost all of these complaints came from the vicinity of large cities,
such as Tokyo and Osaka. The need to understand the mechanisms of the
photochemcal smog phenomena in these areas is urgent. Since the areas
are so large, aircraft observation is necessary to this understanding.
A series of aircraft observations were conducted in the Los Angeles
Air Basin to understand the photochemical smog behavior (Edinger et al.,
1972; Edinger, 1973; Gloria et al., 1974; Husar et al., 1977; Blumenthal
et al., 1978). For another area, Lyons and Cole(1976) and Keen and
Lyons(1978) investigated the relation between the land-lake breeze at
Lake Michigan and the vertical distribution of aerosols, and showed that
there was the possibility of re-circulation of pollutants under the
land-lake breeze condition. Sexton and Westberg(1980) observed the
region downwind of the Chicago-Gary urban complex using aircraft and
showed that the transport of pollutants from the Chicago urban complex
is the main reason for local high ozone levels. From these investigat-
ions, the following has been made clear : 1) there exists a strong
relation between the vertical ozone concentration profile and the
temperature inversion height (lid), 2) there is a possibility that the
layer which contains rich ozone moves along with the terrain or slope,
3) ozone above the lid(i.e., aged ozone) remains a long time compared
with that below the lid and affects the higher ozone formation into the
next day, 4) a strong vertical and horizontal gradient exists in the
concentrations of both primary and secondary pollutants, 5) the sea-
breeze is an important factor for the transport and spatial distribution
of high concentration pollutant profiles.
These observations are mostly restricted to the qualitative under-
standing of pollution. As mentioned above, the transport process of
photochmical pollutants is very complicated under the local wind circu-
lation system, so it is important to study the relationship between the
behavior of the polluted air mass and the three-dimensional wind system.
National Institute for Environmental Studies (NIES) has been conducting
PROCEEDINGS PAGE 366
-------
summer aircraft surveys since 1978 to understand the photochemical
pollutant formation and transport process in the Tokyo Metropolitan
Area(Wakamatsu et a!., 1982a). To follow the three-dimensional behavior
of the polluted air mass, pilot-balloon observation was also conducted
at 19 points covering the Tokyo Metropolitan Area beginning 1979.
Tokyo Metropolitan Area is located in the Kanto Plain, which has a
complicated topography. The Kanto Plain has a horizontal scale of about
100 km in the east-west direction and 200 km in the north-south
direction. Figure 1 shows a perspective view of the Kanto Plain. Figure
2 shows the industrial zone and other topographical information. The
topographical characteristics of the Kanto district are as follows: the
western and northern sides are walled by mountains 1000 to 2000 m high,
the Tokyo and Sagami Bays and the Pacific Ocean are located at the
center, south and east sides of the plain, respectively. On the
southwest side of Sagami Bay, the Izu Peninsula protrudes into the
Pacific Ocean, and this acts as the wall to stop the transport of
pollutants from two industrial zones. Two industrial areas are around
Tokyo Bay, i.e., the Keihin industrial area(the west side of Tokyo Bay)
and the Keiyo industrial area(the east side of Tokyo Bay). A detailed
emission map of NOx can be found in Wakamatsu et al.(1982a).
In this paper, we are concerned with the vertical behavior of
pollutants based on the three-dimensional wind field. To do this,
trajectory analysis is useful. Angel 1 et al.(1972,1973) used tetroons to
determine the trajectory over the Los Angeles Basin, and also compared
these to the indirect trajectory calculated from surface wind data. They
showed that the both methods agreed in their determination of the main
paths. In the Los Angeles Reactive Pollutant Program(LARPP), the three-
dimensional trajectory was obtained by using three tetroons( Feigley and
Jeffries, 1979). In the Tokyo Metropolitan Area, because of the
compilicated topography and the restrictions from air-traffic control,
it is difficult to use tetroons for measurement, so we determined the
three dimensional trajectory from the wind field calculated by an
objective analysis procedure. In the following section, we apply, the
objective analysis technique to the wind data obtained by pilot-balloon
observation. Using hourly wind field data, we analyze the vertical
distribution of pollutants and the spatial behavior of polluted air
mass.
PROCEEDINGS--PAGE 367
-------
2. Observation data
Aircraft observation was conducted over 3 days from 31 July to 2
August 1979. The four instrumented aircraft were : three twin engine
Cessna ( C-402 x 2 , C-404 ) and a twin engine Rockwell Aero Commander
(AC-685). Each of the four aircraft flew at a fixed altitude. Four or
five runs were performed on each day and the four aircraft measured the
vertical cross section at one time, except the runs in the evening and
at midnight. In total, 13 runs were conducted. Measured were NO, NOx,
ozone, condensation nuclei(CM), hydrocarbons, UV intensity, pressure
altitude, temperature, relative humidity and position (Loran-C). Table 1
shows the instrumentation used and the main flight heights. Figure 2
shows the daily flight patterns. The measured data were averaged over 4
sec intervals and recorded on digital cassette tapes. The measurement
system is fully described elsewhere (Wakamatsu et a!., 1982b). Each run
was labeled Run UK, where the first digit I indicates the day, J is run
number in each day, and K means the aircraft identification which
corresponds to the operation height. Occasionally, a run was called Run
IJ for short to indicate the four flights(Run Ul - 104} made at one
time. The daily flight paths were as follows: in the early morning(Run
II) the path was A-B-C-D-E-F-G-H, near noon(Run 12) E-F-G-H-I-J-K, in
the afternoon(Run 13) E-F-M-K-L-G-H, in the evening(Run 14) H-G, and in
the midnight is H-G (See Figure 2 for these paths).
Vertical .wind profiles were observed at 19 points using a pilot-
balloons by the single theodolite method. Data was collected at each 100
m interval above the terrain up to the 3000 m level (Figure 2 shows the
points). Also, at 3 additional points, temperature and humidity profiles
were measured using the radiosonde at about three hour intervals.
To compare the flight data, we also took into account the hourly
surface air quality monitoring stations data ( NO, NOx, oxidants, wind
speed and wind direction etc.) during the period of the aircraft survey.
The ground stations number is about 150 and they are located throughout
the Kanto district to follow the behavior of ground level smog
phenomena.
PROCEEDINGS—PAGE 368
-------
3. Vertical distribution and behavior of pollutants
a. Determination of the three-dimensional wind field
The air pollution phenomena in the Tokyo Metropolitan Area cover
100 - 200 km area in horizontal scale and continues for several days(
Wakamatsu et al., 1982a). To study these phenomena, it is necessary to
understand the transport process of the pollutants. When vertical air
motion exists due to the convectional motion or topographical effects,
this will become an important factor in the transport process of the
pollutant (e.g., see Liu and Seinfeld, 1975)
In general, only the horizontal components of the wind are
measured. In addition, the observation points are generally distributed
irregularly and sparsely in the region of interest. To use these data,
objective analysis is necessary to get the mass-consistent wind field.
According to Goodin et al.(1980), the objective analysis of the wind
field is defined by a two-step processes. First, interpolation of
sparse and discrete measured data to finer mesh data(interpolation
step). Second, adjustment of the wind vector at each grid point so as to
satisfy an appropriate physical constraint (adjustment step).
In this study, we adopt the MATHEW method to get the three-
dimensional wind field. The MATHEW method was proposed by Sherman(1978)
based on Sasaki's variational method(Sasaki 1958, 1970). In our
application, we modified the vertical coordinate system to a terrain-
following one. Using this modified MATHEW method, we could reduce the
total divergence to less than one-tenth of the initial value. In the
following section, we use this wind field for the detailed analysis of
transport process of pollutants.
b. Observation Results
In this section, we show the trajectories and vertical distribution
of pollutants and study the transport process of the pollutants. We
mainly consider the flights from Run 11(early morning flight of 31 July)
to Run 21(early morning flight of 1 August) when the local wind
PROCEEDINGS — PAGE 369
-------
circulation dominated in the Kanto Plain. Figure 3 shows the geostrophic
wind at the 800 mb and 700 mb levels calculated using aerological data
from the Japan Meteorological Agency (JMA). This shows that the air
pressure gradients were relatively stagnant and wind speed was small on
31 July. Figure 4 shows the diurnal change pattern of wind field at the
350 m height observed from pilot-balloon measurements. In this figure,
the typical wind patterns are shown, that is, in the early morning, the
land breeze(N or NE direction) covers the whole Kanto plain; in the
middle morning the sea breezes from Sagami and Tokyo Bays are detected
in the coastal area but in the remaining area the land breezes
dominated; in the afternoon the sea breeze from Sagami Bay covers the
whole plain and this continues until the late evening; and at midnight,
a mountain valley wind and land breeze start and the penetration of the
sea breeze decreases. These wind circulation systems complicate the
transport process of photochemical pollutants.
The vertical crosssection of pollutants and meteorological elements
were observed in several sections(e.g., AB.CD and so on in Figure 2) and
complete representation of all this data is outside the scope of this
paper. The sea breeze from Sagami Bay has a great effect on the
transport process and is a south wind. In addition, in the evening and
at midnight, only section H-G was observed by a single aircraft(AC-685),
so the south-north crosssection is sufficient to show the details of the
pollutant behavior, and we analyzed the section H-G in the following
analysis.
In the early morning of 31 July, to observe the initial
distribution of photochemical pollutants, the first observation Run 11
was conducted from 0430 to 0630JST. Figure 5 shows a portion of the
results(vertical crosssection H-G). The Keihin industrial complex and
the Tokyo Metropolitan Area are located in the southern part of section
H-G (area covering 0 Km to 40 Km from point H in Figure 5) and in the
northern part of this section, no significant NOx emission sources
exist. In this figure, the temperature profile at Otemachi(point 6 in
Figure 2) is also shown. Otemachi is on the east side(15 Km) of the
center of section H-G and this data "point represents the typical 'urban
area temperature profile. The ground pollutant level was determined from
the ground monitoring stations. Note that oxidant was treated as ozone
in the following analysis. The operation heights were at four fixed
PROCEEDINGS—PAGE 370
-------
levels, so that an isopleth line of pollutant concentration can be drawn
by interpolation. The significant temperature inversion layer(lid) was
observed about 500 m height at Otemachi. The well-aged ozone layer (the
area where the ozone concentration exceeded 100 ppb for about 40 Km )
was above the lid and below it , the primary pollutant(N02 ) was
trapped(Figure 5). A strong vertical gradient of the pollutants existed
at the lid layer. Also, a horizontal gradient of pollutants existed at
the industrial area(0-60 Km area in section H-G) and in the rural
area(60Km - G). Under the lid, the N02 and ozone were inversely related.
The N02 had a maximum at the 350 m height and this was due to the
emission from the stacks of the industrial zone. In Figure 6, we show
the prevailing wind direction at 0600JST at the 350 m height and the
shaded area indicates that the ozone concentration was less than 10 ppb
at the 350 m height. The shaded area was estimated and extrapolated from
Run 111(lowest operation height data). The spread of these zones exactly
coincided with the downwind areas of the Keihin and Keiyo coastal
industrial complexes. The ozone precursors are consistenly transported
to the Sagami Bay area along the NE wind(land breeze) and this' brings
the high ozone concentration inland with the penetration of the sea
breeze. The Sagami Bay area plays the part of a storage tank for the
ozone precursors. In Figure 6, the Trajectory A which started from
0600JST at the 350 m height is shown. The trajectory was integrated
using the three-dimensional wind field data obtained by the modified
MATHEW method. In the trajectory analysis, horizontal diffusion is
neglected, so to estimate the mean path of the polluted air mass, the
calculation area was considered as to be a circle of about 10 Km in
radius at start point. The circle line of the trajectory at every hour
was distorted by the horizontal wind shear. Trajectory A started at
0600JST coincided with the behavior of pollutants measured in Run
111(the early morning observation). Trajectory A shows the re-
circulation of pollutants from the sea area to the inland area. The
vertical change of elevation in Trajectory A is not so very significant.
Figure 7 shows the surface sea breeze front from 1000JST to 1700JST and
the surface oxidant isopleth at 1200, 1300 and 1500JST. On the west side
of the Tokyo Metropolitan Area, the oxidant concentration increased
rapidly in the afternoon, and its increase exactly coincided with the
penetration of the sea breeze front. The Trajectory A in Figure 6
•indicates that the early morning's primary pollutant around Tokyo Bay
area is the main source of the afternoon's high oxidant pollution.
PROCEEDINGS—PAGE 371
-------
From 1000JST, the sea breeze was observed at Miura(point 10 in
Figure 2) and Chigasaki(point 9), but still the N or NE wind dominated
the whole plain(Figure 4). Wind flow at the 350 m level showed no
difference from the early morning observation. So until this time, the
primary pollutants of two industrial zone were transported into the
Sagami Bay area. Run 12 was performed from 0945 to 1130JST(31 July) to
observe the middle stage concentration of the photochemical reaction.
Figure 8 shows a portion of the results of Run 12(section H-6: N02 ,
ozone and potential temperature profile). A lid was observed at about
the 1,100 m level at Otemachi. The lid rose about 600 m compared with
the level at 0600JST and following this, the high N02 area ascended due
to mixing. Two high ozone concentration layers were observed at 300 m
and 900 m heights. In particular, the upper ozone layer is considered to
be an aged layer from trajectory analysis(not shown in the figure)
which was entrained into the mixing layer in accordance with the elevat-
ion of the lid. The ozone at the lower elevation is considered as a
fresh ozone. This shows that aged and fresh pollutants are mixed in the
same mixing layer. The potential temperature profile has a relatively
high value above the Tokyo Metropolitan Area(10-60 Km area) and the
potential temperature has a convex form. The N02 profile also has a
convex shape and it coincides with the vertical diffusion pattern of
N02. In Figure 5, Trajectory B started from 1000JST at the 350 m height
and shows the backward and forward path. This area is just above Yoko-
hama(point 7 in Figure 2) and above this area, the highest N02 (exceed-
ing 60 ppb) was observed in this flight. Trajectory B (backward direct-
ion) shows that this air mass has its origin in the Keihin industrial
zone. So this high N02 area is considered to be the plume from this
industrial zone. The vertical section of Trajectory B in Figure 6
indicates that the height of this marked air mass changed rapidly. At
the front of sea breeze zone, a small updraft wind exists(Simpson et
al., 1978, Mitsumoto et al., 1982), and from 1000JST in the coastal
area, the sea breeze began to be observed. So the sudden change of
height of Trajectory B was due to the updraft wind of the sea breeze
front.
In the afternoon, the sea breeze from Sagami Bay covered the Kanto
plain(see Figure 4), and it brought the polluted air mass to the inland
area. Figure 9 shows a portion of the results from Run 13 performed
PROCEEDINGS—PAGE 372
-------
from 1450 to 1630 JST. The period of this observation(Run 13) corre-
sponded to the maximum concentration stage of the photochemical
reaction. From the potential temperature profile depicted by the dashed
line in Figure 9, the sea breeze layer is easily distinguished. The
southern area(0-50 Km from point H) has low potential temperature and
is considered to be the sea breeze layer which has a relatively low
temperature, while the northern area is considered to be a well-heated
and the highly turbulent area. In the sea breeze layer, N02 is not
diffused vertically, on the other hand, in the well-heated area, N02 is
transported to more than 1,000 m, which shows that the sea breeze layer
became more turbulent as it penetrated inland due to thermal convection
and became the cause of high ozone formation in the high altitue layer.
In Figure 9, the vertical wind profile(v and w components) at 1500JST
obtained by modified MATHEW method are also shown. From this figure, the
v component of the wind dominated the whole vertical crosssection. A
small updraft wind was detected in the inland area. The distribution of
pollutants is very interesting. At the 350 m height, there existed two
maximum ozone air masses(10-20 Km area C'and 70-80 Km area D in Figure
9). Above the sea breeze layer(south area) there existed a relatively
clean air mass whose ozone concentration was less than 100 ppb. It is
considered that there the border of the polluted air mass exists near
the 10 ppb isopleth of N02. This explains that the borderline of the sea
breeze layer ascended with penetration into the inland area. At about
the 1,000 m height, high ozone air masses were also detected and
belonging in the mixing layer. The highest ozone concentration
(exceeding 350 ppb) measured during airborne survey was observed in this
run. This high ozone concenration area is labeled as C in Figure 10,
which is just to the west side of C1 area in Figure 9. The horizontal
distance between C and C1 is about 10 Km, so the transport pattern is
believed to be similar. For the analysis of the transport process of the
polluted air mass at 350 m, we selected the two air masses C and D.
Figure 10 shows the results of this trajectory analysis. Following this,
there exist two types of high ozone concentration air mass. One has an
indirect route, that is, its origin is the early morning emissions from
the Tokyo Bay coastal industrial zone, which are transported to the
Sagami Bay area by the land breeze and penetrate inland due to the sea
breeze. The other is almost directly transported from the two industrial
zones(Trajectory D) and is well-mixed vertically by thermal convection.
These two types penetration pattern are the main reason for the separat-
PROCEEDINGS—PAGE 373
-------
ion of the polluted ozone air mass. The highest ozone concentration was
observed in the former layer, and it most likely that the long photo-
chemical reaction time contributed strongly.
From evening to midnight, the sea breeze from Sagami Bay covered
the whole Kanto district(Figure 4) and the polluted air masses were
carried inland. The result of Run 14, which was performed in the late
evening(1810-1950JST), showed that the relatively clean air mass from
Sagami Bay was transported inland(not shown in this paper). In Figure
10, Trajectory E, started at 1830JST at the 350 m height, shows the
penetration of pollutants from the Keihin coastal industrial zone in the
evening. The primary pollutants were advected to the middle of the Kanto
area. This phenomenon continued until the land breeze began at midnight.
The sea breeze from Sagami Bay continued until about 0100JST (1
August) and then the land breeze started gradually down to the
southern area. The air pressure gradient became stronger than the
previous day and a strong SW wind zone was formed near the line
connecting the Izu and Boso Peninsulas. A line of wind discontinuity was
formed on this line. Run 20 was conducted from midnight( 0000-0145JST ,
1 August), and its results showed that the surface pollutants were
trapped and diffused slightly to a higher elevation (Figure 11). Sonde
observation was not performed at midnight. These high N02 areas(40-80
Km) are from the Keihin and Keiyo industrial complexes, and the high
ozone concentration layer just covered the N02. The rich ozone
concentration decreased with the dark reaction with NO. The location of
the strong gradients of both pollutants are at that rich N02 area. This
rich ozone layer remained until the next morning and contributed to the
next day's high concentration of pollutants. The vertical wind field at.
0100JST 1 August is also shown in Figure 11 with an ozone concentration
map. The flow pattern is the same as Figure 9, but in the northern
area, the southerly wind decreased.
The early morning flight Run 21 was conducted from 0445 to 0620JST,
1 August 1979. The flight path was the same as Run 11. Figure 12 shows a
portion of the result of Run 21. Tjie temperature profile at Otemachi at
0600JST indicated that a weak temperature inversion was located at the
1,000 m height, and the N02 distribution(5 ppb line) is coincident at
this height. The weak and the strong wind zones are distinguished(dashed
PROCEEDINGS—PAGE 374
-------
line in Figure 12) and the vertical profiles of ozone are separated in
this discontinuity zone(Figure 12). The vertical wind field at 0500JST,
1 August was overlayed on the ozone concentration map. The wind field is
very interesting. Near the discontinuity line, a small downdraft wind
was detected and this wind would entrain the high ozone air mass from
the high altitude to a lower one. The high ozone concentration zone at
the 700 to 800 m level is the aged air mass, which was transported from
the north by the land breeze; the polluted air was sheared near the
strong wind speed area and blown out to the NE direction. The maximum
concentration in the aged ozone exceeds 140 ppb and this value is higher
than the previous day's early morning observation result(Run 11 in
Figure 5). The concentration gradients became stronger. These results
indicate that the secondary pollutant concentration level becomes higher
than the previous day's maximum under the stagnant pressure condition.
On the other hand, the N02 maximum in Run 21 was less than the previous
day's observation. One reason for this is that on the second day, the
wind direction was different so that the emission sources, such as
Keihin industrial complex was downwind. The second reason is that lid
height was much higher on the second day allowing more mixing.
4. Conclusion and discussion
An aircraft investigation was conducted from 31 July to 2 August
1979 covering the Tokyo Metropolitan Area in Japan. Vertical cross
sections of pollutants and wind field were constructed and the
transport process of pollutant was studied.
1) From three-dimensional trajectory analysis, it is clear that the
complicated air flow pattern due to the local wind circulation is very
important in understanding the photochemcial smog phenomena in the Tokyo
Metropolitan Area. In particular, the early morning night time radiation
inversion inhibits the vertical mixing of primary pollutants and forms a
high NO concentration air mass, which is transported to the Sagami Bay
area along with the land breeze. These polluted air massses are
converted to secondary pollutants and transported inland with the
penetration of the sea breeze(Figure 6,10).
PROCEED INGS--PAGE 375
-------
2) There exists a strong relation between the penetration pattern
of the sea breeze and the photochemical oxidant level on the ground.
Blumenthal et al.(1978) reported that the sea breeze transported the
clean air inland in Los Angeles, but in the Tokyo Metropolitan Area,
Sagami Bay plays an important role in the storage of primary pollutants,
which are later converted to secondary pollutants and this leads to the
high ozone concentration in the inland area. These continue to bring
high ozone concentration inland until the late night. On the other hand,
relatively clean air penetrates inland after the midnight. This means
that the pollutants stored in Sagami Bay are finally removed.
3) In the early morning in the upper air(500-l,000 m height), an
aged ozone layer, whose concentration exceeds 100 ppb, was observed on
both days of observation(Figure 5,12). The second day's maximum
concentration was higher than the first day's value. From the trajectory
analysis, potential temperature profile and sonde observation, the aged
ozone air mass was entrained into the mixing layer in accordance with
the elevation of the lid (Figure 8). This effect is believed to
accelerate the formation of secondary pollutants during the next day.
4) From the potential temperature profile(Figure 10), it is easy to
distinguish the marine layer. The marine layer is thermally stable. On
the other hand, the air mass ahead of the sea breeze front was turbulent
and this layer transported the N02 to the 1,000 m level. This is one
reason for the high level ozone layer. Edinger et al.(1972) showed that
the ozone layer ascends along a slope and breaks through the lid to
create a high ozone layer above the lid. However, our results indicate
another possibilty for the high levels of ozone. The difference of the
density between the marine layer and the surrounding also contributes to
these phenomena.
5) The higher concentration of photochemical ozone were observed in
the upper air(300-500 m height, i.e., the sea breeze layer) rather than
at the ground. In general, except in the midnight and early morning, the
oxidant level on the ground was proportional to that in the upper layer.
Because of the restriction of air-traffic control, we could not measure
below 300 m; however, we car> state qualitatively that the upper layer's
pollutants affect the levels of ground level pollutant concentrations.
PROCEEDINGS—PAGE 376
-------
6) We could follow the behavior of the polluted air mass on an
urban scale by trajectory analysis. In our next project, the construct-
ion of a simulation model of the Tokyo Metropolitan Area, the following
must be taken into account :
i) to simulate the photochemical smog phenomena on a scale
size of the Kanto Plain, it is necessary to take into account
such features, as Sagami Bay as a tank and has a time delay
effect. Also, the treatment of the local wind circulations are
important to the model.
ii) to predict or control the high ozone level in the
afternoon, the early morning's meteorological conditions
are especially important. For example, the existence of a night-
time radiation inversion is an important factor.
Acknowledgements
The authors wish to express their thanks to K.Goi in the Saitama
Institute of Environmental Pollution, Y.Aburamoto in the Toyama Environ-
mental Center, and H.Tsuruta in the Yokohama Environment Research
Institute for their cooperation in the aircraft field study. Thanks are
also given to the staff of Kimoto Electric. Co., Ltd. for their valuable
cooperation during the aircraft study.
PROCEEDINGS—PAGE 377
-------
REFERENCES
Angell.J.K., Pack.D.H., Machta,L., Dickson.C.R. and Hoecker.W.H. (1972):
Three-dimensional air trajectories determined from tetroon flights
in the planetary boundary layer of the Los Angeles Basin. J. Appl.
Meteorol., 11, 451-471.
Angell.J.K., Hoecker.W.H., Dickson.C.R. and Pack.D.H. (1973): Urban
influence on a strong daytime air flow as determined from tetroon
flights. J. Appl. Meteorol., 12, 924-936.
Blumenthal.D.L., White, W.H. and Smith,!.B. (1978) Anatomy of a smog
episode: Pollutant transport in the daytime sea breeze regime.
Atmos. Environ., 12, 839-907
Edinger.J.G. (1973): Vertical distribution of photochemical smog in the
Los Angeles basin. Environ. Sci. Technol., 7, 247-252.
Edinger.J.G., McCutchan.M.H., Miller,P.R., Ryan,B.C., Schroeder,M.J. and
Behar.J.V. (1972): Penetration and duration of oxidant air
pollution in the south coast air basin of California. J. Air
Pollut. Control Assoc. 22, 882-886
Feigley.C.E. and Jeffries,H.E. (1979): Analysis of processes affecting
oxidant and precursors in the Los Angeles reactive pollutant
program(LARPP) Operation 33. Atmos. Environ., 13, 1369-1384.
Gloria,H.R., Bradburn,G., Reinisch.R.F., Pitts,Jr.,J.N., Behar.J.V. and
Zafonte.L. (1974): Airborne survey of major air basins in
California. J. Air Pollut. Control Assoc., 24, 645-652.
Goodin,W.R., McRae.G.J. and Seinfeld,J.H. (1980): An objective analysis
tehnique for constructing three-dimensonal urban-scale wind field.
J. Appl. Meteorol. 19, 98-108
Husar.R.B., Patterson,D.E., Blumenthal,D.L., Warren,W.H, White,W.H. and
Smith,T.B. (1977): Three-dimensional distribution of air pollutants
in the Los Angeles basin. J. Appl. Meteorol., 16, 1089-1096.
PROCEEDINGS—PAGE 378
-------
Keen.C.S. and Lyons,W.A. (1978): Lake/Land breeze circulations on the
western shore of Lake Michigan. J. Appl. Meteorol., 17, 1843-1855.
Liu.C.Y. and Seinfeld,J.H. (1975): On the validity of grid and trajec-
tory model of urban air pollution. Atmos, Environ.,9, 555-574.
Lyons,W.A. and Cole.H.S. (1976): Photochemical oxidant transport:
Mesoscale lake breeze and synoptic-scale aspects. J. Appl.
Meteorol., 15, 733-743.
Mitsumoto,S., Ueda.H. and Ozoe,H. (1982): A laboratory experiment on the
dynamics of land and sea breeze. To appear in J. Atmos. Sci.
Sasaki,Y. (1958): An objective analysis based on the variational method.
J. Meteorol. Soc. Japan, 36, 77-88.
Sasaki,Y. (1970): Numerical variational analysis under the constraints
as determined by longwave equations and lowpass filter. Mon. Wea.
Rev., 98, 875-883.
Sexton,K. and Westberg.H. (1980): Elevated ozone concentrations measures
downwind of the Chicago-Gary urban complex. J. Air Pollut. Control
Assoc., 30, 911-914.
Sherman,C.A. (1978): A mass-consistent model for wind fields over
complex terrain, J. Appl. Meteorol., 17, 312-319
Simpson,O.E., Mansfield,D.A. and Milford.J.R. (1977): Inland penetration
of sea breeze front. Quart. J. R. Meteorol. Soc., 103, 47-76.
Wakamatsu,S., Ogawa,Y., Murano,K., Goi,K. and Aburamoto,Y. (1982a):
Aircrafts survey of the secondary photochemical pollutants covering
the Tokyo Metropolitan Area. To appear in Atmos. Environ.
Wakamatsu,S. Uno,I., Suzukis,M. and Ogawa,Y. (1982b): The Lagragian
observation of polluted air masses using aircrafts. Japan-US Joint
Conference on Photochemical Air Pollution and Air Pollution-related
Meteorology, at Tsukuba, 1-2, December 1982
PROCEEDINGS—PAGE 379
-------
•CEEDINGS-
*o
%
ta
CO
00
o
Table 1. Instrumentation for the aircraft
Parameter
NO/NOx
Ozone
Condensation
Analysis Technique
Chemi luminescence
Chemi 1 umi nescence
Light attenuation
Resolution
2-1000 ppb
1-1000 ppb
#1 n
Manufacturer AC-685 C-404
Kimoto o o
Kimoto o o
Environmental o
#3 #4
C-402 C-402
0 0
0 O
_
Nuclei(CN)
UV Radiation
Intensity
Aerosol size
Distribution
Humidity
Temperature
Altitude
Nitrate
Sulfate
Hydrocarbon
Position
UV radiometer
optical particle
counter
Humicup 0-100 %
Platinum resistence 0-50 °C
Pressure diode
Fluorocarbon filter
Ion Chromatography
Glass vessel
Gas Chromatography
LORAN-C
One
Eppley
Rion
Ogasawara
Kimoto
Nissho
Dionex
0
0
0
0
-
0
0
H
-
0
0
..
-
0
0
—
Shimadzu
Furuno
Operation
Height
350m
650m
900m 1200m
-------
•/. ' ' ~l
Figure 1. Topography of the Tokyo Metropolitan Area viewed from the
south-east.
I'KOCKKDINGS PACK
-------
Objective Analysis Area
KASHIMA
Sea
•I1MAI
IZUOSHIMA'
Island
PACIFIC Ocean
o Pilot-balloon sUtion
O Pilot-balloon & Sonde station
20km
Figure 2. Map of Kanto District, flight paths and objective
analysis area.
Pilot-balloon and Sonde observation points
1 Oyama 2 Kumagaya 3 Satte
5 Inzai 6 Otemachi
9 Chigasaki 10 Miura
13 Urawa 14 Tsukuba
17 Kusashino 18 Hachioji
4 Iruma
7 Yokohama 8 Sodegaura
11 Ichihara 12 Ohara
15 Nagareyama 16 Tsudanuma
19 Atsugi
PROCEEDINGS—PAGE 382
-------
• 700 mb -
o 800mb
9 21 9 21 9 21 9 21 JST
30 JULY ' 31 JULY 1 AUG. 2 AUG.
Figure 3. Variation of geostrophic wind during the
Arrows indicate wind directions.
aircraft survey.
PROCEEDINGS—PAGE 383
-------
31 July 1979 (0500JST)
31 July 1979 (1000JST)
31 July 1979 * (1500 JST)
\
1 Aug. 1979 (0100 JST)
1 Aug. 1979 (0500 JST)
20 KM
Figure 4. Diurnal change of local wind circulation at 350 m height from
31, July to 1, August.
PROCEEDINGS -PAGE 384
-------
31 July 1979 RUN 11 0535-0623 JST
Temp.CC )
20 25
:
H
K
20 30
40 50 60 70
DISTANCE (KM )
90 100
G
Figure 5. Vertical distribution of pollutants in the early morning of
31 July 1979(RUN 11 : 0430-0630 JST , Crosssection H-G) with
vertical temperature profile at Otemachi(0600JST). Night
time temperature inversion is observed at about 450 m
height. Id is the dry adiabatic lapse rate slope.
PKOCKKDINCS PACK :
-------
tit
c.
Vertical motion of
Trajectory B
13
03 s 10 ppb
(350m)
Figure 6. Trajectories of pollutants(31 July 1979): Trajectory
A( starting from 0600JST at 350 m level), Trajectory B
(starting from 1000JST at 350 m level, forward and backward).
The vertical movements of trajectory B are also shown. The
shaded area indicates that the ozone concentration was less
than 10 ppb estimated from Run 111.
PROCEEDINGS -PAGE 386
-------
31 July 1979
Ox>100ppb
Fx^3 1200 JST
E/773 1300 JST
mm 1500JST
prevailing
wind directio
20km
Figure 7. Movement of the heading edge of the sea breeze
level oxidant concentration isopleths.
and ground
PROCEKDINGS - PACK
-------
31 July 1979 RUN 12 1010-1046 JST
remp.( 0 temachi)
1000 JST
40 50 60
DI STANCE (KM )
70
80
90
100
G
Figure 8. Vertical distribution of pollutants in the midmorning of 31
July 1979 (RUN 12 : 0945-1130 JST , Crosssection H-G) with
potential temperature(dashed line) and temperature profile at
Otemachi(1000JST). A temperature inversion is observed at
about 800 and 1100 m heights.
I'KOCKKIJ 1NOS PACK
-------
31 July 1979RUN13
1540-1623 JST
o
H
10
40 50 60 70
DISTANCE(KM )
Figure 9. Vertical distribution of pollutants in the afternoon of 31
July 1979 ( RUN 13 : 1450-1630 JST, Crosssection H-G) with
potential temperature(dashed line) and temperature profile at
Otemachi(1500JST). Vertical wind profile! v and w
components) is calculated from the modified MATHEW method.
P ROC K K D 1NG S I' A (', K 3 8 9
-------
• \ i
Figure 10. Trajectories of pollutants(31 July 1979): C(starting fron
1500JST at 350m level), D(starting from 1530JST at 350 m
level) and E(starting from 1830JST at 350 m level).
Calculated in both the forward and backward directions.
PROCEEDINGS PAGE 390
-------
1. Auq. 1979 RUN 20
0009-OU2 JST
0 10 20
H
AO 50 60 70 80
DISTANCE (KM )
90
100
G
Figure 11. Vertical distribution of pollutants at midnight of 1
August 1979 (RUN 20 : 0000-0145JST , Crosssection H-G).
Vertical wind profile is calculated from modified MATHEW
method.
PROCKKDINGS PAGE
-------
1 Aug. 1979 RUN 21
0551 -0618 JST
Temp. ( Otemachij
0600 JST
40 50 70
DISTANCE (KM )
Figure 12. Vertical distribution of pollutants in the early morning of 1
August 1979 (RUN 21 : 0445-0620JST, Crosssection H-G)
with temperature profile at Otemachi(OGOOJST). The hard
dashed line indicated the discontinuity line of wind and in
the near of this line, the small downdraft wind zone are
detected.
-------
US STUDIES ON STRATOSPHERIC OZONE
presented by H. L. Wiser
Environmental Protection Agency
United States
PROCEEDINGS—PAGE 393
-------
SUMMARY
INTRODUCTION
This report reviews current knowledge about man-made
causes of changes in concentrations of stratospheric
ozone and the effects of those changes. Recent reports
of the National Research Council (NRC 1975, 1976a,b,
1978, 1979a,b) have treated the chemical and physical
aspects of potential reductions of stratospheric ozone in
detail. Part I of this report reviews recent develop-
ments on that subject. Part II deals with the effects of
reduction of stratospheric ozone on humans, other animals,
and plants, independently of what might cause the
reduction.
CHEMISTRY AND PHYSICS OF OZONE REDUCTION
The abundance of ozone in the stratosphere is determined
by a dynamic balance among processes that produce and
destroy it and transport it to the troposphere. According
to current understanding, the most important photochemical
reactions regulating ozone involve molecular and atomic
oxygen and various radicals containing nitrogen, hydrogen,
and chlorine. All of these compounds have natural
H§ sources, but their concentrations in the stratosphere can
^ be significantly altered by human activities. The human
M activities that have thus far been identified as
O potentially influencing stratospheric ozone are as
2 follows:
o
I • The release of gaseous chlorinated carbon
' compounds, mainly chlorofluorocarbons (CFCs) and methyl
> chloroform (CH3CCl3). CFCs are used as foam-blowing
H
CO
ID
01
-------
O
O
ft
W
CO
I
S2
O
w
agents, as working fluids in refrigeration systems, and
as propellants in aerosol sprays. Methyl chloroform is
an industrial solvent. These gases decompose in the
stratosphere providing a significant source of radicals
that contain chlorine.
* The release of nitrous oxide (^0) from
combustion and its enhanced release from soils and waters
as a result of various agricultural and waste management
practices. Nitrous oxide decomposes in the stratosphere,
introducing radicals that contain nitrogen.
* The direct input of nitrogen radicals to the
stratosphere due to nitrogen oxides (NO*) in aircraft
engine exhausts.
* The increased abundance of carbon dioxide (002)
in the atmosphere due to combustion of fossil fuels and
deforestation. Increased carbon dioxide has a subtle
influence, causing the temperature of the stratosphere to
decrease, which leads to increased stratospheric ozone,
and changing stratospheric concentrations of water vapor.
Key Findings and conclusions
Over the past several years, research, driven by
discrepancies between theory and observation, has led to
considerable improvement in our understanding of the
effects on stratospheric ozone of releases of CFCs and
oxides of nitrogen. As a result, previous discrepancies
between the estimates of models of stratospheric processes
and observed concentrations of certain important species
have been reduced. Important discrepancies still remain,
however, which means that there are still uncertainties
inherent in the results of modeling exercises.
Current scientific understanding, expressed in both 1-
and 2-dimensional models, indicates that if production of
two CFCs, CF2C12 and CFC13, were to continue into
the future at the rate prevalent in 1977, the steady
state reduction in total global ozone, in the absence of
other perturbations, could be between 5 percent and 9
percent. Comparable results from models prevalent in
1979 ranged from 15 percent to 18 percent. The
differences between current findings and those reported
in 1979 are attributed to refinements in values of
important reaction rates. Also, as an example, if the
atmospheric concentration of ^0 were doubled in the
absence of other perturbations, total ozone would be
reduced by between 10 percent and 16 percent. Although
atmospheric concentrations of N20 appear to be
increasing, we cannot reliably project the future course
of N20 emissions. Steady state reductions in both
these cases would be reached asymptotically in times on
the order of a century, although the assumption of doub-
ling NiO concentrations is unrealistic on such a time
scale. The effects of perturbations by CFCs and ^0
are not additive, so the estimates of effects of combined
perturbations require investigation of specific cases.
These results should be interpreted in light of the
uncertainties and insufficiencies of the models and
observations. For example, other chemicals released from
human activities are understood to have the potential for
affecting stratospheric ozone. Examples are methyl
chloride (O^Cl),.carbon tetrachloride (CC14), and
particularly methyl chloroform. Observations of critical
species need to be extended and confirmed by a number of
measurements using independent techniques. Important
assumptions in the models about rate constants,
distributions of certain species, and the reactions
taking place need to be tested. Furthermore, three
important discrepancies between models and observations
remain to be resolved: More chlorine monoxide (CIO) is
observed at altitudes above 35 km than is predicted, the
behavior of NOX in winter at high latitudes is
unexplained, and concentrations of CFCs in the lower
stratosphere are lower than the models suggest.
We anticipate that research on these problems in the
field, in the laboratory, and in theory currently under
way, planned, and proposed will lead to continued
improvement in understanding, resulting in further
reduction of the remaining discrepancies between theory
and observation. In particular, simultaneous measurement
of the important chemical species as a function of
altitude and latitude by various methods should prove
critical to improving understanding during the next
several years.
Examination of the historical record of measurements
of ozone does not reveal a significant trend in total
ozone that can be ascribed to human activities. This
observational result is consistent with those of current
models, since no detectable trend would be expected on
the basis of current theory.
Because data on total global ozone cannot be analyzed
to distinguish among causes of ozone changes, total ozone
data alone cannot be relied upon for early detection of
an anthropogenic change. Measurement of the spatial and
-------
O
n
M
W
D
O
00
"0
>
O
H
temporal distribution of critical trace species and
ozone, together with theoretical modeling taking into
account all the major influences on stratospheric ozone,
offers promise of understanding the causes of ozone
changes and the consequences of alternative actions in
response.
Recommendations
1. The national research program, including
atmospheric observation, laboratory measurements, and
theoretical modeling, should maintain a broad perspective
with emphasis on areas of disagreement between theory and
observation. Highest priority in research should be
given to a coordinated program to understand the spatial
and temporal distributions of important species, such as
CIO and the hydroxyl radical (OH).
2. The global monitoring effort should include both
ground-based and satellite observations of total ozone
and concentrations of ozone above 35 km, where theory
indicates the largest reductions might occur. Sound,
satellite-based systems for stratospheric observations
are essential.
3. Potential emissions of N2O, C02, CH3CC13,
and other relevant gases should be assessed and their
consequences for stratospheric ozone evaluated. Models
should be developed to describe the consequences for
stratospheric ozone of future emissions of these gases.
BIOLOGICAL EFFECTS OF INCREASED
SOLAR ULTRAVIOLET RADIATION
Stratospheric ozone acts as a shield to screen out much
of the short-wavelength ultraviolet (UV) in sunlight.
Slight changes in thin ozone layer may result in large
changes in the amount of damaging UV striking the surface
of the earth. Living creatures have adapted to the
present level of UV and to its fluctuations from season
to season and during the day. Part II of this report
gives the current state of knowledge about the effects on
biological systems of an increase in UV resulting from a
decrease in stratospheric ozone concentration.
Each of the findings and conclusions summarized below
has important implications for future research—either in
efforts to decrease the uncertainty in concepts or in
efforts to increase quantitative knowledge. These
research implications are spelled out in our list of
major recommendations. Recent advances in knowledge
since the last NRC report on the subject (NRC 1979a) have
clarified our view of the problem but have also pointed
out scientific areas not emphasized in earlier reports
that confound the simple prediction of the effects of
ozone depletion on biological systems. The unraveling of
these difficulties will be accomplished only by a
research effort directed by knowledgeable scientists,
especially photobiologists. In many instances, we are
still not sure of the scientific questions to be asked.
Similar comments were made in earlier NRC reports (NRC
1975) . The fact that they have not been acted on with
any reasonable financial commitment accounts for a large
part of our inability to make better predictions.
It seems certain that more than 90 percent of skin
cancer other than melanoma in the United States is
associated with sunlight exposure and that the damaging
wavelengths are in the UV-B region (290 nm to 320 nm) of
the spectrum. A decrease in ozone will be accompanied by
a well-predicted increase in UV-B. We estimate that
there will be a 2 percent to 5 percent increase in basal
cell skin cancer incidence per 1 percent decrease in
stratospheric ozone. The increase in squamous cell skin
cancer incidence will be about double that. Where in
this range the value falls depends on which theory is
used to make the estimate and on the appropriate
dosimetric data used. The predicted increases are
appreciably greater at lower latitudes than at higher.
Although the incidence of malignant melanoma increases
with a decrease in latitude, the degree to which sunlight
is responsible is not apparent, and there are few data
implicating UV-B as the only responsible wavelength
region. Therefore it is not appropriate to make
quantitative predictions about the increase in the
incidence of this disease associated with a decrease in
ozone.
Some of the difficulty in making quantitative
predictions about humans comes from uncertainties (even
in simple cellular systems) about the effects of inter-
actions among single wavelengths in a broad band, such as
in the ultraviolet of sunlight, in producing antagonistic
or synergistic effects. Moreover, it has been learned
only recently that rapid repair of sunlight damage to
human skin takes place during irradiation. An appreciable
fraction is photorepair mediated by visible light, and a
-------
13
§
O
w
w
o
OQ
I
I
•fl
O
M
IO
00
similar phenomenon seems to take place in anchovy
populations. The quantitative magnitudes of such effects
ace not known.
The effects of ozone depletion on other animals and
plants in the biosphere are as important as the direct
effects on human health. However, scientists are still
not able to predict quantitative effects on crop plants
or ecosystems.
The details of our findings and recommendations are
spelled out in Chapters 3, 4, and 5. Key findings and
conclusions and major research recommendations have been
extracted from the chapters and are listed below.
Estimates ace given, where possible, of how long the
recommended research might take under ideal circumstances.
Key Findings and Conclusions
Molecular and Cellular Studies (Chapter 3)
1. Deoxyribonuclelc acid (DNA) is probably the primacy
target in animal cells for most deleterious effects of
UV-B, especially effects involving mutagenesis and
neoplastlc transformation. Other targets of possible
biological significance for UV-B effects include
membranes, ribonucleic acid (RNA), and proteins.
2. The spectrum foe absorption of energy by DNA for
wavelengths in the UV-B region and the spectra for
biological damage to DMA as a function of wavelength
(action spectra) are-known. The absorption spectrum and
the action spectra are similar but not identical,
probably because long-wavelength light is absorbed in
some components of this genetic material that are not
effective in changing the structure of DNA. The action
spectra in the UV-B region for affecting mammalian cells
(killing, mutation, and neoplastic transformation) are
similar to those for damaging DNA.
3. The formation of pyrimidine dlmers (bonds between
pyrimidine residues in one of the two strands of DNA that
distort the normal DNA helical structure) appears to be
the major injury to DNA from UV-B irradiation.
4. There are major interactions between the effects
of UV-A (320 nm to 400 nm) and those of UV-B on DNA in
cells. Some of these are antagonisms, whereby UV-A
effects significantly reduce or repair the UV-B damage.
Except for photoreactivation, which involves enzymic
splitting of pyrimidine dimers back to normal single
residues mediated by UV-A and visible light, these
interactions are still poorly understood.
5. In excision repair, dimers are removed from one
strand of a DNA double helix by enzymes that work in the
dark, leaving the unaltered strand as a template for
reconstitution of a new normal strand. Photoreactivation
and excision repair of pyrimidine dimers occurs rapidly
in human skin.
Ecosystems and Their Components (Chapter 4)
6. Both UV-A and UV-B have been reported to be
detrimental to plant growth and development and to a
number of physiological processes of plants, when
examined under non-field conditions. The adaptability of
plant species appears to be sufficient, under current
ambient levels of UV-B, to maintain food crop yields.
The potential for further adaptation to predicted
increases in ambient UV-B is not known.
7. Ambient UV-B at present levels or similar levels
in the laboratory can damage sensitive aquatic organisms
or stages in their lifecycles that occur at the water's
surface. Natural populations of aquatic organisms have
adapted to current UV-B levels so as to maximize
reproduction potential. In the case of anchovy larvae,
it has been demonstrated that photorepair of UV-B damage
is effective even at UV-B levels significantly higher
than those that would result from predicted ozone
depletions. Photorepair may be a general adaptive
mechanism of organisms evolving in the presence of UV-B.
Currently, there is no information from which to predict
the magnitude of adverse effects of enhanced UV-B on
aquatic organisms.
8. From limited field experiments on terrestrial
plants and laboratory experiments with captured or
cultured aquatic organisms, it appears that different
species of both plants and animals have different
sensitivities to increases in UV-B above current levels.
Changes in species compositions and abundances of
organisms have been observed in simulated aquatic
ecosystems subjected to enhanced UV-B. Mathematical
models show that in systems subject to large natural
oscillations in the size of the population, there are
severe limitations on the minimum population density
needed to maintain a species. However, the data
currently available on food chains in the natural
-------
O
n
M
M
O
i
CO
O
ID
to
ecosystem are not precise enough or complete enough to be
used to predict population dynamics or the displacement
of an individual species under current environmental
conditions. It is doubtful therefore that a statistically
significant causal relationship between increased UV-B
levels and food chain success can be predicted in the
near future.
9. Only minor effects of increased UV-B levels are
predicted for animals used for human food.
Direct Human Health Hazards (Chapter 5)
10. A reduction in the concentration of stratospheric
ozone will not create new health hazards, but will
increase existing ones.
Effects Other Than Cancer
11. There is evidence that direct acute effects of UV
on humans, such as sunburn (acute erythema) and corneal
inflammation (photokeratitis), are linked more strongly
to UV-B than to UV-A.
12. Acute erythema and photokeratitis can be
predicted accurately for a given dose and spectrum of
UV-B, since the action spectra, dose-response curves, and
intensity-time reciprocity relationships are known.
13. Ultraviolet radiation affects many aspects of the
immune system of animals and humans. Allergic contact
dermatitis, skin graft rejection, tumor susceptibility,
and function and viability of individual circulating and
noncirculating cells of the immune system can be altered,
primarily by UV-B.
Skin Cancer Other Than Melanoma
14. Data on the relative incidence rates of basal and
squamous cell cancers in highly pigmented (black) versus
lightly pigmented (white) persons indicate that more than
90 percent of skin cancers other than melanoma in U.S.
whites are attributable to sunlight.
15. Molecular, cellular, and whole animal data all
implicate UV-B as the major carcinogenic component of
sunlight for skin cancers other than melanoma. The
evidence is stronger for squamous than basal cell cancers
because animals rarely get basal cell cancers. In
humans, basal cell cancers are virtually all related to
sunlight.
16. Based on animal studies, UV-B is implicated not
only as an initiator of carcinogenesis but also as a
promoter (in the general sense and via indirect effects)
of chemical carcinogenesis. With the current state of
knowledge, it is not possible to assess the extent to
which increasing exposures to chemicals would result in
increases in skin cancers due to synergism, over and
above any increase because of increased UV-B exposure
alone.
17. A 1 percent reduction in the amount of strato-
spheric ozone is predicted to give an approximate 2
percent increase in biologically effective UV-B.
Epidemiological data suggest that a 2 percent increase in
UV-B would give a 2 percent to 5 percent increase in
basal cell skin cancers. For squamous cell skin cancers
the increase would be about twice these values (4 percent
to 10 percent).
18. The risk of developing skin cancers other than
melanoma and the increased risk due to increased exposure
to UV-B could be mitigated by individuals through changes
in lifestyle that would reduce exposure.
Melanoma
19. The incidence of skin melanoma appears to depend
on latitude, an indication that sunlight is a contributing
factor. Circumstantial evidence such as occupational
differences and location of the cancers on the body
suggests, however, that exposure to sunlight is only one
of several factors. The association between sunlight and
melanoma is not strong enough to make a prediction of
increased incidence due to increased exposure to UV based
on epidemiological data.
20. The only evidence that suggests UV-B causes
melanoma in humans comes from studies of people with the
inherited disease xeroderma pigmentosum. These people
have a known defect in the mechanism that would repair
UV-B damage to DNA, and they also have a very high
incidence of skin cancers, including melanoma.
21. There are no reliable animal models for
light-induced melanoma. The only models currently
available are animals with chemically induced,
preexisting pigmented lesions that can be made to look
like melanoma after UV irradiation.
-------
o
o
w
ts
o
en
I
I
O
W
o
o
Major Research Recommendations
The estimates following each recommendation of how long
the research might take are educated guesses based on the
experience of individual committee members. The estimates
provide only a rough idea of how long the research might
take under ideal circumstances.
Molecular and Cellular Studies (Chapter 3)
1. An understanding is needed of why broad bands of
UV (heterochromatic radiation) often do not act on DNA in
vivo and on in vitro cell systems as a simple sum of
monochromatic wavelengths.
(a) Studies of interactive effects between UV-A and
UV-B are fundamental to understanding the mechanisms
of cancer induction by sunlight. Such studies,
employing bacteria or cultured mammalian cells, would
take about two to five years.
(b) An understanding is needed of UV-A-induced repair
systems in bacteria, as a first step in understanding
possible similar systems in higher organisms. This
would take about two to five years.
(c) Experiments should be conducted to determine the
rate and extent of photoreactivation in humans in
sunlight. Data are needed on how the level of diraeca
depends on >the relative amounts of UV-A and visible
light compared with the amount of dimer-producing
UV-B. These experiments would take about two to five
years.
2. Data are needed on the rates of repair, in the
dark and in laboratory light, of UV-irradiated human skin
cells as a function of UV dose. The differences, if any,
between acute and chronic irradiations should be deter-
mined. One might be able (with informed consent) to
study individuals who are exposed to high levels of UV-B
as part of phototherapy for psoriasis. The aim of such
experiments would be to determine whether the kinetics of
dark repair of damage from pyrimidine dimers in human
skin show two components, a slow one and a fast one, as
is true for human cells irradiated in vitro. The two
components represent repair of DNA in different regions
of the DNA strands. Equally important questions are,
what other types of biologically important damages occur
in skin, what are their lifetimes, and are any of them
persistent? These data could be obtained in about four
or five years.
Ecosystems and Their Components (Chapter 4)
3. Techniques must be developed for simulating
changes in UVrB under natural ambient conditions. Only
in this way can dose-response relationships be obtained.
If these techniques cannot be developed for studies at
temperate latitudes, they might best be achieved in a low-
latitude (subtropical), minimal-cloud-cover, multiuser
facility, which would provide UV-B radiation corresponding
to reduced ozone concentrations at more northern lati-
tudes. Priority should be given to screening representa-
tive species of important food plant systems for identify-
ing possible adverse effects on crop productivity.
Dosimetry and environmental regulation techniques must be
developed to ensure optimum experimental conditions--
conditions equivalent to the higher latitude ambient
field conditions of the plants being tested. Without
strict attention to these control conditions, studies
will have limited potential for extrapolation or
prediction. It would take about three years to develop
the facility and another three years to conduct the
species screening experiments.
4. The effects of UV dose on elements of aquatic food
chains cannot be determined unless (a) the underwater
spectral irradiances are integrated over the varying
positions of organisms in water columns to obtain the
exposures that simulate spectral intensities in the
natural systems, and (b) damage to individuals can be
related to population dynamics in the natural ecosystem.
This would require an integrated research approach
involving physical hydrography, physical optics, and
organism physiology. It would take about five years to
develop this approach and obtain results. Unless UV-B
studies are made as a part of an ecosystem study, effects
on populations and interactions among populations cannot
be predicted. (Testing for whole ecosystem effects is
addressed in another NRC report, Testing for Effects of
Chemicals on Ecosystems (NRC 1981).)
An attempt to incorporate such an integrated approach
was made for anchovy larvae. The interdisciplinary
approach used in the anchovy study to assess UV-B damage
to food chains, together with the specific laboratory
-------
measurements, should serve as a model for future research
proposals.
o
o
H
I
5
o
01
o
M
Direct Human Health Hazards (Chapter 5)
5. Studies (animal and human) should be conducted in
the developing field of photoimmunology to determine the
magnitude of UV effects on the human immune system, the
effective wavelengths, and the dose-response relationship.
Results may increase understanding of skin cancer
mechanisms, other effects of UV on skin, and certain
other diseases. These studies would take about two to
five years.
6. Animal studies of UV-induced skin cancers other
than melanoma are needed to understand interactions among
parameters such as intermittent exposures, different
wavelengths, dose rates, chemical carcinogens and
promoters, and agents that modify cellular responses to
irradiation. These studies would take about two to five
years.
7. The Surveillance, Epidemiology, and End Results
program of the National Cancer Institute routinely
collects data on incidence of melanoma. The incidence of
skin cancers other than melanoma should be surveyed every
decade at a time coinciding with the population census,
so as to determine trends in time. Only a few locations
are necessary, but these should be the same as past
survey locations. Data should be collected in a way that
permits cohort as well as cross-sectional analysis.
8. Animal models for UV- or light-induced melanomas
are needed. They would allow studies of action spectra,
dose-response curves, waveband interactions, and other
parameters. It is not possible to predict how long it
would take to develop such models.
9. To determine the association between UV and
melanoma, it would be useful to determine the incidence
of the various subtypes of melanoma and their dependence
on latitude. Although this will be difficult because the
majority of melanomas are of the superficial spreading
type, the methodology is available. Careful epidemio-
logical studies that are based on reliable clinical and
histological studies of subtypes of melanoma are needed.
PART I: CHEMISTRY AND PHYSICS
OF OZONE REDUCTION
-------
Chapter 1
W
S CURRENT STATUS
13
n
w
*.
M INTRODUCTION
This chapter reviews recent changes in the state of
understanding of the chemical and physical processes that
determine the effect of human activities on concentrations
of stratospheric ozone. The report is motivated by a
continuing need to assess the potential effects on
stratospheric ozone of chlorofluorocarbons (CFCs) and
other chemicals, as prescribed in the Clean Air Act, as
amended (42 USC 7450}. The topic has been the subject of
intense study during the past decade; our report builds
on that work, most notably on previous studies by the
National Research Council (NRC 1975, 1976b, 1977, 1978,
1979b) and the National Aeronautics and Space
Administration (NASA) (Hudson and Reed 1979). To prepare
our assessment, we relied on our professional knowledge,
on a concurrent technical review prepared under the
auspices of NASA, the Federal Aviation Administration,
the National Oceanic and Atmospheric Administration, and
the World Meteorological Organization (WHO) (Hudson et
al. 1982), and on a series of topical reviews prepared at
our request by technical consultants. The consultants'
reports are contained in Appendixes A to F.
PROCESSES DETERMINING OZONE CONCENTRATIONS
Ozone (O3) is formed in the stratosphere by reaction of
atomic oxygen (O) with diatomic molecular oxygen (02).
The process is initiated by photolysis of 02, that is,
the dissociation of 02 into atomic oxygen by absorption
of solar ultraviolet radiation at wavelengths below 240
nanometers (nm). Photolysis of 02 occurs mainly at
altitudes above 25 km.
-------
O
n
H
K)
13
>
a
According to current understanding, approximately 1
percent of the ozone created in the stratosphere is
removed by transport to the troposphere? the remaining 99
percent is destroyed by chemical reactions in the
stratosphere that re-form ozone into 02. The net
effect of these chemical reactions is either the
combination of ozone with atomic oxygen to form 02,
represented by the equation
O + 03 + 202, I
or the combination of two ozone molecules represented by
°3 "*" °3 * 302- IJ
These equations represent the net results of a number of
complex sets of reactions catalyzed by a variety of gases
and chemical radicals present in the stratosphere in
trace amounts.
Important examples of sets of reactions summarized by
process I are
(la)
Cl + 03 *
CIO + O -»
NO + O3 *
N02 + 0 -»•
OH + 03 +
O + H02 +
CIO + 02
Cl + 02,
N02 + 02
NO + O2,
H02 + 02
OH + 02.
Process I may also proceed by the direct path
O + 03
2O
(2a)
(2b)
(3a)
(3b)
(4)
These reactions are limited by the availability of oxygen
atoms and therefore occur mainly at altitudes above 25
km. The reactions that limit the rates at which chains
1, 2, and 3 proceed are (lb), (2b), and (3b),
respectively.
Process II summarizes reaction schemes in which atomic
oxygen is not limiting, for example,
OH
H02
03
O3
H02
OH
02
202.
(5a)
(5b)
Reactions (5) account for most of the ozone lost below 25
km in current models. The chemistry of the lower strato-
sphere is complex, however (Appendix A), and one cannot
exclude additional reaction schemes involving oxides of
nitrogen and chlorine (NOjj, C10X) and oxidation
products of hydrocarbons such as methane (CH4>.
Ozone removed from the stratosphere by transport to
the troposphere is ultimately lost by chemical reactions
in the gas phase or at the earth's surface.
The spatial and temporal distribution of the
concentration of ozone reflects a dynamic balance among
the processes that form and remove ozone (Figure 1.1).
According to current understanding, photolysis of 02
provides a global source of ozone of 50,000 million
metric tons per year, with more than 90 percent of this
amount formed above 25 km. Most of this ozone is removed
by reactions represented by process I. At altitudes
between 25 km and 45 km, reaction (2b) accounts for
roughly 45 percent of the ozone removed while reactions
(lb) and (4) each account for about 20 percent and
reaction (3b) for 10 percent (S.C. Wofsy, Harvard
University, private communication, 1982). About 1
percent of stratospheric ozone, 600 million metric tons
per year, is removed below 25 km by process II, with a
similar amount being lost by physical transport to the
troposphere.
Only 30 percent of global ozone is stored at altitudes
above 25 km, reflecting the relatively short chemical
lifetime of ozone at high altitudes. The rest is
contained in the region below 25 km, and more than 70
percent of the amount below 25 km is found at latitudes
above 30°. The abundance of ozone below 25 km is
determined by the balance between transport from the
chemically more active region at higher altitudes and
losses to the troposphere; its distribution is regulated
by atmospheric motions.
Adding to the stratosphere substances that destroy
ozone has the effect of creating a new balance between
production and removal processes in which the total
abundance of ozone is reduced. For example, stratospheric
concentrations of chlorine monoxide (CIO) and nitrogen
dioxide (N02) may be increased as a result of emissions
of CFCs and nitrous oxide (N20) from human activities.
The effects are persistent. A typical CFC molecule,
CF2C12 for example, survives for approximately 75
years in the atmosphere before it is decomposed by
sunlight releasing its constituent chlorine atoms in the
-------
8
M
W
D
i
I
"8
>
O
tfl
a
a
n
is
§
I
a.
o
as
I
O
E
stratosphere. A chlorine atom can affect recombination
of between 104 and 105 ozone molecules during Its
lifetime in the stratosphere (on the order of two years)
before it returns to the troposphere, mainly as hydro-
chloric acid (HC1). A similar situation holds for N20.
Approximately 10 percent of N2O molecules released to
the atmosphere decompose by paths leading to production
of stratospheric nitric oxide (NO), and subsequently
N02, by reaction (2a). The average NOX molecule also
removes between Id1* and 10^ ozone molecules before it
returns to the troposphere, after its typical two-year
residence in the stratosphere. Current theoretical
models lead us to. conclude that the dependence of ozone
concentration on altitude will also change, the net
effect being a redistribution of ozone from higher to
lower altitudes. Quantitative estimates of these effects
have varied somewhat over the past decade (Appendix A).
Perturbations by Chlorine
Currently, approximately 3 parts per billion (ppb) of the
lower stratosphere consists of chlorine bound in organic
molecules such as methyl chloride (CH3C1), carbon
tetrachloride (CC14), and CFCs (Hudson and Reed 1979,
Hudson et al. 1982). Table 1.1 indicates the abundances
of the more prevalent species; only methyl chloride is
known to have natural origins. The table also shows
estimates of current rates of release of man-made
compounds found in the lower stratosphere.
Halocarbons decompose under the influence of sunlight
at altitudes above 20 km; the fractional abundances
(mixing ratios) of halocarbons (in ppb) are observed to
decrease with increasing altitude (Appendix C). The
chlorine produced by decomposition of halocarbons is
converted to inorganic species, including HC1, chlorine
nitrate (CINO-j), CIO, and atomic chlorine (Cl). Hydro-
chloric acid is the major reservoir for chlorine at
altitudes above 25 km (Appendix C). Concentrations of
Cl, CIO, and HC1 have been observed in the stratosphere;
observations and predictions of theoretical models are in
general agreement, although some difficulties remain
(Appendix D), as we shall see.
Computer calculations using current understanding and
incorporating new data on rates of several important
reactions (Appendixes C and D) suggest that continued
release of the CFCs, CF2Cl2 and CFC13, at rates
-------
TABLE 1.1 Concentration in the Lower Stratosphere and Release Rates of Major
Sources of Chlorine in (lie Stratosphere
"0
w
o
n
H
M
O
3
o
CO
I
I
"0
>
a
M
o
01
Concentration (ppb)"
Rate of Release
Compound
Methyl chloride (C1I3 Cl)
F.12(CF,C1,)
F.lt(CFClj)
Carbon tetrachloride (CCU)
Methyl chloroform (CHjCClj)
Molecular
0.62
0.30
0.18
0.13
0.11
Chlorine
0.62
0.60
0.54
0.52
0.33
tons of Cl per year)
2b
0.1 9C
0.20C
0.053
0.35*
"Hudson et al. (1982).
''About 85 to 90 percent of CI^Cl is naturally produced, the remainder being attrib-
uted to industrial sources (Cicerone 1981). The total release rate varies slowly in time
because of the large contributions of natural sources.
C1980 release rate from "World Production and Release of Chlorofluorocarbons 11
and 12 through 1980," Chemical Manufacturers Association Fluorocarbon Program
Panel, July 29, 1981. Release rate has decreased by about 20 percent from the peak
rate of 1974.
^1976 release rate (NRC 1979b). The release rate is apparently relatively constant,
although somewhat uncertain.
«Neely and Plonka (1978).
prevalent in 1977 would ultimately cause a net decrease
of total global ozone roughly between 5 percent and 9
percent assuming no other perturbations (Hudson et al.
1982). We regard a representative result to be 7 percent
(Appendix C). This would result in a smaller steady
state reduction in ozone than reported in NRC (1979b),
which was 16.5 percent with a 95 percent probability that
the true value lies between 5 percent and 28 percent.
(Other models current in 1979 gave reductions ranging
from 15 percent to 18 percent (Hudson and Reed 1979) .
Estimates have fluctuated between roughly 5 percent and
20 percent over the past eight years as models have been
refined (Appendix A).) The steady state reduction would
be reached asymptotically in times on the order of a
century. Calculations now indicate that the reduction
would occur almost entirely at altitudes above 35 km, in
the region of the stratosphere where the ozone concentra-
tion is determined primarily by chemical processes, with
a smaller, partially compensating increase in ozone
concentrations at lower altitudes. The current result
obtains for both 1- and 2-dimensional models and further
differs from that prevalent in 1979 in that earlier
calculations showed regions of reduction both above and
below 35 km.
The differences between current findings and those
reported in 1979 are the result of refinements in the
values for the rates of several reactions affecting the
concentration of the hydroxyl radical (OH) (Appendixes C
and D). The refinements are the result of improved
laboratory measurements (Hudson et .al. 1982). OH is
important because the concentration of CIO in the lower
stratosphere is particularly sensitive to it. Results of
model calculations using current values for these reaction
rates are in good agreement with observations of CIO for
altitudes below 35 km (Appendixes C and D), whereas
models using the reaction rates favored in 1979 give
concentrations of CIO a factor of 3 higher than observed
values in this range. The new reaction rates have not
changed greatly the results of calculations for altitudes
above 35 km, however, so that the amounts of reduction in
ozone above 35 km obtained in the 1979 and current models
are about the same. The models continue to indicate lower
concentrations of CIO in the stratosphere above 40 km
than are observed. We shall return to this discrepancy.
Increased attention to effects of releases of methyl
chloroform (CH3CCl3) on stratospheric ozone is
warranted because of the growing use of this compound, an
industrial solvent. The release rate increased by a
factor of about 50 between 1958 and 1978 (Neely and
Plonka 1978).
Perturbations by Oxides of Nitrogen
The chemically active oxides of nitrogen in the strato-
sphere (such as N02) are thought to arise mainly from
photooxidation of N20. N20 is formed naturally by
bacteria in soil and water. As indicated earlier,
reactions involving N02 account for about 45 percent of
the ozone removed in the stratosphere between 25 km and
45 km.
The human influence on the global cycle of fixed
nitrogen is thought to be significant and increasing (NRC
1978). The global atmospheric concentration of N2O
appears to have increased by 2.7 percent (from 292 ppb in
1964 to 300 ppb in 1980) over the past 16 years (Weiss
1981, Weiss and Craig 1976) . The concentration of N2O
in the atmosphere is likely to continue to increase with
increases in emissions associated with agricultural
practices, disposal of human and animal wastes, and
possibly combustion; but we cannot say how or on what
time scale.
-------
An increase In N20 concentrations of about 30
percent in the absence of other perturbations could cause
a reduction in global ozone of an amount comparable with
the 7 percent reduction currently estimated due to
continued emissions of CF2C12 and CFC13 at 1977
rates, also taken as the sole perturbation. This estimate
is based on current model calculations that indicate
that, should the concentration of N20 double in the
absence of other perturbations, total global ozone would
decline between 10 percent and 16 percent (Hudson et al.
1982).
Early attention to human influences on the strato-
sphere focused on effects of NOX released by high-
flying aircraft (NRC 1975). Models then and now suggest
that an input of NC^ at altitudes above about 20 km
should lead to reduction in stratospheric ozone. A
source of NOV at lower altitude, associated for example
A
with subsonic commercial aviation, can modify local
chemistry such as to cause an increase in tropospheric
ozone, it has been suggested that reductions in the
column of ozone above the earth's surface due to
reductions in stratospheric ozone may be masked to some
extent by increases in tropospheric ozone attributable to
subsonic jets and urban smog.
Assessment of the impact on stratospheric ozone due to
a combination of perturbations requires investigation of
specific cases since the effects are not simply
additive. Hudson'et al. (1982) Deport the results of
several studies of the effects of doubling atmospheric
N20 concentrations and continuing releases of CFCs at
1977 ratesi both separately and in combination. The
Lawrence Livermore National Laboratory (LLNL) model, for
example, indicated a reduction of 12.5 percent due to
doubling N2O with a reduction of 12.9 percent due to
the combination of perturbations. The LLNL model gives a
reduction of 5.0 percent for CFC releases alone. Another
model, from Atmospheric and Environmental Research, Inc.,
gives reductions of 9.5 percent for doubling N20, 6.1
percent for continuing CFC releases, and 13.0 percent for
the combination. The results may be misleading, however,
since current trends suggest a considerably longer time
scale for doubling atmospheric concentrations of N2O
than for reaching the steady state reduction due to
continued emissions of CFCs.
Perturbations by Other Species
Stratospheric ozone may be affected by human activity in
a number of other ways. Of greatest potential concern
are changes in concentrations of carbon dioxide (C02),
water vapor (H20), and perhaps methane (CH4).
The well-documented increase in atmospheric
concentrations of CO2 is directly attributable to
combustion of fossil fuels and wood. This increase is
expected to lead to a global warming of the atmosphere
near the surface of the earth but is expected to cause a
reduction in the temperature of the stratosphere (Fela et
al. 1980).
Lower stratospheric temperatures would have at least
two effects. First, the chemical removal processes
affecting ozone that were described earlier are sensitive
functions of temperature, being less efficient at lower
temperature. Consequently, with lower temperature the
equilibrium concentration of ozone would be higher.
Current models incorporating this effect suggest that the
steady state reduction in total ozone due to continuing
emissions of CFCs at 1977 rates would change from 5
percent to 9 percent to between 4 percent and 6 percent
if global C02 were doubled concurrently (Hudson et al.
1982). (Global C02 has increased by about 6 percent in
the past 22 years.)
The possible second effect of lower statospheric
temperatures resulting from increased C02 is a thermally
driven change in stratospheric water vapor (H20) caused
by a change in the temperature of the tropical tropopause,
Dissociation of H20 provides the source of hydrogen
radicals, and these radicals play a key role in strato-
spheric chemistry regulating abundances of both active
NOx and Clx species in addition to their contributions
to reactions (3) and (5). A complete model for strato-
spheric chemistry should include a description of H20
interactions, a requirement beyond current capability.
Stratospheric ozone may also vary in response to
changes in concentrations of 014, which plays an
important role in reaction (Ib) by regulating the
partitioning of chlorine between HCl and CIO (Hudson et
al. 1982). Recent reports (Rasmussen and. Khalil 1981)
suggest increases in global concentrations of Glfy, but
likely future changes and their consequences are unknown.
-------
CURRENT STATUS OF MODELS OF THE STRATOSPHERE
O
n
O
GO
Tl
!»
O
M
Theoretical models of stratospheric chemistry cannot be
validated by measurements of total ozone only, owing to
the diversity of factors, natural and man-made, that may
affect ozone concentrations. Comparison of calculated
and observed values for the concentrations of important
trace species and radicals—such as OH, CIO, HO2' and
atomic oxygen—must play a central role in any orderly
strategy for validating models.
In general terms, agreement in detail between the
predictions of theoretical models and observations is
excellent. For example, changes in reaction rates since
1979 have resulted in substantial agreement between
theory and observation for CIO below 35 km. There are,
however, three areas in which discrepancies remain. The
discrepancies may or may not point to significant
difficulties in modeling. Similarly, agreement between
modeling results and observation of CIO, while
encouraging, need not imply validity of the model at
lower altitudes.
The improved agreement between observed and calculated
concentrations of CIO in the lower stratosphere may be
attributed mainly to changes in rate constants for
reactions affecting OH. Concentrations of OH in current
models are lower than values obtained in 1979, with the
result that a larger fraction of Clx is now found as
HC1. The chemistry of the lower stratosphere is complex,
however. Agreement between model and observed values of
CIO in the lower stratosphere should be considered
necessary but not sufficient for validation. A more
extensive and demanding test would require comparison of
theoretical and observed profiles of other radicals,
particularly OH.
To improve understanding of stratospheric chemistry
also requires that attention be directed to the
assumptions of the models and to the measurements against
which models are tested. High-quality measurements are
obviously prerequisite to validation of models.
Confidence in observations of critical species is
enhanced by using a number of independent, inter-
calibrated techniques, each relying on different physical
properties. Validation of measurement technique is
difficult since concentrations of the important
atmospheric species may vary in time and space on scales
that are not well understood. Validation procedures
involve coordinated studies in the field requiring
considerable logistical support.
The assumptions of models are of two types: (1) input
data on environmental conditions, reaction rates, and
other parameters, and (2) the reaction schemes incor-
porated into the model. There are still uncertainties
about the appropriateness of some assumptions common in
current models. For example, the rate for reaction of OH
with H02» an important path for removal of hydrogen
radicals, remains uncertain despite extensive and
continuing efforts in the laboratory. There are other
reactions in need of similar clarification. Models are
sensitive to assumptions about the abundance and
distribution of stratospheric H20; the underlying
physical and chemical processes that regulate this key
parameter are not well understood. It is difficult to
rule out the possibility of an important role for species
not now included in models, and, if history is a guide,
there may well be future surprises in this area. Models
for the stratosphere have been adjusted over the past
decade in just this manner to include gases such as CIO
(1974), C1N03 (1976), and HOC1 (1978) (Appendix A,
Figure A.I), and there is current discussion of a
possible participation of sodium (Kolb and Elgin 1976,
Murad et al. 1981). Progress in recognition of missing
species or reactions occurs through a combination of
laboratory, field, and theoretical studies, the normal
practice of validating models and resolving discrepancies.
As was noted earlier, there is reasonable agreement
between model calculations and observations for CIO in
the lower stratosphere. Currently, however, there is a
discrepancy between theory and observation for CIO in the
region above 35 km, where chlorine-mediated catalysis is
most important. The average value for the concentration
of CIO measured by Anderson and co-workers (see Appendix
D) near 40 km is almost a factor of 2 larger than the
value calculated from models. Furthermore, theory and
experiment give different dependences of the concentration
of CIO on altitude in the upper stratosphere. The CIO
discrepancy is particularly important because it occurs
at altitudes where ozone is most sensitive to perturba-
tions caused by CFCs (Appendix D, Figure D.52).
Extensive ground-based observations of N02 have been
made over a range of latitudes by J. Noxon (see Appendix
C), revealing a sharp spatial discontinuity in concentra-
tion in the winter with very low concentrations poleward
of the discontinuity. Thus far, no theroretical model
has been able to explain this phenomenon.
-------
o
o
w
w
o
I
I
"0
H
o
oo
A third area of discrepancy between current models and
observations Is in concentrations of CFC13 and
CF2Cl2 at altitudes above about 20 km (see Appendix
C, Figure c.11). Observed values are substantially lower
than predicted values. The difference could be due to
errors in model simulation of ultraviolet radiance in the
lower stratosphere, which, if true, would imply that the
CFCs have shorter residence time in the lower strato-
sphere. The issue is not resolved and requires
continuing attention.
Nevertheless, the extent of agreement between
measurement and theory is encouragingly good.
MONITORING AND ASSESSMENT OP TRENDS
Measurements of the total amount of ozone above a unit
area of the earth's surface (called total column ozone)
are essential for assessing the human influence on ozone
(as well as the potential effects of changes in ozone on
humans and other organisms). As detailed in Appendix A,
total column ozone fluctuates on a variety of spatial and
temporal scales owing to natural causes; these fluctua-
tions tend to mask possible systematic changes due to
man-made perturbations. For example, current models for
single and combined perturbations predict a reduction of
total column ozone over the past decade of less than 1
percent, but a change of this magnitude cannot be
distinguished from fluctuations due to other causes
(Appendixes D and E).
Models of the stratosphere predict that the largest
reductions in ozone due to releases of CFCs should occur
near 40 km. Reductions should therefore be most readily
detectable at this altitude. Current models suggest that
ozone concentrations at 40 km should have decreased by
several percent over the past decade. There have been
reports in the press that an effect of this order has
been detected in data from satellite experiments (see,
for example, Science, Sept. 4, 1981, pp. 1088-1089). The
community of atmospheric scientists has not yet had the
opportunity to scrutinize this evidence, which must
therefore be regarded as preliminary (Appendix F).
Our ability to detect trends in ozone in the future
will depend on the availability of consistent, high-
quality data taken over long time intervals. Improvements
in the current monitoring systems are feasible and clearly
needed. For example, it is vitally important to improve
and enhance systems for monitoring ozone profiles in the
upper stratosphere that could provide a valuable early
indication of systematic changes in ozone due to emissions
of CFCs or N20, but existing data in the upper strato-
sphere are inadequate for this purpose. It is also
imperative to continue, and desirable to expand, the
high-quality monitoring of total ozone by Dobson
spectrophotometers.
THE QUESTION OF EARL* DETECTION
A notable feature of the ozone issue is that a reduction
due to increases in the tropospheric concentrations of
CFCs or N20, once it has taken place, is expected to
persist for more than 100 years even if the practices
that caused it are stopped immediately. It is therefore
important to detect an anthropogenic effect at the
earliest possible time. Three methods currently exist
for this purpose.
1. Measurement of total ozone. Relying on
measurement of total ozone has the following advantages
(Appendix E): There exists a relatively long historical
base (30 to 50 years) of data. Ground-based
instrumentation is available and may be readily
complemented by observations from satellites. Finally,
total ozone is most directly related to one of the
consequences of depletion that is of concern, the
possibility of enhanced exposure to ultraviolet radiation
at the ground (Part II). Since, however, the reduction
due to CFCs is expected to be concentrated at high
altitudes, measurements of total column ozone are less
sensitive indicators of an anthropogenic effect than are
measurements of ozone profiles.
2. Measurement of ozone at high altitudes. The
advantages of this method derive from the theoretical
result that changes in ozone due to CFCs are predicted to
be largest at high altitudes. Changes in the spatial
distribution of ozone may be important for understanding
the second major consequence of depletion that is of
concern, the possibility of climate change (Appendixes B
and C). The disadvantages stem from the difficulty of
making the measurements, whose quality and stability are
inferior to those of total ozone (Hudson and Reed 1979,
Hudson et al. 1982). satellite data are particularly
subject to changes in calibration of instruments, which
-------
"C
PC
o
n
w
w
o
cannot be refurbished; ground-based measurements by the
Umkehr method give poor height resolution and are subject
to perturbations by hazes and stratospheric particulate
matter. Partly because of these difficulties, the data
base is relatively small and somewhat fragmented (Appendix
F). Ozone measurements using satellites would have the
desirable attribute of obtaining temporal and spatial
distributions that would be useful in validating 2- and
3-dimensional models.
3. Measurement of key radicals involved in chemical
removal processes. Measurements of spatial and temporal
profiles of important species such as CIO and OH may be
combined with chemical models for assessment of trends
and their causes, such that the dependence on specific
models can be relatively slight. This method is in
principle the most sensitive, but it is also the least
direct.
The last approach is regarded by many experts as
having already shown the effect of chlorine of human
origin, mainly connected with emissions of CFCs. But
this conclusion would be more firmly established with
more direct confirmation, as discussed in the previous
section. Ideally, all three types of measurement should
be integrated (with due regard to their sensitivity) in a
strategy for early detection of anthropogenic effects.
UNCERTAINTY
O
o
to
Quantitative estimates of the uncertainties inherent in
current estimates of reductions in ozone due to emissions
of CFCs and N^O are difficult to obtain. The ability
to make quantitative estimates of uncertainty depends
both on what we know and on what we do not know. Such
estimates employ professional judgments about the
importance of various factors and the sensitivity of the
results to potential changes in understanding.
Our major concern in estimating uncertainties in our
understanding of stratospheric ozone is with the
possibility that some key process or processes may be
missing from current models. In an orderly scientific
strategy, continuing development of models on the basis
of an ongoing comparison with observational data is
expected. Progress is stimulated by the existence of
discrepancies or uncertainties and tends to occur in more
or less discrete steps rather than uniformly. Our
understanding of the lower stratosphere has improved over
the past two years as a result of developments that may
be attributed at least in part to efforts to resolve
earlier (and larger) discrepancies between observed and
computed values for the concentration of CIO. Agreement
between observed and computed values of CIO is now
satisfactory below 35 km, but, as noted earlier, there
continues to be a serious discrepancy at higher altitudes.
This disagreement illustrates the difficulty of estimating
limits of uncertainty for current estimates for reduction
in ozone due to CFCs.
For example, observed values of CIO at higher altitudes
are larger than calculated values, suggesting that the
long-term reduction in ozone could be correspondingly
larger. One can, however, conceive of speculative
chemical schemes that could suggest a stratosphere less
vulnerable to perturbations.
In circumstances such as this, the usual ways of
estimating uncertainty (using mathematically rigorous
procedures) are not applicable. Instead we rely on
professional judgment. The predictions of the current
chemical scheme have been cross-checked against observed
atmospheric data in many ways, and the agreement in
general is quite good. As stated earlier, a representa-
tive estimate of potential steady state reduction of
global ozone due to continued releases of CFCs at the
1977 rate in the absence of other perturbations is 7
percent. There continue to be, however, important
discrepancies between theory and observation.
Our opinions are divided on whether there are
sufficient scientific grounds to estimate the effect of
resolving one of the discrepancies, that of CIO in the
upper stratosphere, on calculations of ozone reduction.
We agree that we do not know enough at this time to make
a quantitative judgment of the uncertainty associated
with the other major discrepancies, N02 at high
latitudes and lifetime of CFCs in the stratosphere above
20 km.
Those of us who believe there are grounds to judge the
effect of resolving the CIO issue conclude that our
estimate of ozone reduction from CFC emissions should not
change by more than a factor of 2.
Those of us unwilling to offer quantitative estimates
of uncertainty hold the conviction that no rigorous
scientific basis exists for such statements. We are
concerned by implications of the discrepancies noted
earlier. These discrepancies should be resolved in the
-------
o
o
00
I
I
1)
Q
next few years by orderly application of the scientific
method with appropriate interaction between theory and
observation. We see no reason to prejudge the result of
this process.
Research during the past several years has enhanced
our understanding of the factors affecting stratospheric
ozone. Development of the field is progressing rapidly.
We anticipate further developments in both observation
and modeling in the next few years that will result in
considerable improvement in ouc understanding, both
clarifying and reducing uncertainties.
FINDINGS
1. Our understanding of the stratosphere has advanced
considerably in the past two years. Progress is
significant in all areas with improvements in our ability
to model the system in more than 1 dimension, with
impressive achievements in techniques for measurement of
chemical reactions in the laboratory/ and with major
advances in our ability to measure concentrations of
important trace species in the atmosphere. We note here
that the success of the research is due in no small part
to the breadth of the scientific effort involving
scientists from many countries with support from both
private and governmental sources. We expect continued
improvement in,understanding of the chemistry and
dynamics of ozone reduction to result from research
currently under way,-planned, and proposed.
2. The concern regarding the possibility of reduction
in stratospheric ozone due to CFCs remains, although
current estimates for the effect are lower than results
given in NRC (1979b). The change in estimates of ozone
reduction reflects improvements in our understanding of
chemical processes in the stratosphere below 35 km.
There has been no significant change in results obtained
by models for the stratosphere above 35 km. The major
impact of CFCs is predicted for the height range of 35 km
to 45 km.
3. The chlorine species Cl and CIO participate in a
series of chemical reactions that destroy ozone. The
radical CIO has been measured in the stratosphere in
significant amounts and is believed to be primarily of
human origin. Our current understanding indicates that
if production of CFCs continues into the future at the
rate existing in 1977, the steady state reduction in
total Ozone, in the absence of other perturbations, would
be between 5 percent and 9 percent. Previous estimates
fluctuate between roughly 5 percent and 20 percent, with
those current in 1979 ranging from 15 percent to 18
percent. Latest results also suggest that CFG releases
to date should have reduced the total ozone column by
less than 1 percent.
4. According to current understanding, increases of
N20 in the stratosphere would result in reductions in
total ozone, with the largest effects occurring in the
lower stratosphere. Although concentrations of N2O in
the stratosphere appear to be increasing, we cannot
reliably project the future course of N20 sources. If,
however, the concentrations of ^O in the atmosphere
were to double, in the absence of other perturbations,
current models suggest that the steady state reduction in
the total ozone would be between 10 percent and 16
percent.
5. On the whole, there have been substantial
improvements in the agreement between model predictions
and observed profiles of trace species in the past
several years. Three exceptions are still a cause for
concern: Above 40 km, more. CIO is observed than is
predicted by current theory; the behavior of NOX in
winter at near-polar latitudes is unexplained; and
concentrations of CFCs in the stratosphere above 20 km
are lower than predicted by the models.
6. Examination of historical data (extending back 30
to 50 years) has not yet shown a significant trend in
total ozone 'chat can be ascribed to human activities.
Current models of combinations of pollutants suggest that
a reduction of total ozone to date from human activities
would be less than 1 percent. No detectable trend would
be expected on the basis of these results.
7. Data on total ozone should not be used alone to
guide decisions on whether to take action to prevent
future changes in stratospheric ozone. Although an
important guide, analysis of trends in total ozone cannot
by itself reveal causes of ozone reductions or
increases. Such analysis, together with measurement of
altitude profiles of trace species and ozone and
theoretical modeling, offers promise of understanding
causes of ozone changes and the consequences of
alternative actions in response.
8. The impact of CFCs should be assessed in-the
context of a broad understanding of the variety of ways
in which human activity can alter stratospheric
-------
composition. Ozone may be reduced by increasing levels
of CFCs and ^0, but reductions might be offset in part
by higher concentrations of C02 and perhaps CH4.
Human activities have already increased the amounts of
COj and CFCs in the atmosphere, and from the known
release rates, further increases can be confidently
expected. In addition, there is evidence that ^O and
CH4 concentrations are also increasing. A special
reason for concern about perturbations potentially caused
by CFCs and N20 is the long lifetime of these gases in
the atmosphere, of the order of 50 to 150 years. Even if
the releases of these gases were reduced, the atmosphere
would not recover until far in the future.
RECOMMENDATIONS
In light of our findings, we believe it is important
to maintain a competent, broadly based research program
that includes a long-term commitment to monitoring
programs. The research effort should extend over at
least two solar cycles (of 11 years each) to distinguish
between changes induced by variations in the sun from
those associated with man. Accordingly, we make the
following recommendations:
1. The national research program, including
atmospheric observations, laboratory measurements, and
theoretical modeling, should maintain a broad perspective
with some focus on areas of discrepancy between theory
and observation. A coordinated research program to
understand the spatial and temporal distributions of key
species and radicals merits highest priority.
Observations should be extended to include studies of the
equatorial and polar regions.
2. The global monitoring effort should include both
ground-based and satellite observations of total ozone
and of concentrations of ozone above 35 km, where theory
indicates the largest reductions might occur. We also
need data to define the variability of stratospheric
temperature and water vapor. We regard sound,
satellite-based systems for stratospheric observations as
essential.
3. Potential emissions of a number of relevant gases,
in addition to CFCs and N20, and their consequences for
stratospheric ozone should be thoroughly evaluated and
assessed. It is important that we understand current and
potential rates of emissions of these compounds and the
effects these emissions might have on ozone in addition
to understanding emissions and effects of CFCs. There is
observational evidence that atmospheric concentrations of
N2O and CO2 are increasing. Models should be
developed to describe the combined effects on strato-
spheric ozone of future changes in releases of all
relevant gases, such as CFCs, N2O, CO2, CH4,
CH3C1, and CH3CC13.
-------
Ref. for Dr. Wisers Presentation
40
1.0
0.1
>
t-
o
g
en
LLJ
>
0.01
0.001
0.0001
-,-1.0
Sunlight
Throuah Ozone
UV-A -
0.1
0.01
0.001
c/)
2
I-
X
2
cn
0.0001
290 300 310 320
WAVELENGTH (nm|
330
FIGURE 2.2 The relative intensity of sunlight (solar elevation of 60') reaching the
surface of the earth for different amounts of stratospheric ozone (the normal amount
is close to 3.4 atmosphere • mm). The shapes of two biological sensitivity curves are
also shown: (a) damage to DNA multiplied by the transmission of human epidermis,
and (b) human erythema or sunburnt Curve (c) is the response of the Robertson-Berger
meter (discussed in Chapter 5). (Source: The.three curves of sunlight intensity are from
U.S. Congress, Senate (1975); the two biological sensitivity curves are from Setlow
(1974) and Scott and Straf (1977); the Robertson-Betser meter curve is from Berger
etal. (1975).)
PROCEEDINGS—PAGE 412
-------
INTRUSION OF STRATOSPHERIC OZONE
INTO THE TROPOSPHERE
presented by H. Muramatsu
Meteorological Research Institute
Japan MA
PROCEEDINGS—PAGE 413
-------
Intrusion of Stratospheric Ozone into the Troposphere
H.Muramatsu,Y.Makino,M.Hirota and T.Sasaki
Meteorological Research Institute, Tsukuba, Japan.
1. Trajectory analysis
1.1 Vertical distribution of ozone
Thin layers which have the maximum ozone mixing ratio are frequently
observed below the tropopause especially in spring over Tateno,
Tsukuba. The example of the vertical profiles of ozone and tempera-
ture obtained by the ozone -sonde is shown in Fig.l for 28 May 1969.
The ozone maximum layer at about 500 mb corresponds to the tempera-
ture inversion layer.
Fig.2 shows the vertical cross section along 140° E for the
same day of Fig.l. Two jet cores are observed over Sendai and Akita
about 250 km and 400 km north of Tateno,respectively. The jet stream
front is observed at 550 mb and 430 mb over Tateno, corresponding
to the maximum layers of ozone in Fig.l. The relative humidity is
below 30 % in the frontal zone (below 10 % in the central part of
the frontal zone).
1.2 Isentropic trajectory
The potential temperature at the ozone maximum layer is 325°K.
A three-dimensional trajectory on the 325°K surface was traced from
the adiabatic assumption. Geostrophic wind was obtained from the
isentropic stream-function( Montgomery stream-function), that was
calculated from the radiosonde data. From the wind field obtained
above, the air parcel was traced backward in time and its location
was determined every three hours.
Fig.3 shows the isentropic trajectory on the 325°K surface from
26 May.OOZ to 28 May,06Z. The a-ir parcel (which arrived over Tateno
on 28 May,06Z) crossed the jet axis over Osaka at 03Z (three hours
befre) from the cyclonic side toward the anticyclonic side.
The air parcel is within the stratosphere before 12Z on 27 May,
PROCEEDINGS—PAGE 415
-------
moving south-westward along the jet axis. After that the air parcel
at 300 mb begins to descend from the stratosphere to the troposphere
on the southern edge of the cyclone by changing the direction eastward.
The descending velocity of the air parcel shows the maximum value
of 17 cm/sec (three hour average) at OOZ on 28 May, when the air parcel
crosses the jet axix (Fig.4).
1.3 Frequency of the event
Time" cross section "of the relative humidity over Tateno for May
1969 is shown in Fig.5. Arrows show the day on which ozone profiles
were obtained. Ozone maximum layers (filled circles) are found in area
where the relative humidity is less than 20 % (marked by slanting lines)
It is also found that the ozone maximum layers are frequently observed
around 500 mb. This result is related to the fact that frequency of
the appearance of the dry air over Tateno is maximum in late spring
at 500 mb level (Fig.6).
2. Aircraft observation
In order to know the mechanism of the intrusion process that
was shown in the previous section, the spatial distributions of ozone
and other constituents were observed from an aircraft around the tropo-
pause gap and the frontal boundaries.
2.1 Case study : 15 March 1981
(1) Intrusion around the jet axis
There were two jet streams over Japan as shown in Fig.7; subtropi-
cal jet at 150 mb and 250 mb and polar front jet at 350mb. Observations
were made around the polar front jet along 135°E. Horizontal flight
courses are shown by a(5.1 km altitude),b(7.1 km) and c(8.8 km).
The vertical cross section along 135°E is shown in Fig.8. The dis-
tribution of ozone(solid lines) around the polar front jet(J) are shown.
Strong downward wind (4- ^ ) was observed at the layers of the maximum
ozone concentration and upward wind ( f ) was observed on thr north.
Low water vapor concentration is observed in the layer of high ozone
concentration. Thin broken curves show the range of relative humidity
below 20 %; lowest in the frontal boundary shown by heavy broken curves.
PROCEEDINGS—PAGE 416
-------
It is known that the stratospheric ozone is transported into the
troposphere below the polar front jet along the frontal boundary.
(2) Diffusion of ozone
The horizontal distribution of ozone at the altitude of 5.1 km
obtained from the south to north flight course is shown in Fig.9. On
the north of the high ozone region, the layer of the constant ozone
concentration is observed. There is the strong gap of the ozone con-
centration between them. This gap coincides with the boundary of down-
ward and upward winds. In the region of this constant ozone concent-
ration the strong vibration of the aircrat was recognized.
It is known that the ozone diffuses on the north of the maximum
ozone layer by the turbulent air motions whose frequencies are 80 to
180 sec (Fig.10).
(3) Potential vorticity
The variation of the potential vorticity that has the conservative
property in the stratosphere was analyzed. The region of the high
potential vorticity of stratospheric origin (hatching in Fig.11) is
transported downward; southward on 14 and eastward on 15 March. The
region of the high potential vorticity decreases rapidly in the tropo-
sphere. The total air mass of the high potential vorticity( the air
14
mass of the stratospheric origin) was 11x10 kg at 12 Z on 14 May,
14
and decreased to 5x10 kg at 00 Z on 15 May.
This shows that the potential vorticity is not conserved in the
troposphere and has the half-value period of a half day. The decrease
of the potential vorticity occurs on the north of the ozone maximum
layer as shown in the previous section.
PROCEEDINGS—PAGE 417
-------
OZONAGRA_M
~ TA_TENO_
OAI« 28MAYB691 I
( Ifc30 JST
Fig.1 Vertical distributions of ozone and temperature
'over Tateno,28 May 1969,14:30 JST
- 50
Fig 2. Vertical cross-section for 28 N.ay 1969, 002 al0ng 140'E trom
Thin lines sho«- po«nt:al temperatures fK; ; broken lines. ,so«chs .ousec,
lines, tropopauses and frontal boundar.es. The area wUh relac.ve n«»,a.ty less tl
30% is marked by hatching.
PROCEEDINGS—PAGE 418
-------
-" X •• * /"y
V- iv \X-V' X'l \ /76-
y< -\ * ' 'A '
xC xv /***''< \/\*27-/o°Tz^:.5mb''
"\X V, IRKUTSK yW=^ -"/-'A
' I/O Tj"^™ i f
PEKING ;^"-r—*d—
r
\ / • 1' A . l^'^JJ \280Q
/ ^'x. ;jr~"-^".
Fig.3 Isencropic trajectory on the 325°iC surface
fron 26 May,OOZ to 28 May,06Z, 1969.
Jet axes on 200iab surface are indicated for
27 May,OOZ (dash-dotted line) and 28 May,
OOZ (dashed line)
5 i IU
in
E
i
S-io
UJ
c
1 1 1
\/
r V-
1 1 1 1 1 1 1 1 1
0 12 00 12 00 I!
MAY 26 27 28
UNIVERSAL TIME
Fig 4 Vertical velocity (three hour average) of the
air parcel from 26 May to 28 May 1969
CROC BED INGS—PAGE 11 9
-------
1000
MAY 1969
Fig.5 Tine section of the relative humidity for
Tateno in May 1969. Area with relative
humidity less than 20% are marked by slanting
lines, those with less rhan 50% but over 20%
are stippled. Arrows show the days on which
ozone sonde observation'was aade. Filled
circles show the height of the maximum ozone
concentration.
RH (•/.)
12
Fig.6 Distribution of the relative humidity(%) over
Tateno averaged for the period 1966 to 1970.
PROCEEDINGS -PACK r_'n
-------
Fig.7
Flighc courses and
the jec streans
10
9
'"- 8
6
JC
— 7
liJ
Q 6
I I
4080
1 T
-
:
OZONE CONCENTRATION-
(ppbv)
15 MARCH 1981
09- 13 JST
:
:
20' 30' 4Q' 50' 00'
34 »
10' 20' 30' 40' 50' 00'
35°
LATITUDE (NORTH)
!0' 20' 30' -C1
36°
Fig.8 Meridional cross section along 140°E.
Solid curves show the ozone concentration
(ppbv); heavy broken curves, frontal boundaries
and tropopause; solid line with arrows, flight
course; /f- and I- show the position of upward
and downward winds; M shows the posotions where
the heavy vibration of aircraft was observed; c
shows the position of no vibrations; thin broken
lines show the boundary of relative humidity
below 207..
PROCKEDINGS 1'ACK l.M
-------
I6O
fV\ A AA />
9.25
9.3O
9:35
TIME {JST)
9:40
9:45
Fig.9 Horizontal ozone distribution and its deviation
at 5.1 km across the frontal boundary.
DOWNWARD and UPWARD show the region where the
downward and upward winds was observed.
I i
03 DEVIATION
3:28-9:ii JST
15 MARCH 198J
5.1km —
400200100 50 40 30
PERIOD (sec)
20
Fig.10 Power spectrum of the ozone deviation
on the north of the frontal boundary.
PROCEEDINGS—PAGE 422
-------
C^'_! ICC
, I ^
/ i
3IO
15 MARCH
308
I9SI, CO Z
306
3IO°K
308
120
Fig.11 Deviation of the pocncial vorricicy on the
310°K isencropic surface.
Solid curves show the Montgomery screatn funccion
(10 erg/g) ; broken curves, pressure; region of
che potential vorticity higher than 6xlO"rand
3x10 "K/Tib/sec are shown by heavy and thin
slanting lines, resoectively.
PROCEEDINGS -PACK
-------
THE VERTICAL DISTRIBUTION OF CF2C12, CFC13
AND N20 OVER JAPAN
presented by M..Hirota
Meteorological Research Institute
Japan MA
PROCEEDINGS—PAGE 425
-------
THE VERTICAL DISTRIBUTIONS OF CF2C12,. CFC1, AND N" 0 OVER JAPAN
Michio KIROTA, Hi sa fund. MURAMATSU, Yukio MAKING
and Tom SASAKI
Meteorological Research. Institute
1-1, Nagamine, Yatabe-aechi, Tsukuba-gun,. Iharaki-ken,- JAPAN
The ataospheric CF Cl , CFC1, and N_0 are thought as
the major sources of the stratospheric CIO and NO respectively.
•i Jt
If such species are released in large quantities by manrs
activity, the natural balance of the stratospheric ozone
will be damaged-
It is known that these compounds, especially CF Cl
and CFC1,, are recently being accumulated in the tropsphere-
Cn order to assess the influence of these compounds on. the
natural ozone balance, the ataospheric concentrations and
distributions of these compounds must be examined..
In this paper, our observations over Japan, since
1978 will be reported.
EXPERIMENTAL
Sampling of air and the gas-chroaiatographic analysis
are described in the previous paper(Kisaki et al~, 1980)»
In order to return the sampling can on land, collections
of the stratospheric air samples were performed only in summer*
PROCEEDINGS—PAGE 427
-------
RESULTS AND DISCUSSION
Typical vertical profiles of CF Cl r CFC1 and NO are
shown in Fig. 1- These compounds were well mixed in: the
troposphere except over the urban, area up to the altitude
of 2 km where the higher mixing ratios of CF Cl? and CFC1,
were observed- Mean tropospheric mixing ratios for- each.
observation period are summarised in Table 1» All three
compounds show increasing trends from 1978 to 1981.,
Increasing trends of CF Cl and NO,, however, are not
clear because only one observation was performed in 1982.-
Vertical profiles of CF Cl , CFC1, and N_0 in the
stratosphere are .shown in Figs. 2 - k- These were observed
in summer from 1973 to 1982,- and annual variations could
not be observed in the range of the experimental error.
Vertical profiles of these compounds were compared with.
those obtained by a one-diaentional photochemical-diffusive
snodel(solid lines in Figs. 2-4)- In. the =odel calculation,
aa eddy-diffusion coefficient profile was obtained using
vertical profiles of N_0. Table 2 shows the estimated
fractional change in 0, due to CF Cl and CFC1, in 1980-
Total 0, depletion was 0.6%.
PROCEEDINGS—PAGE 428
-------
Table 1. Mean volume mixing ratios of CF Cl , CFC1 and NO in
the troposphere
Observation
period
Oct. '78
- Mar. '79
Oct. '79
- Feb. '80
Dec. '80
- Mar. '81
Feb. '82
Altitude
-* 7.k km
-7 13-6 km
-» 9-3 km
-> 8.1 km
2 2' *
282(5D*2
SD=2,-
o
29^(62)
306(17)
! -i
CFCl,/ppt*
j
162(53)*2
179(H)
SD=7
18^(12)
•SD=1
I8g(2l)
N.O/ppb
31Q(90)
32g(88)
SD=1
1C*
33jt35)
:
* :
SD:
Values over the urban area up to the altitude of 2 ka are excluded.
*f values obtained in Mar. '78 are included.
standard deviation, number in ( ) : number of samples.
Table 2. The fractional change in 0 due to CF Cl and CFC1 in 1980
<- t-
Altitude
(km)
55
51
^7
^3
'ZQ
S7
35
31
27
Fractional change in 0 (%)
CF2C12
-0.59
-1.13
-2.32
-5.30
-2.71
-0.93
-0.09
fO.Oif
23 i +0.05
19 ! +0.15
15
total
+0.13
-0.29
CFC1
3
-0.3^
-0.67
-i.kk
-2.38
-1.81
CF;C1_+CFC1_
22 3
-0.93
-1.76
-3-63
-6.11
-if. if 3
-0.70 j -1.63
-0.16
-O.07
-0.03
+0.09
+0.07
-0.26
-0.26
-0.0k
+0.03
+0.23
+0.23
-0.55
In the calculation of the total 0 depletion, its effect on the
tropospheric 0 is neglected.
PROCEEDINGS—PAGE 429
-------
Figure captions
Fig.1 Vertical profiles of CF Cl , CFC1, and NO in. the
troposphere
Air samples were collected on. Cessna k
Flight cource: Eaneda - Eachijo jina(Feb.. 14-1
and Haneda - Sendai(Feb» 18)
Fig.2 Vertical profile of CF Cl in: the stratosphere
In 1978, air sample nas collected-.iit.a plastic
bag of 250 1-
- : rertical profile calculated from a 1-D
model
XvVJi : range of the tropospheric mixing ratios
( ): value for which large experimental error
was considered
Fig.3 Vertical profile, of CFCl^ is. the stratosphere
. - • • ./
- , ttXX, ( ): saae as la. Fig. 2
Fig.it Vertical profile of N_0 in the stratosphere
(o) : Ran air was saapled on DC-9, TDA over
Osaka on Jan. 8, 1979.
- , XXV, ( ): sane as in Fig. 2
PROCEEDINGS—PAGE 430
-------
Fig. 1
8-
7-
6-
4-
-a
3"
< 2-
1-
0-
CFCt3
CF2Ct2
14-16.18 Feb 1982
CFC13
N20
o 0Q o
o
o
O * O5O O
N20
-iS.
t-a-
0-2 0-3 ' 0-4 ,
Volume Mixing Ratio / ppb,
0-5 0-3 0-4
/ pp m
OD O> O —
co OD 03
CD
rc5 -a
A/V V \A/VV V V
—I—
o
—i—
O
Q.
CL
•^ O
O —
o en
c
x
m —
"P E
o
E
(V4 U
S "o
p
,o
PROCEEDINGS—PAGE 431
-------
o
n
w
w
o
Kl
I
I
"d
>
O
H
w
to
30
20
(U
TJ
D
Fig. 4
(V)
(7)
N20 in summer
+• 1978
o 1979
V 1980
G) 1981
A 1982
(V)
0-
0-01 0-02 0-05 0.1 0-2 0-5
Volume mixing ratio / ppm
1-0
Fig. 3
CFCi3 in summer
o
0-01 0-02 0-05 0-1 0-2 0-5
Volume mixing ratio / ppb
------- |