EPA-AA-TEB-81-9
Summary Report of Several Ambient Carbon Monoxide Studies
                           by
                      Mark Wolcott
                     November  1980
               Test and Evaluation Branch
          Emission Control Technology Division
      Office of Mobile Source Air  Pollution Control
           Office of  Air,  Noise, and  Radiation
          U. S. Environmental Protection  Agency

-------
         Summary Report of Several Ambient Carbon Monoxide Studies

The  Federal Clean  Air Act  assigns  to  the  United  States  Environmental
Protection  Agency  (EPA)  the  responsibility to promulgate National Ambient
Air  Quality   Standards   (NAAQS)  regarding   carbon  monoxide (CO).   On
August 8, 1980  EPA  proposed CO  standards of  25 parts  per  million  (ppm)
maximum allowable one  hour expected  concentration level and 9 ppm maximum
allowable expected  eight  hour average  concentration level  (Reference i).
In an effort to achieve these  standards the  Clean Air Act also assigns to
EPA  the  responsibility to promulgate regulations  regarding  the amount of
CO new highway vehicles may emit from the tailpipe.

To  properly fulfill this  second mandate,  it  is necessary  to understand
the  conditions  associated with  high  ambient  CO  concentrations.   The EPA
Office  of  Mobile   Source  Air Pollution Control  (OMSAPC)  has  conducted
several studies concerning ambient CO  concentrations and the relationship
between those  concentrations and motor  vehicle emissions.   The purpose of
this  report is  to  provide a  single  concise  summary  of  several  such
studies.  Four of the  included studies  were  conducted by OMSAPC.  Anotner
related  study,  sponsored by  EPA Region  X,  is  included in  this  summary
report for  completeness.

The  five  studies  summarized  in  this report, provide  information on recent.
CO emission inventory  estimates, the  meteorological  conditions associated
with  high  ambient   CO concentrations,  the  distribution  of  CO  concen-
trations within  selected  cities, the  effects  on CO concentration levels
of both  meteorological conditions  and  traffic  characteristics at  typical
hot  spot locations  and the estimated  amount  that in-use vehicle emissions
must be reduced to achieve the NAAQS for CO.

This  report presents  a  synopsis of  how  each  study  was conducted,  the
questions addressed by the studies and the major results from each study.

Synopsis of Studies

The  first  study is summarized  in  the  Mobile  Source  Emission  Inventory
Report (Reference 2).  That  study  focuses on  the  development  oc emission
inventory  estimates  for  a  variety  of  mobile  pollution  souices.   The
purpose of  the  study was  to  evaluate the contributions  of various mobile
sources  to  the  total emission  burden.    The  study   was  performed  to
provide one input  into the OMSAPC  Program Assessment  Group's  evaluation
and  direction of mobile source control  programs.   Both  current and future
year (up  to 2005)   carbon monoxide  county inventories were  estimated and
presented in  this  report.  The  estimates were  based on  the  1978 Mobile
Source Emission  Factors  document  (Reference  3)  and  the  1977  National
Emissions Data Systems (Reference 4).   Eighteen  different  types of mobile
sources were  considered.  The  contribution  of  unregulated  as  well  as
presently  regulated   mobile   sources   were   included.    Inventories  for
non-methane hydrocarbons  and oxides of  nitrogen  were also  estimated.   The
sensitivity of  the  emission inventory  estimates to  low ambient  tempera-
tures,  different  stationary  source  control and  retirement rate  assump-
tions,  different driving modes and different mobile  and stationary source
growth  rate assumptions  were   considered.   Finally,   the  results  were
interpreted in terms of their likely air quality effect.

-------
The  purpose  of  the  second  study  (Reference  5)  was  to  determine  the
temporal  and  meteorological  characteristics  that  are  associated  with
deteriorated  air  quality.   The  study  used meteorological  observations
obtained  from the National  Weather Service  (NWS)  for  60  U.S.  cities in
calendar  year 1975-1977.  Corresponding  ambient  concentration  data  were
provided  by  the  EPA Storage  and Retrieval  of Aerometric  Data (SAROAD)
system (Reference  6).   In  general,  sites were ranked and chosen according
to the number of times the  NAAQS  was  exceeded  in the 1975-1977 period.*
The parameters studied included:

              1.  Month of the year
              2.  Hour of the day
              3.  Observation temperature
              4.  Rush hour  temperature
              5.  Pre-rush hour temperature
              6.  Absolute humidity
              7.  Barometric pressure
              8.  Wind speed
              9.  Insolation
              10. Atmospheric stability

Ambient  CO  concentration levels  were  related  to  each  of  these temporal
and meteorological parameters.

EPA Region  X sponsored  the  third  set  of studies  in Seattle,  Washington
and Boise,  Idaho (References  7  and 8)  to find out  how high  CO  concen-
trations  were  distributed  in  those  two  cities.   The  Seattle  study
involved  monitoring  CO concentrations  at 36 outdoor sites,  five  indoor
sites and two pedestrian walking routes.  Data were  collected  for twenty
days  in  the  fourth  quarter  of 1977.   The fourth  quarter  corresponds to
the season  when  high carbon monoxide levels  frequently  occur.   A similar
study  methodology  was  established  for   Boise.    The   purpose  of  these
studies  was  to  determine the  magnitude and spatial  extent  of  the carbon
monoxide  problem  and  the   representativeness  of  the  central  business
districts'  permanent monitors.   Although  this  set  of studies was  not
sponsored by  OMSAPC, they have  been  included  in  the   report  because of
their relevance.
*CO  10 mg/nj3 (8 hour average)
Ozone   160 ug/m^ (1 hour average)

An exceedence  of one  of these levels  is  not necessarily  a  violation of
the  standard.   The  standard may,  for  example, recognize only  one viola-
tion per  day.   If  the concentration  level  was higher  than  the standard
for  an  entire  day,  then  while  there would  be only one  violation,  there
would be 24 exceedences.

-------
EPA also  collected  data in the vicinity  of congested intersections - hot
spots  -  in  another  study  to  find  out  more  about  the  meteorological,
traffic  and  CO  background   conditions   in  specific  areas  that  have  a
history  of high  CO concentrations  (Reference  9).   One  intersection was
chosen  in  each  of  the  four cities:    San  Jose,  Phoenix,   Seattle  and
Chicago.   Vehicle and traffic characteristics,  as well as meteorological
parameters  and  ambient  concentration  levels,  were measured.   Several
techniques were used to  collect  these  data.  Postcard  survey and driver
interviews were used to  collect  vehicle  characteristics  information.   A
traffic  analyzer  and instrumented pace car were  used  to  collect  traffic
speed  and volume data.   Ten CO sampling  devices  were   placed  in  the
vicinity  of each intersection,   some  in close  proximity  to  the inter-
section to measure  the  effects of local  traffic and  some  further  away to
measure  urban  background  concentration levels.   Each of  the vehicle and
meteorological  variables  were related  to  the ambient  CO  level.  Further,
an attempt  was made  to  ascertain the degree to which vehicles operating
in the  immediate  vicinity of the study intersection  were  responsible for
the observed ambient concentrations.

Another  study  was  performed  to  estimate  the  amount  in-use  vehicle
emissions  must be  reduced to achieve  the  NAAQS  for  CO.   This  work  is
described  in  a memorandum entitled "Percent  Reduction Needed  in Mobile
Source  Emissions" memorandum (Reference  10).  Data  from  over 50 of  the
worst CO  problem  cities  in  the  U.S. for the three  year  period 1975-1977
were  included  in the  analysis.   Ambient  CO concentration data  from the
SAROAD  system  were  combined  with meteorological  data from  the National
Weather  Service  and  the   1980 draft mobile  source  emission  factor  data
from EPA.  These  data are the same  as  those that  were used in the second
study of  this  series.   Cities were  chosen  on the  basis of the  number  of
hourly  concentration values  in  one year   that  exceeded   10 mg/nH.   In
general,  the   cities  with   the   greatest   number  of  exceedences   were
included.  The  percent  reduction required  from mobile  sources to  achieve
the  NAAQS  was  calculated as   function  of ambient  temperature.   This
required percent  reduction was estimated  for each  calendar year from 1978
through  1995.   In the  years  that the  expected  percent reduction needed
became zero, no violations of the CO standard were  predicted.

Each of  these  studies was designed  to  address  a different aspect of  the
high  CO  concentration  problem.    Together,  they  attempt  to  answer  the
following questions:

    1.   How are CO emissions distributed  throughout the country?

    2.   What  is the mobile source proportion of total CO emissions?

    3.   During what  seasons and   times  of day   are CO  concentrations
         highest?

    4.   What   meteorological  conditions   are  associated   with  high  CO
         concentrations?

    5.   What   vehicle and traffic  conditions  are  associated with  high
         ambient CO concentrations?

-------
    6.   To  what  extent  are  high CO  concentrations  related  to  activity
         within a few hundred meters of  the  CO monitoring  instruments?

    7.   How  does  the degree of  emission reduction  required  from mobile
         sources to meet  the NAAQS change over time?

    8.   How  does  the degree of  emission reduction  required  from mobile
         sources change with temperature?
The  following  five  sections of  this  report present  the  study results  as
they pertain to these questions.  One  section  is devoted  to each  study.

Mobile Source Emission Inventory Report

To  answer  the  first  two  questions,  the Mobile  Source Emission  Inventory
Report  focused  on the development  of emission  inventory estimates for a
variety of  mobile pollution sources.  In  1977,  the latest year  for which
data were available at the  time  the  report was written,  CO emissions  from
all  sources  totaled  113,715,311  tons.  On  average,  36 thousand  tons  were
emitted in  each  of  the  nations"s  approximately 3200  counties.   In  most
counties,  however,  fewer than 12  thousand tons were  emitted (Figure  1).
Obviously,  the  total  amount of  CO emitted is  only one  aspect  of  the  CO
inventory.   The  distribution  of  emissions  within  counties  is   also
important.  If emissions  are uniformly distributed within counties and  if
meteorological  conditions  are  the same  for  all  counties,  then emission
densities would be better  predictors  of  ambient concentration differences
among counties than would  total  tonnage.   (The density of CO  for a county
is the total amount of CO emitted annually divided by  the county  area.)

Like  the  total annual  CO  emission  levels  shown  in  Figure   1,  Figure 2
shows that  emission  density in  most,  counties  is  comparatively  low.   For
the  majority  of  counties  CO  density  is   less  than  19   tons  per  square
mile.  However,  in  five percent  of  counties  CO density  is  greater  than
230  tons  per  square  mile.  Since the  distribution of CO  emissions within
counties is not. uniform  and meteorological conditions across  counties are
not  constant,  it  is  also  important  to relate  CO  concentrations to popu-
lation density.   Most CO  emissions are caused  by  motor.vehicles and  tend
to  be  concentrated  in  urbanized areas.    This assertion  is  supported  by
Figure 3.   The high  emission  density  counties are  also the  high popu-
lation density  (urbanized)  counties.   The  relationship is significant for
two  reasons.  First,  it  indicates  that a  large  portion of the population
may  be  exposed to harmful CO levels.   Second, since  most  stationary  CO
sources are in   rural  areas,   it  supports  the hypothesis   that  the   CO
emissions  of greatest concern are caused primarily by mobile sources.

Of  the  113  million  tons  of CO  emitted in 1977,   transportation  sources
accounted   for  approximately  85  percent.   Figure 4  shows  the  relative
contribution of  mobile  source emissions  to   total  county emissions  for
each  county.   The   horizontal  axis   indicates  the  percent  of  total
emissions   contributed  by  mobile   sources.    (The  percent  of  emissions
contributed by  stationary  sources  can be  computed  as 100 percent minus

-------
 this number.)  The  vertical  axis indicates the number of counties at  each
 mobile  source  contribution  level.   As  Figure 4 indicates,  mobile sources
 account  for  a  relatively large  portion  of  CO emissions in most  counties.
 In  5  percent of counties, mobile  sources  contribute more than 98 percent
 of  all  the  CO  emitted.  The median  contribution  is 92 percent.  However,
 as  the  portion of  the  curve at the left hand  end of the horizontal  axis
 indicates, there are  a few  counties in  which mobile sources contribute a
 small  percentage of  county  emissions.   In  most  of these  counties  the
 dominant  emissions  sources  are  forest  wild  fires  and  forest managed
 burning.  This  is  particularly  true in  Alaska,  Idaho,  Montana and Puerto
 Rico.   However, there  are  a  few  counties  in  the  southwest  in  which
 chemical  manufacturing is the  dominant  CO source.   These  counties  often
 contain  carbon black  manufacturing plants,  which,  when uncontrolled,  emit
 large amounts of CO.

 In  order to narrow  the  focus  to  the  counties  with  a CO  problem,  143
 counties  showing   violations   of   the   eight  hour  average  NAAQS    were
 analyzed.   (At  the time  the report was written  the  primary CO, standard
 was  10 mg/m .   Also,  the National  Emission  Data  System  data  for   Con-
 necticut was erroneous,  so that  state's  data  were excluded.)  Most of the
 143 counties include  a large urban area.  Figure 5 shows the distribution
 of  total CO emissions  for  these  counties.   As would  be  expected,  total
 average  annual CO emission levels  tend  to be  greater among these counties
 than among  all  the nation's counties.   Similarly,  emission  densities in
 these 143 counties  also  tend to be higher (Figure 6).  Further,  as can be
 seen  from Figure 7,  the  general  association  of  high  emission  densities
 with high population  densities  also prevails  among  the  counties showing
 NAAQS violations.   Finally,  as Figure  8 shows, most  county  CO is emitted
 by  mobile pollution sources.  The general  shape  of Figure  8 follows  that
 of  Figure 4.  However, while in Figure  4 several  counties are represented
 in  which mobile sources contributed a  negligible  portion  of  total  CO,
 among  143 counties of Figure  8 which  represent  areas  with  ambient CO
 violations,   mobile  sources  contributed at  least  37  percent  of  total
 county CO.  The median CO contribution was 95 percent.

 To  summarize, three conclusions can be  drawn  from these  figures:  1) the
 highest  CO  emissions  levels are  found  in the nation's urban  centers;
 2)  in  these areas  mobile   source  emissions   predominate;   and   3)  large
 segments of the population are exposed to high CO emission levels.

 Identification of Meteorological Conditions

 The second  study  summarized  in  this report  focused  on the  temporal  and
meteorological   characteristics   that, are associated with deteriorated CO
 air quality.   This  study  addresses  the third  and  fourth  questions   that
were  previously listed.   Figure  9  shows  the  relationship between CO
 emission  rates  and  ambient  temperature  developed  for  the  draft   1980
mobile  source   emission  factor  document.    The  specific  curves  shown
 represent the  1975-1977  calendar  year  period.    Since  mobile source CO
 emissions tend  to  increase  as temperature  decreases,  the  absolute CO
 ambient  concentration  levels in the  nation's  urban  centers  also  increase
during  cold  weather months.    In Figure 10 the  total number  of  CO  eight
hour moving  averages  greater  than the  10  mg/m^  NAAQS are  displayed  for
 each month of  the  year.   As  indicated,  most  exceedences of  the  standard

-------
occur  during the  winter months.   Figure 11  shows  the  number  of NAAQS
exceedences  as  a  function  of  pre-rush hour  temperature.   Pre-rush hour
temperature was arbitrarily defined as  the average temperature during the
three hours  preceeding  the  rush hour associated with the exceedence.  For
many  vehicles  the  temperature during  this  period  is  the  temperature
during  the  time the vehicle  was parked.  It  is  this temperature that is
related to CO emissions.  For a given length of parked time, CO emissions
tend  to  increase  as ambient  temperature  decreases.   Therefore,  one would
expect  that   an  increased  number   of  exceedences would  occur  at  colder
temperatures.   This would  skew  Figure  11   to  the  left.   Instead,  the
distribution  in Figure  11  is  more  or  less  centered around  50°F.   The
confounding  factor  is  that  there are comparatively few eight hour periods
anywhere  in  the  country during  which  the  average   temperature  is  below
0°F.  Normalizing  the  data  for the total number of hours in each tempera-
ture  interval,  Figure  12 presents  the  fraction of  time  that  the ambient
CO concentration  level  exceeded the NAAQS, given that the ambient temper-
ature was  within  the range  indicated.   Thus, within  a  given  temperature
interval, the likelihood  of a high ambient CO concentration is greater at
low temperatures than at high  temperatures.

The  likelihood  of high  ambient CO concentrations is  also  greater  at low
wind  speeds   and   stable*  atmospheric  conditions  (Figures 13  and  14).
These  are conditions  under  which the  CO  generated  from  motor  vehicle
operation  is not   readily  dispersed.   The   low  wind  speeds  and  stable
conditions occur  most  frequently   during  the late  Fall  than during  the
February  -  March  winter months.   That is  why more  standard  violations
occur during  the  fourth  quarter  of the year  than occur  during  the first-
quarter.  The average  temperature  difference between the two  quarters  is
not great.   However,  there  is  a difference  in the prevalence  of unstable
atmospheric conditions.

Collectively, Figures  10-14  indicate that CO violations occur  most fre-
quently  during  the  October -  January  season and  that,  in general,  the
colder  the  temperature  during low  wind  speed  and stable  atmospheric
conditons, the greater will be  the  likelihood of a NAAQS violation.

The frequency of  CO violations also  changes  during  the  day.  This  varia-
tion  is  associated  with both  traffic  and  meteorological   patterns.   As
Figure 15  shows,  the likelihood of a NAAQS  CO violation  increases  from a
minimum at 6:00 hours (6 a.m.)  to  a maximum  at 17:00 hours  (5  p.m.).  The
hour  indicated  in  the  figure  is the  ending  hour  of  the  eight  hour  moving
average.  Thus,  a  5 p.m.  ending   hour  represents  the 10  a.m.   -  5 p.m.
eight hour average.  Traffic  volume  generally  follows a  bi-modal pattern
with  peaks during  both  the morning and evening rush hour  period.   Since
in most localities  the  morning rush hour  is  over by 10 a.m., one hypothe-
sis  that  may  explain  the  pattern  of  exceedence  frequencies  given  in
Figure  15  is that  the  emissions  from  the morning traffic  rush  mix with
emissions from  vehicles  operating  during the  day.   That  hypothesis  would
explain the  increased  frequency of  a NAAQS  exceedence in  the  1-2 p.m.
period.   The same  sort of hypothesis can be set forth for  the relatively
*Stability classes range from unstable (A) to stable (G).

-------
                                      8

high  likelihood of  an exceedence  during the  late evening  hours.   High
levels  of CO  seem to  disperse slowly,  over a  period of  three  or more
hours.   For  example,  if  this  hypothesis is  correct,  then emissions from
vehicles  operating  after  the evening  rush hour  (8 p.m.) would contribute
to the eight hour average ending at  6 a.m.  the next morning.  (The 8 p.m.
emissions  would contribute  to the  11  p.m.  ambient one  hour  CO concen-
tration  level  and  that  level would then be  averaged with  the levels from
the  subsequent seven  hour   period,  thus forming  the   eight  hour average
ending  at 6 a.m.  the  next  morning.)   If  CO  emissions  dispersed  imme-
diately,  then  one would  expect the  eight  hour average  CO  concentration
level  to fall  to  its  minimum  value  eight hours  after the  close  of the
evening  rush hour, about 2 a.m.

A  second  hypothesis  that  can  be  supported by  Figure  15  is  that  the
citywide  distribution  of high  CO  concentrations  may   be  more  widespread
than  is  sometimes thought.   If this  hypothesis  is correct, the  high  CO
emission  levels orginating at  one  congested  intersection have time to mix
with high  CO emissions  originating at other  congested  intersections.  If,
indeed,  high ambient CO levels  are widely  distributed  over an urban area,
then  all segments  of the urban population are exposed to high  CO levels
and not  just those segments  located in a one  or  two block  radius of a few
congested  intersections.

Seattle, Washington and Boise, Idaho Carbon Monoxide Studies

Evidence  supporting the  proposition  that  high  ambient   CO levels  are
widely distributed  over an  urban  area was  collected in Seattle, Washing-
ton  and -Boise, Idaho.   The  study   results  indicate  that  the  carbon
monoxide   problem  is  widespread  and  not  restricted  to  the  downtown
commercial  district.   Figure   16   shows  the   locations   of  the  study
monitoring  sites  within  Boise  and  the  maximum  8-hour  values   over  the
entire  sampling  period.   The  bar   heights  are   proportional  to  those
maximum  values.  The  central business district is  detailed in  the cutout
section  of Figure  16.   At   several   sites  within  the central  business
district the maximum 8-hour  value  exceeded  9 ppm CO.   (The 9 ppm level  is
approximately  equivalent  to the  10  mg/m^  NAAQS  in   the  effect at  the
time  the study was  conducted.)   However, as Figure 16 shows,  the  NAAQS
exceedences  were  not  restricted  to   the  district.   In several  outlying
areas the maximum 8-hour values also exceed the NAAQS.

To  paraphrase  the  report,   the  spatial extent  of  the Boise CO  problem
encompasses not only the  downtown  commercial district  but  also  locations
along  traffic   corridors  outside  the  core  area.  When   higher  concen-
trations were observed in the  core  area,  elevated  CO levels also occurred
elsewhere.  Altogether  about  70  percent of  the  study sites experienced
one or more days when the eight-hour average exceeded 9.0 ppm.

The study  also found that pedestrians were,  at times,  exposed  to eight-
hour  average  CO concentration  levels  exceeding   the  standard.   This  is
inferred  from   the  fact  that   for   nearly   sequential sampling  periods
totaling seven or more hours,  the  average  exposure was  equal to  or  above
9.0 ppm  on four  days.   Also,  for sampling  periods  between two  to  four
hours, pedestrians were on one occasion exposed to 14 ppm.

-------
This  study,  conducted  by  EPA Region X, also measured CO concentrations  at
indoor  sites  such  as  large  department  stores  and   office  buildings.
Changes in  the  level  of CO concentrations at indoor sites often coincided
with  changes in  CO  concentrations at  outdoor sites.   At  several  sites,
late  afternoon  increases  in indoor CO were observed, suggesting an  influ-
ence  from the afternoon traffic peak.  On one or more days the eight-hour
average CO  concentration  exceeded  9.0 ppm at four of the indoor sites.

Finally,  although the  state monitor  in  Boise is  located  in the central
business  district, the magnitude of Boise carbon monoxide levels measured
as  part  of the study  was somewhat  greater  than indicated  by  the  state
monitor.  On 19 of the 20 study days the eight-hour standard was exceeded
at  one  or  more sites.   Exceedences at the  state  site  occurred  on only  9
days.   Further,  at other  locations,  eight-hour  concentrations  above the
standard  were  up  to  three times  greater than those  measured  at the
permanent  monitor.   Since  control strategies  for  an  area are  based,  in
part, upon  the  levels of  CO measured by that  area's permanent monitoring
network,  it is  important  that  the network  accurately  reflect  the  levels
of  CO concentration  to which the  population is  exposed.  If the measure-
ments made  by the permanent monitor  are  too high,  then expensive control
measures  may be  undertaken  unnecessarily.   On  the other  hand,  if the
measurements are  too  low,  then  the area's population may  be  exposed to  a
significant health risk.

CO  Hot Spot Study

The CO hot  spot  study was designed to characterize the meteorological and
traffic conditions at  congested intersections.   At  each of the four  study
sites,  ambient  CO concentrations  were measured  in  the  immediate vicinity
of  an intersection.    CO  monitors were  also placed  on roof  tops,   alley
ways  and  parks  away  from  the intersection  in  an attempt to measure  back-
ground   concentration    levels.    The   meteorological   parameters  were
measured  within a  few blocks of the study intersection.  These parameters
included  ambient   temperature,  barometric  pressure,  humidity,  windspeed
and  direction  and  insolation.   (Insolation  is  a  measure  of  sunlight
intensity.)  A traffic  counter  was placed on the roadway leading into the
intersection while  an instrumented pace  car monitored  traffic  speed.  As
.vehicles  stopped  for  the  red  light  at   the intersection  itself,  drivers
were  asked  a series  of  questions pertaining to their vehicles  and the
length of time they had been  driving.

Since ten concurrent  CO measurements were made  for  each sampling period,
and  since  at least  a few  of  the  measurements  were made  away  from the
traffic  flow,   the minimum  of those  ten  measurements  was  used  as  an
indicator   of the background  CO   level  for  the  sampling   period.   The
difference   between  the  maximum  and  minimum  values  was  used  as  an
indicator  of the  CO  concentration attributable  to traffic at  the inter-
section.  The difference  was then  divided by the maximum value to express
the local  contribution as a  percentage.  Thus,  a high  local  contribution
indicates  that  vehicles  at  the  study  intersection  are  responsible for
most  of  the measured  CO.   Conversely,  a  low  local  contribution   level
indicates  that  vehicles  operating throughout  the urban area are respon-
sible.   If  vehicles  removed from the  intersection  contribute  signifi-
cantly  to  the   CO concentrations at  the  intersection,   then   CO   is   a
mesoscale,  not a microscale, problem.

-------
                                     10

In San Jose  the  eight-hour average CO NAAQS  was  exceeded 10 times  during
the study week.   On all 10 occasions the sampler recording  the highest CO
level was  directly downwind of the  intersection.   Average wind speed was
4  knots;  traffic  volume  was high.   The local contribution during  these
periods  -  as  we have  defined it  -  ranged   from  62  to 98  percent and
averaged 80  percent.

The  local  contribution during the NAAQS  exceedences  in  Seattle  ranged
from  36  to  76 percent  and averaged 60  percent.   Average wind  speed was
2.8 knots.   Twice in Seattle high  CO  concentrations were recorded  by all
of the sampling  devices in the immediate vicinity of the intersection and
by  the  "background" devices  situated  away   from  the  intersection.   In
these  cases   it  is  probable that  the  high  CO  concentrations were not
representative of  a typical "hot  spot"  situation  but were widespread and
not restricted to the  immediate study area.   These particular violations
were  likely  to  be  the  result of  traffic throughout  a  fairly  wide  area.
Three  other  exceedences   in  Seattle  were   associated   with  high   local
contributions and followed the pattern seen in San Jose.

Four  exceedences of  the  NAAQS occurred during  the  test  week  spent  in
Phoenix.   High  CO  concentrations  were   recorded  simultaneously  by many
sampling devices  there, including  those  positioned  to measure background
concentrations.   Since  the  NAAQS  exceedences occurred  during  eight hour
periods that ended  between  1  and 3  a.m.,  the local contribution  averaged
only  35  percent.   Thus,  the Phoenix  location  can  not  be described as a
true hot spot, at least during the  study  period.

In Chicago  there were  two  eight-hour  NAAQS  exceedences  during  the  study
week.   They  occurred after  periods  of heavy traffic  and were associated
with relatively high local contributions  - from 79 to  97  percent.

The  study   findings  indicate  that  the  local  CO  contribution   varies
directly  with  wind  speed  and  insolation  and  varies  inversely  with
barometric pressure.   It  seems  that  high  wind  and insolation,  and low
barometric pressure are conducive   to  higher  CO dispersion  rates.   And a
high rate of CO  dispersion means  that CO from  adjacent traffic corridors
tends to be  diluted  before it reaches the study intersection.

Traffic  conditions  also  influence  the  local contribution level.   The
level varies directly  with  the  proportion of  vehicles  that  are  started
after having been parked  for  several hours.   It  also varies  directly with
the total  number of vehicles passing through  the  intersection  and with
the  age  of  those  vehicles.  Under  each  of  these  conditions,  vehicle
emissions tend to  increase so that the  greater  the total emission  level,
the greater  will be  the  local contribution.   This  is  a typical  CO hot
spot situation.

High CO  concentrations, at  times,  were  also recorded  at all  monitoring
locations.    This  phenomenon was observed most  frequently in  Phoenix and
Seattle.   On several occasions when the  highest  CO concentration recorded
was greater  than 10 mg/m ,  the local  contribution was  between  15 and 40
percent.   Such a low local  contribution indicates  that  several  segments
of the  urban area  experienced high concentration  levels simultaneously.

-------
                                      11

On such occasions,  the  CO problem is not strictly a  hot  spot problem but
is a mesoscale problem affecting  the entire urban area.

Percent Reduction Needed  in Mobile Source Emissions

The  purpose  of each  of  the studies discussed  so  far has  been  to gather
information about  the conditions associated with high  ambient  CO concen-
trations  in  an  attempt  to determine  the  level  of  emissions  reduction
needed to  meet  the CO National  Ambient  Air  Quality  Standard.  Currently,
many  areas of the  country do not  attain  the NAAQS.  With the  principal
exception  of  the  inspection and  maintenance  strategy, the Federal  Motor
Vehicle  Control  Program has  been  applied   to  new  vehicles.   Thus,  as
older,  generally  dirtier  vehicles   leave  the vehicle  pool, the  average
level of  emissions  has declined.  As emission  rates  continue to decline,
so  will   ambient  CO  concentration  levels.    However,  the  reduction  in
average emissions  and the resulting improvement in  air quality  will take
time.

In  the  final study  being summarized the  concept  of  a percent  reduction
needed from mobile  source emissions was constructed  to estimate when and
how  many  of  the  nation's  counties will meet the  NAAQS.  The  percent
reduction  needed  is defined in  this work as  the  amount  by  which  mobile
source  emissions  must  be  reduced   in  order  to  eliminate   NAAQS  exceed-
ences.  As an example,  reducing a  40  mg/m^ ambient  concentration  to 10
mg/nH  requires a  75 percent  emission  reduction.    If  the maximum  CO
concentration  measured   anywhere  in   the   country   is  40  mg/nP,   then
reducing  emissions  throughout  the country by 75 percent would  eliminate
all  exceedences.    Reducing  emissions   by   less  than  75   percent  would
eliminate  only a portion  of the exceedences.

Figure  17 shows  the  percent  by which  the  highest  eight  hour  average
ambient CO levels  must be reduced in order  to meet  the CO NAAQS  in each
calendar  year.   Figure  17  assumes   that FTP  conditions prevail,  that  no
inspection maintenance  program is instituted  and  that VMT  growth  is two
percent.    Data  from over 50 cities  with the  highest  1975-1977  CO concen-
trations  were  included  in  the analysis.  The  dotted line indicates the
level  at  which 100 percent of   ambient CO  standard exceedences  will  be
eliminated.  The solid  line  indicates  the level at which only  90 percent
of exceedences will  be eliminated.   For example, to  eliminate  99 percent
of the  total  number  of  eight  hour   periods that  presently   exceed the  CO
NAAQS, each eight  hour average would need to be reduced  by 51  percent in
1982 and by 32 percent by 1987.

Several  assumptions  concerning  vehicle miles  traveled   and  stationary
source  impact  underlie  the construction  of  Figure  17.    Vehicle  miles
traveled by mobile  sources are  expected to  grow at  a  two  percent annual
compound rate.  Stationary  source emissions  are assumed not  to  influence
urban  concentrations,  but  to  completely disperse  before  reaching  urban
centers.    The  CO  emission  factors   for  the  entire  highway  vehicle  fleet
without inspection maintenance are derived from MOBILE2 (Reference 11).

Two points should  be  made with  respect  to  Figure  17.  First, it  appears
that 90  percent  of NAAQS  CO  exceedences will be eliminated around  1985.
(The logic here  is that  if no  further reduction is  needed to meet  the

-------
                                     12

standard, then the  standard  must have been  met.)   Elimination of 90 per-
cent of  the  CO exceedences is  a major  achievement.  However, eliminating
90  percent   of  the  standard  exceedences  still  leaves  many eight  hour
periods  each  year  during which  the  public  is exposed to unhealthy levels
of  carbon  monoxide.   The  remaining  10  percent   of   exceedences  would
generally occur  in  large  urban areas where  a  significant  segment of the
population would  be exposed.   Second,  Figure  17   suggests  that,  at least
one  percent  of  the exceedences  will never  be eliminated,  since the  99
percent  line  begins to  turn  upward in 1995.   (Inspection  maintenance  or
other  control  programs  not  currently in  place would have  the  effect  of
lowering each of the curves in  Figure 17.)

It  should be remembered while  interpreting Figure, 17  that  it represents
only one set of  conditions.    Other  sets of  conditions would  result  in
different curves.    Specifically,  as we  found  out in  previous   studies,
both  emissions  and ambient  concentrations  tend  to  increase   at  lower
temperatures.   Changing   the   assumptions  concerning   stationary  source
influence, vehicle  miles  traveled,. vehicle  control programs  and driving
conditions would  also  result  in  different  curves.  (Temperature  was  of
prime concern in these  studies because CO emission rates  are very sensi-
tive to  temperature  changes.)

The  70  -  75°F  temperature  interval was  used to construct Figure  17.
Figure 18 displays  the  percent reduction needed from mobile sources  as a
function of  temperature.   Individual vehicle emissions  increase at colder
temperatures.  At  extremely  cold  temperatures,  however,   other   factors,
such  as  increased  wind   speed  and  decreased   atmospheric  stability,
disperse the greater per vehicle emissions.  As a  result,  a smaller  per-
centage  reduction  in mobile  source  CO  emissions is  required  at.   tempera-
tures below  0°F  than at 0°F.   (At  the  temperature  extremes comparatively
few data points determine  the position  of the curves.   Fairbanks, Alaska,
for  example,  accounts  for most  of the  very  low  temperatures.   Caution
should therefore be  used before making  any inferences from these  sections
of the Figure 18 curves.)

Both the 90  and  95 percent  curves  of Figure  18   show  that the  greatest
level  of control  is needed  in  the 0  -  5°F  temperature  interval.   A
decreasing level of control  is needed  over  the 0  - 30°F  interval.   From
30 - 75°F an  increasing  level  of  control  is  needed.    Finally,  from
75 - 100°F a decreasing level of control is  needed.  The  required levels
of control are summarized  in Table 1.

The initial  70 - 75°F  interval was chosen as one aspect of study in this
analysis  to  be  consistent  with  the  Federal  Test  Procedure.   Motor
vehicles are tested  in  that  program at  a  temperature within the 68 - 86°F
range.    However, since  the likelihood of a  standard exceedence  increases
as  temperature  decreases  and  since  the  percent reduction  required  from
motor  vehicles  is  greater at.  colder  temperatures, Figure  19   was  con-
structed with the  zero  to five degree temperature  range.   In this colder
range,  the elimination of  90  percent  of the  NAAQS exceedences will noi:  be
achieved under  the  study  assumptions.  Around  1995 growth  in  the number
of  vehicles  increases  total   emissions  more than   the   reduction   in
emissions from each  individual  vehicle  decreases total  emissions.  Beyond
1995  sufficiently   more  vehicles  enter  the  vehicle  pool  to more  than
compensate for the reduced emissions from each of them.

-------
                                      13
Conclusion

Each  of  these  studies  has  addressed  a different  aspect  of  the  high
ambient CO concentration problem.   The  National  Emission Inventory report
showed that the  highest  CO concentration levels  are found in the nation's
urban  centers,  that in  these areas  mobile  source  emissions predominate
and that  large segments  of the population are  exposed to high CO emission
levels.   The  meteorological  study  showed  that  the likelihood  of  a  CO
National Ambient Air Quality  Standard exceedence tends  to increase at low
temperatures,   low  wind speeds  and  stable atmospheric  conditions.   These
conditions are  most prevalent  during the late  fall months.   Within any
season,  the   likelihood  of an exceedence also  changes  during  the  day,
reaching a maximum during  the evening rush hours.   Evidence from the EPA
Region X  Seattle,  Washington  and Boise,  Idaho studies  indicates  that the
citywide distribution  of high CO concentrations may  be  more uniform than
has sometimes  been thought.   The  study  results  indicate  that  the carbon
monoxide problem is widespread and not  restricted  to downtown commercial
districts.  Further,  both  pedestrians and  those individuals  indoors are
at  times  exposed  to  eight hour  average CO  concentrations  exceeding the
standard.  The  "hot   spot"  study  characterized the  meteorological  and
traffic   conditions   at   congested   intersections.    Under  conditions
conducive to  high  dispersion  rates,  the  local contribution  to  CO concen-
tration  levels  tends  to  be   high.   Under  conditions  conducive to  low
dispersion  rates,  the  local  contribution  tends   to   be   low.   At  low
dispersion rates,  the CO  problem  may be more of  an areawide  than  a hot
spot problem.   The  percent  reduction  needed study  focused on when and how
many of  the nation's  counties will meet  the NAAQS.   That study indicated
that 90  percent  of warm weather standard exceedences would  be  eliminated
by  1985.  However,  during  cold   weather,  when  the   likelihood  of  an
exceedence is  greater, an  additional 20  percent emissions  reduction from
mobile  sources  beyond that  which  is  already   accounted  for  by current
control measures will  be required  to  eliminate even  90  percent  of current
exceedences.

-------
                                  Distribution of  CO Emissions For All Counties
       FIGURE t
                                                                                FIGURE 2
 goo

 720

 610

SS60
 320
  160

  80
                                                                            1000

                                                                             900

                                                                             800

                                                                          u! 700

                                                                          \ 600
                                                                          o
                                                                          ° 500
                                                                          u.
                                                                          o
                                                                          c 100
                                                                          u
                                                                          r 300
                                                                          z
                                                                             200

                                                                             100
      <5     <20    <80    <320   
-------
                                Distribution of CO  Emissions  For Worst  Counties
  FIGURE  S  •
                                                                          FIGURE 6
ns.o
uo.o
35. 0
-30.0
x
o2S.O
o20.0
|.5.0
a
Z10.0
s.o
0.0






•
	 ^n




^™*









^^























 IO    <160   <610   <2S60
      TOTflU CO  EMISSIONS 11000 TONS)
27
2M
221
1"
Sis
u.
o
UJ
" 9
3
6
3
0

'•




__
n

-


-


—






—




~



















n







In
                                                                          0













__
1 1
_r-innl 1
•••

































































—

















IS
tu
i
o
°IO
Ik
0

flC
UJ
A
X
x S



0





















'






._/
I ir
h W
. AAA AAA AA/W
















<100   <100   <1600   <6>IOO  <25600  <80000
    <200    <800   <3200  <13800  
-------
                           FIGURE 9  MOTOR VEHICLE EMISSION TEMPERRTURE EFFECTS
                     155
                     ISO
                     125
                    £120
                    §110
                      95
                    S BO
                    I 75
                    •" 70
                    o
                    u 65

                      60
                      55
                               \
                                                                       LDV CO
                                                                       EMISSION
                                                                       FflCTOR

                                                                       MOBILE
                                                                       SOURCE FLEET
                                                                       CO EMISSION
                                                                       FRCTOR
                                10  20   SO
                                              MO   50   BO  70
                                              TEMPERRTURE  IF)
                                                                80   90  100  MO
 en oo


 5600


I "4800


! uooo
i
i

, 3200


' 2UOO
I

I 1600


  800
FIGURE 10 EXCEEOEHCES OF TME 8 HOUR CO NRROS
                             &

JRN.    MRR.    MRT   JULT  SEPT.   NOV.
   FEB.   RPR.   JUNE    RUG.   OCT.   DEC.
                  MONTH
                                                                   FIGURE 11  EXCEEOENCE3  OF  THE 8 HOUR CO HRR03
                                                              2800.
                                                                      -65  -US  -25  -5   15   35   55   75   95
                                                                        -55  -35  -15   S   25   145   65   85  105
                                                                             PRE-RUSH  HOUR  TEMPERRTURE  (F)

-------
                                         Likelihood of a 8  hour  CO NAAQS Exceedeiice
     ricunt 12
    -65  -US  -25   -5   IS   35   55   75   95

      -SS  -35  -IS  5   25   US  65   85   IDS

          PRE-RUSH HOUR  TEMPERATURE  (F>
                                                                          .12


                                                                          .11


                                                                          .10


                                                                          .09


                                                                          .08


                                                                          .07


                                                                          .06


                                                                          .05


                                                                          .04


                                                                          .03


                                                                          .02


                                                                          .01


                                                                          .0
                                                                             FIGURE 13
0    4    8    12   16   20   24    28
   2     6    10    14    18    22   26   30
            WIND  SPEED  (XT)
  .18



  .14



g.12
e
x.

~.10
u
X

^ .08
ul
Cl

* .06
o
UJ
u

x -011
UJ


  .02



  .0
     FIGURE 14
                                                                             FIGURE  IS
             8    C     D     E     F
              ATMOSPHERIC STABILITY
.UOU
.055
.050

S .045
O
" .040
u
5 .035
u .030
£.025
w
S.020
X
".OIS
.010
.005
MI







|
^
^
^
0
1






i.
i
i
2
]






bull
\\i\\ii
4 6 8 10
S 7 9 II
MO





1
1
2

Jfl




I
\
1
S




ii

II
M
i|
1 \
'$
ii
«
IS


IP
03 ^
j|

H
! 1
i I
Ii
( f
li
6 1
17



^
Mi
f
s 1 1
I
ill
III
t 20
19




J
I
1
\

M




»
1
|
1
I
12
23


-------
                                                             Figure 16
   Scale in PPM CO
                                                                                                                      T
  Denotes 20-Day Sites (11/25/77)

1 Denotes First 10-Day Sites (11/25/77-12/8/77)
na
I Denotes Last 10-Day Sites (12/9/77 - 12/22/77)


U Black Denotes Values Greater than 9 PPM CO


-------
                           19
       Reduction Needed In  Mobile Source CO  Emissions
   flGURE 17
  80
HJ70
060
u
o

wSO



2»o
>-
u

§30
5"
u
u

SlO
o o
     1978  1980  1982  1981  1986  1988  1990 1992  1994  1996

                 VMT  CROMTMi 2X  70-7SF  NO 1/M
   FIGURE 18
  80
  70
  60
§50

t-
u
u
•

i-SO

u
u
Sto
      -so
              -20       10      40      70

                 PRE-HUSH  HOUR TEHPERRTURE
                                             100
                                                     130
                                          (Fl
   FIGURE  19
w70
a

uSO




§«0
 20
  lO
O
u o
     1978 1980  1982  198U  1986  1988  1990  1992  1994  1996

                  VHT GROHTHi  2X  0-SF  NO 1/M

-------
                                         20
                                  Table 1

                 Level  of  Control  Required  to  Meet  CO NAAQS
Temperature
Range (F)
0-5
25-30
70-75
95-100
Exceedences
90%
51(a)
36
42
23
of CO NAAQS
95%
56
41
50
24
Eliminated
99%
64
51
64
24
(a) Percent  reduction  in ambient CO concentrations  required  to eliminate
90% of  NAAQS  exceedences in  the  0-5°F  temperature  range.    This  is  a
reduction  from the  emissions  levels  observed in  the  1975-1977  vehicle
fleet.
Source:   Reference 10

-------
                                    21
                                 References

1.  Federal  Register,  Part IV,  "Environmental Protection  Agency,  Carbon
    Monoxide;  Proposed  Revisions  to  the  National  Ambient Air  Quality
    Standards  and  Announcement  of Public  Meetings," Monday,  August  18,
    1980.

2.  M.  Wolcott,  Mobile  Source  Emission  Inventory Report,  U.S.  Environ-
    mental Protection Agency, Discussion Draft, May,  1980.

3.  Mobile  Source  Emission  Factors:   Final Document,   EPA 400/9-78-006,
    U.S. Environmental Protection Agency, Washington, D.C., March, 1978.

4.  1977 National  Emissions Report, EPA-450/4-80-005,  U.S. Envii-onmental
    Protection  Agency,   Research  Triangle  Park,   North  Carolina,  March,
    1980.

5 •  Identification  of  Meteorological  Condition's^  During Oxidant, CO  and
    Nitroge	Dioxide  Ambient  Violation  Conditions,  U.S.  Environmental
    Protection Agency,  Discussion Draft,  April, 1979.

6.  Storajye^ and Retrieval  of  Aerometric  Data, U.S.  Environmental Protec-
    tion Agency, Research Triangle Park,  North Carolina.

7.  Carbon Monoxide  Study, Seattle, Washington,   EPA 910/9-78-054,  U.S.
    Environmental  Protection  Agency,   Region  X,  Seattle,  Washington,
    December, 1978.

8.  Carbon Monoxide  Study,  Boise, Idaho, EPA  910/9-78-055, U.S.  Environ-
    mental Protection  Agency, Region  X,  Seattle, Washington,  December,
    1978.

9.  Analysis  of  Pollutant  and  Meteorological  Data  Collected  in  the
    Vicinity of Carbon Monoxide "HoJ^ Spots",  Stanford Research  Institute,
    International,  Discussion Draft, May, 1979.

10. M.  Wolcott,  "Percent  Reduction  Needed  in  Mobile  Sources",   U.S
    Environmental  Protection  Agency,   personal   communication  to   Karl
    Hellman,  October, 1980.

11. MOBILE2 is a highway mobile  source emissions  computer  model  based  on
    data released in November, 1980.

-------