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 . ,\
/
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
-------
   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

-------