United States
                     Environmental Protection
                     Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
                     Research and Development
EPA-600/S3-83-042  July 1983
&EPA          Project  Summary
                     Meteorological   Factors  in  the
                     Formation  of  Regional  Haze
                     James G. Edinger and Timothy F. Press
                       The purpose of this research project
                     was to determine the role of mete-
                     orological factors in the formation of
                     widespread areas of haze in the eastern
                     United States.
                       Three  case studies were made: A
                     summer haze episode, an off-season
                     haze episode and a non-haze episode.
                       Results showed that over the course
                     of 2 or 3 days emissions from widely
                     separated sources such as St Louis,
                     Chicago, Cincinnati and Pittsburgh are
                     leafed together by vertical and horizontal
                     shears and mixing by daytime convec-
                     tion to form a dilution volume many
                     hundreds to well overathousand km in
                     extent and 2 or 3 km in depth.  Almost
                     all stations reporting haze during an
                     episode were confined to this dilution
                     volume and most of these in that part
                     of the plume containing emissions that
                     were 2 or 3 days old.
                       The dilution volume associated with
                     the off-season episode was of about
                     the same  magnitude as that of the
                     summer case, but was shallower and
                     horizontally more extensive.  Both of
                     these 3-day haze volumes were much
                     smaller than the dilution volume as-
                     sociated with the non-haze case which
                     blanketed  almost the entire eastern
                     United States.
                       This Project Summary was developed
                     by EPA's Environmental Sciences Re-
                     search Laboratory,  Research Triangle
                     Park, NC, to announce key findings of
                     the research project that is fully doc-
                     umented in a separate report of the
                     same title (see Project Report ordering
                     information at back).

                     Introduction
                       Extensive areas of hazy air covering
                     major portions of the eastern United States
                     have become a frequent summertime oc-
currence during the last few decades. This
regional haze, though well described by
satellite observations and the conventional
surface synoptic network, is not as well
understood as the  smaller,  meso-scale
visibility blight associated with plumes
from individual point or area sources such
as power plants and urban complexes.  A
variety of models have been developed for
these  meso-scale  plumes.  But  as  is
reported in a recent study by the National
Academy  of Sciences, the modeling of
regional haze is not nearly as far advanced
  The present study is directed toward
improving our understanding of the mete-
orological  mechanisms involved  in the
formation  of such large volumes of more
or less uniformly polluted air.  It is  a
diagnostic enterprise.   Its purpose is to
provide information useful for constructing
models, of the regional haze formation
process.
  Air quality simulation models describ-
ing the long-range transport of air pollu-
tion have been devised. Usually they are
based on a requirement of mass balance.
The processes and mechanisms included
are: emission, chemical transformation,
physical removal, turbulent diffusion and
transport
  The air  motions used in these models
are observed atmospheric motions, not air
movement calculated from fundamental
physical principles. Typically they are the
observed wind averaged over the depth of
the polluted layer or the wind observed at
some standard level interior to the polluted
layer, often the 850  mb level.  A certain
amount of observed detail in the wind field
is lost in the process of constructing
vertically averaged wind fields or  in ac-
cepting the wind at any one level as being
representative of the vertically averaged
flow. These neglected (dispersive) motions

-------
are parameterized in the models by eddy
diffusion terms.  Some air quality simulation
models have taken vertical wind shear
within the polluted layer into account for
travel times out to 24  hours. In the work
reported here the winds at all levels within
the polluted layer are  used to determine
the long-range transport and dispersal of
pollution out to 3  day's travel from the
sources. It is done for a number of haze
episodes chosen from 30 years of record
(1948-1978). The purpose is to discover
the meteorological mechanisms associated
with  the formation of these  extensive
volumes of hazy air.
  The analysis  consists of constructing
the previous 3-day history of the hazy air. It
is assumed that the responsible pollutants
came from the major air pollution sources
in the eastern United States (large cities,
power plants, etc.).   A  3-day dilution
volume is  constructed for each major
source. These individual dilution volumes
overlap to form a conglomerate  volume
which encompasses all of the air which in
3 days passed  over the major sources.
Because of the overlap it may be assumed
that this large conglomerate volume would
also include air from any other source,  of
whatever size, located in the area inter-
mediate in location to  the major sources.
  In reconstructing the extensive volume
of hazy air the major sources of emissions,
i.e.  the large urban centers, are taken as
continuous point sources. Their nighttime
emissions are treated  as plumes  moving
with the surface winds as reported on the
3-hourly surface synoptic maps.  Their
daytime emissions are assumed to mix
vertically by thermal convection through
the lowest 2 km of the  atmosphere and to
move with  the  observed wind fields  at
each level (50 mb intervals) within that
layer.  The shearing  motions and transla-
tion combined  with the vertical  mixing
produce a dilution volume associated with
the particular source.  The effects of
horizontal motions on a scale smaller than
those described by the streamline analyses
are neglected.
  The results of the investigation indicate
that the widespread volumes of hazy air
examined correspond closely with the
conglomerate dilution volume of the air
which during  the  previous three  days
passed over the area that includes  St
Louis, Chicago,  Cincinnati and Pittsburgh
and points  in between. They also show
that most of the haze is reported in air that
contains blended emissions that are 2 or 3
days old.  Sources closer to the Atlantic
coast probably contribute to the haze but
because of the  absence of data over the
 adjacent ocean their influence could not
 be determined.

 Procedure
   The meteorological analysis makes use
 of the U.S. National Weather Service's
 surface and  upper air data.  Horizontal
 wind fields were constructed and analyzed
 for the 950,900,850 and 800 mb levels.
 Figure 1  is an example. Surface flow was
 determined from the photostatic records
 of the U.S. National Weather Service's 3-
 hrly surface synoptic maps.  The vertical
 profiles of temperature and humidity were
 generated by computer graphics.
   The dilution volume developed during
 the first 24  hrs for any one source is
 constructed as follows. The initial instant
 is taken as 00 GCT.  During the subsequent
 twelve hours, nighttime, a plume is  laid
 down at or near ground level.  Its location
 and configuration is determined by con-
 structing a streakline for 12 GCT from the
 3-hrly surface wind maps.  It is assumed
 that vertical mixing  motions  are absent
 during this twelve hour interval.
   During the subsequent twelve hour in-
 terval, 12 to 00 GCT, daytime, mixing
 motions distribute the emissions vertically
through the layer from the surface up to
the 800 mb level. It is assumed that at a
time not far removed from 12  GCT the
surface plume develops vertically into a
curtain which  is subject to winds at all
levels up to 800 mb.  Shears in the wind
subsequently carry the emissions at various
levels off in different directions, deforming
and tilting  the curtain.  Streaklines con-
structed at each level for the time 24 hours
after emission began depict the location of
this deformed curtain.  Vertical mixing
motions from the surface up to 800 mb
transform the  curtain into a volume, its
horizontal projection being the area over
which the  emissions  have  been spread
during the 24  hour interval.
  The volume which contains  the first
day's emissions from the source in question
continues to grow during the next day due
to shearing motion in the vertical and
another 12 hours of vertical mixing, and is
transported (usually) away from the source.
To determine  the shape of the dilution
volume and its location at the end of the
second day, trajectories(24 to48hr) origi-
nating at  significant  points around the
perimeter of the 24 hr volume at each level
are constructed.  The end points of these
Figure 1.   Streamline map, 00 GCT 25 February 1973 at 850 mb level.

-------
trajectories define the surface of a volume,
the horizontal projection of which is the
area covered by the dilution volume at the
end of the second day.
  To construct the  dilution volume that
contains not  only emissions  during the
first 24 hours but emissions during the
second day as well, one needs to construct
the dilution volume for air moving over the
source during the second day. This volume
when added to the  other constitutes the
two  day polluted wake of the source (its
48 hr dilution volume). In the case studies
carried out these procedures were repeated
for a third day's  emissions so that the
dilution volume  containing  three  day's
emissions from a  given source  were
obtained.
  The selection of the 800 mb level as the
upper  limit of the  vertical stirring and
mixing  followed  unsuccessful efforts  to
define the top of  the mixing  layer by
inspecting the plotted radiosonde tem-
perature and humidity profiles at all stations
in the eastern United States at both the00
and 12 GCT times.  Careful analysis of the
profiles together with the associated sur-
face weather maps suggested that during
typical summer haze episodes in the eastern
United States there is no extensive daytime
stable layer in the lowest few kilometers
placing an upper limit on corrective mixing.
To the contrary reports of towering cumulus
frequently appear here and there within
the hazy volume suggesting that the mixing
in some places reaches and exceeds the
500 mb  level.  The  upper limit  to the
vertical  stirring and mixing is at best ill-
defined and irregular.  In all of the case
studies the top of the stirred layer was
assumed to be at the 800 mb level. In one
case, in an attempt to improve the  fit
between the constructed haze volume and
the actual haze observations the level was
raised to 700  mb with some improvement
of the fit with the observations.
  To select the major sources of emissions
in the northeastern United States the sta-
tistics appearing in  EPA reports (1978)
were used.  In preliminary case studies
nine major urban industrial areas were
chosen to represent the sources of emis-
sions. Each was treated as a point source.
Several considerations resulted in the re-
vision downward in the number of these
sources to just four: St Louis, Chicago,
Cincinnati and Pittsburgh. They were: (1)
preliminary results suggested that it was
emissions two or three days old that were
the major contributors to the haze and
some of the  emissions from the cities
along the Atlantic coast moved into the
data void over the ocean in less than three
days preventing  the completion of the
construction of their dilution volumes and
(2) the two and three day dilution volumes
from individual point sources overlapped
each other to such  a  large extent  that
including  sources  more closely  spaced
than the four chosen would have only a
small effect on the size of their combined
dilution  volume.
Results
  Four haze episodes were selected for
analysis from the historical record, 1948
to 1978. In addition, one non-haze episode
which was meteorologically similar to the
haze cases was examined.
  Figures  2,3, and 4 show the results of
the study  of an off-season (not summer)
haze episode, OOGT  1  March   1973.
Figure 2  shows the 00 GCT 1   March
1973 location of one-day-old emissions
(released during the last 24 hrs) from St
Louis, Chicago, Cincinnati and Pittsburgh.
The black  dots indicate stations reporting
haze at 00 GCT I March 1973. Figure 3
gives the location at 00 GCT I March 1973
of two-day-old emissions (released between
24 and 48 hrs ago), and Figure 4 gives the
location of the corresponding three-day-
old emissions (released between 48 and
72 hrs ago).
  Examination of these figures and others
not reproduced have shown that: (1) only
a small fraction of the haze reported occurs
in one-day-old  emissions, less than one
half occurs in two-day-old emissions, but
almost 90% occurs in the three-day-old
emissions, and (2) the majority of the haze
reports are at locations where a superposi-
tion of emissions from different sources or
different times  occurs.   These  results
suggest that much of the haze forms in
blended, aged  emissions and at  a  con-
siderable distance from the sources.
  The corresponding diagrams for the
non-haze case revealed one-day-old plumes
that are noticeably larger  than  those
generated during the haze episodes.  And
the two-day-old emissions cover an  area
so large that it compares favorably with the
largest of the  three-day-old emissions
volume for the haze episodes.  It is not
possible to determine how much larger
the three-day-old emissions volume is for
the non-haze case since its boundaries
extend beyond the map boundaries on the
north and in the southeast.  It is apparent
that much larger  dilution  volumes are
generated in this non-haze case than during
haze episodes.
 Figure 2.    One-day-old emissions from St. Louis. Chicago. Cincinnati, and Pittsburgh. 00 GCT
            1 March 1973.

-------
                                                                                   Recommendation
                                                                                     The analysis of the haze episodes sug-
                                                                                   gests that it might be instructive to devise
                                                                                   haze formation  models  that  move the
                                                                                   emissions  with  the winds  at  all levels
                                                                                   within  the lowest 2 km layer and that
                                                                                   include vertical eddy diffusion but neglect
                                                                                   horizontal eddy diffusion.
Figure 3.    Two-day-old emissions from St. Louis, Chicago, Cincinnati, and Pittsburgh, 00 GCT
            1 March 1972.
Figure 4.    Three-day-old emissions from St. Louis, Chicago, Cincinnati, and Pittsburgh, 00 GCT
            1 March 1973.

-------
James G. Edinger and Timothy F. Press are with the University of California, Los
  Angeles, CA 90024.
George C. Holzworth is the EPA Project Officer (see below).
The complete report, entitled  "Meteorological Factors in  the  Formation  of
  Regional Haze," (Order No. PB 83-209 742; Cost: $10.00, subject to change)
  will be available only from:
       National Technical Information Service
       5285 Port Royal Road
       Springfield, VA 22161
       Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
       Environmental Sciences Research Laboratory
       U.S. Environmental Protection Agency
       Research Triangle Park, NC 27711

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
                                                                                                            Fees^aid"
                                                                                                            Environmental
                                                                                                            ProtectTn
                                                                                                            Agency
                                                                                                            EPA 335
Official Business
Penalty for Private Use $300

-------