SHORT-TERM N00 STANDARDS

                   VOLUME I

             SOURCE ASSESSMENT AND

           PROBLEM CHARACTERIZATION
                 DRAFT REPORT
                    DRAFT
              DO NOT MUTE OR CITE
                Submitted to:

  Office  of Air Quality Planning and Standards
         Environmental Protection Agency
  Research Triangle Park, North Carolina   27711
                Submitted by:

Dale L.  Keyes,  Bharat Kumar,  Robert D.  Coleman
             and Robert 0. Reid
     Energy and Environmental Analysis,  Inc.
        1111 North 19th Street, 6th Floor
            Arlington, Virainia  22209
                December 26, 197

-------
                       TABLE OF CONTENTS
       TITLE                      .                       PAGE
PART A:  INTRODUCTION	 . .	   1

PART B :  GENERAL BACKGROUND	   2
         1.   Emission Sources	   2
         2.   Ambient NO- Formation	   3
         3.   Current Status of N02 Control	:	   5

PART C:  SHORT-TERM N02	   10
         ,1.   Point Sources	:	   10
         2.   Area Sources	   45
         3.   Alternative Ways to Express the
              Short-Term Standard	   55
APPENDIX A	   57

-------
                            STATEMENT
     This Draft Report is furnished to the Environmental Protec-
tion Agency by Energy and Environmental Analysis, Inc., Arling-
ton, Virginia.  The contents of the report are reproduced herein
as received from the contractor.  The opinions, findings, and
conclusions expressed are those of the authors and not necessari-
ly those of the Environmental Protection Agency.

-------
                          VOLUME I

                SOURCE ASSESSMENT AND PROBLEM
                      CHARACTERIZATION
A.   Introduction

     This report is the first volume of a three-volume report
which attempts to identify the conditions leading to high
short-term concentrations of nitrogen dioxide  (N0_) and. the  types
of sources likely to cause these levels, and to determine the
approximate additional cost of controlling emissions to a
level consistent with attainment of the various ambient air
quality standards under consideration.

     This volume deals with the sources of NO  emissions, the '
                                             Ji
mechanisms by which nitrogen oxide  (NO) is converted to N02
and ambient data on observed NO  and N0~ concentrations.
                               Ji       £•
Volume II provides a preliminary assessment of the sources
which, through the mechanisms identified in Volume I, might
cause or contribute to high short-term  (i.e., one-hour averaging
period) concentrations of NO-.  Once the sources were identified,
control strategies were developed and used to estimate control
costs.  Volume III is a detailed case study of the interactive
impact of multiple point and area sources on the short-term
NO- concentrations in the Chicago AQCR.  The results of the
case study provide both a basis for the assumptions made in
the nationwide analysis, and a means of gauging the degree of
over- or underestimation of control cost impacts.

     The formation of N02 in the atmosphere is a highly complex
process.  NO, formed primarily (i.e., 98 percent) as a result of

                            -1-

-------
                           -2-
fuel combustion, can be converted to NO- through oxidation
with either atmospheric oxygen or oxidizing agents such.as
ozone.  The degree to which NO is converted to N02 in any
particular location is dependent on many variables.

     Since N02 is a largely derived pollutant, some way of
capturing the reaction kinetics had to be devised.  The approach
used here is totally empirical.  The maximum NO- concentrations
for typical point sources were estimated from NO  modelling
                                                a
results and NO -to-NO_ conversion curves derived from actual
              *i      £•
plume sampling.  Alternative approaches have also been pro-
posed.  Cole suggests setting maximum NO- equal to the ambient
                                      I/
0, level plus 0.1 times the NOV level. '  This appears to work
 j                            X
well for summer conditions, but should underestimate NO- under
conditions of low 0- concentrations characteristic of winter.

     High N02 concentrations associated with area sources are
primarily the result of high NO  emission levels and can occur
                               X
with or without high ozone levels.  Emissions from these
sources should be dispersed fairly evenly over a given geographic
area.  Data taken from the Washington Metropolitan Area and
other studies appear to support a peak  (second highest hour)
to mean (annual average) ratio of 6:1 or less for ambient NO-.
Specific one-hour observations greater than six times the
annual average are believed to represent the influence of
point sources.

B.   General Background

     1.   Emission Sources

     By far the most significant source of NO  emissions is
                                             Ji
the combustion of fuel.  High temperatures and rapid mixing,
which accompany fuel combustion, are conducive to the oxidation
of both atmospheric and fuel-bound nitrogen.  The predominant

-------
                          -3-
oxidation product is nitrogen oxide  (NO), with small  amounts  of
nitrogen dioxide  (N02), and much smaller amounts of other oxides
(N_0-; etc.) also formed.  Collectively, these are known as
nitrogen oxides  (NO ).
                   j£
     Table 1 summarizes the estimated nationwide NO   emissions
                                                   J^
in 1976.  Of the estimated 23 million metric tons, mobile
sources accounted for  44 percent and stationary source  fuel
combustion for 51 percent.  The remaining five percent  were due
to industrial processes  (three percent), solid waste  disposal
(0.5 percent), and miscellaneous sources, such as forest fires
(one percent).  The mobile source category was dominated by
highway vehicles, primarily autos and small trucks.   Stationary
source fuel combustion emissions were divided between electric
utilities (56 percent) and industrial fuel combustion sources,
such as boilers, process heaters, furnaces, and kilns (38 per-
cent) .

     2.   Ambient N02 Formation

     As noted above, nitrogen oxidized  during fuel combustion is
emitted primarily as NO.  Only about five percent ' of  the NO
                                                              j±
in the combustion products is in the form of N0_.  In general,
the oxidation of NO to NO- in the atmosphere occurs by  two
mechanisms:   (1) oxidation of NO to N02 in the presence of
atmospheric oxygen, mixed into the plume; and (2) oxidation of
NO to N02 in the presence of oxidizing  agents such as ozone,
hydroxyl radicals, or organic peroxy radicals, again  introduced
as the plume mixes with the ambient air.

     Chemically, the first mechanism is:

-------
                              -4-
                             TABLE  1
              NATIONWIDE EMISSION ESTIMATES,  1976

                       (10  metric tons/year)
        Source Category                           NO     Percent
                                                    x    ~~^"""^™~~^~


Transportation                                    10.1       44

  •  Highway Vehicles                             7.8       34
  •  Non-Highway Vehicles                         2.3       10

Stationary Fuel Combustion                        11.8       51

  •  Electric Utilities                           6.6       27
  •  Industrial                                   4.5       20
  •  Residential, Commercial and
       Institutional           ..                  0.7        4


Industrial Processes                              0.7        2

     Chemicals                                    0.3        1
     Petroleum Refining                           0.3        1
     Metals                                       0          0
     Mineral Products                             0.1        0
     Oil & Gas Production and Marketing           0          0
     Industrial Organic Solvent Use               0          0
     Other Processes                              0          0


Solid Waste                                       0.1        0

Miscellaneous                                     0.3        1

  •  Forest Wildfires and Managed Burning         0.2        1
  •  Agricultural Burning                         0          0
  •  Coal Refuse Burning                          0.1        0
  •  Structural Fires                             0          0
  •  Miscellaneous Organic Solvent Use            0          0
TOTAL                                            23.0      100
SOURCE:  Reference 3.

-------
                          -5-
In this reaction, the rate of N02 production is proportional to
the square of nitric oxide concentration.  Therefore, nitrogen
dioxide production by this mechanism is immediate for NO concen-
trations greater than about 100 ppm and significant 0- concen-
trations.  These high concentrations only occur very near the
point of exhaust.  As the nitric oxide is diluted to concentra-
tions below 100 ppm, the rate of conversion decreases to a point
where at 1 ppm, the direct reaction with oxygen becomes unim-
portant.  It is believed that the conversion of NO to NO-
through this mechanism is limited to roughly ten percent of the
initial NO  concentrations.  However, for high initial NO con-
          a
centrations and high excess air or moderate levels of mixing,
Calvert suggests that this mechanism could result in an initial
                                  A/
ratio of NO, to NO  of 25 percent. '
           £      *t
     Oxidizing agents, such as ozone and hydroxyl radicals,
result in the rapid conversion of NO to NO-.  The degree of
conversion, under certain conditions, is directly proportional
to the level of oxidants present.  In the absence of photochem-
ical activity, NO- formation due to ozone is on a one-to-one
basis.  The same does not seem to be true in the presence of
ultra-violet light accompanied by photochemically reactive
precursors which result in an equilibrium set of reactions
between NO- and ozone.
      NO
+ 0- (or other oxidizing agents) 	»-NO- + 0-
                       N02  hv -MO+ 0
                  (hv   =   ultra-violet light)

-------
                          -6-
During days when UV radiation levels are high, the afternoon
levels of N02 are set by these competing reactions.  The forma-
tion of organic nitrates .from oxygenated hydrocarbon compounds
also serves to diminish the concentration of NO-.

     Meteorological conditions also play an important role.
Weather conditions obviously determine the amount of sunlight
available to drive the photochemical reactions.  In addition,
mixing depth and wind speed will influence the degree of atmo-
spheric dispersion.

     The various factors affecting the formation of ambient NO-
can be narrowed down to:

     (1)  NO  emission levels;
            Jt
     (2)  Meteorological conditions:  atmospheric
          stability and wind speed;
     (3)  Ambient levels of ozone, hydroxyl radicals,
          and organic peroxy radicals; and
     (4)  Solar ultra-violet radiation and temperature.

     The degree to which the above factors affect the NO- and
NO  relationship is discussed in detail under Section C.I.a.
  X
     3.   Current Status of NO- Control

     Nitrogen dioxide is one of the original six criteria pollu-
tants for which EPA National Ambient Air Quality Standards were
promulgated in 1971.  Based on the health and welfare effects
information available at that time, an annual standard was set
at 100 yg/m  (annual average) to protect the population against
the effects of long-term low-level NO- concentrations.

     In 1975, only seven AQCR's recorded annual N02 averages
above the annual standard.  However, many AQCR's do not have
sufficient data on NO- concentrations to determine a valid

-------
                           -7-
annual average.  Thus, the  actual number  exceeding  the  annual
standard could be greater.*

     Monitored NO- concentrations for  sites  with  several  years
                                       6/
of data do not show  consistent  trends.  '   In the  Los  Angeles
Basin, the annual average of hourly  data  and the  annual average
of daily one-hour maximum increased  from  1971 to  1973 and de-
creased thereafter.   In  Chicago, NO- concentrations have  fluctu-
ated since 1969 and  no clear trend is  discernible.  New Jersey
and Denver NO, concentrations increased except during the fuel
shortage in  1973 and 1974.

     Since only a few areas have shown violations of  the  current
NO- standard, most state implementation plans (SIP) limit NO
  ft                                                         X
emission regulations  for "stationary  sources  to large  utility and
industrial boilers and nitric acid plants.

     Most of the emphasis on reducing NO  emissions has come
                                        ji
from the Federal motor vehicle control program and  the develop-
ment of new source performance standards for  selected stationary
sources.   Table 2 shows  the emission limits  imposed by these
programs.   Interest in NO  emission  controls  is increasing, and
                         X>
subsequent sections of this report will detail  the  available
controls for stationary  sources.

     Since nitrogen oxide contributes to the  formation of ozone
and other oxidants,  the  control of NO  emissions  has been used
                                     a
by some states, noticeably California, to reduce  levels of
photochemical smog.   However,  the chemistry of  oxidant formation
*0ne recent study ' concluded that there is greater than 50 per-
 cent chance that 34 AQCR's would be in nonattainment status if
 sufficient data were available.

-------
                            -8-
                         Table 2
               FEDERAL LIMITATIONS FOR NO
Vehicle Exhaust Emission Standards (non-California)
1.  Light-duty vehicles

        Model year

           1973-76
           1977-80
           1981+

2.  Light-duty trucks
No  Emission Standard
    3.0 gin/mi.
    2.0 gm/mi.
    1.0 gm/mi.
        a.  Vehicle weight less than 6,000  Ibs.
        Model year
No  Emission Standard
           1975-78
          *1979-84
          *1985+
    3.1 gm/mi.
    2.3 gm/mi.
    1.4 gm/mi.
          b.  Vehicle weight 6001-8500 Ibs.
        Model year

           Pre-1979
           1979-82
          *1983-84
          *1985+
3.  Heavy-duty gasoline vehicles

        Model year

           1974-78
           1979-84
           1985+

4.  Heavy-duty diesel vehicles

        Model year

           1974-78
           1979-82

          *1983+
No  Emission Standard
    Same as HDG vehicles
    2.3 gm/mi.
    2.3 gm/mi.
    1.4 gm/mi.
                                  Emission Standard

                                 15.3 gm/mi.
                                 13.3 gm/mi.
                                   5.35  gm/mi
                              HC plus NO  Emission Standard
                                 16 g/bhp-hr.
                                 1.5 g HC and 10g. NO
                                 5gHC plus NO
                                 Same as HDG Vehicles
                      x
                        or

-------
                           -9-
5.  Motorcycles

        Model year           NO  Emission Standard

          *1985                  0.14 g/km

Stationary Source Emission Limits

     1.  Fossil-fuel fired steam generators

            Coal-fired boilers   0.7 Ibs. of NO /lo!j Btu
            Oil-fired boilers    0.3 Ibs. of NO;Y10; Btu
            Gas-fired boilers    0.2 Ibs. of N0^/10b Btu

     2.  Gas turbines            0.3 Ibs. of NO^IO6 Btu

     3.  Nitric acid plants      3 Ibs. of NO /ton of 100% acid
*  Predicted standards


Sources:  References 8, 9 and 18.

-------
                          -10-
is not well characterized at the present time, and the effec-
tiveness of such a strategy is open to question.  Evidence from
smog chamber experiments indicates that reducing NO  emissions
                                                   X
may actually increase 0^ concentrations if the initial ratio of
nonmethane hydrocarbons to NO  (both measured as ambient con-
centrations) is low.  '  Empirical studies in Los Angeles, on
the other hand, indicate that levels of 0, downwind of Los
Angeles are correlated with NO  emissions.

C.   Short-Term NO,,

     This section reviews selected ambient monitoring data
around stationary sources and in the vicinity of area sources.
Diurnal, seasonal, and spatial variations in the NO and NO-
concentrations at these sites are discussed.  Finally, the re-
lationship between the NO, and NO  levels is further examined.
                         ^       X
     1.   Point Sources

          a.   Observed N02 and NO  Concentrations

     The observations reported in this section are based on
maximum hourly NO  and NO- concentrations measured at continuous
monitoring stations around selected American Electric Power
Service Corporation (AEP) plants.  '   In addition, continuous
NO  data in the vicinity of Potomac Electric Power Company
                             127
(PEPCO) plants were analyzed.  /

     Continuous NO  data at each of these sites have been col-
lected using the chemiluminescent measuring technique.  The
monitoring stations were sited to measure the maximum ground-
level impact of nonreactive emissions from the sources based on
diffusion modelling.  Simultaneous NO- measurements were not
made at the PEPCO sites.

-------
                          -11-
     Figure 1 shows the location of AEP plants and the monitor-
ing network around each power plant cluster.  Stations where NO-
and NO  were measured are marked NO .   (The location of these
      X                            X
sites is approximate.  Appendix A contains detailed maps of the
plants and the monitoring sites around them.)  Figures 2,  3, and
4 show the location of PEPCO plants and the respective NO
monitoring sites.  Although the power plant in each of these
areas is the largest source of NO  in the immediate area,  there
generally are other sources which can contribute to the observed
NO  and NO, levels.
  X       £»
               1)   Diurnal Variations

     Figures 5 through 10 show the diurnal variation in the NO,
N0_, and NO  concentrations observed in the vicinity of two AEP
plants for different periods of the year.  The values  in  these
curves are averages  for each hour over  a period  of  three  months.

     It is difficult to generalize the diurnal variations  in the
ground-level concentrations of a pollutant on any given day.
But with concentrations for an hour of the day averaged over an
extended period, anomalies are smoothed and the emerging trends
can then be generalized.  A pronounced NO  bulge in the mid-
morning, significantly higher than that in the evening, can be
seen in almost all these curves.  Another interesting observa-
tion is the occurrence of N02 bulges, less pronounced than those
for NO , but again, in the mid-morning and evening hours.  Even-
      X
ing peaks in this case are slightly higher than the morning
levels.  Overall, NO  appears to be considerably more variable
than H02.

     The highest hourly NO- and NO  concentrations at these
sites have been further analyzed to check for any deviation in
their occurrence from the general findings discussed above.

-------
                                        FIGURE 1
                                   AEP MONITORING SYSTEM
     UOHI10IIHO ItATIOMI
     flLCPllOMI Llllll
Forl Wayne o
               /n d la n a

                      o Indianapolis
 IIKEEI)
 ROCKPOR
           NO,
                                         .NO.,
                                                                             o Canton
                                O/iio

                          Columbus o
                                         PIIILO
                                                              GAVIN
                                                           Kff« Creek
                                                            IDVIC)
                                Kentucky
                                                      BIQ SANDY
                                                               UUSKIHGUM
                                                                 NOy
                                                                           Cliarleslon
                                                     West
                                                  Virgini a
                                                                        ro
                                                                        I
SOURCE;  American Electric  Power Service Company

-------
                            -13-
                            FIGOEE 2
  NO,, MONITORING SITE LOCATION IN THE DICKERSON AREA,
    A                                           '
MARYLAND
                                    LILYPONS

                                    (THOMAS)
                          BEALLS^

                         (HUNTER
SOURCE;   Potomac Electric  Power Company

-------
                                    -14-
                                  FIGORE 3
             NO  MONITORING SITE LOCATION IN THE MORGANTOWN AREA
               2£


                               (PEPCO) MARYLAND
                                                                      SHILOH

                                                                      OULBY
C1NG OROROE

  COUNTY
                                          •>r/    /
                                      MOIltiANTOWN
 U..S. NAV.M.



tlOVINC; GHOLIND
                                                                      O'DEE
                                                                      (LLOYD)
    mile
      SOURCE;   Potomac Electric Power Company

-------
                                 -15-

                                 flGUKE 4
   NO  MONITORING SITE LOCATION IN THE CHALK POINT AREA -
                                                               MARYLAND
                                                       .   .

                                             Ann Arundel  County
  Prtnce fieorges  County
\!i '
'»\ CHESAPEAKE1 BAY
                                           Calvert County
                                  Lower Marlboro
    CTarles County     ^f *m

           ' AQUASGO (Jdv)-'""
                                          X Stoakley
                                                   Frederick
                      4 oi   «  C
                            Benedict«l^ERIDAN KT,  (WATSON)
                                  * .'f~' \ \
                           VICINITY OF CHALK POINT

                           POWER GENERATING STATION
SOURCE:   Potomac  Electric Power Comoanv

-------
                                -16-

                                  FIGURE 5


          AVERAGE DIURNAL VARIATIONS IN THE AMBIENT NO, N02 AND NOX
                   CONCENTRATIONS FROM STATIONARY SOURCES
     o.oo
         GAV!W.KYG£R CREEK.SPQRM
             LAKIM C33
         C  1   3   S  «  C  0  T  C   8  tO  II  IS II 1C  IV  16  IT  1C 19  IO  *t t* »•
         KOJ3  &  DAY      CCTQ8GR   !S7S  - DSCCHfccR  IS76
         HOURLY AVERAGE   K32  K3 AFO WUX COCEKTRAT1QK CPPHJ
   X  ••«•
         GAVIN.KY6E1? CREEK.
             LAKIK
   S  «•••


   |  ••-
   hJ

   §  «•*»
      *.«


s

^
--*





mm***
^




^^
**.
"-*




X
NO
1MMM
NO




---'
r..



x

,^^


I
1
A
7


ox

&




\
Sv
**"S




\






immmm
• •





mm^mm
^•VM





— *
— .





•**.






. i*






--^






^•W
— —




/
>
/*




•*s

•«^.















NO:
NO
















i

-------
                                -17-
                                   FIGTIKE 6


       AVERAGE DIURNAL VARIATIONS IN THE AMBIENT NO, NO2 AND NOX

                CONCENTRATIONS FROM STATIONARY SOURCES
  „     GAVIN.KYGER CREEK.SPORN
  r o.w    LAKIN (33
  Q.
  Q»
  £_  O.O4
  U
  O  o.o»
  X
  o
     O.Oi
  a
  . <    -
  O  o.o«
  (N
  2  O.O
                 I
                                         13 ||  |4  It |*  17
        HOUR OF DAY      APRIL     1977  -  JUNE     1977
        HOURLY AVERAGE   N02  NO AND NOX CONCENTRATION CPPM3
NDX
SOURCE:   American Electric Power Service  Corporation

-------
                                   -18-
                                      FIGORE 7



            AVERAGE DIURNAL VARIATIONS  IN THE AMBIENT NO, N02  AND NOX


                     CONCENTRATIONS FROM STATIONARY  SOURCES
     C.
     a
     QC
     >-

     IU
     u
     2'
     C
     U

     X
     o
     2
                   CREEK
        e.o,     PETFRSSURG C1D
        e.et
   o.ei
        O.OJ
        e.et
            O   I
                                7  (  «  10 11  11  II  14  IS 1C IT  If  I*  10 II  II  II
           HOUR 0? DAY      SEPTEMBER  1S7G  - KJVEM3ER   1976
           HOURLY AVERAGE   N02  «0 AMD KOX CONCENTRATION  CPPH)
     &
     0.
§
                   CREEK
        e.o«    EL12ASETHTOW C5)
        O.OT
     >  0.04

     UJ
     ^
     O  e.e.
     x
        o.os
     O  0.0

                              &*

           HOUR OP DAY       SEPTCHSER 1S76  - WOVEHSCR  1876
           HOURLY AVERAGE    W02  NO MQ NOX COKCENTRAT10N CPPM3
SOURCE;   American  Electric Power  Service Corporation

-------
                                 -19-
                                      FIGDRE 8


           AVERAGE DIURNAL VARIATIONS IN THE AMBIENT NO,  N02 AND NO

                    CONCENTRATIONS PROM STATIONARY  SOURCES
          TAWCRS C3SEK
       •••K
    I
    o ...
    C3

    §

    S
                         C13
          raua o?
           !•  it ta u  i«  ic  u if te  i* *o at
          IS78  - PggftlKRy  IS77
K32  W3 A*Ci FOX COPCEKTR/VTIOM CPWJ
                S CRSEX
           »•"  I   *  •  *  9   «   7  •  » IO tl  IS  !• 1» 10 !•  II1  IS 19 ti»  Ct ti! »•
           KXC3  C?  OAT      DETCEWSS?   1S7S  - r5B^!/*.RY  1S7T
           KUm-Y AV^RAC-   M02 N3 AM3  K3K CCSCS4TRKTIQN (f.^-n
SOURCE;   American  Electric  Power Service  Corporation

-------
                                -20-
                                 FIGURE 9



       AVERAGE DIURNAL VARIATIONS IN THE AMBIENT NO, N02 AND NOX

                CONCENTRATIONS FROM STATIONARY SOURCES
        TANNERS CREEK
           PETERSBURG CD
 u
 o
 u
 g
    «.ot
    0.04
0.0»
    o.oa
 O  «.0|
          WOX
        o   i
                                   10 It  It  It  14 l»  It  17 |*  t»  ft -II  li  11
       HOUR 0" DAY      KARCH     1377   - MAY       1977
       HOURLY AVERAGE   N02  NO AND NOX  CONCENTRATION CPPM)
 _     TANNERS CREEK
 S  •.o«;7   ELIZABETHTO^N C53
KQX CONCENTRATION (PP
? ? f ?
3 3 t S
S ••«•
A*
2 «.o
(•
e



•u «»





— "^





"^^^k


i i



S,






—
/


/
f
^^^^
/
/




/


J^J

^


!OX

22
,'0
\



V
V
^V
\
\



\
^^



\
>
	 j

*«»,







1*^fc



























\^









****
-




^=*
-^-





^
'





^











~*~



OU3 OF DAY Kf^CH 1S77 - KAY 1S7T
OJRLY AVftlRAIsZ K02 K3 AfO NOR CONCENTRATION CPPM3
SOURCE:   American Electric Power  Service Corporation

-------
                               -21-
                                       10
      AVERAGE DIURNAL VARIATIONS IN THE AMBIENT NO, N02 AND NOX

               CONCENTRATIONS  FROM STATIONARY SOURCES
~  8.89
a.
u.
 U*
 tl

               CRSHK
           PSTSRS5UR3 C13
                                                                     TZ^J
       HOUR OF DAY
       HOURLY AVSRAIS
                                  1S77  - AUGUST    1ST7
                       N02  NO AND NOX CONCSHTRATION  CPPM3
 a.
 a.
 a
 S
       TANKERS
           ELZZASETHTDKN  £53
 n  e.es
 ut
 u
    t.ts
        i;  t   2  3
       HOJR  OF DAY
       HSJKLY  AVERAGE
                                 &  u)  tt  i£ si i-» 10  te  iv  tw
                       JUKE      1S77 - AU8UST    1ST?
                       KC2  NO  AND NOX CONCENTRATION  CPPM3
                                                                Jai  U  ii
SOURCS:   American Electric  Power  Service  Corporation

-------
                           -22-
Tables 3 through 5 shew the five highest one-hour WO, and NO
                                                    ^       j£
concentrations observed over a period of one year at four AEP
monitoring sites influenced by NO  emissions from power plants
                                 X
and the date and hour of their occurrence.

     As can be seen, the NO, and NO  maxima at these sites are
                           ^       X
in general agreement with the monthly average diurnal curves.
That is, NO, peaks occur in the evenings and NO  peaks in the
           b                                   J±
mid-morning.  A closer look at the NO, maxima reveals several
peaks as well, sooner than the consistent evening bulge shown in
the average diurnal curves.

     The meteorological conditions and the NO, conversion mecha-
nisms responsible for these trends are briefly outlined here and
are examined in more detail in the sections that follow.

     The high NO  concentrations occurring in the morning can be
                X
understood in terms of fumigation conditions such as inversion
break up or plume trapping.  But the NO, concentrations in the
morning are mainly governed by the initial oxidation of NO,
i.e., the direct oxygenation of NO.  As can be seen from the
diurnal curves, the NO,/NO  ratio for this time of the day is at
                      ^   X
its lowest.

     The observed peak N02 excursions in the afternoon, upon
analysis of the corresponding meteorological data, were found to
be associated with unstable atmospheric conditions.  These
conditions result in occasional high ground-level NO  concentra-
                                                    Ji
tions in the afternoon.  The high NO, concentrations are a
result of both high NO  levels and a greater fraction of NO in
                      X
the plume converted to NO, ,  The latter is driven by good  atmo-
spheric mixing , which  creates good conditions  for oxidation by
both atmospheric oxygen and oxidants.  The highest NO,/NO
                                                     ^   X
ratios are found in the late afternoon or evening, with the

-------
                                                 TABLE 3


                                   MAXIMUM NO  AND NO  CONCENTRATIONS AT
                                            ft       X
                                       AEP - CARDINAL-TIDD PLANTS

                                       MONITORING SITE:  POWER (6)
      Date

July 30,  1976

July 30,  1976

October 1,  1976

March 9,  1977

January 5,  1977
Hour Ending

    13

    12

    19

    20

    17
  5 Highest
  1-hr.  NO
Concentrations
    (ppm)	

    0.180

    0.128

    0.087

    0.079

    0.079
                                                                                                   5 Highest      ,
     Date

February 23,  1977

February 23,  1977

December 15,  1977

March 10, 1977

July 30, 1977

Hour Ending
9
10
9
8
13
1-hr. NOX
Concentrations
(ppm)
0.352
0.319
0.295
0.246
0.229
    SOURCE;   Data from American  Electric Power Service Corporation

-------
                                           TABLE 4


                             MAXIMUM NO  AND NO  CONCENTRATIONS AT
                                       ft      X
                                   AEP - TANNERS CREEK PLANT

                               MONITORING SITE:  PETERSBURG (1)
                                    5 Highest
                                    1-hr.  NO
                                  Concentrations
  5 Highest
  1-hr.  NOX
Concentrations
Date
., 1977
15, 1977
:3, 1977
i, 1977
ir 3, 1976
Hour Ending
16
13
14
12
14
(ppm)
0.087
0.074
0.066
0.062
0.062
Date
March 4, 1977
May 10, 1977
October 30, 1976
December 5 , 1975
February 19, 1977
Hour Ending
14
11
9
11
6
(ppm)
0.23
0.142
0.126
0.119
0.114
SOURCE:   Data  from American  Electric  Power  Service  Corporation

-------
                                           TABLE 5


                            MAXIMUM NO  AND NO  CONCENTRATIONS AT

                            AEP - GAVIN, HYGER CREEK,  SPORN PLANTS

                                  MONITORING SITE:  LAKIN  (3)
5 Highest
1-hr.  NO^
                                                                                           5 Highest
Date
February 27, 1977
October 12, 1977
March 10, 1977
December 1, 1977
April 20, 1977
Hour Ending
3
14
21
20
13
Concentrations
(ppm)
0.075
0.069
0.069
0.062
0.062
Date
February 27, 1977
December 15, 1976
November 2, 1976
October 12, 1976
March 10, 1976
Hour Ending
3
11
13
14
9
Concentration!
(ppm)
0.278
0.245
0.189
0.189
0.172
3 ^
1




SOURCE;   Data  from American Electric Power Service  Corporation

-------
                          -26-
upper limit to this ratio determined by the level of oxidants
.and the photochemical steady state of N02.

     From this analysis, the following conclusions emerge:

     •    NO  and NO, maxima occur during different
            *V       £»
          periods of the day; whereas the NO  maxima
                                            X
          show a trend of mid-morning excursions,
          the N02 peaks occur in the afternoon-
          evening most frequently.

     •    Diurnal variations in NO  concentrations
                                  Ji
          are much more pronounced than those in NO- levels.

     •    The NO,/NO  ratio varies from a low in
                <£   a
          mid-morning to a high in the early evening.

               2)   Seasonal Variation

     Figures 11 through 16 show plots of the monthly highest
one-hour NO  and NO, concentrations at the AEP and PEPCO moni-
           X       ^
toring sites.  The following observations have been made:

     •    NO  concentrations show significant seasonal
            X
          trends, with higher peak values occurring
          in the colder months; and
     •    No consistent seasonal variation is observed
          with respect to NO, concentrations.

               3)   Background Levels of NO,

     The AEP Rockport station was a pre-construction monitoring
site and thus provides data on background levels.  The diurnal
variation with hourly values averaged over a quarter and the
monthly peak one-hour NO  and NO, values are shown in Figure 17,
                        X       £,

-------
                                -27-
                                 FIGOKE 11
                          AEP -  TANKER CREEK PLANT
0.3 •-
0.2 -^
o.r--
                                                   PETERSBERG
                                                   MONITORING SITE (1)
        Sept. Oct.  Nov.
        1976
0.3 ••
0.2 "
0.1 "
Dec.  Jan.   Feb.  Max. April  May June   July  Aug.
      1977
                                                   MONITORING SITE (5)
         Sent. Oct.  Nov.  Dec.   Jan.  Feb.  Mar.  April May   Jnne  July  Aug.
         1976                    1977

                ' -  MONTHLY HIC3EST 1-HOUR NOjj AND NO2  LEVELS

SOURCE;  Data from American Electric Power Service  Corporation

-------
                                     -28-
     0.31
2-
04
     0.2'
     O.-lf
                                     FIGURE 12
                       AEP - GAVIN,  KYGER CREEK, SPORN PLANTS
                                                       MONITORING SITE (3)
                                                                         NO.
            July  Aug.   Sept.  Oct. Nov.  Dec.   Jan.  Feb.  Mar. April  May  June

            1976                                1977
X    0.2 "
cu
cu
     0.1 •
                                                       MONITORING SITE (7)
   SOURCE:
July  Aug.  Sept.  Oct.  Nov.   Dec.  Jan.  Feb.  Mar.  April

1976                              1977




               MONTHLY HIGHEST 1-HOUR NOX AND N02 LEVELS




 Data  from American Electric  Power Service Corporation

-------
                                 -29-

                              FIGOHE  13
                       AEP - CARDUCa/TIDD PLANTS
   0.3
   0.1- •
                                                      DTOUSTRIAL AREA SITE (2)
                                                                             NO,
                                                                           -t"
   0.3...
ST  0.2 t
dl
   0.1"
            May June  July   Aug.  Sept. Oct.  Nov.  Dec.  Jem.   Feb.  Mar.  April
            1976                                          1977
                      trox
                                                                   INDUSTRIAL A5EA
                                                                     SITE (6)
         ""May  June   July  Aug   Sept. Oct. Nov.  Dee.    Jan.  Feb. Mar.  April
           1976                                            1977

                       MONTHLY. HIGHEST 1-HODR NOX AND NO2 LSVEI£

  SOURCE;   Data  from  American Electric  Power Service Corporation

-------
                                   -30-
                                        FIGORE 14
                                     AEP - AMOS  PLANT
£
0.
0.1!
                                                      SITE: FISHER RIDGE  (2)
                                                                  NO-
            Sept. Oct.   Nov.  Dec.  Jan.  Feb.   Mar.  April May  June
            1976
      0-
            NO
                                    AEP - MITCHELL S RAMMER PLANTS

                                                      SITE: MCCLAIN  RDN  (5)
            Nov.  Dec.   Jan.  Feb.  Mar.   Apr.  May    June  July
                              MONTHLY HIGHEST 1-HOUR NOX AND N02 LEVELS

   SOURCE;   Data from American Electric Power  Service Corporation

-------
    0.1. "
z
    0.05 •'
                                   -31-
                                     FIGURE 15
                                PEPCO-- CHALK POINT
Total Capacity i  1260  MW

      3 Stacks:  400 Ft.
                400 Ft..
                700 Ft.
                                            MAXIMUM 1-HOOR CONCENTRATION
             Mar. April May    June-  July  Aug.  Sept. Oct.  Nov.  Dec.
                   Higfaea
    0:20 r
    0.15 "
*   0.1  f
a.
    0.05
                                PEPCO'- DIOCERSON  PLANT
                                                       Total Capacity: 540 MW

                                                             2 Stacks: 400 Ft.
                                                                       400 Ft.
                                            MAXIMUM  1-HOUR CONCENTRATION
             Mar. April  May   June  July   Aug.  Sept. Oct.  Nov.  Dec.

                              MONTHLY HIGHEST  1-HOUR NOX LEVELS


   -°URCE:   Data fr°m Potomac Electric Power Company

-------
                                 -32-
                                     FIGOFE 16
                                PEPCO -  MORGANTOVm
 o.i
 0.05 T
                                                    Total Capacity: 1000 MW

                                                         2 Stacks: 700 Ft.
                                                                   700 Ft.
          Mar.  Apr.  May  June  July   Aug. Sept.  Oct.  Nov.  Dec.
                             MONTHLY HIGHEST 1-HODR NOX LEVELS
SOURCE;   Data from Potomac Electric  Power Company

-------
   0.-1--
I
o.
                                    -33-


                                       P1GURE  17


                                    SOCK PORT SHE
                                                       MONITORING SITE  (1)
             t     t	1	j	1	1     <	1	1	1	'	k—
           May   June  July  . Aug. Sect. Oct.  Nov.   Dec.   Jan.  Feb.  Mar.  Apr.
           1976                                       -    1977
—  0-.2-..
   O.I--
cu
cu
                                                       MONITORING SITE  (2)
            •4-
            May  June  July  Aug.  Sept. Oct.   Nov.   Dec.   Jan.  Feb.
            1976                                           1977

                      MONTHLY HIGHEST 1-HOOR NCX AND N02  LEVELS

  SOURCE;   Data  fron  Potomac Electric  Power Company
Mar.   Acr.

-------
                           -34-
      Peak  NO-  concentrations  at  this  site  averaged around 80
     3                                                 3
 yg/m  and  have been  observed  to  be  as high as  140  yg/m .   Annual
 average  NO-  concentrations were  on  the order of 20 yg/m  .
 However, there are several small power plants  west and southwest
 of  these receptors within a distance  of 15 to  30 miles,  thus
 making these values  suspect as true measures of natural  back-
 ground NO-.
b.   Modeled Relationships Between NO- and
                                     £»
                                                      NO
      Simple  approaches  have  recently  been  used to translate the
 ambient  concentrations  of  NO due  to point  sources,  as  estimated
 by  dispersion modelling, into NO- levels.   One of these ap-
 proaches assumes  that the  maximum NO- observed at any  receptor
 in  a  field about  a  point source approximates  a fixed fraction of
 maximum  NO   of  any  receptor, irrespective  of  the various factors
          Ji
 affecting the formation of NO-.   The  conversion factor is de-
 rived from either the ratio  of measured  NO- to NO  at  a receptor
                                           ^      X
 influenced by a point source,  or  by relating  estimated NO
                                                          J^
 concentrations  at a receptor to the NO-  concentrations measured
 there.   This approach may  be sufficient  for estimating the
 maximum  impact  of individual point sources at a single location.
 But the  degree  to which multiple  sources impact a receptor
~v"aTie~s~with  plume travel time to  the  receptor.   The N00/N0
                                                       ^   H
 relationship in a plume changes continuously, especially during
 the first 30 to 60  minutes after  plume discharge.  It  is, there-
 fore,  not valid to  use  a fixed translation factor while assess-
 ing the  air  quality due to several interactive point sources.

      A second approach  for estimating the  NO- concentrations
 from  isolated point sources  suggests  that, except for  the ini- .
 tial  conversion of  NO to NO- (about ten  percent)  due to thermal
 oxidation, further  conversion of  NO is limited by the  ambient

-------
                          -35-
levels of ozone.  The NO. concentration at a receptor due to an
isolated source is expressed, therefore, by the relationship
below:

     NO- .      =         0.1 N0v        +         0.         (1)
       ^                       X                   O
(concentrations at  (concentrations at   (concentrations in the
 a receptor, ppra)     a receptor, ppm)        vicinity, ppm)

The above approach has several limitations.  High ambient N02
concentrations due to point sources are common in winter when
the ambient 0, levels are at their lowest.  This certainly
downplays the role of ozone as the only limiting factor in the
conversion of NO to NO-.  It also seems likely that N02 forma-
tion from thermal oxidation accounts for a major portion of NO-
concentration when ground-level NO  maxima occur during poor
                                  X
atmospheric mixing and/or short travel times.  Under these
conditions, the conversion of NO in the plume due to ozone is
limited by the extent of plume mixing with the surrounding
atmosphere rather than by the ambient level of 0_.

     The initial conversion of HO to NO- could be higher than
ten percent of the initial NO  levels.  It is, for the most
                             Ji
part, a function of the NO  concentrations in the stack gas.
                          a
The conversion of NO to NO- in the flue gas before exhaust is
not limited because of oxygen deficiency,  but due to unfavor-
ably high temperatures.  The concentration of oxygen in most
flue gases is in the order of 30,000, while the NO  levels range
                                                  a
only up to 1,000 ppm.   The NO to N02 conversion can be explained
by the following two-step reaction mechanism:

          NO   +    02  	^ N03                            (2)

          N03  +    NO     *"" 2N02                           (3)

-------
                          -36-
The effect of high temperature on reaction  (3) is to decrease
the equilibrium concentration of NO,, thereby limiting N02
formation.  Once the plume discharges into the atmosphere,
lowering of temperature and the continued availability of addi-
tional oxygen results in rapid conversion of NO to N02 at a rate
proportional to the square of NO concentration.  As mentioned
earlier, at NO concentrations above 100 ppm, the oxidation of NO
by oxygen is reported to be rapid.  But it slows down to a less
significant rate as the NO concentrations reach 1 ppm.  Plume
centerline NO- concentrations in the range of 25 percent of the
NO  concentrations within one to two minutes of plume discharge
  a
have been reported in the Four-Corners Power Plant plume stud-
ies.   "/

     As a plume disperses further, or its time travel increases,
the rate of NO conversion is controlled by the level of oxi-
dants.  But given sufficient time, almost all NO in a plume
should be converted to N02 due to recurring encounters with
fresh parcels of ozone-(or other oxidant-)containing air.  The
depletion of ozone is complete (as suggested by equation  (1))
during the absence of ultra-violet radiation.  In the presence
of sunlight, however, a photochemical equilibrium between NO,
NO-, and ozone is established.  Thus, the use of the highest
daily 0- concentrations in equation (1) is not justified since
they occur during periods when the solar UV-flux is maximum—
total conversion of NO to NO- is implied at the same time that
photochemical dissociation of N02 is maximized.  Further, the
presence of other oxidizing radicals in the summer, as well as
in the winter, are not accounted for by the measured 0., concen-
trations.

     We may be able to predict the order of magnitude of N0~
concentrations due to isolated point sources in the summer using

-------
                           -37-
this approach.  And assuming little seasonal variation in the
peak NO- concentrations, the predicted NO. concentrations could
be considered a "worst case" estimate of the impact of a point
source during a year.  The approach has been tested only on a
limited basis and its correlation with observed N0_ concentra-
tions is yet to be shown.  Furthermore, this approach is only
suggested for application to isolated sources; it cannot assess
the interactive impact of multiple sources as it does not ac-
count for the variation in the NO,/NO  ratio with travel time in
                                 j£   X
a plume.

     A third approach and the one suggested here involves pre-
dicting the NO-/NO  relationship in individual plumes at differ-
              "   J»
ent points in time for a given set of meteorological conditions
and background oxidant levels.  The relationships are developed
for those meteorological conditions and for oxidant levels at
which highest NO- concentrations,, due to point sources, have been
observed.  Simultaneous measurements of NO, N0~, and 0, across
power plant plumes in different parts of the country have re-
cently been made.  '  ' The time for 50 percent conversion
of NO to NO, in plumes with varying initial NO  concentrations,
           ^                                  X
and different meteorological conditions and background ozone
concentrations, has been found to range between 10 to 60 min-
utes.  The range of values can be explained in terms of the
observed variations in the ambient ozone concentrations, solar
ultra-violet radiation, wind speed, atmospheric stability, and
initial NO  concentrations in the plume.
          J£
     The general shape of the curve relating N02/W0  to plune
travel time based on plume studies is as shown in Figure 18.  As
discussed earlier, the rate of conversion during the first one
to two minutes of plume discharge is dependent on the initial NO

-------
                              -38-
                             FIGURE  18
                THE RELATIONSHIP OF NO  AND NO.  IN
                                      x        2

            A POINT SOURCE PLUME AS A FUNCTION OF TIME
        0.9
NO,
  i

NO"
0.5
       0.05
                              B
Curve B reflects

a higher initial

NO., concentration
                           60
                                  120
       180
                                        Minutes
  SOURCE;   Energy and Environmental Analysis, Inc.

-------
                          -39-
concentration in the plume.*   It has been  generally  accepted
that further conversion of NO  to NO. in a  plume is controlled  by
the rate of mixing of the plume with the ambient air, rather
than by the kinetic rate constant of the reaction of NO with
                         147
ambient ozone (oxidants).  '   Given sufficient time, almost  all
NO in a plume should be converted to NO. with continuous  in-
fusion of fresh oxidant-containirig air into the plume as  it
travels.

     Expressed in mathematical terms:

     NO. concentration at time t = f (initial NO  concentration,
       «                                        Ji
     plume diffusion, ambient  oxidant concentrations, solar  UV-
     flux, and plume travel time t)

For a given set of meteorological conditions and ambient  oxidant
concentrations, the NO-/NO  ratio at time  t in individual plumes
                      ^   X                               -~
can be assumed to be mainly a  function of  the initial NO  con-
                                                        4V
centration and diffusion characterized by  travel time:

         /jV\     =    f  (NOX);.
         01077  t            x if t                         (4)
         V  A /   w

     In Figure 18, curve B approximates the impact of higher 110
                                        -""" ™ •" *—                 jt
emissions on the NO,/NO  relationship.  Initial NO to NO-
                   4b   X                                fc
conversion is faster since the rate is proportional to the
square of NO concentration.  But later on  as the plume diffuses,
the amount of NO converted to NO- is limited by the rate  of
mixing with the surrounding atmosphere.  Allowing sufficient
*Under normal diffusion conditions, NO  concentrations in most
 power plant plumes should drop below 100 ppm levels within one
 to two minutes of discharge.

-------
                           -40-
time, almost all the NO in a plume is converted to NO- by
repeated mixing with fresh parcels of oxidant-containing air.
In an urban atmosphere characterized by high levels of reactive
hydrocarbons, however, competing reactions will limit the
resulting concentrations of NO-.

     Lack of sufficient NO,, and NO  data for individual plumes
                          £»       rf»
under similar meteorological conditions did not allow the
specification of separate N02 conversion curves for various
initial NO  concentrations.  Instead, a base curve for an
          Ji
initial NO  concentration of SOOppm was derived from power
          H        *l *3  T A f
plant plume studies  '   
-------
                           -41-


                           FIGURE 19

          N02/N0x RELATIONSHIP IN POINT SOURCE PLUMES
           WITH DIFFERENT INITIAL NOx CONCENTRATIONS
0.0
20
                                      I
                                     40
I
50
                               30
                               TIME IN MINUTES

SOURCE:   Energy and Environmental Analysis,  Inc.
T
60
70

-------
                        -42-
     Further conversion of NO, as the plume travel time increases,
is considered to be mainly dependent on plume diffusion (mixing).
For a certain level of ambient 03 concentration and under
similar meteorological conditions, only a fixed amount of NO is
assumed to convert to N02 per unit time.  This amount was
estimated directly from the base curve and added to the N02
levels estimated after five minutes, to generate the other
curves.

     Because of time constraints, a composite equation express-
ing all curves could not be developed.  Instead, the specific
curves were used to represent five ranges of initial NO  con-
                                                       Jt
centrations.  A piece-wise linear approximation of each was
used for modeling purposes.

     A limited test of this approach on power plants with
short-term ambient N02 data showed good correlation between
predicted and observed NO- concentrations.

     The reactive photochemical models presently being developed
are sinply too sophisticated and expensive to use for studies
involving a large number of sources.  The approach presented
here is simple and theoretically  justifiable.  For the limited
use for which it was intended (i.e., to predict the N02 concen-
trations from point sources under adverse meteorological condi-
tions) , it adequately represents two of the important variables
(the initial NO  concentrations and the travel time)  affecting
               X
the N02 concentrations.

     Constraints on time and lack of experimental data on point
source plumes prevented proper validation of these curves.
Further validation and generalization of  the approach is intended
as more ambient N02 data around point sources and data on point
source plumes become available.

-------
                           -43-
      The generalizability of the curves in Figure 19 is a
 function of the range of environmental conditions prevailing
 during the plume studies (on which the curves are based).  In
 general, the plumes studied were influenced by a fairly good
 mixing environment (typical summer afternoon condition) and 03
 concentrations  around  0.10 ppm.  For conditions considerably
 different than these (e.g., morning inversion break-up or high
 0-  levels), the shape of the curves may be altered.

           c.   Expected Point Source Impacts

      Several model plants  (sources of NO  emissions) were de-
                                         •A
 veloped to estimate the types of point sources which might be
 expected to produce high short-term HO, concentrations.  These
 plants  were modeled using  PTMAX to estimate maximum NO .
                                                        Ji
'(Details of  specific point  source modelling appear  in Volume  II
of this  report.)  The overall results are  shown  in  Table  6.

      As can be seen from this table, the facilities most likely
 to produce high N02 levels are characterized by a  combination
 of short stacks and low exhaust flows and/or multiple  sources
 at a single plant.  These combinations are likely  to be char-
 acteristic of large industrial facilities such as  iron and
 steel mills, petroleum refineries, or large numbers of industrial
 boilers at a single plant.

      Power plants, due to higher stacks and high flue gas ex-
 haust rates, are not likely to cause high levels of short-term
 N02.  The exception to this could occur where the plume inter-
 sects  a hill or valley  wall nearby.   Although terrain  im-
 paction typically  reduces plume  transport time and  thus the
 degree of  NO/NO- conversion,  high  NO   values  can lead  to
                •                    Jt
 significant levels  of N02.

-------
                                                 -44-
                                                  TABLE  6


                 ESTIMATES OF MAXIMUM AMBIENT NO,  CONCENTRATIONS FOR SINGLE STACK FOSSIL FUEL  USER MODEL PUNTS1
PLANT SIZE(MW(e)]
(Fuel Type2)
2000 (c)
2000 (o)
2000 (g)
400 (c)
400 (o)
400 (g)
200 (c)
200 (o)
200 (g)
100 (e)
100 (o)
100 (g)
50 (c)
50 (o)
50 (g)
20 (c)
20 (o)
20 (g)
10 (c)
10 (o)
10 (g)
5 (c)
.5 (o)
5 (g)
0.1 (c)
0.1 (o)
• 0.1 (g)
STACK HEIGHT [ft]
Ranqe of
Estimates
430-660
430-660
430-660
200-430
200-430
200-430
110-430
110-430
110-430
110-200
110-200
110-200
6S-200
65-110
65-110
30-110
30-110
30-110
30-110
30-110
30-110
30-110
30-65
30-65
30-65
30-65
30-65
Best Estimate
660
430
430
430
200
200
430
200
200
200
110
110
200
110
110
110
110
110
110
65
65
£5
65
65
65
30
30
STACK TEMPERATURE [°F]
Range of
Estimates
280-330
280-330
280-330
280-330
280-330
280-330
280-330
280-330
280-330
280-530
280-530
280-530
280-530
280-530
280-530
280-530
280-650
280-650
280-650
280-800
280-800
280-800
280-800
280-600
280-800
280-800
280-800
Best Estimate
300
300
300
300
300
300
300
300
300
330
330
330
330
350
350
400
470
470
470
530
530
530
530
530
530
530
530
MAXIMUM NO, CONCENTRATION fug/m3!
Range of
Estimates
130-220
85-140
75-130
110-270
85-250
80-230
110-350
40-220
40-180
70-300
30-140
20-75
80-360
90-250
50-130
110-290
30-330
15-180
20-180
20-200
10-50
20-140
30-150
10-50
5-30
20-60
5-15
Best Estimates
150.
130
120
130
220
210
130
120
110
120
120
60
110
130
70
160
35
20
50
80
20
45
50
15
10
40
10
1.  Assuming:
                (1) Constant ratio of flow rate to Btu's fired developed from combustion calculations, assuming
                   1 MM(e) equivalent to 10 million Btu/hr,  for each fuel type as  follows:
                     (a)
                         1.00 srn^/s per MU(e) for moderate sulfur bituminous coal (-721 C. 4.51  H, 51 0?,
                         1.5S NZ, 2X S. 61 H20,  & 9S ash) burned with 22X excess air (average of 3 cases);
                    (b)  0.66 smVs per MH(eT for California Bunker "C" fuel oil (87. 9X C, 10.3S H2, 1.3 S,
                         0.51 0;. S 0.1Z Nj) burned with 191 excess air;
                    (c)  0.73 snrVs per MW(e) for natural gas (-881 CH4, 81 C2H6, 81 N2)
                         excess air (average of  2 cases);
                                                                                       burned with 141
                (2)
                  Emission rate  calculated for total  capacity of unit as firing rate and for maxima emission
                  factor given for any type of boiler of  the fuel type in the  size range as given in AP-67
                  (1977):                         .    ,
                    (a)  Coal  S-2000 MW(e): 1.66 lb/10°  Btu for wet-bottom cyclone utility or industrial
                         boiler;                     ,
                         Coal  0.1 NW(e):  0.25 lb/10  Btu for packaged firetube firebox stoker;
                         Oil   200-2000 MW(e):  0.75  lb/106 Btu for horizontally opposed or front wall, vertical.
                         or cyclone firing field erected  water tube utility boilers;
                         Oil   20-100 MW(e):  0.573 lb/106 Btu for horizontally opposed or front  wall, vertical,
                         or cyclone firing field erected  water tube utility boilers;
                         Oil   O.l-lOMM(e):  0.43 lb/106 Btu for packaged firetube or water tube conoereial
                         boilers  firing residual;
                         Gas   200-2000 MW(e): 0.70 lb/106 Btu for horizontally opposed or front  wall firing
                         field erected water tube utility, boilers;
                         Gas   20-100 MW(e):  0.301 lb/106 Btu for field erected water tube Industrial boilers
                         firing  natural aas;           ,
                         Gas   0.1-10 MW(e): 0.103 lb/106 Btu for packaged fire tube or water tube conercial
                         boilers.

              (3) Maximum ambient concentrations generated for any meteorological conditions and  distance downwind
                  using the PTMAX model with "reasonable" stack heights and temperatures for the  unit size and
                  fuel type and a single stack for one unit.  As an incomplete set of PTMAX runs  were nade for the
                  numerous combinations of stack height,  temperature, and flow rate considered, many of the extreme
                  and best estimates were by extrapolation or interpolation, to that the concentration* given are
                  very approximate.

2.  fuel  type: (c)- coal, (o)-  oil, (g)- Natural gas.
                     (b)
                     (c)

                     (d)

                     .(e)

                     (f)

                     (g)

                     (h)
SOURCE:     Energy  and  Environmental  Analysis,  Inc.

-------
                           -45-
     2.   Area Sources

          a.   Ambient Concentration Patterns

     The spatial and temporal patterns of ambient N02 levels
resulting from area source emissions differ somewhat from those
due solely to point source emissions.  Our knowledge of these
patterns is based on the continuous record of ambient N02 at
those monitors which, due to their location, are unlikely to
reflect point source contributions.  In urban areas, these
include most continuous monitors.

     Washington, D.C., an area source-dominated urban area
(over 70 percent of total NO  emissions come from area sources),
                            X
provides an exceptionally good opportunity to profile ambient
air pollution arising from area source emissions.  Tables 7
through 9 show monthly peaks (highest and second highest hourly
values), and annual averages for three stations in the Washing-
ton, D.C. area.  These stations reflect a variety of area
source settings—office complexes, high-vdlume transportation
intersections, suburban commercial centers—yet they display a
remarkable uniformity in recorded concentrations.  This is due
to  (1)  a relatively even spread of area source emissions,
and/or (2)  a spatial smoothing effect due to the relatively
slow rate of NO- formation from NO.  The lack of correspondence
between NO and NO- levels at single points in time supports the
latter interpretation.

     Figures 20 through 22 provide further evidence for the
similarity of ambient NO- levels at these monitors, and suggest
possible explanations for the patterns observed.  Figures 20
and 21 show diurnal variations during typical summer days at
two of the stations.  Values shown are concentrations for each
hour averaged over an entire month.  These are days when ozone

-------
                                   -46-
                                   TABLE 7
             MONTHLY TIME TRENDS  TN HOURLY NO AND NO  CONCENTRATIONS
                            LEWINSVILLE STATION
                      WASHINGTON, D.C. AREA IN 1977
                Highest NO     Corresponding       Highest       2nd Highest
   Month          (yg/m3)       N02  (ug/m3)      N02 (ug/m3)      N02 (yg/m3)

  January          680              65              130             125
  February         630              65              290             225
  March
  April            615              95              190             180
  May              380            120              290             255
  June             290            180              225             205
  July             290              75              170             140
  August           650              10              280             265
  September        515              75              280             235
  October          580              40              160             150
  November         700              45              140             130
  December         680              75              130             120
  Annual Average NO.:   56 yg/m
  Second Highest N02:   280 ug/m3
  Peak/mean - 5.0
SOURCE:   Data  from local pollution control  agencies,

-------
                                -47-
                                  TABLE  8-1.

          MONTHLY TIME TRENDS IN HOURLY NO AND N0_ CONCENTRATIONS
                              MASSEY  BUILDING
                     WASHINGTON,  D.C. AREA  IN  1977
               Highest NO     Corresponding       Highest       2nd Highest
  Month          (Pg/m3)        NOo  (yq/m3)      NO?  (ug/m^)     NO? (yq/a3)
 January          655             140               160             150
 February         650             105               190             170
 March            350             105               140             130
 April            230             120               160
 May               40              20               180             130
 June              80              85               170             160
 July              20              65               85              75
 August            70                              95              75
 September        125  •            95               115           .  105
 October          420              75               225             130
 November         680              43               115             105
 December         645             120               120             115
 Annual Average N02t  40 yg/m
 Second Highest IK^s  190 yg/m
 Peak/mean =4.8
SOURCE:   Data  from  local  pollution control agencies,

-------
                                -48-
                                 TABLE 9

          MONTHLY TIME TRENDS IN HOURLY  NO AND NO  CONCENTRATIONS
                           SEVEN CORNERS STATION
                     ' WASHINGTON, D.C. AREA IN 1977
               Highest NO     Corresponding       Highest      2nd Highest
  Month          (tig/m3)       N02  (ug/m3)      NO? (yig/m3)     NO^

                                                   150           120
                                                   205           145
                                                   120           110
                                                   265           195
                                                   190           170
                                                   140           130
                                                   170           150
                                                   235           150
                                                   205         •  170
                                                   160           150
January
February
March
April
May
June
July
August
September
October
630
620
540
655
240
170
185
320
505
420
10
50
0
0
40
20
55
105
105
85
 Annual Average N02:   46
 Second Highest NO2:   235
 Peak/mean =5.1
SOURCE;  Data from local pollution control agencies

-------
    NO,  NO.,
     (ppm)

  0.035-

  0.030-

  0.025-

  0.020

  0.015

  0.010

  0.005-
                                             FIGURE 20
                      AVERAGE HOURLY NO AND NO  CONCENTRATIONS FOR JULY  1977
                            (Engleside Station, Washington, D.C. Area)
                       NO
                       NO,
   -I	r
-i	1	-,	,	1	p-
    6)8      10
                                                  12
                                                  Hours
~1	•	T
 14      16
                                                              18
                                                                                r~
                                                                                20
24
                                                                                                  VO
                                                                                                  I
U"'
ta ,c---m ?.-"--il    'lui'* i  c'  ' rol  -

-------
                                         FIGURE  21
                     AVERAGE HOURLY NO AND NO2 CONCENTRATIONS FOR JANUARY 1977

                            (Engleside Station, Washington, D.C. Area)
        NO, NO,
              4

         (ppra)



   1.0000 -




   0.0875 -



   0.0750 -




   0.0625 -




   0.0500 -




   0.0375 -




   0.0250 -




   0.0125 -
                             NO
                   	NO,
                        I
                        4
I
6
I
8
10     12


     Hours
14
16
18
I
20
 I
22
24
                                                                          o
                                                                           I
SOURCE:   Data from  local pollution control agencies

-------
                                                         FIGURE 22
                                 AVERAGE HOURLY NO AND NO  CONCENTRATIONS FOR JUNE 1977
                                       (bewintrvilla Station, Washington, D.C. Area)
     NO, NO
      (ppm)
0.05-1
0.04-
0.03-
0.02-•
0.01 •
                                                                                                                   \
                                                                                                                     \
                                                                                                                      \
                                                                                                                          I
                                                                                                                          ui
10
12
                                                                      14
16
                                                                                         18
20
                                                                                                            22
                                                                   24
           SOURCE:   DaLa  from  Tonal  noTJnHon

-------
                          -52-
levels are likely to be high.  NO and NO-  levels both  fall
after midnight due primarily to low NO  emissions  and  to
                                      X
atmospheric dilution.  If 0_  (or other oxidant) levels are  ini-
tially also low, NO will build up rapidly  in the morning  in
concert with traffic levels, while the rise in N02 lags behind
the HO buildup.  The resultant N02 peak in early to mid-morning
then subsides as photo-dissociation becomes a balancing factor.
In both Figures 20 and 21, afternoon NO levels are shown  to be
fairly lov; (due to high 00 , while NO- has reached a steady
state level.  With the reduced UV insolation in late afternoon
and evening and the increased NO  emissions from rush-hour
                                J^
traffic, NO- begins a rapid rise followed  by or simultaneously
with an NO increase as 0., is depleted.

     Wintertime diurnal variations are similar, with NO levels
showing the same daily variation corresponding to  traffic fluc-
tuations.  Figure 22 shows that the peaks  in N02 are somewhat
muted, though still obvious.  W,ith reduced UV radiation,  lower
temperature, and low 03 levels in the winter, NO.  concentrations
are limited by slower conversion rates and atmospheric dilution.

     The EPA  (OAQPS, MDAD) has recently reviewed patterns of
                                         197
hourly NO,, levels in several urban areas.  /  They have
labeled the morning peak "photo-chemical synthesis" and the
afternoon peak "ozone titration", though the processes  which
lead to each are both obviously related to photochemistry.
In addition,  they stress the importance of "carry-over" NO-—
high levels of NO- formed one day which are not depleted
when new NO- production begins the next.   The highest  hourly
NO- levels observed are likely to be a result of NO- "build-
up" over several days.

-------
                           -53-
          b.   Potential Violations of Short-Tern Standards

     If continuous 110- monitors were common in all AQCR's, the
likelihood of any AQCR violating the various suggested  short-
term N02 standards could be determined with ease.  Unfortunate-
ly, continuous monitors are not common.  As a consequence, the
potential for violation must be judged from 24-hour  readings or
annual averages.  Evidence is now available to suggest  that the
relationship between one-hour peak  (second highest hourly)
readings and annual averages for area source-dominated  monitors
is less than six to one.  Figure 23 shows the distribution of
peak to mean NO- .ratios averaged over the years  1972, 1973,
      .--. .._	 .. _ .£	_.	._	  1 A -T 7 X
and 1974 for 120 urban sites.  ''  It is further reported
that for continuous monitors in central city commercial or
residential areas, the average peak (maximum) to mean ratio
lies between six and seven.  '  Area sources are undoubtedly
the major contributors to the N02 concentrations at  these
sites, although point sources in the region also have some
impact.  The ratio due to the area source impact alone  should
thus be below this average value.  The peak to mean  value of
six, therefore, seems to be associated with sites impacted
predominantly by area sources.  The urban area monitors reporting
peak to mean ratios of over six are believed to  be impacted by
point sources significantly and should not be used to characterize
short-term N02 problems from area sources.

     In order to estimate the fraction of AQCR's likely to
violate alternative N02 one-hour standards as a  result  of area
source emissions alone, short-term NO. concentrations were
estimated for each AQCR.  This estimate was set  equal to the
highest second high one-hour concentration recorded  at  any
continuous monitor, from 1972-1974, the peak to  mean ratio of

-------
                              -54-
                             FIGURE  23
  30% -
  20% -
u.
o
u
oe
  10% -
            T

            2
 8    10     12   14



MAXIMUM/MEAN RATIO
16
18   20
  Distribution of maximum/mean NOj  ratios for 120 urban locations


  averaged over  the years  1972,  1973,  and 1974.
  SOURCE:  Reference  16

-------
                            -55-
which was 6:1 or less.  If no continuous monitor in the AQCR
fits this criterion, then the short-term estimate was set
at six times the highest annual average of all monitors with
a sufficient number of data points.  An annual average is
calculated only if at least  165 percent of.'.the  total  possible
hourly values' are recorded  at a~-site- during  a given year.
     The results are shown below.  Only at a one-hour standard
of 250 yg/m  is it probable that  an  appreciable number  of AQCR
would be in violation due to area source emissions alone.
    . Estimated Number of AQCR's Not Attaining the Standards

                        250 yg/m3:  95
                        500 yg/m3:  17
                        750 yg/m3:   2
                      1,000 yg/m :   0

      3.    Alternative Ways to Express the Short-Term Standard

      Short-term NAAQS are normally expressed as a concentration
 (averaged over the specified sampling time) which cannot be ex-
 ceeded more than once per year.  The one allowed excursion is
 justified on the basis of unusual unrepresentative meteoro-
 logical or emission conditions that may be experienced in any
 AQCR.  However, it is unlikely that these conditions would
 occur just once or not at all each year.  Fumigation episodes,
 inversions aloft, or similar events which can lead to high
 ground-level concentrations of airborne materials tend to
 behave like random variables.  That is to say, the average
 occurrence over a long period of time may be once a year, but
 multiple yearly occurrences could, and in fact would  be expected
 to occur occasionally.  Thus, the standard should be expressed
 in statistical terms  (e.g., "cannot be exceeded on the average
 more than once a year").

-------
                           -56-
     Determining when a violation of a statistically-based
standard has occurred is more problematic since a violation
will only occur if the number of excursions above the standard
is higher than would be expected for that AQCR.  Thus, a frequency
distribution of pollutant concentrations nust be derived from
recorded data in each AQCR.  The distribution is then adjusted
so that the standard concentration will occur no more than once
a year on the average  (is likely to occur no more than once in
3,760 hours per year for a one-hour standard).  The probability
that two, three, or more excursions could occur can then be
read directly from the frequency distribution.  If the number of
excursions  actually observed over several years is considerably
higher than the expected number  (one in this  example), then the
AQCR can be said to be in nonattainment.

     EPA has estimated the frequency distribution of  ozone
concentrations for several AQCR's using the Weibull function.
Unfortunately, a similar analysis of N02 data has not yet been
made, so that little can be said about the nature of  these dis-
tributions.

     In addition to using statistical terminology in  expressing
the N02 standard, the number of allowable times the specified
concentration is most likely to be exceeded can also  be varied.
If occasional high concentrations do not pose a health risk
(e.g., if chronic, rather than acute, exposure to levels above
the standard is the primary health concern),  then more than one
expected excursion per year can be allowed.  The choice must be
based on a  combination of health effects and monitored/modelled
ambient air quality days.

-------
APPENDIX A

-------
o
a
H-
H
o
00  hr)


Z  H

«  2
5
H,
r»
tr
o

n
»
H
P.
H'
          NEW ALEXANDRIA
              EL. 1260*
                                                                                  TIDO
                                                                                   CARDINAL
                                                              3FOWLERSTOWN
                                                                 EL.1230     '
            WELLSBU^G
                                                                                           °4
                                                                                MEADOWCROFT
                                                                                 , VILLAGE
                                                                                           '
CARDINAL/HDD PLANTS

    PLANT ELEV.  675*
  >  STACK HEIGHT ELEVS.
    STK HO. CARDINAL TlOD
BEECH
 BOTTOM
   EL. 730*
                                                                            West
                                                                          Virginia
      Legend

D  POWER PLANT
                                                                          O


                                                                          O  SO, -f NOB


                                                                          O~ SO, -f PARTICULATES



                                                                          Q  SO, + COM


                                                                              SO, -f WIND


                                                                          A  VISIOMETER


                                                                              TOWER
                                                                            O5
                                                                          BETHANY
            All EUvalloni

         Are Abova Mean Sea Level
                                                                                                               i
                                                                                                               en

-------
                                 -58-
                                                         10
                           SPORN
                                      COOLVILLE
                                       EL, 800*
                  KYGER CREEK
                      (OVEC)
            Ohio
                        O-WOLFPEN
                            " EL. 900'
                     8
                   O-FfVE POINTS
                        EL. BOO'
                                    M'ASON
                            LAK1N
                              EL. GOO'
                           NEW HAVENO-
                             EL.800'  V»
        GAVIN
     POWER PLANT./
      PLANT ELEV. 567'  ''
STACK HEIGHT ELEV. 1667*
                  •KYGER CREEK
                     POWER PLANT
                       PLANT ELEV. 578'
                       STACK HEIGHT ELEV. 1116*
                                  9
                               . -O-OORCAS

                        • SPORN    B-720'
                        POWER PLANT
                         PLANT ELEV. 586's*
                         STACK HEIGHT ELEV. 1186'
                                                                   NEPTUNE
                                                                     EL. 910'
                                                MOUNT ALTO
                                               7  EL. 930'
                        O-HENOERSON
                         O   EL. 558'
                     WCst
                   Virginia
   All Ctoovtlem

Ar« Ak*v* Mean
       3  4 MILES
                                                               Legend

                                                             POWER PLANT
                                            PART1CULATES

                                       S02'+ RAIN GAUGE

                                       S02 -t- VISIOMETER

                                       S02-t-WINO
Gavin -
                                  FIGURE  A- 2

                             Kyger - Spom Monitoring Network.

-------
                                 -6B-
 MITCHELL & KAMMER
                                        McCLAIN RUN
                                          EL. 1300'
                                                     BLAIRS RIDGE
                                                       EL. 1340'  .
                                                           0-6
                              2
                        GRAVE CREEK
                          EL. 1230'
PLANT ELEV.
STACK HEIGHT ELEV.
 NOS.I&2  1245'


GERMAN RIDGE
 EL. 1240'  .40?
'KAMMER
MITCHELL
        PLANT ELEV. 6S6
        STACK HEIGHT ELEV. 1866
                                             O-B
                                            4. L.U BOWMAN
                                             t  f    PI  iAnn'
                                                     EL, 1400
                                          West Virginia
      Legend

     POWER PLANT
  O.so2

  (I)  SOZ + NO,
           TSP


           WIND


     TOWER
                                                                N
                                FIGURE  A-3
                                                          All Elevations

                                                      Art Above Mean Sea Level

                                                        •^*

                                                                     3MH.C3
             Field Monitoring  Network  in the Vicinity of the Mitchell and
             Kamraer  Power  Plants.  McClain Run  (5) started operation on
             December 30,  1974 and German Ridge  (1) started operation on
             January 7,  1975.

-------





















r»
0
rt

rt
M
rt
t»

>
n
Ul

H*
JJ
H-
n

n
£
n
o.
H-
H
n
n
rt

o
a
M
0
M»

s.
Jj
0-











!«••«•>
•I***

-------
                              -61-
               OH*
 I

(I
I I

' I.
                                                KILBY ROAD
                                                 EL. 515'
       ^TANNERS CREEK
                                                     1  5 ELIZA8ETH70WN
                                                     Q~    EL. 500'
         Indiana
                      LAWRENCEBURG
                            EL.465*  °>
         WILSON CREEK
           EL.8IO'
                 BO-
                                                          Kentucky
                                        TANNERS P. P.
                               -?/>TANNERS CREEK
                                     PLANT ELEV. 497'
                                     STACK HEIGHT ELEVS.
                                          1-3 770'
                                           4  1047'
                              1 PETERSBURG
                                 EL. 510'
              DUTCH  HOLLOW
                EL. 810'
    All Elevations
Are Above Mean Seo Level
        i
              ? MILES
                                    Legend
                             C3  POWER PLANT

                             O  so,
                             O" SOj^ PARTJCULATES


                             W  WIND INSTRUMENTS

                             R3  TOWER
                           FIGURE A-5
          Field Monitoring Network in the Vicinity of the Tanners Creek.

-------
     •-• "••
   !•*!••»
/    ••
                  \
  ^ROCKPORf SITE
       N
    All Elavollont
Ara Above Meon Sea Laval


 0    I    ( MILES
                         ;i
                         / i
                                                                                      EVANSTON
                                                                                     o EL. 480* •

                                                                                     30
       ROCKPORT
           EL. 405r
                                                             CHESTNUT GROVE
                                                                                       SO, + PARTICULATES
              .ROCKPORT
                                                                                                          I
                                                                                                          cr\
                                           FIGURE A-6
                      l;icld Monitoring litjulpinont in .tho Vicinity of  the Rockport  Site.

                      Directions arc tlioso of tlio wind that would blow from  the plant

                      tf   e s. .Ion.

-------
                            -63-
                       REFERENCES
1.  Memorandum from Joseph A. Tikvart (EPA Source-Receptor
    Analysis Branch) to Edward J. Lillis (EPA Air Management
    Technology Branch), February 16, 1978.

2.  Control Techniques for Nitrogen Oxide Emissions From
    Stationary Sources—^Revised Draft Second Edition. Aerotherm
    Report TR-77-87, December 1977.

3.  The National Air Monitoring Program;  Air Quality and Emis-
    sion Trends. EPA, December 1977.

4.  Calvert, J.G. (1973), Interactions of Air Pollutants. Pro-
    ceedings of the Conference on Health Effects of Air Pol-
    lution , National Academy of Sciences, 19-101, October 3-5.

5.  National Academy of Science, Air Quality and Stationary
    Source Emission Control, prepared for the Committee on
    Public Works, United States Senate,  Serial No. 94-4, March
    1975.

6.  U.S. EPA National Air Quality and Emissions Trends Report,
    1975, EPA No. 450-11-76-002, November 1976.

7.  Roger Morris, "Nitrogen Dioxide Problem Areas", March
    1978.

8.  Code of Federal Regulations, Section 40, Part 60.

9.  Section 202(b)(1)(B), Clean Air Act of 1977,  (42 U.S.C.
    1857, et seq.).

10. EPA, '"Uses, Limitations, and Technical Basis of Procedures
    for Quantifying Relationships Between Photochemical Oxidants
    and Precursors," EPA-450/2-77-021a,  November 1977.

11. Ambient Monitoring data obtained from American Electric Power
    Service Corporation, Canton, Ohio.

12. Ambient Monitoring data obtained from Potomac Electric
    Power Company, Washington, D.C.

-------
                          -64-
13.  "Ozone and Nitrogen Oxides in Power Plant Plume," D.  Megg,
    P.V.  Hobbs, L.F. Radke, and H. Harrison,  International
    Conference on Photochemical Oxidants,  August 1977.

14.  Davis, D.D., G.  Smith, and G. Klauber  (1974),  "Trace  Gas
    Analysis of Power Plant Plumes via Aircraft Measurement:
    0-., NO , and SO, Chemistry," Science 186;  733-736.
     O    X        ^                   ™  —"

15.  Thuillier, R.W., W. Viezee, Air Quality Analysis in Support
    of a Short-Term Nitrogen Dioxide Standard.  Discussion Draft
    prepared for EPA; SRI International, December 1977.

16.  Trijonis, John,  Empirical Relationships Between Atmospheric
    NO., and its Precursors, EPA, Research  Triangle Park,  North
    Carolina, February 1978.

17.  Health Effects for Short-Term Exposure to Nitrogen Dioxide,
    Final Draft, March 31, 1978, U.S. EPA.

18.  Mobile Source Emission Factors, Final  Document, EPA-400/9-
    78-005, USEDA, Washington, D.C., March 1978.

19.  U.S.  EPA, "Status and Implications of  Analyses of Ambient
    N02 and Other Air Quality Data"  (Draft),  EPA,  OAQPS,  MDAD,
    Research Triangle Park, N.C., May 26,  1978.

-------
           SHORT-TERM N02 STANDARDS

                 VOLUME II
          ESTIMATED COST OF MEETING

           ALTERNATIVE STANDARDS


               DRAFT REPORT
                    DRAFT**-
              DO  NOT QUOTE OR CITE
               Submitted to:

Office of Air Quality Planning and Standards
       Environmental Protection Agency
Reserach Triangle Park, North  Carolina   27711
               Submitted by:

Dale L.  Keyes,  Bharat Kumar, Robert D.  Coleman
             and Robert 0. Reid
  Energy and  Environmental Analysis,  Inc.
     1111 North 19th Street, 6th Floor
        Arlington, Virginia  22209

             December 26, 1978

-------
                           STATEMENT
     This Draft Report is furnished to the Environmental Pro-
tection Agency by Energy and Environmental Analysis, Inc.
(EEA), Arlington, Virginia.  The contents of the report are
reproduced herein as received from the contractor.  The opin-
ions, findings, and conclusions expressed are those of the
authors'  and not necessarily those of the Environmental Pro-
tection Agency.

-------
                        TABLE OF CONTENTS
             Title                                    ,     Page
A.   Introduction and Caveats	   1"
B.   Air Quality Assessment Methodology	   4
     1.   Point Source Modelling	   5
     2.   Area Source Modelling	  17
C.   Air Quality Modelling Results	  24
     1.   Point Source Analysis.....	  24
     2.   Area Source Analysis	  29
     3.   Point and Area Source Analyses Together	  29
D.   Control Options and Cost Analysis	  34
     1.   Point Source Control Options and
          Unit Costs	,	  34
     2.   Point Source Cost Analysis Procedure	  44
     3.   Point Source Control Costs	  45
     4.   Impact of Growth on Point Source
          Control Costs	  47
     5.   Area Source Control Options and Costs	  52
     6.   Area Source Costing Procedure and
          Results	  56
E.   Comparison of the Nationwide Cost Analysis With
     the Chicago Case Study.	.	  60
F.   Economic Impact	  61
     1.   General Comments	.•	,	  61
     2.   Point Sources	  64
     3.   Area Sources	  66
G.   Summary and Conclusions	  5-7

-------
                           VOLUME II
            ESTIMATED COST OF MEETING ALTERNATIVE
                   SHORT-TERM K02 STAWDARDS
 A.  Introduction and Caveats '

     This study is the second volume of a three-volume report
which attempts to identify the causes of high short-term
concentrations of nitrogen dioxide (NO-) and the additional
cost of controlling sources of nitrogen oxides•(NO ) emissions
                                                  X
to levels consistent with attainment of the ambient standards
under consideration (that is, the costs in addition to those
required to meet the Federal motor vehicle emission standards
and the NSPS's for stationary sources).  Volume I focuses on
the mechanisms by which high ambient concentrations of N02 are
believed to occur.  Volume II provides a preliminary assessment
of the sources which, through the mechanisms identified in
Volume I, might cause or contribute to high short-term (i.e.,
one-hour averaging period) concentrations of N02-  Control
strategies are developed for these and used to estimate nation-
wide control costs.  Volume III describes a detailed case
study of short-term 110- concentrations in Chicago.  An area
source and multiple point source model was used to capture the
interactive effects of all NO  emission sources in a region.
                             2C
The results shed light on the accuracy of the nationwide
study.

     It should be emphasized at the outset that actual monitor-
ing data which could be used to estimate the expected contribution
of point sources are extremely sparse.  Lacking those data, a
modelling technique based on a nonreactive Gaussian dispersion
model and empirical NO -to-NO_ conversion curves was used to
                           -1-

-------
estimate maximum ground-level concentrations of NOj.
Although it is believed that the results are reasonable, this
technique should be validated with data on observed ambient
NO- concentrations.

     Any short-term study which attempts to assess the national
dimensions of what is, in effect, a multiplicity of localized
problems must necessarily impose simplifying assumptions on the
analysis.  The most significant of these assumptions are de-
scribed below:

     •    Though ambient concentrations of N02 are
          created by a mix of emissions from both
          point and area sources, the two categories
          of sources are treated separately in this
          analysis.  The area source analysis is
          based on proportional  ("rollback") modelling
          of monitored air quality and estimated area
          source emissions in urban areas.  This is
          a reasonable approach if, as appears true
          from the available evidence, most monitors
          in urban areas reflect contributions pri-
          marily from area sources of NO  emissions.
                                        f±
          Point sources are assessed by estimating
          theoretical maximum contributions to
          short-term N©2 levels with the dynamic
          modelling technique noted above and de- -
          scribed in Volume I.  As noted, these
          results are largely unverified by empiri-
          cal observation.

     •    Interaction between point and area sources
          is captured in a limited way by adding area
                           -2-

-------
source background concentrations to the es-
timated point source contribution.  The high-
est annual average value recorded in an AQCR
is used to estimate  the area  source back-
ground for all point sources located in that
AQCR.  Based on the results of the Chicago
case study, a background value SOper^gjat,
higher than the highest observed annual
average is used to represent area source
contributions to ambient N0_ levels under
conditions which maximize area and point
source interaction.

The need to verify our N02 modelling
technique has already been noted.  Alterna-
tive approaches for representing the secondary
(derived) nature of N02 have also been dis-
cussed.  Cole suggests setting maximum NO-
equal to the ambient 0, level plus 0.1 times
              I/
the NO  level. '  This appears to work well
      jf*
for summer conditions, but should underesti-
mate N02 under conditions of low 0, concen-
trations characteristic of winter.  In addi-
tion, empirical analyses of relationships
between N02 and its precursors suggest only
limited correspondence between levels of N0_
       2/
and 0.,. '  However, this approach is used
here for comparison purposes.

Plumes from individual point sources are
assumed to be non-interactive.  This is a
reasonable assumption for sources separated
by tens of kilometers, dependent, of course,
                -3-

-------
          on stack height.  Unfortunately/ time limi-
          tations did not allow an interactive analysis
          for sources where this assumption does not
          hold.  This leads to some underestimation of
          control costs, though in combination with
          the following simplification, the degree of
          underestimation may be small.

     •    Ambient levels due to NO   emissions  from
                                 Jt
          multiple stacks in single plants are assum-
          ed to be additive.  This  is  the  same as
          assuming that the points  of  maximum  ambient
          impact from every source  have the same loca-
          tion, irrespective of differences in stack
          and emission characteristics. Ambient levels
          and control costs are consequently overesti-
          mated.  However,  the Chicago results show that
          the overestimation in this case  tends to balance
          the underestimation introduced by the previous
          assumption.

     •    The point  source  modelling  assumes  every
          source is  located in flat terrain;  where
          this does  not hold, ambient concentrations
          may be underestimated.

B.   Air Quality Assessment Methodology

     Since both point and area sources emit NO , an assessment of
                                              X
the causes of high short-term NO- concentrations must consider
each type of source  and. their respective  concentrations.   Studies
have shown that either can  lead to relatively high concentrations.
However, the nature  of their impacts  are  different.  Point sources
tend to produce infrequent  and spatially  confined NO- peaks,
though the slow formation rate of NO-  smooths out these  "hot
spots" to some extent.  Area sources,  on  the  other hand, are  lass
                             -4-

-------
varied in their impact on peak NO- levels in both time and
space.

     Due to the placement of NO- monitors in urban areas and to
the relatively small variation in recorded levels over time
(for continuous monitors), a case is built in Volume I for the
use of monitoring data in assessing area source contributions
to NO- problems.  The impact of point sources, however, can
only be captured through dispersion modelling exercises, given
the nature of existing monitoring networks.

     The following sections of the report discuss the separate
treatment of point and area sources.  The results of the two
analyses are then brought together in a final discussion and
assessment.

     1.   Point Source Modelling

          a.   General Approach

     Detailed multiple source modelling of each AQCR was far
too ambitious for a nationwide study.  Instead, each source of
NO  emissions in EPA's NEDS file of jpoint sources was modelled
individually with a single source dispersion model.  This
approach may be adequate for isolated sources, but for urban
areas, a means was needed to assess the interaction of multiple
point sources and the degree to which area source emissions
exacerbate point source problems.  To assist in this regard, a
separate, detailed study of a single AQCR was undertaken.  The
results of this case study can then be used to judge the
adequacy of the nationwide assessment.

     The NEDS modelling analysis was initiated by specifying
model plants which ranged in size and operating parameters corres-
ponding to combustion processes in various source categories  (e.g.
utility boilers, industrial boilers, and furnaces).  The model
                          -5-

-------
plants were analyzed individually using a simple Gaussian dif-
fusion model  (PTMAX) to assess the meteorological conditions
associated with ground-level maximum N02.  These conditions
were then used with a simplified version of PTMAX to model the
air quality impacts of all NO  sources in the NEDS file.  Once
                             J^
the concentrations  of NO  around each point source had  been
                        X
characterized by the diffusion'model, two alternative approaches
were employed to translate NOX into N02-  Emission control
methods and their associated costs were then estimated  for each
source.

     The case study involved a detailed analysis of the multiple
point and area source contributions to peak short-term  NO-
concentrations in the Chicago region.  An EPA-approved  point
and area source interactive model, RAM, was used for this
analysis.  A detailed point and area source emission inventory
combined with the versatility of RAM provided a unique  character-
ization of NO- concentrations in a large urban area.  The
reasons for selecting Chicago as the case study AQCR and details
of the study are described in Volume III of this report.  The
results of the Chicago case study provide both a sensitivity
check on the assumptions used in the nationwide study,  and a
means of gauging the degree of over- or under-estimation of
control cost impacts.

     It is important to re-emphasize that the nationwide point
source analysis focused on achieving alternative short-tern H02
standards through point source controls alone.  Area source
emissions were treated as a background influence.  Results of
the Chicago study were extrapolated to estimate the area source
contributions in each AQCR.

          b.   Processing of Point Source Emissions Data

     Samples of raw data were retrieved from the NEDS point
source sub-file to determine the extent of erroneous and missing
                          -6-

-------
information.  Upper and lower limits and default values for the
stack height, temperature, air flow rate, hours of operation,
and fuel heating values were established for each source  category
based on standard operating practice and soibstituted for  erroneous
or missing data in the file.  These are shown in Table 1.
Default values for flow were calculated for each type of  fuel
at the excess air rate normally associated with it.  For  combus-
tors fired with dual fuels, estimates were made assuming  100
percent firing of the fuel which had the highest emission rate.

     The most recent emission factors from the Control Techniques
        3 /  •                                                 4 /
Document ' , California Air Resources Board ARE 2-1471  Report '
                         5/
and AP-42 (Third Edition) ' were used for this analysis.  Data
on the emission factors for industrial combustion processes are
limited.  The variability in the emissions with fuel.type is
not as significant as it is among processes in the same fuel
category.  The emissions factors used for industrial combustion
processes are, therefore, average values applicable to the
greatest number of processes in a fuel category.  Revised emission
factors for cyclone boilers, which are significantly lower than
'those listed in AP-42, have recently been reported by Aerothern
and were used in this analysis  .  For example, the revised
emission factor for coal-fired cyclone boilers is 1.3 Ibs.
NOx/MM]
AP-42.
NO /MMBtu, approximately 60 percent of the value reported in
  X
     Emissions were calculated for operation at full  load.   In
the absence of information on rated capacity or maximum operating
rate, the hourly heat input rate was calculated using the
annual fuel consumption and the hours of operation.

     All existing nitric.acid plants were assumed to  meet the
5.5 Ib. IJ02/ton nitric acid regulation applicable to  old sources.
                           -7-

-------
                                                             TABLE 1

                                                         DEFAULT VALUES
               Source
       Boiler
            Acceptable  range
            Default
                                                 Stack
                                              Temperature
                                       320 < T <800
                                            410
                                                                  Stack Height
                                                                      (m)
10  £ H £ 500
     30
                              Hours of
                              Operation
                                                                                            0 < hours < 8736
                                                                                                 5400
i
00
i
Internal Combustion
     Acceptable range
     Default

Chemical Processes

     Acceptable range
     Default

Petroleum Industry

     Acceptable range
     Default
                                              320 < T i 800
                                                  410
                                             290 < T  £900
                                                 13 500
                                              320 £ T £1100
                                                   500
 5 £ II <  500
     20
 5 i II  £ 500
     15
 5 £  II £  500
     15
                                                                                            0 < hours £ 8736
                                                                                                 5400
                                                                                            0 < hours £8736
                                                                                                 8000
                                                                                            O < hours<8736
                                                                                                 8000
In-Process Fuel

     Acceptable range
     Default

Incinerators

     Acceptable range
     Default
                                              310 £ T£l300
                                                  500
                                              310 < T£1100
                                                  370
                                                                  5 <  II £  500
                                                                      15
                                                                  0 £  H £  500
                                                                      15
                           0 < hours £8736
                                8000
                           0 < hours Is 6736
                                8000
       SOURCE:   EEA,  Inc.

-------
This regulation is in effect in most states with nitric acid
plants.

          c.   Dispersion Modelling Methods and Assumptions

     The Gaussian diffusion models used in these analyses are
non-reactive models.  Maximum NO  concentrations from the
                                Jt
single or multiple point sources were first estimated and then
translated to N02 levels using both of the approaches described
in Volume I of this study.  The dynamic translation model de-
veloped by. EEA is the preferred approach for isolated point
sources.

     For sources with multiple stacks and each stack serving
multiple combustors, diffusion calculations were made on a
stack basis.  The maximum impact of all stacks in a source was
assumed to occur at the same point and to be additive—clearly
a conservative approach.  Finally, the list of sources that
were estimated to violate.a 250 yg/m  ambient N02 standard was
screened manually.  Data -for sources with unusually high
estimated NO  concentrations were reviewed, corrected if necessary,
            J^
and ambient concentrations re-estimated.

          d.   Influence of Meteorological Conditions on Point
               and Area Source Interaction

     As noted earlier, meteorological conditions play a major
role in the short-term build-up of N02 concentrations from both
point and area sources of NO  emissions.  Meteorological condi-
                            Ji
tions, in addition to controlling the diffusion of N02, also
have a binding influence on its formation rate.  A detailed
discussion of this subject is presented in Volumes I and III of
this report.
                           -9-

-------
     A summary of the conclusions derived from the Chicago
case study is presented here.  These results were used to select
the meteorological setting for the nationwide study.

     •    NO  emissions from either point or area
            Ji
          sources can cause high short-term NO-
          conclusions.

     •    Two distinct groups of point sources can be
          identified in terms of their, response  (dilu-
          tion and N02 formation rate) to different
          meteorological conditions:  (1) plants with
          tall stacks such as utilities, and  (2) plants
          with a large number of short stacks such as
          steel mills and refineries.

     •    The diffusion characteristics of the second
          point source group seem to be similar  to
          those of the area sources.

     •    The meteorological conditions that maximize
          the impact of sources with high effective
          stack heights are at an opposite extreme
          from the conditions that result in high con-
          centrations from area sources or point
          sources with short effective stack heights.

     •    In the Chicago case study, an intermediate
          set of meteorological conditions, closer to
          those diffusion conditions which maximize
          area source contributions, resulted in the
          highest short-term HO, concentrations.
                          -10-

-------
     The following summary of the test case analyses from the
Chicago case study should be helpful.  The results of the point
and area source interactions under three different meteorologi-
cal conditions are presented in Table 2.  The table shows the
receptors with the five highest estimated concentrations for
each set of meteorological conditions and the average concen-
tration at all receptors estimated to be above 200 yg/m .
This sensitivity analysis was performed on a portion of the
Chicago AQCR which included Cook, Dupage, and portions of Will,
Lake, and Porter Counties.  As can be seen, total N02 concen-
trations are highest for the intermediate case.  In addition,
the degree of point source control required is higher in the
intermediate case due to higher area source background concen-
trations than in the point source "worst case".

          e.  Selection of Scenarios

     Since the point sources in NEDS include rural, isolated
sources as well as those  located  in  dense  urban  areas,  a  single
set  of meteorological  conditions  was deemed  insufficient.   For
sources in urban areas, as noted  in  the  previous section,
the  multiple point and area  source influence is  overriding.   An
intermediate set of meteorological conditions  would be  more
appropriate for assessing the point  source impact.  But for
large point sources located  in isolated  areas, the  "worst case"
for  point sources alone is the appropriate choice.

     Ideally, point sources  in isolated  areas  should  be
analyzed separately from  those in urban  areas.   An  attempt
at such an analysis was not  made  because of  the  large number  of
sources and lack of information  on their location.  Instead,  all
point sources were analyzed  under the two  sets of conditions.
Furthermore, the two alternative  approaches  for  translating the
modelled NO  from each point source  into N09 were applied to
           Jt                               £•
each set of meteorological scenarios.  This  produced  four test
cases which are summarized in Table  3.
                         -11-

-------
                                                    TABLE 2


                 COMPARISON  OF ESTIMATED NO  LEVELS FROM POINT AND AREA SOURCES UNDER DIFFERENT

                                METEOROLOGICAL CONDITIONS IN CHICAGO3'' 
I
Receptors
with the
Five Highest Worst Case - Point
Concentrations Total Point
i 1 509 428
ft 2 589 409
* 3 348 209
# 4 348 225
* 5 342 219
Average of
all receptors
above 200
pg/m 277 165
Number of Receptors
Above 2OO |ig/in-* 47
Percentage of Power Plants
With Significant Contribu-
tions, i.e., Effective Stack
Height Less Than Mixing
Height 17
Source Worst Case - Area Source
Area Total .
81 568
81 479
139 472
123 472
123 472
111 371

Point
549
279
272
272
272
142
68
Area
19
200
200
200
200
199

Intermediate
Total Point
603 493
602 434
600 407
598 430
553 383
316 142
67
0 3
Chicago Case Study - Cook, Dupage, and portions of Will, Lake and Porter Counties.
The receptors used in the analysis were selected to record maximum total concentration.
receptors may reflect higher area source contributions, but lower total concentrations.
Case
Area
110
168
193
168
170
174

Other
      UOURCE:   EEA, Inc.

-------
                             TABLE 3

                          LIST OF CASES
            Meteorological
              Conditions
  Translation
    Approach
  "Worst Case"
  Point Source
 Intermediate
     Case
  Wind Speed =
   5.0 m/sec
Stability Class 3
 Wind Speed =
  1.5 m/sec
Stability Class
     EEA's Rate Curves
                     Worst Case Point Source

     Use of meteorological conditions maximizing the point
source influence alone implies that all sources are located in
isolated areas.  Under these conditions, the impact of area
source contributions is estimated to be a minimum.  This scenario,
therefore, results in an underestimation of the short-term prob-
lem for point sources in an urban area.  The degree of control
required and the cost of control is lower.  The selection of
meteorological parameters for this case is based on the model
plant analysis.
                        Intermediate Case

     This scenario implicitly assumes that all point sources are
located in urban areas.  The area source background is estimated
to be higher in this case than in the previous one.  For sources
in isolated areas, this may overestimate the need for and cost
of controlling them to achieve a given ambient level.

               (0.1 NO  + O.J Translation Approach
                      X    +j

     The approach and its limitations are discussed in detail in
Volume I of this report.  This approach assumes that, except for
the initial conversion of NO to NO- (about 10 percent), conver-
sion of NO is limited by the ambient levels of ozone.

                    EEA's Translation Curves

     NO  in each plume was translated into N02 concentrations as
a function of initial NO  level in the plume and plume travel
time.  A series of time-dependent curves relating NO  to N0_ for
different initial N0x levels were developed based onxplume track-
ing studies.  Details of the approach also are presented in
Volume I of this report.
                             -13-

-------
     For a conservative  (high side) estimate of the cost of im-
plementing point source control strategies to achieve a given
short-term standard, the intermediate meteorological case was
chosen.  In addition, EEA believes that the translation curve ap-
                                   *
proach is theoretically more valid.   Thus, the final cost
and economic impact analysis was based on the results of Case
Four, while Cases One, Two, and Three were used to test the
sensitivity of the results to assumptions about meteorology
and the NO-to-NO- conversion process.

          f.   Area Source Contributions as M02 Background

     Referring back to Table 2  (results of the Chicago case
study), one can see that the average area source concentrations in
each of the three cases are very close to the area source com-
ponents at the receptors with the highest total concentrations.
This implies a low spatial variation in the component of peak
NO- concentrations due to area sources.  Some loss of spatial
variation in the area source concentrations could have, resulted
from limitations in the way RAM models area sources, and is ex-
plained in Volume III of this report.  Regardless, observed am-
bient data jy a_regionwide level support the above conclusion.
The use of a regionwide average value to account for the area
source contribution at all receptors in an AQCR is therefore
reasonable.

     As noted earlier, the short-term area source background in
an urban area varies with meteorological conditions.  To quantify
the area source background level for the two sets of meteorological
conditions used in the national study, the average area source
concentrations for the same conditions in the Chicago study
* However, due to.the lack of readily available ambient NO-
  data around point sources and to time constraints, the
  translation curve approach could not be properly calibrated.
                          -14-

-------
were first related to the highest observed annual average
N0_ concentration in the Chicago AQCR.*  Table 4 shows a
comparison of the estimated average one-hour area source
concentration and the highest observed annual average con-
centration in the Chicago AQCR.  For the first case, maximiz-
ing the point source impact, the average estimated one-hour
NO- level from area source NO  emissions alone is almost
  £»                          H
equal to the highest observed annual average value at any
monitor; for the intermediate case, the estimated area
source background is higher by about 50 percent.

     Assuming Chicago to be a reasonable representation of
area source problems in other AQCR's in the country, these
findings were used to estimate area source background levels
upon which point source contributions are superimposed.  The
highest observed annual average ND.- in an AQCR was used for
the area source background in Cases One and Three, and 1.5
times the highest annual average for Cases Two and Four.
These background values were used uniformly across each AQCR
for all point sources irrespective of their actual location.

          g.   Changes in Point and Area Source Emissions

     Growth in point sources was not, considered explicitly
in the analysis for two reasons.  First, the new source
performance standards for NO  are sufficient to prevent any
                            A,
new source affected from violating the most stringent ambient
standard considered.  Secondly, though clusters of new
sources or expansions at existing sites may combine to
create a violation, the modelling approach used does not
consider multiple point source interactions.
*Studies have shown that the annual average NO- concentrations
 in urban areas are mainly due to area source influence and
 are relatively less sensitive to point source impacts.
 Use of observed annual average concentrations to quantify the
 area source influence is therefore reasonable. (See Volume I)

                          -15-

-------
                         TABLE 4

      COMPARISON OF ESTIMATED AVERAGE ONE-HOUR N02

CONCENTRATION DUE TO AREA SOURCES ALONE  (FOR ALL MODELED

  RECEPTORS) WITH THE HIGHEST OBSERVED ANNUAL AVERAGE

                  IN THE CHICAGO AQCR

                          (ug/m3)
    Case

 Worst Case

  Point
  Source
             Estimated
              One-Hour
             Average N02
                 111
                           a/
 Highest
 Observed
Annual NO.
b/
   109
        Ratio
         1.0
 Intermediate
     Case
                 174
   109
         1.5
a'  Average area source contribution for all receptors  above
           . j
b/
200 yg/ra .  (Data from Table 2.)


1975 annual average values at the Chicago Camp Station.
SOURCE:  EEA, Inc.
                          -16-

-------
(^
      Changes in area  source emissions were considered  indi^ctly.
 A  20 percent reduction in  area source ambient background  levels
 would be expected in  1982  assuming a 1.0 percent annual increase
 in travel and a 30 percent decrease in composite mobile source
 emission factors due  to new exhaust standards, no  increase  in
 stationary area source emissions, and an average ratio of
 mobile to stationary  area  source emissions of 3:1.   In 1990, the
 reduction would be about 35 percent.  For the high growth
 scenario (i.e., 3.0 percent increase in travel per year and a
 1.0 percent increase  in stationary area source emissions) , the
 reductions in area source  background levels are 10 percent
 in 1982 and 16 percent in  1990.

      2.   Area Source Modelling

           a.   General Approach

      The point source analysis represents a situation  where the
 total NO- concentrations from the combined impacts of  point and
 area sources are at a maximum.  The meteorological conditions in
 this situation tend to maximize the point source influence  more
 than that of the area sources.  Therefore, the orientation  in
 the point source analysis  is toward point sources  as the  cause
 of, and the sole means to  prevent, violations of a short-term
 standard.  However, there  are other meteorological situations
 under which the area  source impact is enhanced to  a  level where
 they alone may cause  a violation of an NO- standard.

      As one way to capture the maximum impact of area  source
 emissions directly, a simple rollback modelling analysis  was
 made of monitored NO- concentrations and current NO  emission
                           -17-

-------
levels in those AQCR's which may experience future short-term
problems.*  As noted in Volume- I. the monitoring jietworks  in.
most AQCR's are believed to reflect 'the impact of area, as  op-
posed to point, source emissions.  Annual average concentrations
tend to show relatively little variation among stations,  and
most continuous monitors show peak  (second highest hourly
values) to mean (annual average) ratios of 6:1 or less.   Using
ambient air quality data to analyze short-term NO- pollution
from area sources would thus appear reasonable.  However, where
point sources do contribute significantly to peak NO- levels at
monitors which would qualify under the above definition as  area
source-dominated,  the burden of meeting a short-term NO-
standard would fall on both point and area sources.  To the
extent that point sources may be more amenable to control,  the
area source control requirements estimated by this method would
be overstated.

          b.   Model Description

     Rollback is based on a simple proportional relationship
between emissions and ambient air quality:

     Allowable NO   _    N02 standard     x   1975 NO  Emissions
      Emissions
No background NO,, level was assumed in the analysis since the
over al 1 approach^ (i.e., totally ignoring point sources) may
somewhat overstate the need for area source controls.

     The basic procedure involved calculating allowable emission
levels for each AQCR and comparing them to current and future
emission levels.  The percent control needed to meet the standard
was then obtained directly from these estimates.
*A totally comparable approach to the point source assessment
 would be based on area source dispersion modelling in every
 AQCR.  However, this is impossible without detailed knowledge
 of the spatial variations in area source emission levels with-
 in each AQCR.   Even if this were possible, dispersion model-
 ling may not capture the worst case impact of area source
 emissions due to the generalization of emission levels over
 space (i.e., the spreading of emission evenly over a geo-
 graphic area).
                           -18-

-------
          c.   Current Air Quality and Emissions

     The sample of AQCR's on which the analysis is based, in-
cludes all those estimated by SR1/X  to have current one-hour
NO,, concentscrtirons above 200 yg/m .  These include 150 out of
a possible^ 243.AQCR's.  The area source emission patterns for
these AQCR's are shown in Table 5.  In over three-fourths of
the sample,  area sources account for at least 50 percent of
all NO  emissions.  Highway vehicles vary appreciably in their
      X
contribution, but in most regions, the emissions are above 20
percent (and in over one-fourth, above 50 percent) of total
NO  loadings.  External combustion sources (residential,
  X
commercial,  and institutional space heating; space and process
heating in small industrial plants) comprise the other major
category,  but are much less important.  The remaining area
sources are solid waste incineration, internal combustion
engines (e.g., gas turbines used to generate electricity), the
residual mobile sources (off-highway vehicles), and miscellaneous
sources such as forest fires.

     Base year one-hour ambient N02 concentrations for each
AQCR were set equal to the higher of  (a) the second highest
hourly concentrations recorded for all continuous monitors
with peak to mean ratios of 6:1 or less; or  (b) six times the
highest annual average for any 24-hour monitor in the AQCR.
This procedure was designed to eliminate point source influences
while capturing the worst area source N02 problem in each
AQCR.

          d.   Change in Emissions and Impact on Attainment

     Those AQCR's which currently would be unable to attain
one of the short-term NO- standards under consideration could
only achieve it at some future point if controls placed on
area sources brought sufficient net reductions in NO  emissions.
                                                    Jv
Conversely,  growth within those AQCR's which currently attain the
                          -19-

-------
                          TABLE  5
                       ABE*. SCOHC5S TO TOTAL

                       TSZ AQSl'l
                           (1975)
                                               LQAD33IG
                     BtiMlOM Of HO,
          (P«rc«nt Contribution To Total Eaisaions)

AOC*
Ito.
2
3
4
f
7
12
13
14
IS
17
18
19
20
21
22
24
25
28
29
30
31
33
36
38
' 41
42
43
44
45
46
47
49
SO
52
53
54
55
56
58
59
64
65
uii
67
08
69
70
71
72
73
«/
Total
ATM
_ 3oure««
99
100
46
54
86
52
20
28
77
99
SO
51
71
97
55
75
63
95
77
30
75
18
6S
43
78
74
$1
94 •
56
S3
71
48
51
SO
56
45
48
65
39
78
95
18
7U
$0
39
35
32
52
12
65
Thas* an csa
one-hour NO? t
external
Coofiuseion
Area Sourest
08
07
04
06
04
04 "
01
02
07
14
02
07
07
11
OS
07
04
OS
06
07
OS
01
06
03
07
12
14
08
08
03
07
01
01
01
03
02
02
04
02
03
09
02
06
08
03
03
03
04
01
07
AQOt's which !I»VB
sanontrationa com
                                         Light Ouey
                                          Highway
                                          Vehicles

                                             46
                                             S2
                                             21
                                             2S
                                             4S

                                             30
                                             13
                                             IS
                                             41
                                             39

                                             23
                                             22
                                             30
                                             43
                                             27

                                             37
                                             36
                                             56
                                             43
                                             42

                                             41
                                             11
                                             31
                                             20
                                             54

                                             45
                                             31-
                                             66
                                             26
                                             19
                                               t
                                             41
                                             23
                                             23
                                             29
                                             32

                                             24
                                             27
                                             33
                                             20
                                             45

                                             34
                                             08
                                             37
                                             22
                                             IS

                                             IS
                                             17
                                             26
                                             OS
                                             30
Kaavy Duty
 Highway
 Vahiclea

    20
    19
    09
    09
    16

    17
    04
    04
    13
    20

    10
    07
    11
    14
    07

    11
    09
    14
    12
    14

    U
    02
    12
    08
    08

    08
    07
    10
    11
    20

    10
    09
    09
    09
    09

    08
    08
    11
    07
    12

    16
    02
    08
    08
    07

    06
    06
    95
    02
    09
SOURCE;   iPA's NEDS file,  January 9.  1978.access.
                           -20-

-------
                  TRBLE  5  (Continued).


                        Emissions of N0x
            (Percent Contribution of Total Emissions)


            Total          External         Light Duty       Heavy Duty
AOCR        Area          Combustion         Highway          Highway
 Mo.       Sources       Area' Sources        vehicles         Vehicles


 74          69               04                37               08
 75          22               02                11               03
 76          81               09                40               16
 77          25               02                11               04
 78          27               03                12               OS

 79          40               03                21               07
 80          67               08                31               14
 81          96               13                44               18
 82          57               07                29               10
 83          38               04                19               07

 84          36               04                18               07
 85          50               06                19               12
 92          68               OS                30               14
 94          17               02                07               04
 95          S3               03                19               IS

 98          37               02                IS               07
 99          66               OS                26               14
101          94               04                SI               18
102          56               04                28               09
103          24               03                10               03

10S          63               04                29      .   '      10
106          28               02                10               04
107          55               OS                32               09
109          SO               OS                32               09
112          69               03                45               09

113          64               03                41               09
114          71               04       -        46               09
US          56               04                33               08
116          26               01                IS               03
117          80               20                43               09

118          83               23                43               09
119          68               18                33               08
120          63               11                36               08
121          61               10                37               08
122          58               OS                34               08

123          60               07                34               09
124          28               03                15               04
125          77               07                44               11
127          94               07                47               14
128          79               06                37               12

129          69               02                15               07
130          92               12                33               18
131          58               09                26               10
136          59               03                30               13
144          75               18                21               10

145          62               06                22               15
149   '       88               11                56               11
151          67               09                34               12
152          65          -     08                25               11
153          78               OS                35               13
                           -21-

-------
                TA3UL5  (Continuari)


                      SaiMiOM of NO^
          (Pcrcmc Contribution of Tot»l R«i»»iong)
                           Bxc«rn*l         Light Duty       HMvy Oaty
                          Coototwcioa         Higtnwy          Highway
                         Ar»« Soure««        veh'ieiaa         Vchiei«»


155          SI               03                27               10

15»          57               Ot                35               Ot
1*0          38               OS                24               04
1(1          SO               OC                32               OS

112          CO               12                34               06
165          37               02                17               07
144          SI               03                28               12
167          49               02                29               10
16»          79               02                20               09

1«»          17               03                40               IB
171          S9               02                21               25
173          73               06                41               11
174          58               06                31               09
175          96               07                S3               13

176          91               08                49               14
178          36               07                29               09
180          94               07                52               13
111          10               .01                05               01
182          64               03                38           "   09

1*3          22               02                12               03
184          80               OS                37               25
188          SI               03                20               16
1»9          59               02                22               18
193          76               07                37               18
193          64               09                37               41
196          38               05                19               07
197          29               04                14               05
198          82               03                47               12
199          36               02                19               06
200          42  -            02                22               06

201          65               01                37               09
202          76               02                48               11
203          82               03                48               12

208          40                0                22               07

311          38               02                18               05
212          55               02                27               06
214          13                0                05               01
215          79               07                24               27
21«          41               OS   ••            16               05

217          7b               OS                32               10
218          19                0                09               03
220          70  •            16                20               10
223          58               03                30               09
225          41               03                23               07

226     "S3               03                32               08
22'          S6               OS                28               09
234          20               Q8                08               02
237          44                o                20               07
239          «2               07                28               11
243          48               06                13               06
                          -22-

-------
standard could outweigh emission reductions achieved by con-
trols and thus lead to future violations.  Projections of
future attainment status are consequently sensitive to the
assumed growth rates for emission sources and the assumed
effectiveness of emission controls.

     Time trends in emissions were projected separately for
stationary and mobile sources.  Population growth is the
logical driving force for growth in emissions from many station-
ary area sources.  Based on a projected annual population
                                                  8/
growth rate of 0.9 percent for the entire nation,    a 1.0
percent growth rate was set as the upper bound for increases
in stationary area source emissions.  Of course, some area
sources such as solid waste incineration are not expected to
grow at all and could, in fact,  u..*/ decrease  over time.  Con-
 sequently,  an annual  growth rate of zero was  set as the  lower
 bound.   For both growth  scenarios,  no additional NO  emissions
                                                    J^
 control  was assumed.

     A national growth rate for mobile sources  (light and
heavy duty highway vehicles) of between two and three percent
per year in vehicle miles travelled  (VMT) is a reasonable
	~~	*/9/)
expectation.\j/ However,  due to monitor location and the
dispersal and transformation characteristics of NO  , recorded
                                                  X
peak N02 concentrations may be most responsive to emissions
from vehicles on the most heavily travelled roads.  VMT growth
rates for these highways (and thus, the effective growth
rates)  may be much lower than growth in total VMT.  We used
rates of 1.0 and 3.0 in order to bracket this average and thus
test the sensitivity of the results to this assumption.  With
an increase in VMT per year comes a concomitant decrease in
emission rate (gm/VMT) as new vehicles meet the increasingly
stringent Federal emission standards and, over time, become a
larger fraction of the vehicle fleet.  Composite emission
factors for a weighted national average of all highway vehicles
                          -23-

-------
were provided by EPA for each year between 1975 and 1990.  '
The fractional reductions in emission rates were thus combined
with the fractional increase in VMT to obtain changes in
emissions between the years 1975 and 1982, and 1975 and 1990.
Table 6 Summarizes the results.  Improvements in emission
controls as mandated by the 1977 Amendments to the Clean Air
Act are seen to bring significant overall reductions between
now and 1990.  However, examination of the emission changes on
a yearly basis reveals that the lowest emission level occurs
in 1989 (for an annual VMT rate of 3.0 percent) and the level
turns upward in response to increases in VMT thereafter.

     The nationwide fractional changes in stationary and mobile
rea  source  emissions  were  then applied uniformly to all
AQCR's.  Implicit in this procedure is the assumption that the
composition of area sources within the stationary and mobile
categories is everywhere the same.  This is obviously incorrect,
but is reasonable when applied in the framework of an initial
assessment and when compared to the approximate nature of the
ambient air quality estimation procedure.

C.   Air Quality Modelling Results

     1.   Point Source Analysis

     Table 7 presents the results of the modelling analysis in
terms of the following:

     •    Numbers of point sources requiring control
          and the combustors or process facilities  (i.e.,
          SCC's) associated with these point sources.

     •    The number of AQCR's in which these facili-
          ties are located.
                          -24.

-------
                                TABLE 6
         FRACTIONAL CHANGE IN MOBILE SOURCE EMISSIONS OVER TIME
Composite Emission Factor'
   (gm/VMT)
                         a/
                                      1975
                                  4.64
                                                1982
3.25
             1990
2.15
Ratio of Emission Factor to
  1975 Emission Factor
                                  1.00
 .70
 .463
Ratio of VMT to 1975 VMT

  •  1.0% annual growth
  •  3.0% annual growth
                                  1.00
                                  1.00
1.07
1.23
1.16
1.56
Ratio of Emissions to 1975
  Emissions*3^

  •  1.0% annual VMT growth
  •  3.0% annual VMT growth
                                  1.00
                                  1.00
 .749
 ..862
 .538
 .723
a/
b/
Composite emission factors are based on the current level of Federal
tailpipe emission standards, the estimated age distribution of the
vehicle stock, the estimated national distribution of vehicle types
(weighted by annual VMT), and assumed deterioration in emission
controls.  (These values were provided by EPA.)

Computed by multiplying together values in the second-and third rows.
SOURCE:   EEA,  Inc.
                            -25-

-------
                                      TABLE  7

                           POINT  SOURCE MODELLING  RESULTS
                                  (1975 Conditions)
Translation
Approach
Ambient Standard
yg/m
One -Hour Average

Case 1
Case 2
EEA ' s Rate Curves
Case 3
Case 4
• • Point Sources Required to Control
1000
750
500
250
9
25
53
732

1000
750
500
250
183
377
641
5330
6
19
66
638
1
3
10
101
18
35
79
408
SCC's Requiring Control
103
314
705
4774
10
65
284
1185
324
636
1113
4069
AQCR's in Violation
1000
750
500
250
4
6
14
144
4
7
18
143
1
3
6.
39
6
11
30
119
a/
    0  levels are the 1975 second high hourly values in each AQCR.
Cases 1 and 3:  Meteorological conditions correspond to maximization
                of point source influences.   (Area source background =
                highest observed annual NO. average at any monitor
                in each AQCR.)

Cases 2 and 4:  Intermediate meteorological conditions.   (Area source
                background » 1.5 x highest observed annual NO. average
                in each AQCR.)

See Table 3 for a further description of the  four cases.

SOURCE;  EEA, Inc.
                                      -26-

-------
     The cases refer to the matrix presented in Table 3.  For
Cases One and Two, where the  (.1 NO  + OO approach was used,
                                   H    *J
the results are similar.  The ozone limiting component is a
major portion of the N0_ concentration and is unaffected by
the change in meteorological conditions.  Higher NO  levels
                                                   a
in Case One (due to higher wind speed) over-compensate for the
low background level used in this scenario and result in
overall higher concentrations than in Case Two.

     Use of the translation curve approach showed a larger
variation in the control requirements.  Ground-level NO-
concentrations are much higher in Case Four  (intermediate
meteorological conditions) than that in Case Three (worst
case-point source conditions).  This is consistent with the
results of the Chicago case study described previously.  A
much higher degree of point source control is thus required in
Case Four.  The number of AQCR's with modeled receptors showing
one or more violations  (before point source control)  of a 250
yg/m
Three.
yg/m  standard is 119 for Case Four, as opposed to 39 for Case
     For the final cost and economic impact analysis, only
those point sources estimated to cause violations of the 250
and 500 yg/m  standard in Case Four were considered.

     The estimated number of AQCR's not attaining the alternative
standards for Case Four and for each of two years is shown in
Table 8.  These were calculated using the assumptions about
growth of area source emissions described in Section B.l.g.,
and summarized at the bottom of Table 8.

     A total of 4,069 sources associated with 408 industries
located in about 119 AQCR's are estimated to be in violation
                          -27-

-------
                      TABLE 8
   NUMBER OF AQCR'S NOT ATTAINING THE ALTERNATIVE
    SHORT-TERM AMBIENT N02 STANDARDS BASED ON THE
                 POINT SOURCE ANALYSIS
One-Hour Standard                                       ,
     (yg/m j)             Number of AQCR's in Violation '
     1000

      500

      250

SOURCE:  EEA, Inc.
 '  Changes over time are caused by changes in area source
   emissions and thus ambient background levels.   (Mobile
   sources (i.e., travel) are assumed to increase by 3.0
   percent per year, new vehicle emission factors correspond
   to the Federal Motor Vehicle Control Program, and sta-
   tionary area source emissions increase by 1.0 percent
   per year.)  Point source growth is not considered since
   new sources considered individually should not lead to
   violations.
                         -28-

-------
of a 250 yg/m  standard in 1975.  For the 500 yg/m  standard,
the number of affected sources and AQCR's is reduced signifi-
cantly.  Seventy-nine industries with about 1,113 processes and
approximately 30 AQCR's are estimated to be in violation of
the standard.  A breakdown of the types of processes (source
classification codes—SCC's)  and industries (standard indus-
trial classifications—SIC's) that are likely to contribute to
violations of the 250, 500, 700, or 1,000 yg/m  standards is
shown in Tables 9 and 10.

     2.   Area Source Analysis

     Table 11 summarizes the results of two growth scenarios
(low:  0 percent stationary and 1.0 percent mobile source
growth  rate;  high:   1.0  percent stationary  source  and  3.0
percent mobile  source  growth rate),  assuming mobile  source
emission  standards  will  remain  as  currently mandated.
Except  for  the  250  yg/m   standard,  only  a few AQCR's are
estimated to be in  nonattainment status  due to  area  source
emissions.   For the 250  yg/m standard,  the current  Federal
motor vehicle control  program is seen  to effect a  considerable
improvement  in  attainment  status over  time, though more
than 70 AQCR's  may  still experience  violations  in  1990.
     3.   Point and Area Source Analyses Together

     Table 12 lists the number of AQCR's estimated to be in
nonattainment in either of the two analyses.   (The data are
taken from Tables 8 and 11.)  The two analyses represent two
separate air quality situations.  As noted earlier, the point
source analysis is designed to reflect the maximum expected
ambient N02 levels due to the influence of both the point and
area sources together  (though point sources predominate);
these levels are probably not captured by the existing monitor-
ing networks in most urban areas.  However, the control of

                         -29-

-------
                       TABLE 9

    TYPE OP SOURCES  LIKELY TO EXCEED SPECIFIED
            N02 LEVELS (CASE FOUR ANALYSIS)
                             NO.,  Levels (yg/m3)
Source Category
Utility Boilers - Coal
Utility Boilers - Oil
and Gas
Industrial Boilers -
Coal
Industrial Boilers -
Oil and Gas
Gas Turbines
Reciprocating 1C Engines
Industrial-Combustion
Processes
Nitric Acid
Municipal and Industrial
Incinerators
250
350
599
300
742
268
698
1,045
61
11
500
42
7
72
207
19
516
235
19
1
750
15
0
10
108
10 -
376
114
3
0
1000
0
0
0
21
5
278
17
3
0
TOTAL            4,069       1,113        636         324
                   -30-

-------
                             TABLE 10
TYPE OF INDUSTRIES LIKELY TO EXCEED SPECIFIED ONE-HOUR  N02  LEVELS
                       (CASE FOUR ANALYSIS)^
                                        NO.,  Levels (ug/m3)
Utility
Boilers
Internal Combustion - Reciprocating
Industrial - In Process Fuel Use
Gas & Oil Pipe Lines (I.e. Engines)
Refineries
Steel
Metal Melting
Asphalt
Lime Kiln
Glass
Cement
Food
Automotive
Waste Water Treatment
^Miscellaneous
Non-Combustion
Chemical
(Nitric Acid, etc.)
Total
250
136
26
48
37
23
15
10
4
4
3
4
4
3
67
24
408
500
6
14
33
8
5
' 2
1
0
-
-
-
-
-
10
6
79
750
2
7
17
5
2
0
0
-
-
-
-
-
-
-
1
34
1000
0
4
9
3
0
0
0
-
-
-
-
-
-
-
1
17
                               -31-

-------
                                                    TABLE 11  ,


           CHANGES IN AREA SOURCE EMISSIONS FOR 150 AQCR's AND AQCR ATTAINMENT STATUS FOR ALTERNATIVE

                                       GROWTH RATES AND ONE-HOUR NO2 STANDARDS3/
     Area Source Emissions Summed
     For All 150 AQCR's
     (tons x 106/year)
                                     1975
                                                            1982
                                           High
                                          Growth
                                                         b/
 Low
Growthb/
                                                                            1990
 High
Growth
 Low
Growth
i
Ol
• Mobile Sources
• Stationary
Total
Number of AQCR's Not
Attaining the Standard
• lOpQ yg/m
• 750 pg/m3
• 500 pg/m
• 250 ug/m
6.3
3.2
9.5


"""p""
2
17
94
5.6 4.9
3.4 3.2
9.0 '8.1


0 0
2 0
10 4
84 68
4.9
3.7
8.6


0
0
7
73
3.6
3.2
6.8


0
0
2
45
     a/
     b/
Based on 150 AQCR's recording (or estimated to exhibit) second highest one-hour NO_ levels of
200 |ig/m3 or more in 1975.

"Low growth" assumes a 1.0 percent annual growth rate for VMT and a zero percent annual growth
rate for stationary area sources.  "High growth" assumes a 3.0 and 1.0 percent annual growth rate
for VMT and stationary area sources,  respectively.   Statutory mobile source emission standards
are also assumed.

-------
OJ
I
                                              TABLE 12


                 NUMBER OF AQCR'S ESTIMATED TO BE IN NONATTAINMENT IN BOTH THE AREA

                  AND POINT SOURCE ANALYSES (NO ADDITIONAL SOURCE CONTROLS ASSUMED) a//
     One-Hour Standard
               One-Hour Standard      '       Number of AQCR'S in Violation
                         ):                   1975        1982         1990

                    1000                         666
                     500                        45          36           29
                     250                       158         145           133
      '  High area source growth assumed:   3.0 percent increase in travel and 1.0 percent
        increase in stationary area source emissions.

-------
point sources alone does not assure attainment of the various
standards under all meteorological conditions in all AQCR ' s ,
though control of point sources alone is estimated to attain
the standards under the conditions used in the point source
       -                              s
analysis.  Under certain other extreme meteorological conditions,
the area source impact alone  (even though not as severe  as  the
combined point and area source impacts) may  exceed one or more
of the various standards.  The area source analysis  attempts  to
capture this possibility.  Thus,  the  two together provide a basis
for estimating the number of AQCR's in nonattainment under  all
conditions.
     It would appear that about 160 AQCR's would currently
    ate a 250 ya/m  standard and about 45 would violate a 500
      standard if sufficient monitors were available to record
them.  By 1990, without additional emission controls, the' num-
bers would decline to 133 and 29, respectively.

D.   Control Options and Cost Analysis

     1.   Point Source Control Options and Unit Costs

     Each plant identified as having a potential to exceed any
of the alternative standards was analyzed to determine the
percent reduction required to reach each level with the types
of control available.  Tables 13 through 19 show the kinds of
control available for each source type.  Controls and efficien-
cies are taken from EPA's. control techniques  document  for nitro
gen oxides.    These tables also show the date of availability,
capital and annual costs  (expressed as $/MMBtu) , and the
effect on fuel consumption.   (Annualized costs are in 1976
dollars and assume a capital recovery factor of 16 percent.)
                          -34-

-------
                                                    TABLE  13
                              1 I '


                       COST AND EFFECTIVENESS OF NOX CONTROLS FOR UTILITY BOILERS - COAL-FIRED
                                                           Differential Control  Costs
Ol
01
I

Control Techniques*'
LEA
LEA + OSC
Retrofit: Low
NOX Burner
Retrofit: Dry
SCR (only NOX)

Control
Potential
11%
22%

40%

90%
Earliest
Year
Available
Present
Present

1980

1985

Initial Investment
(102$/106Btu/hr.)
0.60
1.25

2.00

60.00
Annual
Costc
(f/106Btu)b/
0.2
0.5

0.8

26.0
Effect On
Fuel
Consumption
.5% Decrease
0.5% Increase

0

3.0% Increase
'
       a/
        'LEA = Low Excess Air
        OSC = Off-Stoichiometric Combustion
        SCR => Selective Catalytic Reduction

        Annual Cost = Initial investment  annualized at 16 percent plus operation and maintenance  costs.

             6                                                                    4
        14/10 Btu annual heat input =  0.54 $/kW 8 5,400 hours of operation and 8 10 Btu =  1  kWh.
                                    =  0.1 nil/kW hr.
        Nominal heating value of coal  = 28 KJ/KG (12,000 Btu/lb).

       SOURCE:   Constructed from data in Reference  3

-------
                                                  TABLE 14
                  COST AND EFFECTIVENESS OP NOX CONTROLS FOR UTILITY BOILERS - GAS- 4 OIL-FIRED
                                                            Differential Control Costs
Ol
I
a/
Control Techniques
LEA
LEA + OSC
LEA + OSC + FGR
LEA + OSC + Nil 3
Injection
Retrofit: Dry SCR
(NOX only)
Control
Potential
17%
40%
59%
70%
90%
Earliest
Year
Available
Present
Present
Present
1981
1982
Initial Investment
(102J/106Btu/hr.L
0.3
0.8
9.0
8.0
60.0
Annual
Cost
(f/106Btu)b/
Neg.
0.28
2.8
14.0
26.0
Effect On
Fuel
Consumption
0.5% Decrease
1% Increase
1% Increase
1% Increase
3% Increase
      a/
        FGR - Flue Gas Recirculation
        SCR = Selective Catalytic  Reduction
      b/lf/106Dtu annual heat  input « 0.54 $/kW for 5,400 hours  of operation 9 10 Btu - 1 kH hr.
                                  - 0.1 mil/kWh
      SOURCE:   Constructed from data  in Reference 3

-------
                                            TABLE 15
                COST AND EFFECTIVENESS OP NOX CONTROLS  FOR  INDUSTRIAL BOILERS - COAL-FIRED
                                                     Differential Control Costs
Control Techniques
LEA
LEA + OSC
Retrofit: Low NOX
Burner
Retrofit: Dry SCR
(NOX Only)
Control
Potential
10%
20%
50%
90%
Earliest
Year
Available
Present
Present
1985
1985
Initial Investment
(102$/106Btu/hr.)
0.7
1.8
3.0
50.0
Annual
Cost •
0.29
0.66
1.10
23.0
Effect On
Fuel
Consumption
1% Decrease
1% Increase
0
3% Increase
 a'l
-------
                                                TABLE 16
                  COST AND EFFECTIVENESS OP NOX CONTROLS FOR INDUSTRIAL  BOILERS - GAS- $ OIL-PIRED
                                                           Differential Control Costs
01
00
I
Control Techniques


l.liA

LCA + OSC

Retrofit: Low NOX
Burner

Flue Gas Treatment
Retrofit: Dry SCR
(NOX Only)
Control
Potential
10%
16%
Earliest
Year
Available
Present
Present
Initial Investment
(lo2|/l06Btu/hr.)
0.7
1.4
Annual
Cost
(»/106Btu)a/
0.26
0.6
Earliest
Fuel
Consumption
1% Decrease
1% Increase
                              50%
                              90%
1981
3.0
                   50.0
1.1
                 23.0
0
             3% Increase
       M/10 Dtu annual heat input »'0.1 mil/kW hour.
   SOURCE:  Constructed from  data in  Reference 3

-------
                                                TABLE 17

                   COST AND EFFECTIVENESS OP NOX CONTROLS FOR GAS TURBINES AND 1C ENGINES
        Control Techniques
                               Differential Control Costs	
               Earliest                              Annual
 Control         Year        Initial Investment       Cost
Potential     Available      (102|/106Btu/hr.)      (j/106Btu)^
                                                   Effect On
                                                     Fuel
                                                 Consumption
        Gas Turbines
CM
to
I
          Water  Injection
           Engines

          Fine Tuning Q
          Changing A/P

          Retrofit Dry SCR
   60%
   30%

   90%
Present
Present
                 1985
12.5  ,
 0

30.0
 4.6
 2.0

17.0
 2% Increase
10% Increase

 1% Increase
       a/lf/106Btu = 0.1 mil/kWh 8 104Btu/l kWh.
  SOURCE:   Constructed from data  in Reference  3,

-------
                                             TABLE  18





                   COST AND EFFECTIVENESS OP  NOX CONTROLS FOR INDUSTRIAL PROCESS FURNACES
                                                         Differential Control Costs
-k
Earliest
Annual Effect On
Control Year Initial Investment Cost Fuel
Control Techniques Potential Available (102$/106Btu/hr.)a/ (f/106Btu)V Consumption
l.liA 25% Present 1.03
LliA + FGR 40% Present 5.80
Advanced Design
Burner New Equipment/
Retrofit 50% 1981 3.00
Retrofit Dry SCR
(NOX Only) 90% 1985 .50.00
0.4 1% Decrease
1.9 1% Increase
1.1 0
20.0 2% Increase
     al kWe = 10Btu/hr.




     b/U/106Utu annual  heat  input -0.1 mil/kWh.
   SOURCE:  Constructed from data in Reference 3,

-------
                                           TABLE  19
                      COST AND EFFECTIVENESS OF N0y CONTROLS FOR NITRIC ACID MANUFACTURING
  Control Techniques
              Earliest         Differential Control Costs
 Control         Year       Initial InvestmentAnnual Cost
Potential     Available        lp3|/Ton/daya/        I/Ton*/
                                                 Effect On
                                                   Fuel
                                                Consumption
  Chilled Absorption
   90%
Present
2.0
2.0
0
a/ Darned on a model plant of  300 tons/day capacity operated  for 8,000  hours per year,
SOURCE:   Constructed from data in Reference 3.

-------
     For each source category, up to five levels of control were
established:
     •    Low Excess Air (LEA)—Operation of the source
          burner at close to theoretical air.  This is
          the easiest and least costly control techno-
          logy and can provide a 10 to 17 percent re-
          duction in uncontrolled NO  emissions.
     •    Low Excess Air Plus Off Stoichiometric
          Combustion (LEA & OSC)—This includes any
          of a series of combustion modifications
          which reduce peak flame temperature and
          suppress NO  formation.  This technique is
                     J£
          applicable to most source types, is imme-
          diately available, and provides an estimated
          20 to 40 percent reduction in NO  emissions.
                                          X
          Flue Gas Hecirculation (FGR)—In addition to
          LEA and OSC, FGR can be used to provide an •
          extra 20 percent reduction in NO  emissions
                                          Jt
          for cas- and oil-fired sources.

          Advanced Burner Designs—Special burners
          designed to operate at very low excess air
          and use advanced combustion techniaues are
          being developed by EPA and industry.  If
          these proarams are successful, they offer
          about a 50 percent reduction in NO  emis-
                                            Ji
          sions.  For this study, it has been assumed
          that the burners will be developed and will
                          -42-

-------
          be available for new and existina sources.
          Retrofit costs are assumed to be considerably
          higher  than new  installations.

     •    Selective Catalytic Reduction (SCR)—The
          reduction of ammonia and NO  over a catalyst
                                     X
          can be used to reduce NO  emissions by up to
                                  X
          90 percent.  SCR systems are operating on several
          laroe oil- and oas-fired boilers in Japan, but
          have not yet been demonstrated on U.S. boilers.
          In this study, it has been assumed that SCR
          will be developed and available for use in
          1982.  Costs for SCR are taken from EPA studies
          and ad-iusted to allow for retrofit (includina
          installation of a flue gas heater).

     The above control techniques were assumed to be applicable
to boilers and industrial furnaces.  Specific techniques
such as water injection, fine tuning, and chilled absorption
were assumed for gas turbines, internal combustion engines,
and nitric acid plants.

     Cost and effectiveness values for each control type are
reasonably well-established for boilers and process sources,
but limited investigation of NO  reduction has been carried
                               X
out for furnaces such as cement kilns, petroleum heaters,
and glass melting furnaces.  Based on the NO  control techni-
             3/                                   ll/
cues document / and communications with Aerotherm,  ' we
have assumed that combustion modifications and FGR are as
effective for these sources as for boilers.

     It should be noted that, due to time and data constraints.
all control costs are assumed to be linear with size (i.e.,
a constant cost per unit size regardless of capacitv).  This
will introduce some error in the costing.  The overall
                          -43-

-------
magnitude of the error introduced is not clear  since  the
linear assumption will tend to overstate some costs while
understating others.

     2.   Point Source Cost Analysis Procedure

     Each plant identified as being capable of causing an
N02 concentration greater than one of the specified standards
was evaluated individually to determine the type of control
required and the cost and energy penalty of the control
technique.  The procedure followed is outlined below:

     •    The maximum concentration from each source  in
          the plant (estimated by the dispersion model)
          was summed to determine the total concentration
          of N02.  This concentration was then compared
         .-to each of the specified standards (250 to  1,000
          ug/m }  to determine whether emission control was
          needed.

     •    The type of control used by each source within
          the plant to meet each alternative standard was
          determined from the combination of controls at
          all sources which would achieve required reductions
          in ambient N02 at least cost.  For each source,
          each applicable control technology was considered.
          Combinations of sources and technology were ranked
          by annual cost of reducing the plant's air quality
          by 1 yg/m .  This measure ($/ug/m ) was used since
          it combines the air quality impact of the source
          with the cost of controlling its emissions.  Con-
          trol technologies were selected one by one, start-
          ing with the lowest cost (highest cost-effective-
          ness) option until the air quality goal was met.
                         -44-

-------
     o    After technologies were selected for each
          source within a plant, the capital costs and
          the effect on source and plant fuel consump-
          tion were computed.

     3.   Point Source Control Costs

     Control costs only for standards at 250 and 500 yg/m  are
presented here.  For levels above 500 yg/m , the costs were in-
significant.  As noted in section B.l.e, a conservative but rea-
listic set of meteorological conditions and a theoretically more
justifiable approach used for estimating the NO- concentrations
form the basis for selecting case four for the final economic
impact analysis.  But for the purpose of comparison, the capital
and annualized costs for all four cases are presented in Table
20.

     The capital cost of point source control to meet a 250
yg/m  standard could range anywhere from about 600 million
to over 3 billion dollars, depending on the estimation
approach.  For the 500 yg/m  standard, the capital cost is
reduced significantly, and ranges from 14 to 46 million
dollars.  The annualized cost for both the 250 and 500
yg/m  standards is significantly affected by the impact on
energy consumption.  Frequent use of combustion modification
results in a net savings in the fuel cost in almost all cases.
The savings in some cases are substantial enough to completely
balance the annual cost of control.  The use of more stringent
controls incurs an energy penalty.  The variation in the fuel
impact among the different scenarios is dependent on the
severity of the air quality impact in each case and the level
to which individual sources have to be controlled.
                           -45-

-------
                                                       TABLE  20

                                   POINT SOURCE CONTROL COSTS FOR DIFFERENT CASES
Case Number
and Description
CASE 1: Worst Case - Point
Source (low background)
.1 NO + 0
x 3
CASE 2: Intermediate Case
(high background)
.1 NO +0,
x 3
CASE 3: Worst Case - Point
Source (low background)
EEA's Translation Curves
CASE 4: Intermediate Case
(high background)
EEA's Translation Curves
250 |ig/m Standard
Capital Cost9/
(106 dollars)

3,073

2,169

589

1,496
Annual Cost"/
(106 dollars)

630

480

66

363
500 pg/m Standard
Capital Cost9/
(106 dollars)

37

14

4

46
Annual CoatP/
(106 dollars)

11

-1

-1

-1
a/ 1976 dollar basis.
b/ Annual Cost = capital cost annualized @  16  percent  +  operation  and maintenance cost based on capacity
   and hours of operation + fuel penalty.
 SOURCE:  EEA, Inc.

-------
     Table 21 lists further details of the Case Four analysis.
The plants and combustors shown are those which the costing
algorithm selected as being least costly to control.  The
number of sources is smaller than shown in Table 8 because
not all sources within a plant have to be controlled in
order to achieve the necessary emission reduction.

     Tables 22 and 23 show a breakdown by source type of
capital and annual costs (Case Four) for the 250 and 500
yg/m  standards.  Table 24 provides a legend for this table.

     4.   Impact of Growth on Point Source Control Costs

     No attempt has been made at this point in the cost
analysis to account for the impact of changes in emissions
from point sources over time.  As noted in the methodology
section, a proper treatment of growth would consider both
point and area sources and both emission-increasing and
emission-reducing factors.  Factors which may lead to reduc-
tions in emission levels include increasingly stringent
emission requirements for mobile sources and the displacement
of fossil fuel by electricity for space heating purposes.
Both of these factors should lead to a decrease in NO
                                                     x
emissions from mobile and stationary area sources.

     On the other hand, increases in electrical demand and a
shift to coal from oil and gas in major fuel burning installa-
tions, will lead to increased emissions from point sources.
(Emission factors for coal-fired equipment run two to four
times higher than gas-fired equipment, and one to two times
higher than oil.)  In addition, the use of flue gas desulfuri-
zation may increase the ambient concentration of N02/ by
reducing flue gas temperature and thus plume rise.  Since
the Clean Air Act Amendments of 1977 require that new fossil
fuel-fired stationary sources meet fixed emission removal
standards, FGD systems are likely to be employed in most
utility and industrial combustors constructed after 1980.
                            -47-

-------
                                                         TABLE 21

                                DETAILS OF POINT SOURCE OCMmOL COST /JJALYfIS FOR CASE POUR
NO Standard
(One-Hour Average)
ug/m3
500
250
Number of
Plants
Affected
79
408
Number of
Sources
Controlled
794
3,628
Approximate
Capital Costs
(106 dollars)
46
1,496
Approximate
Annual Cost
(106 dollars )b/
-1
363
Approximate
Fuel Penalty
(Barrels of
oil/day
Equivalent)0
-2,331
-4,096
00
I
      a'    Mostly combustors-.


            Including fuel with coal assumed at  $1.60/MMBtu and oil  and gas at $2.60/MMBtu.


            A negative  sign  indicates  a fuel  savings.
       SOURCES:   EEA,  Inc.

-------
   TABLE  22 - COST BY  COMBUSTOR TYPE FOR ATTAINING A 500 yg/nf5 STANDARD
                                                a/
 Source Category  1
                              ^

 NUMBER OF USES                12
 EMISSIONS REDUCED(G/S>       524.
 INITIAL COST (1000'S*)      1826.
 ANNUAL COST (1000'S*>      -1776.
 FUEL COST (10E+9BTU)       -1329.
 Source Category  2L	,  _
                   I..        -. 1
 NUMBER OF USES                 2
 EMISSIONS REDUCED(G/S)      •  75.
 INITIAL COST (1000'S*)       149.
 ANNUAL COST (1000'S*)       -514.
 FUEL COST (10E+9BTU)        -207.
.Source ..Category-1
 NUMBER OF USES
 EMISSIONS REDUCED(G/S)
 INITIAL COST  (1000'SS)
 ANNUAL COST (1000'S*)
 FUEL COST (10E+9BTU)
. Source.. _Ca tegory_4	

 NUMBER OF  USES
 EMISSIONS  REDUCED(GXS)
 INITIAL COST  (1000'S*)
 ANNUAL COST (1000'S*)
 FUEL COST  (10E+9BTU)
 Source  Category 6

 NUMBER OF USES
 EMISSIONS REDUCED  1
   42
  122.
  696.
-1038.
 -758.
-  1
  168
  130.
 2397.
-6873.
•2849.
   1
   23
   77.
   0.
 1013.
  362.
 Source "Category 7^

 NUMBER OF USES
 EMISSIONS REDUCED(G/S)
 INITIAL COST  (1000'S*)
 ANNUAL COST (1000'SS)
 FUEL COST (10E+9BTU)
   1
  123
  307.
 1714.
-3002.
-1308.
.Source  .Category 8

 NUMBER OF USES
 EMISSIONS REDUCED(G/S)
 INITIAL COST (1000'S*)
 ANNUAL COST (1000'SS)
 FUEL COST (10E+9BTU)
_Source  Category. 9.

 NUMBER OF USES
 EMISSIONS REDUCED(G/S>
 INITIAL COST (1000'S*)
 ANNUAL COST (1000'S*)
 FUEL COST (10E+9BTU)
    6
   57.
 1152.
  304.
    0.
    1
    2.
    0.
    0.
    0.

•^
0
0.
0.
0.
0.

2
0
0.
0.
0.
0.

2
0
0.
0.
0.
0.

2
0
0.
0.
0.
0.

2
309
1322.
8559.
3509.
202.

2
1
. 49.
816.
301.
61.

2
0
0.
0.
0.
0.

*»
0
0.
0.
0.
0.
TECHNOLOGY
3
14
320.
1546.
346.
0.
TECHNOLOGY
3
0
0.
0.
0.
0.
TECHNOLOGY
3
7
23.
148.
31.
0.
TECHNOLOGY
3
0
0.
0.
0.
0.
TECHNOLOGY
3
0
0.
0.
0.
0.
TECHNOLOGY
3
9
27.
, 326.
53.
0.
TECHNOLOGY
3
0
0.
0.
0.
0,
TECHNOLOGY
3
0
0.
0.
0.
0.

4
1
45.
2898.
748.
66.

4
0
0.
0.
0.
0.

4
3
11.
445.
172.
23.

4
21
1444.
7650.
3245.
•379.

4 '
0
0.
0.
0.
0.

4
52
190.
9055.
2709.
242.

4
0
0.
0.
0.
0.

4
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
TOTAL
27
889.
6270.
-681.
-1263.
TOTAL
o
75.
149.
-514.
-207.
TOTAL
52
156.
1289.
-835.
-735.
TOTAL
189
1574.
10047.
-3628.
-2470.
TOTAL
332
1399.
8559.
4524.
564.
TOTAL
185 .
573.
11911.
62.
-1004,
TOTAL
6
57.
1152.
304.
0.
TOTAL
1
•? ^
0.
0.
0.
   a/
      See Table 24 for  legend
                                      -49-
                    SOURCE:  EEA,  Inc.

-------
TABLE -23
_OQSI_.RY_COMmiS!ro& TYPE- 23R...A1T£IHISG_A  250._U qZmJ. STANDARD'
Source Category 1

NUMBER OF USES
EMISSIONS REDUCED (G/S)
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
"source" 'Category ~Z ~

NUMBER OF USES
EMISSIONS REBUCEJHG/S)
INITIAL COST <1000'S»)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
Source Category 3
NUMBER OF USES
EMISSIONS REDUCED < G/S >
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
Source Category 4

NUMBER OF USES
EMISSIONS REDUCED < G/S >
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
Source Category 5

NUMBER OF USES
EMISSIONS REDUCED (G/S)
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)

... . Source _Cateao.ry_6_
"NUMBER OF USES
EMISSIONS REDUCED (G/S)
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
Source Category /

NUMBER OF USES
EMISSIONS REDUCED (G/S)
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
Source Category a
NUMBER OF USES
EMISSIONS REDUCED (G/S)
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)
	 Source Category 9
NUMBER OF USES
EMISSIONS REDUCED (G/S)
INITIAL COST UOOO'S*)
ANNUAL COST UOOO'S*)
FUEL COST UOE+9BTU)

1
139
3213.
14943.
-13342.
-10111.

1
273
3438.
7084.
-21866.
-8854.
— 1
159
636.
4478.
-6390.
-4685.

1
467
573.
8743.
-24569.
-10195.

1
132
1622.
79311.
24472.
3863.

1
40
342.
0. -
5051.
1804.

1
490
6921.
45468.
-79221 .
-34530 .
~ 1
42
580.
15072.
3966.
0.
1
8
20.
0.
0.
0.

2
12
369.
2586.
1771.
730.

2
26
705.
1591.
3531.
1233.
2
7
64.
349.
315.
150.
^

2
9.
196.
372.
122.

2
0
0.
0.
0.
0.

2
566
4219.
29445.
12254.
712.

^
13
76.
1263.
434.
S3.
2
0
0.
0.
0.
0.
2
0
0.
0.
0.
0.
TECHNOLOGY
3*
4
104 65
5549 . 5595 .
21611. 281676.
5213. 92835.
0. 11023.
TECHNOLOGY
3*
4
163 20
3670. 641.
64147. 8212.
23885. 5277.
4990. 279.
TECHNOLOGY
3 4
61 58 •
1003. 1213.
5578. 65355.
1165. 23175.
0. 2935.
TECHNOLOGY
3 A
**
27 218
133. 2607.
1523. 139190.
300 . 55444 .
0. 6220.
TECHNOLOGY
3 4
0 0
0. 0.
0. 0.
0. 0.
0. 0.
1
TECHNOLOGY
3 4
0 0
0. 0.
0. 0. —
0. 0.
0. 0.
TECHNOLOGY
3 4
78 366
1030. 1766.
16152. 110010.
2626. 34869.
0. 3321.
TECHNOLOGY
3 4
0 0
0. 0.
0. 0.
0. 0.
0. 0.
TECHNOLOGY
3 4
0 0
0. 0.
0. 0.
0. 0.
0. 0.
e-
if
0
0.
0.
0.
0.


75
7006.
525564.
198040.
21366.
5
0
0.
0.
0.
0.


0
0.
0.
0.
0.


0
0.
0.
' 0.
0.

5
0
0.
	 0.
0.
0.

5
0
0.
0.
0.
0.
5
0
0.
0.
0.
0.
5
0
0.
o.
0.
0.
TOT — 1
1 U 1 itU
320
14926.
320815.
86476.
1642.
THTAI
( U 1 ML.
359
15460.
606597.
208847.
19014.
TOTAL
285
2916.
75759.
18264.
-1600.
TOTAL
i w i nw
714
3322.
149652.
31547.
-3853.
TrtTAI
TOTnL
132
1622.
79311.
24472.
3863.

TOTAL
606
4361.
29445.
17305.
2516.
TflTAI
TOTAL
947
. 9792 .
172893.
-41293.
-31125.
TOTAL
42
580.
15072.
3966.
0.
TOTAL
3
20.
0.
0.
0.
    a/
       See Table 24  for legend
                           SOURCE;  EEA, Inc.

-------
                              TABLE  24
                    LEGEND FOR TABLES 22 AND 23
 Source
^Category  Description
      Technology
           Utility Boiler
              (coal)


           Utility Boiler
            (oil and  gas)


           Industrial Boilers
                (coal)

           Industrial Boilers
             (oil and gas)


           Gas Turbines
           1C Engines
           Industrial  Pro-
           cess Combustion

           Nitric Acid
1
LEA
LEA

LEA
LEA
Water
Inject
Engine
Modifications
LEA
Chilled
I
LEA+OSC
LEA+OSC

LEA+OSC
LEA+OSC
N/A
sea
LEA+FGR
N/A
I
Low NO
Burner
LEA+OSC
+FGR
Low NO
Burner31
LEA+OSC
+FGR
N/A
N/A
Low NO
Burner
N/A
£
SCR
LEA+OSC
+NH3
Injection
SCR
SCR
N/A
N/A
SCR
N/A
5_
-
SCR

S/A
N/A
N/A
N/*
N/A
B/A
                               Absorption
           Municipal
           Incinerators
-no control  possible—-—	
   NOTE;    LEA = Low Excess  Air
           OSC - Off Stoichionetric Combustion

           FGR • Flue Gas  Recirculation
           SCR - Dry Selective Catalytic Reduction

           N/A - Not Available


   SOURCE;   EEA,  Inc.
                                   -51-

-------
               As noted in Section B.l.g., growth in point source
          emissions was not considered since the methodology used for
          the national assessment could not account for interaction
          among point sources, and since major new sources of NO  will
                                                                X
          be adequately regulated by new source performance standards.
          However, the impact of mandated mobile source controls (perhaps
          the most obvious factor impacting future NO  emissions) was
          considered.  Control costs were calculated for 1982 assuming
          an approximately 20 percent reduction in background N02
          concentrations due to a 1.0 percent increase in travel, a 30
          percent reduction in the composite mobile source emission
          factor, and no growth in stationary area sources.  (This low
          growth scenario leads to an even greater reduction in control
          requirements than the high growth scenario shown in Table
          12.)   A comparison of the control costs with and without the
          change in area source emissions is shown in Table 25.
                "
   0?    __-•     The change in both capital and annual costs are substan-
  f>                                      3
  _^       tial for a standard of 250 yg/m , but small for a standard of
>/•'        500 yg/m .  However, these savings may be counteracted to an
          undetermined extent by the location of new point sources at
          sites where plume interaction is significant.  In this case/
          the assumption of plant isolation is not satisfied and control
          costs would consequently rise.

               5.    Area Source Control Options and Costs

                    a.   Stationary Sources

               Current available techniques for the reduction of emissions
          from space heating units are:  (1) tuning — the best adjustment
          in terms of the smoke-C02 relationship that can be achieved
          by normal clean up, nozzle replacement, and simple scaling
          and adjustment with the benefit of field instruments;  (2)
          burner replacement — installation of a new low emission burner;
          and (3)  unit replacement — installation of new advanced low
          NO  unit.
            X                       -52-

-------
                          TABLE 25

     IMPACT OF CHANGES IN AREA SOURCE EMISSIONS ON POINT

                    SOURCE CONTROL COSTS
                           Point Source Control Costs   ($10 )

                                               1975 Point and
                                             Stationary Area Source
One Hour           1975 Point and Area       Emissions, 1982 Mobile
Standard           Source Emission Base      Source Emissions
(yi'g/m )             Capital       Annual      Capital"         Annual

  500                 46           - 1          34             ~1


  250              1,496           363         809            124
 '   Low growth mobile source scenario: 1.0 percent growth in travel
    and zero percent growth in stationary area sources.
SOURCE;   EEA, Inc.
                           -53-

-------
      Burner maintenance  or replacement typically has a bene-
 ficial  impact  on all  pollutants  except NO.   Therefore, new
 fuxaacss  with  advanced design or low NO  emissions hold the
 most promise for the  control  of  NO  from residential/commercial
 space heating.   These advanced designs are based on one, or a
 combination of,  combustion modification techniques described
 briefly under  the point  source section.   Low NO  systems for
 installation in  new homes  and stores are available at a cost
 of  ten  percent or more above  conventional systems.  Replacement
 of  existing furnaces  is  obviously much more expensive (about
 $700 for  a  residential furnace.    }   The NO  emission reduction
 potential of these new systems is up to 80 percent,  and the
 increase  in operating efficiency is  about ten percent.

     Small gas- and oil-fired firetube boilers are used  to
heat some commercial/institutional and many relatively  small
industrial buildings.   Existing units can be retrofitted (LEA
and FGR) at a capital cost of about  $60 per million Btus/hr
capacity  (annualized cost of about $.01 per million Btus),
while new boilers with improved burners would cost about
$40 per million Btu/hr ($.007 per million Btu/year) more
than conventional boilers.*  Emission reductions would  be
about 40 percent.  However, many existing facilities would
require boiler replacement or supervision by full-time
                                                   18/
operators.  A new boiler would cost  up to $200,000  /.

          b.   Mobile Sources

     The. Clean Air Act Amendments of 1977 mandate  an emissions
standard of 1.0 g of NO  per VMT for light duty vehicles
 (LDV's)  by 1981,  which we have assumed extends through  1990  in
the baseline projections.  The next  step in a schedule  of
increasingly stringent NO  standards is 0.4 g/VMT.  Based on
                                         12/
information in the "Three-Agency Report,"  ' this  could  be
achieved for light weight cars by installing a three-way
'Assuming a 16 percent capital recovery rate,
                             -54-

-------
 catalyst,  together with an oxygen sensor in the exhaust stream,
 a mechanical fuel injection system,  and an upgraded electronic
 control unit.   These would replace the improved fuel metering,
 air injection,  start catalyst,  and oxidation catalyst facilities.
 For heavy  autos (greater than 3,000  pounds), the incremental
 control package includes electronic  spark control,  mechanical
 fuel injectors, a switched air aspirator, a switched-start
 catalyst,  and an upgraded electronic control unit.   (A three-
 way catalyst is required, but is also needed to meet the 1.0
 g/mile standard.)   These replace a less exacting package of
 fuel,  ignition, and exhaust controls necessitated by the less
 stringent  standard.   Though these specifications apply, strictly
 speaking,  only to autos, we will assume they apply to all
 vehicles less than 6,000 pounds.

      Additional capital costs are estimated to be approximately
 $80 in 1985  (in 1977 dollars)  for the control configuration
 which is fuel-optimal.   '   (The less expensive "cost-optimum"
 configuration would incur substantially more in lifetime fuel
 costs than the initial  savings.)  Lifetime maintenance expenses
 are estimated at $30 '  or as a first approximation at a rate of
 $3 per year.   Though these are  rough approximations of costs
 for unproven technologies, they serve as useful initial esti-
 matesT '""

      Inspection and maintenance programs in individual AQCR's
 could be instituted or  expanded to cover NO  emission controls.
                                            X
 Since I&M  programs for  NO  currently do not exist,  it is
 	„ ^                      Ji
KJLmpossible to accurately gauge  their costs and effectiveness.
 Cost estimates for hydrocarbon and carbon monoxide I&M programs
 which employ a dynamometer (a necessary ingredient for NO
                                                          X
 emission tests)  suggest that the inspection costs are approximately
 $5 per test.   '   Repair costs may also be incurred by vehicle
 owners.  However,  in the absence of  information on the size of
                          -55-

-------
'
"
        these costs or on  the possible  fuel  savings  which may be
        generated as a side benefit,  repair  costs  are  assumed to be
        perfectly offset by fuel  savings.
       i
       >_     Initial , first order estimates  of  the effectiveness of an
       '                                            147
        I&M program for NO have  been made by EPA.   '   A very optimistic
                           ^
        assessment is that all  failures and  maladjustments could be
 •?
 \> = :,"   detected and corrected  for all  autos manufactured after 1981.
   ^T    EPA's mobile source emission  model was  run with the correspon-
 • -"\* \
  ^     ding adjustments to the deterioration factors,  and the results
\
^ .     inputted to the area source model used  here.   A 28 percent
        reduction in mobile source (highway  vehicle) emissions is
        produced in 1990.

             6.   Area Source Costing Procedure and  Results

             Based on the  relative share of  total  emissions and the
        unit costs of control,  mobile sources appear to be much more
        attractive candidates to  focus  on in strategies for achieving
        alternative short-term  NO- standards than do other area sources .
         Stationary  area  sources emit a small fraction of total area
       ..  source  NO   (see  Table  5)  and are expensive to retrofit.
                 Jv
         Replacing  all  existing residential,  commercial, institutional,
         and small  industrial boilers and furnaces with low NO  units
                                                              J^
         would  seem to  be the only way to extract a sizable reduction
         from these  sources.  A simple,  first order assessment of
         the costs  and  effectiveness of replacing all residential,
         commercial/institutional and small industrial boilers in
         each of the 150  AQCR's in this study indicated that the
 «-,_ _.    capital cost may be  as high as $50 billion while only
 -'-  >    three  additional AQCR's would be able to achieve the 250
    S. .-
    X,
      ug/m  -^standards.
       This analysis assumed that a replacement residential
       furnace or hot water boiler would cost $700, replacement
       units for commercial/institutional firetube boilers
       would cost $120,000 for oil- or gas-firing and $200,000
       for coal (spreader stoker), and replacement units for
       industrial boilers would cost $150,000 (oil- or gas-
       fired fire tube)  and $300,000 (stoker watertube).  Details
       of this analysis were provided in a memorandum to EPA.  '

-------
Consequently, the focus here is on mobile sources.  Projec-
tions of the effectiveness of mobile.source controls are
generated by changing the composite mobile source emission
factor input to the rollback model, as noted before.  Control
cost estimates are made separately.

The total additional cost of introducing a 0.4 g/mi.
emission standard in 1985 for those cars purchased between
1985 and 1990 can be estimated from the unit cost figures.
Assuming for purposes of this report that the California
and Federal regulations are identical  (and thus, a 0.4
g/mi. standard would impose additional costs on Californians
as well), the total number of LDV's purchased between 1985 and
1990 can be projected knowing (a) the total VMT for each year
during this period of time; (b)  the fraction of this total
driven by newly purchased LDV's during this period; and  (c)
the average miles driven per year by new LDV's.

     Estimates of annual nationwide VMT in 1985 to 1990 come
from the assumed annual growth rates (1.0 and 3.0 percent)
applied to the 1975 base year total:  1.17x10   miles for
light duty vehicles.  '  Using EPA's estimates of .106 for the
fraction of total annual VMT accounted for by new LDV's, '  and
EPA's .estimate of 15,900 miles for the average VMT per new
LDV,   total new LDV's were estimated for each year between 1985
and 1990:                                         ---.•• . • '   -.,-./  ..'J---'
 jl) VMT19g5  >199o = ^^1975 x  Ccompounded annual growth rate)
     LDV's Purchased1985_1990 = VMT19B5_ >199Q x

                (New LDV VMT as a fraction  of total VMT.)
                        (Average VMT per new LDV)
                           -57-

-------
     This computation produces an estimated total capital
outlay for mobile source NO  controls of from four to over
                           H
five billion dollars (5.3 - 6.7 million cars § $80), dependent
on the assumed mobile source growth rate.  Annual costs in
1990 may be from $0.8 to over 1.0 billion, and include the
annualized capital charged  (.16 capital recovery factor) and
$3 per year maintenance charge for each vehicle.  Again, these
numbers should be viewed as rough, first approximations only.
Most significantly, the uncertainty attached to the unit cost
estimates is high.

     The estimate of aggregate control costs for inspection
and maintenance programs is straightforward.  The number of
autos currently registered  (projected to 1990) can be used to
set an upper bound to these costs.  The 1990 projection is in
the neighborhood of 150 million cars nationwide.*  At about $5
per auto, this is $750 million.  Thus, the annual costs for
vehicle owners in those AQCR's not attaining the standard will
likely be less than $750 million-,. The actual cost is dependent on
the fraction of all cars registered in these AQCRs.  Capital
expenditures are related to the number of testing facilities
needed and thus to the number of AQCRs implementing I&M programs.

     Projection of the effectiveness of mobile source control
programs and the total costs of control appear in Table 26.
It is readily apparent that going to a .4 g/mi. emission
standard for new light duty vehicles is extremely cost ineffective
 There were 106 million registered autos in 1975.  '
                          -58-

-------
                              TABLE 26
    THE IMPACT OF ADDITIONAL MOBILE SOURCE NO  CONTROLS ON THE
                                             X
      NUMBER OF NONATTAINMENT AQCR'S FOR ALTERNATIVE ONE-HOUR

                     AND THE COST FOR APPLYIN

                     (High Growth Assumption)
N02 STANDARDS AND THE COST FOR APPLYING THE CONTROLS
                        250 yg/m  Standard
   Strategy

Baseline
New Exhaust
  Standard (.4g/mi.)
New Exhaust
  Standard plus I&M
                               I's Not
                       Attaining in 1990

                                73


                                68


                                49
                        500 yg/m  Standard
                                                   Control Cost
                                                   (billions of $)
                                                Capital
                                                 4.2-5.4
                                                 4.2-5.4
                                                        a/
                                                      Annual
                                                       0.' 8-1.1
   Strategy

Baseline
New Exhaust
  Standard (.4g/mi.)
New Exhaust
  Standard plus I&M

I&M Alone
                           AQCR's Not
                       Attaining in 1990
                                0

                                0
                                                   Control Cost
                                                   (billions of $)
                                         Capital
                                                 4.2-5.4

                                                 4.2-5.4
                                                  — a/
                                                        a/
a/Does not include capital costs for I&M.

b/Based on estimates of LDV's in those AQCR's which have to
  implement I&M Programs.

SOURCE;   EEA, Inc.
                                                             Annual
                                                              0.3-1.1
                                                       <.5
                                                          b/
                             -59-

-------
 Whereas the present mobile source control schedule will
effect major reductions in fleet-averaged emissions  (emission
factors for early 1970 model years were greater than 5.0
g/VMT) , the change from a 1.0 to a 0.4 g/VMT standard brings
but a small additional improvement in AQCR attainment, but
levies a heavy cost burden on all vehicle owners.  An I&M
program for NO  is seen to be much more cost effective, though
              a
the cost estimates are only rough approximations.

     Even with both control programs in effect, perhaps as
many as 50 AQCR's will be unable to achieve a 250 yg/m
standard in 1990.  Again, this assumes a high growth rate for
area sources, that area sources are primarily responsible for
the violation, and that hourly peaks are at least six times
the recorded annual average NO- level in all AQCR's.  In these
AQCR's, point source controls beyond those required as per the
point source analysis would likely be necessary. For a 500
ug/m  standard, an I&M program in fewer than ten AQCR's would
be needed.

E.   Comparison of the Nationwide Cost Analysis With the Chicago
     Case Study
          Multiple Point Source Interaction
     v
     Point sources were  nedelled individually as isolated
plants in the nationwide analysis.  Further, the estimated
maximum concentrations due to each source in a plant were
considered additive irrespective of the location of maximum
impact.  These locations could vary due to  differences
among sources in the same plant in terms of operating charac-
teristics, stack location, and stack characteristics.  The
first simplification results in an underestimation of the
short-term NO. problem where multiple plants are located close
                         -60-

-------
enough for plume interaction to be significant.  The second
simplification should result in an overestimation in all
cases.  The two opposite errors tend to compensate each other
to some extent where multiple plants are involved.  But the
degree of over- or underestimation is a multi-variate function
of the relative location of different plants and sources
within a plant.  A uniform adjustment across all plants to
compensate for these simplifications is not justified.

     As noted previously, an interactive point and area source
model was used for the Chicago case study.  A comparison of
the modelling results from the two studies  (Chicago case study
versus Chicago portion of the National study) provides a basis
for ascertaining the degree of error introduced by the above
simplifications.   The  two  studies  differed  in many  respects
as described  in Table  27.  The major difference was  the
treatment of  ambient contributions from  sources within each
plant and the interactions among plants.

     Table 28 shows a comparison of the point source control
costs to achieve the 250 yg/m  standard for the 1982 area
source low growth case.

     The results are very similar,  even though there are
significant differences in the analytical approaches and the
data bases used.  The over-and under estimations introduced in
the  nationwide analysis  by (a) ignoring  point  source interac-
tions., and "(b) summing the maximum concentrations of all sources
within a plant appear  to have balanced out.  It may,  there-
fore,  be fair to  say that  the point source  costs  estimated on
a nationwide  basis are not seriously biased by the  simplifi-
cation employed in the analysis.

F.   Economic  Impact

     1.  General Comments
                         -61-

-------
                                                      TABLE 27

                                 MAJOR DIFFERENCES BETWEEN THE NATIONWIDE ANALYSIS
                                            AND THE CHICAGO CASE STUDY
            Treatment of Point Sources
K>
I
            Treatment of Area Sources
            Data Sources
         4.  Dispersion Model
Nationwide Analysis

Each plant was considered
individually.

The maximum ambient conferir
bution from all sources
within the plant were
summed, irrespective of where
the points of maximum impact
were located.

Ambient contributions were
estimated from the regionwide
highest observed annual average
NO , and incorporated as a
background value.

NEDS Point Source File was used.
PTMAX, a single source model,
was used with no mixing
height limit.
Chicago Case Study

Interaction of sources
within and among plants
was considered by estima-
ting the ambient contribu-
tion of each at each recep-
tor in a specified network
Emissions from area source
grid cells and their impact
on ambient levels were
modelled explicitly
Combination of NEDS and Illinois
EPA files were used, with  the
latter updated and allocated on
an hourly basis by Radian  Corp.

RAM, a multiple area and point
source model, was used with a
mixing height limit.
         SOURCE;  EEA, Inc.

-------
                        TABLE 28

  COMPARISON OF THE CHICAGO AQCR RESULTS OBTAINED IN THE

         NATIONAL AND CHICAGO CASE STUDY ANALYSIS

                 (250 pg/m3 N02 STANDARD)
         National Analysis
                          a/
                                 Chicago Case

                                Study Analysis
              b/
         Capital
           Cost
                Annual
                 Cost
          (106 $) (106 $)
           123
                 34
Capital  Annual
 Cost     Cost
(10° $)

 131
(10°  $)

  21
b/
Case 4 with a 20 percent reduction in area source emissions.

Intermediate meteorological conditions with a 20 percent
reduction in area source emissions.
SOURCE:  EEA, INC.
                              -63-

-------
     A detailed assessment of economic impacts is beyond the
scope of this analysis.  The nationwide scope of the study and
the diverse nature of affected sources precludes an investiga-
tion of impacts on product prices and financial conditions
within specific industries.  Instead, a qualitative discussion
of likely impacts is drawn from the cost analysis.

     2.   Point Sources

     It would appear from the preceding discussion that the
expense of additional NO  emission control needed to meet a
                        A
short-term NO- standard will be borne by a variety of sources.
Table 29 shows a distribution of control costs by industry for
the 250 ug/m  standard.  At this level, plants in 14 different
categories have to institute NO  controls.  However, over one-
                               3C
third of all plants are utilities.  More importantly, required
controls at power plants account for-almost 75 percent of
total control costs.  This results from the large size of
utility boilers and the fact that several must be controlled
to advanced levels  (i.e., retrofit of low NO  burners or the
                                   t         *•>
use of dry SCR).  This unequal distribution of costs combined
with the low expenditure total for all point sources indicates
that utilities are the only category of point sources likely
to experience significant economic impacts.

     A closer lool: at the total cost burden accruing to utilities
indicates that the estimated cost levels are indeed modest.
The slightly more than $1 billion capital cost will be spread
over approximately 150 plants, making the per plant cost less
than $10 million on the average.  To put this cost in perspec-
tive, the estimated capital expenditure for NO  controls on
                          -64-

-------
                           TABLE  2-9
          COST OF N0x CONTROL BY  INDUSTRY  FOR A 250
                     yg/m3 N02  STANDARD
                                              Capital   Annual
No. of Cost
Plants (106 $)
Utility
Boilers
Internal Combustion - Reciprocating
Industrial - In Process Fuel Use
Gas & Oil Pipe Lines (I.e. Engines)
Refineries
Steel
Metal Melting
Asphalt
Lime Kiln
Glass
Cement
Food
Automotive
Wastewater Treatment
Miscellaneous
Non-Combustion
Chemical (nitric acid, etc.)
Total

136 1065
26 14
18 12
37 91
23 142
15 13
10
4
4
3
4
4
3
67 114
	 24 	 45_
408 1,496
Cost
(106 $)

334
3
6
-13
16
1
• -
-
-
-
-
-
-
10
	 6_
363
SOURCE:  EEA, Inc.
                              -65-

-------
            existing  utility  sources  is  less  than five percent of the
            estimated capital expenditures  between 1976 and 1990 to meet
            the  NSPS  for  SO-  now under consideration,  and about 0.1 percent
            of the total  estimated capital  expenditures by utilities for
            the  same  period.

       !          We can conclude, with some confidence, that the economic
       I     impacts are unlikely to be large for point sources even for
            the  most  stringent standard  assessed—250  yg/m .
          1
                 3.   Area Sources

                 The  cost analysis for area source controls points to a
            much more significant level  of  expenditure to meet a 250 yg/m
            standard   ($4-5 billion capital—1985 to 1990, and $1 and $2
            billion annual for a 0.4 g/mi.  exhaust level and I&M program).
            In addition,  the  impacts are likely to be  experienced only
            within one industry  (auto and truck manufacturing! and directly
            by consumers.

                 It seems likely that the cost of additional control
            equipment needed  to meet a 0.4  g/mi. exhaust standard (about
            $80)  would be passed along in total to auto and truck purchasers.
            (Car prices would rise less  than five percent.)   This cost
 r~_> V,      plus the  estimated $30 lifetime maintenance cost and $5 annual
.£ ^      '  inspection fee may discourage some potential buyers, thus
x-N-  ^
\ V        depressing sales  industry-wide.   More likely, the cost of
  -!         producing the necessary control equipment  or engine modifications
            may  vary  among manufacturers, especially between domestic and
            foreign companies.   To the extent that these lifetime cost
            differentials are substantial,  certain companies may be disadvan-
            taged.  Beyond this cursory  assessment,  little more can be
            said at this  time.
                                    -66-

-------
G.   Summary and Conclusions

     •    Either point or area sources alone can cause
          violations of a 250 yg/m  short-term NO, standard.
                                                         3
          Short-term concentrations greater than 500 yg/m
          are normally a result of their combined impacts.  Of
          the four standards tested, only a 250 yg/m  one-
          hour level would be difficult to attain.  Perhaps
          as many as 50 AQCR's would be in nonattainment in
          1990 after area and point source controls were
          in place.

     •    Emissions from point and area sources together
          (under conditions that result in peak N02 concen-
          trations overall) may lead to violations of a
          250 yg/m  standard in about 120 AQCR's.  Under these
          conditions, control of point sources alone can
          result in attainment of the standard at a capital
          and annualized cost of about $1.6 billion and
          $340 million, respectively.

     •    Area sources alone under meteorological condi-
          tions that maximize their impacts may lead to
          current violations of the 250 yg/m  standard in
          about 95 AQCR's.  Mobile source controls beyond
          the mandated exhaust standards can bring 10 to 20
          AQCR's into attainment of a 250 yg/m  standard by
          1990, beyond the 20 to 30 which may reach attain-
          ment through the turnover of the vehicle stock.
          Inspection and maintenance programs  (less than
          $750 million annual cost) appear to be much more
          cost-effective than a .4 g/mi. exhaust standard
          (about $4-5 billion capital and about $1 billion
          annual).

-------
Point sources with multiple short
stacks seem to behave similarly to area
sources in terms of the meteorological
conditions which lead to maximum impact.
Their combined effects seem to be the
cause of high short-term N02 concentra-
tions in urban areas.

Point sources with high effective dis-
charge heights do not seem to present a sub-
stantial problem taken alone.  But in the
presence of other point sources and/or area
sources, their contribution is significant
enough to require some degree of NO  con-
                            3
trol.  However, for 250 yg/m  N02 stan-
dard, the cost of controlling these poinf
sources constitutes a large portion of the
total point source control costs due pri-
marily to the large size of their combus-
tors (boilers).
For short-term NO- levels of 500 ug/m  and
above, the magnitude of the problem is
diminished significantly.  Controls on
point sources alone should bring attain-
ment of the standard at all times in most
AQCR's.  The capital costs of point source.
control to attain a 500 yg/m  level are
about $46 million, while annual costs are
close to zero due to fuel savings.  Between
five and ten AQCR's may still be in viola-
tion of this standard in 1990 due to area
source emissions.  These could be brought
into attainment if an aggresive I&M program
were instituted in each AQCR at an annual
cost of a few hundred million dollars.
                -68-

-------
                        REFERENCES
1.  Memorandum from Joseph A. Tikvart (US EPA, Office of Air
    Quality Planning and Standards)  to Edward J. Lillis (US
    EPA, Office of Air Quality Planning and Air Standards),
    February 16, 1978.

2.  John Trijonis, Empirical Relationships Between Atmospheric
    Nitrogen Dioxide and its Precursors, EPA-600/3-78-018,  EPA,
    Office of Research and Development, Research Triangle Park,
    North Carolina, February 1978.

3.  Acurex Control Techniques for Nitrogen Oxide Emissions From
    Stationary Sources—Revised Draft Second Edition, Aerotherm
    Report, TR-77-87, Aerotherm Division, December 1977.

4.  California Air Resources Board,  Control of Oxides of Nitrogen
    From Stationary Sources in the South Coast Air Basin, ARE
    2-1471, September 1974.

5.  US EPA, Compilation of Air Pollutant Emission Factors,  AP-42,
    3rd Edition, 1977.

6.  Radian Corporation, "Impact of Point Source Control Strategies
    on NO- Levels," Discussion Draft prepared for EPA, February
    1978.

7.  Thuillier, R.W., W. Viezee, "Air Quality Analysis in Support
    of a Short-Term Nitrogen Dioxide Standard," Discussion Draft
    prepared for EPA, SRI International, December 1977.

8.  Bureau of Economic Analysis, U.S. Department of Commerce,
    "Tracking the BEA State Economic Projections," April 1976.

9.  Personal communication with Paul Stolpman, US EPA, Office
    of Policy Analysis.

10.  The estimates of future emission factors were computed
    with EEA's mobile source emission model (Mobile 1).

11.  Personal communications with Mike Evans of ACUREX (Aerotherm
    Division).
                           -69-

-------
                       REFERENCES
12. U.S. Department of Transportation, U.S. Environmental
    Protection Agency, U.S. Federal Energy Administration
    (now the Department of Energy), An Analysis of Alterna-
    tive Motor Vehicle Emissions Standards, May 10, 1977, as
    revised on April 13, 1977, Tables A-7 and A-8.

13. B.F. Kincannon and A.H. Castoline, Information Document
    Automobile Emissions Inspection and Maintenance Programs,
    EPA, Washington, D.C., February 1978.

14. Personal communications from Paul Stolpman, US EPA,
    Office of Policy Analysis.

15. Bhatt, K., M. Beasely, K. Neels, Analysis of Road Expendi
    tures and Payments by Vehicle Class, 1956-1975, The Urban
    Institute, Washington, D.C., March i§77.  The figure for
    total VMT by LDV's in 1975 (1.17x10   miles) was approxi-
    mated from reported VMT for autos and trucks up to 12,000
    pounds registered weight, the latter adjusted by the
    percent of registered weight class.  (This corresponds
    approximately to the 0-6,000 pound chassis weight'class.)

16. Motor Vehicle Manufacturers' Association, MVMA Motor
    Vehicle Guide, Facts and Figures, 1977.

17. Personal communication from Richard Jenkins, EPA, Office of
    Air Quality Planning and Standards, Data on S02 scrubber
    costs come from Paul Lashotopf, Temple, Barker and Sloan.

18. Based on EEA's experience in costing boiler systems.

19. Memo from Dale Keyes (EEA) to Kenneth Lloyd (US EPA,
    Office of Air Quality Planning & Standards),
                          -70-

-------
          SHORT-TERM N02 STANDARDS
                VOLUME III


     AN INVESTIGATION OF SHORT-TERM

         CONCENTRATIONS IN CHICAGO



               Draft Report
                   DRAFT
              DO NOT QUOTE OR CITE
               Submitted to:

Office of Air Quality Planning and Standards
        Environmental Protection Agency
Research Triangle Park, North Carolina   27711
               Submitted by:

       Bharat Kumar and Dale L.  Keyes
    Energy and Environmental Analysis,  Inc.
      1111 North 19th Street, 6th Floor
           Arlington, Virginia  22209
             December 27, 1978

-------
                      TABLE OF CONTENTS
                  SHORT-TERM N02 STANDARDS
                         VOLUME III

                                                           Page
A.   Introduction	   '   1
B.   Air Quality Assessment	      2
     1.   Selection of a Case Study AQCR	      2
     2.   Ambient NO- Concentrations in Chicago ....      3
     3.   Modelling Approach	      3
     4.   Modelling Results and Discussion	     13
C.   Assumptions and Limitations	     18
D.   Relationship of Estimated Area Source One Hour
     Concentrations to Observed Annual NO- Levels ...     20
E.   Control Options and Cost Analysis	     23
G.   Summary and Conclusions	     26

REFERENCES	     31
APPENDIX A	•	     32
REFERENCES TO APPENDIX A	     39

-------
                       LIST OF TABLES

                  SHORT TERM N02 STANDARDS

                         VOLUME III
Table 1;


Table 2:


Table 3



Table 4;
Table 5:
Table 6:
Table 7;
Yearly Highest and Second-Highest            4
One-Hour NO- Levels  (ug/m )

Meteorological Conditions Used in           11
The Analysis

Comparison of Estimated NO- Levels From     16
Point and Area Sources Under Different
Meteorological Conditions in Chicago

Comparison of Estimated Average One-Hour    22
NO- Concentration Due to Area Sources Alone
(For all Modelled REceptors) with the
Highest Observed Annual Average in 1975
in the Chicago AQCR

Point Source Control Costs to Meet a   .     25
One-Hour NO- Standard of 250 ug/m  in
Chicago AQCR

Comparison of the Chicago AQCR Results      27
Obtained in the National and Chicago Case
Study Analysis

Major Differences Between the Nationwide    28
Analysis and the Chicago Case Study

-------
                       LIST OF FIGURES

                  SHORT-TERM N02 STANDARDS

                         VOLUME III
Figure 1:      Seasonal-Diurnal Variation in NO       •      5
               and NO- for Plymouth Court Monitoring
               Station

Figure 2:      Seasonal-Diurnal Variation in NO             6
               and NO- for West Polk Monitoring Station

Figure 3:      Seasonal/Diurnal Variation in NO  and        7
               N02 at Joliet Station

Figure 4:      Estimated Highest One Hour Average N02      17
               Concentrations in Chicago - 1975

Figure A-l:    Diagrammatic Representation of Programming  37
               Problem (Two-Receptor Case)

-------
                       VOLUME III
                   AN INVESTIGATION OF
        SHORT-TERM N02 CONCENTRATIONS IN CHICAGO
A.  Introduction

     This is the third volume of a three-volume report which
attempts to identify the causes of high short-term concentra-
tions of nitrogen dioxide  (N02) and the additional cost of
controlling sources of nitrogen oxides (NO )  emissions to
                                          X
levels consistent with attainment of the ambient standards
under consideration. (The costs in addition to those required
to meet the Federal Motor Vehicle Emission Standards and the
NSPS for stationary sources.)  Volume I focuses on the mechan-
isms by which high ambient concentrations of NO- are believed
to occur.  Volume II provides a preliminary assessment of the
sources which, through the mechanisms identified in Volume I,
might cause or contribute to high short-term (i.e., one-hour
averaging period) concentrations of N0_.  Control strategies
are developed for these and used to estimate nationwide control
costs.  This volume describes a detailed case study of short-term
NO- concentrations in Chicago, the results of which are used
in Volume II as input to and for comparison with the point source
analysis.

     The levels of N02 observed on an hourly basis in any
large region are the product of emissions from a myriad of
sources.  In order to fully capture the interactive nature
of these emissions and the relative importance of each
                                -1-

-------
                           -2-
source, a multiple point and area source dispersion model is
required.  A modelling exercise of this type for every AQCR
in the country is obviously prohibitive from a cost standpoint.
Instead, a separate detailed study of a single AQCR was
undertaken to serve as a point of comparison for the simplified
nationwide study.  The results of this investigation shed
light on the contributions of point and area sources to the
short-term NO, problem in an urban area under different
meteorological conditions.  In addition, the relative contri-
bution of area sources to peak hourly concentrations observed
in Chicago provides a rationale for incorporating an area
source background in the nationwide point source analysis.

B.  Air Quality Assessment

     1.  Selection of a Case Study AQCR

     Chicago is a logical choice for this analysis:

          •    Chicago is one of the five AQCR's in
               the nation recording violations of the
               annual average NO- standard, suggesting
               a high potential for short-term viola-
               tions as well.

          •    The point and area source NO  emissions
                                           H
               inventory for Chicago is relatively com-
               plete as a result of several recent
               studies conducted there. '

          •    Chicago, a classic urban-industrial
               area, presents a good opportunity to
               study interactive effects of both
               point and area sources.  The N0_ pro-
               blem in Chicago can be assumed repre-
               sentative of most other urban AQCR's in
               the country.

-------
                           -3-
     2.   Ambient NCU Concentrations in Chicago

     Ambient N02 data in the Chicago AQCR are inadequate, taken
alone, for characterizing and assessing the short-term NO-
problem.  There are only four continuous NO-/NO  monitors in
                                           ^   it
the Chicago region, far below what would be required for an
area of approximately 150 x 80 km.  However, the four sites can
be considered representative of some of the most severely
impacted locations in the region.  The peak one-hour N02
levels normally observed at these sites are among the highest
concentrations observed in other urban areas of the country.
Table 1 shows the highest and second highest one-hour N02
levels measured annually at the four continuous monitors in the
region during the past three years.

     Figures 1, 2, and 3 show the diurnal variation in the N02
and NO  levels for the summer and winter seasons at the Plymouth
      X
Court (Camp), West Polk (Medical Center), and Joliet monitoring
         4/
stations.    These patterns reflect the influence of mobile
source NO  emissions  (morning and evening rush hours) typical
         J^
of urban areas.  The NO- levels show a less pronounced double
peak which is also typical, while the Camp station reflects an
early afternoon bulge characteristic of point source influences.

     3.   Modelling Approach

          a.   Model Description

     A multiple point and area source model, known as RAM, was
used for assessing short-term NO  concentrations. '  RAM is an
                                ji
EPA-approved Gaussian steady-state model capable of predicting
short-term  (averaging times ranging from an hour to a day) am-
bient concentrations of relatively stable pollutants from

-------
                                         TABLE  1

Monitoring
Site Location
Camp
Medical Center
Joliet
La Salle
YEARLY HIGHEST AN,D SECOND
1975 '
Second
Highest Highest
1-hr. N02 1-hr. N02
394 385
678 616
857a 812a

HIGHEST ONE-HOUR N02 LEVELS
1976
Second
Highest Highest
1-hr. N02 1-hr N02
490 , 470
394 354
262 250

                                                                               1977
                                                                         Highest
Second
Highest
                                                                        1-hr. N02   1-hr.  N02
                                                                         1008

                                                                         494

                                                                         604
 968

 488

 524
                                                                                         a
Questionable data.
SOURCE:  Illinois  EPA

-------
                            -5-
                         FIGURE  1
                           WINTER
                                                    I2M
                           SUMMER
                                                  NOV
                                                   13M
         PLYMOUTH COURT  (CAMP STATION)  UTM- (450. 4830)
SOURCE:
   Seasonal-Diurnal Variation in N0x  and N02
   Plymouth  Court Monitoring Station
Reference  1.
                                                         for

-------
                              -6-
                           FIGURE 2
                             WINTER
        13M
                                                         I2M
                             SUMMER
       .21 •


       .it •

       .a*-
      -

 £ 300-.,..
 ui
 u
      .!*•
aoo-
   .10 1





too- •"•

   .04 -


   .03 •


 0-
                                                     NO,
        13M-
                     •*
                                 13N
                                                         ISM
             WEST POLK (MED CENTER)
                                   UTM • (4.45.
             Seasonal-Diurnal Variation  in NOX and N02  for

             West Polk  Monitoring Station
SOURCE:   Reference  1.

-------
                                  -7-
                                 FIGURE 3
    SEASONAL/DIURNAL VARIATION IN NOX AND NOZ AT JOLIET STATION
 CONCENTRATION (PPM)

 0.3 1
 0.2
 0.1
                         HNTER (JANUARY 24,1977)
   12M
6A
12N
 i
6P
12M
CONCENTRATION (PPM)

U
0.2 -
0.1 -
                            SUMMER (JULY 8,1977)
  12M

SOURCE: EEA. lac.
6A
12N
i
6P
12M

-------
                             -8-
multiple point and/or area sources.  Hourly meteorological data
required are wind direction, wind speed, stability class, and
mixing height.  Required data on point sources consist of
source coordinates, hourly emission rate, stack height, stack
gas volume flow, and temperature.  Area sources are specified
in terms of south-west corner coordinates of the area source
grid, grid cell area, total cell emission rate, and the effec-
tive area source height.

     As noted, RAM is designed for modelling nonreactive
pollutants.  Nitrogen dioxide, on the other hand, is primarily
a secondary pollutant formed by oxidation of NO.  The initial
NO concentration in the exhaust gases, the plume diffusion and
travel time, and the ambient concentration of photochemical
oxidants and reactive hydrocarbons are the most important
factors that affect the conversion of NO to N02«  An empirical
representation of NO- formation from point source emissions of
NO , developed by EEA, has been used in conjunction with RAM to
  Ji
translate the predicted NO  into NO, concentrations.  A dis-
                          2C        £
cussion of this approach is presented in Volume I.

     A different approach was used for area sources.  A fixed
NO-/NO  ratio for each period of the day, based on observed
  +*   Xt
data at the continuous monitoring sites, was used to relate
predicted NO  and NO- levels.  This is further discussed in the
            X       «b
Assumptions and Limitations Section.

          b.   Point and Area Source Emission Data

     A partial list of point source emissions, together with
stack parameter data, was obtained from Radian Corporation and
corrected for errors.  Additional sources of NO , along with
                                               Jt
other pertinent data such as source operating rate and certain
classification codes, were obtained from EPA's NEDS point source
subfile and the Illinois EPA.  Hourly NO  emissions obtained from
                                        X

-------
                           -9-
Radian were adjusted to reflect the diurnal and seasonal
variation in the operating loads.  The temporal variations
in the operating loads of power plants and space heating
operations were found to be more significant than those in
the industrial operations.  Emission factors for most utility
boilers and combustion turbines were based on actual stack
test data.  For other point sources, the recent emission
factors from the Oxides of Nitrogen Control Techniques
Document,  (Second Edition)   , California Air Resources Board
ARB2-1471,3' and AP-42  (Second Edition) ' were  used.  The
point source data were screened for erroneous and missing
information.  Default values for the various parameters were
substituted based on what constitutes normal operating practice
in a particular process category.  Data on a total of 813 sta-
tionary processes belonging to about 286 plants were thus
compiled.

     For the final analysis, point sources emitting below ten
pounds NO  per hour were considered too small to be treated indi-
         JL
vidually and were lumped with the area source inventory.  This
was done mainly to reduce the point source input to RAM to a more
manageable size.  The final input to RAM consisted of 457 point
sources  (i.e.,  SCC's).

     Hourly NO  emission estimates for mobile and stationary
              ji
area sources in the Chicago region were provided by Radian Cor-
poration.  Vehicular data from the Chicago Area Transportation
Study (CATS) were used by Radian to convert the 1975 annual
average mobile source emissions (taken from NEDS)  into hourly
emissions in the summer and winter seasons.  The mid-morning
and afternoon hourly emission rates estimated in  this manner
were apportioned to 5 x 5 km grid cells.  Input data to the

-------
                            -10-
area source subroutine of RAM then consisted of 640 cells.  A
finer spatial resultion than a 5 km spacing would have made
the computer time requirements prohibitive.

          c.   Meteorological Data and Selection of Worst Case
               Conditions

     Meteorological conditions play a major role in the short-
term build-up of pollutants from both point and area sources.
High ground level concentrations of NO  (but not necessarily NO.)
                                      X                        fc
from point sources are normally caused by fumigation (inversion
breakup)  or plume downwash where the plume intersects the
ground quickly.  Though these conditions can not be simulated
explicitly by RAM, the following meteorological conditions .
are good  surrogates:  -stability class B or C  (unstable  atmos-
phere) and moderate to high wind speeds.

     On the other hand, the greatest impacts of ground  level
sources,  such  as vehicles and other area sources, occur when the
atmosphere is  quite stable  (stability class of D or E), wind
speeds are low, and mixing heights are small.  The meteorological
conditions that maximize the impact of either point sources  or
ground level area sources are thus at two opposite extremes,
making it difficult to maximize their impact simultaneously.
Instead,  several intermediate sets of conditions were used to
simulate worst case conditions for point and area sources
together.

     The meteorological data for the Chicago region were
extrapolated from the mixing height and wind speed data collected
by the National Weather Service at their Green Bay, Wisconsin,
Peoria, Illinois, Flint, Michigan, and Dayton, Ohio stations.
Typical meteorological conditions, as well as those that
correspond to  stagnant and unstable atmospheres in the Chicago
region, are listed in Table 2.

-------
                                             TABLE  2
                           METEOROLOGICAL CONDITIONS USED IN THE ANALYSIS
 Conditions
                                                       Season and Hour of the Day
Summer Mid-morning
Summer Afternoon
         Winter Mid-morning
   Typical
     Wind Speed
     Mixing Height
     Stability Class
     Wind Direction
     Ambient Temperature
  4.0 meters/sec.
   650 meters
        C
       90°
       70°F
 6.5 meters/sec.
  1600 meters
        D
      225°
       80°F
           6.0 meters/sec.
            450 meters
                 D

               315n
                20°F
    Worst Case -  Point Source
      Impact

    •  Wind Speed
    •  Mixing Height
    •  Stability Class
    •  Ambient Temperature

    Worst Case -  Area Source
      Impact

    •  Wind Speed
    •  Mixing Height
    t  Stability Class
    •  Ambient Temperature

    Predominant Wind
      Directions
  2.5 meters/sec.
   300 meters
        C
       70°F
  0.5 meters/sec.
   300 meters
        C
       70°F
       90
         o
 4.5 meters/sec.
  800 meters
       B
      80°F
 2.5 meters/sec.
  800 meters

      80°F
180
225*
            2.5 'meters/sec.
            30o meters
                 C
                20°F
           1.5 meters/sec.
            200 meters

                20°F
275
215
                                                            o
SOURCE:   EEA,  Inc.

-------
                           -12-
     d.   Selection of Receptor Locations

     A total of 374 receptors were initially selected for
the entire AQCR and were located so that the maximum impact
of all significant point and area sources could be captured.
The first 70 receptors correspond to locations of the con-
tinuous and 24-hour N0_ monitoring sites in the Chicago
AQCR.  The remaining receptors were selected by the model
downwind of significant point and area sources at points of
expected maximum concentrations.  The receptor selection
process was as follows:

     The most significant point and the 100 most significant
area sources (i.e., grid cells) were first identified, based
on their relative contribution to ambient levels.  Two
receptors downwind of each major point source and one for
each major area source were selected.  The first receptor
for each point source was positioned at the estimated -point
of maximum ambient concentration; the second receptor, in
the same direction but twice as far away as the first, was
designed to capture the interaction of overlapping plumes.
The single receptor for each area source reflected the
maximum impact from each area source.

     The 100 receptors corresponding to major area sources
were eliminated after the preliminary modelling runs, indicated
low total NO- concentrations at these receptors under meteorolog-
ical conditions which maximize point source influences.  For
the analysis of maximum area source influences, 27 of the
area source receptors were included.  The model was run for
five predominant wind directions in the Chicago region to
capture the interaction of various sources for the purpose
of developing control strategies.

-------
                            -13-
     4.   Modelling Results and Discussion

     Several preliminary runs of the model were made to
determine the combinations of hourly emissions and meteoro-
logical conditions that would result in the highest short-
term NO- concentrations.*  Maximization of area source
impact resulted in high concentrations at a large number
of receptors.  The point source maximization, contrary to
expectations, resulted in peak N02 concentrations lower than
those estimated in the former case.**  The area source NO-
component under these conditions is significantly lower and
may have been underestimated due to limitations of the
model.  An intermediate set of adverse meteorological condi-
tions was therefore selected to capture large contributions
from both point and area sources which, when combined, could
result in the highest short-term NO- levels.  The meteoro-
logical conditions used for the final analysis are:  wind
speed—1.5 meters/second, stability class C, and mixing
height—300 meters.  The NO  emissions for the analysis
                           a
correspond to the 1975 summer mid-morning levels.

     The results obtained point to the presence of two distinct
types of point sources:  1) large plants with tall stacks
such as power plants and incinerators, and 2) plants with a
large number of smaller sources with short stacks such as the
steel mills and refineries.  The diffusion characteristics of
emissions from the second category were found to be similar
to those of the area source, emission-s.  During meteorological
conditions which cause high concentrations of NO- from both
area sources and point sources with short stacks, the contribution
 * Preliminary runs wer-e made using a portion of the Chicago AQCR
   which included Cook, Dupage, and portions of Will, Lake, arid
   Porter Counties.
** See Table 2 for the meteorological conditions used to maximize
   area and point source influences separately.

-------
                           -14-.
from power plants and other sources with high effective stack
heights was low.  Under other extreme conditions, point
sources with tall stacks could have a greater individual
impact, but these conditions tend to minimize the impact
of other source categories resulting in overall lower concen-
trations.  Table 3 compares the relative point and area source
contributions to ambient NO- levels under three different
meteorological conditions.  High hourly N02 levels can be pro-
duced under conditions favoring either area or point source
influences.  The highest levels are likely to occur when
meteorological conditions allow both types of sources to
exert significant effects.

     Table 3 also shows the percentage of power plants with
significant contributions.  At low wind, speeds, the predicted
contribution due to most of these sources is negligible.  This
is, in part, due to the limitation of the model.  At low wind
speeds, the point source plumes penetrate through the mixing
layer at depths assumed in this study, and the diffusion model
assumes their contribution as zero.  Under certain circumstances,
however, the plume could be trapped below the inversion lid and,
during the morning inversion-breakup, could fumigate to the
ground and cause high NO  concentrations.  Whether this phenom-
                        J&
enon could also lead to very high NO- concentrations is a subject
for further research.  Ozone concentrations in the morning are
generally low and the degree of plume mixing during fumigation
may_be limited.  Both of these factors may lead to low NO,/NO
                                                         ^   X
ratios.

-------
                                    TABLE 3
COMPARISON OF ESTIMATED NO  LEVELS FROM POINT AND AREA SOURCES UNDER DIFFERENT


                METEOROLOGICAL CONDITIONS IN CHICAGO3/
Five Highest Worst Case - Point
Concentrations Total Point
1 509 428
2 589 409
3 348 209
4 348 225
5 342 219
Average of
all receptors
above 200
pg/ra 277 165
Number of Receptors
Above 200 |Jg/m3 47
Percentage of Power Plants
With Significant Contribu-
tions, i.e., Effective Stack
Height Less Than Mixing
Height 17
Chicago Case Study - Cook, Dupage
The receptors used in the analys
receptors may reflect higher area
C/ See Table 2.
' W-inH crloorl — 1 ^m/a o*--,Ki 1 i 4-, . „! —,
C/ r/
Source Worst Case - Area Sourcej^
Area Total
81 568
81 479
139 472
123 472
123 472
111 371

Point Area"'
549 19
279 200
272 200
272 200
272 200
142 199
68
0
, and portions of Will, Lake and Porter
is were selected to record maximum total
source contributions, but lower total c


Intermediate Case^/
Total Point Area
603 493 110
602 434 168
600 407 193
598 430 168
553 383 170
316 142 174
67
3
Counties .
concentration. Other
oncentrations .
i
                                                                                                    Ul
                                                                                                     I

-------
                           -16-
     These findings for large point sources are at variance
with the Radian conclusions. '  It would appear that the
differences are due to  (a) the use of a much higher NO  emission
                                                      H
factor for cyclone boilers by Radian, (b) erroneous emissions
for some point sources in the emission inventory used by Radian
and Radian's use of constant NO-/NO  values as compared with
                               M   X
our time-dependent ratios.  In' addition, EPA reports that the
Radian treatment of summer morning atmospheric dispersions was
somewhat unrealistic leading to high .NO   values. '

     As noted previously, ambient N02 levels in Chicago indi-
cate little variation between the morning and afternoon peaks
at the four continuous monitors.  This is a departure from the
typical situation.  Pronounced morning and afternoon NO- peaks
with low concentrations at noon due to increased vertical mixing
are normally observed in most urban areas where NO- is predomi-
nately an area  (mobile) source problem.  This may be due to the
high noon-time power plant load in Chicago, which is 50 to 80
percent higher  than the mid-morning  load in both the summer
and winter seasons.  Further, most industrial batch opera-
                 s
tions are at their peaks  during the  noon period.  Thus high
NO- concentrations at noon in Chicago can be attributed to
the multiple point sources.

    As noted, the intermediate set of meteorological conditions
produced the highest hourly NO, concentrations.  Figure 4 shows
the five highest  (of the  274) N02 estimates and the relative
contributions of the two source types to them.

-------
                                                FIGURE 4

                                ESTIMATED HIGHEST ONE HOUR AVERAGE NO2
                                    CONCENTRATIONS IN CHICAGO - 1975
                                 (Intermediate Case-Meterorlogical Conditions)
  AMBIENT
N02 LEVELS
  (ug/ro>)

      700
      600
      500
      400
      300
      200
      100
                  AREA
                   75
274
      202

RECEPTOR NUMBERS
254
228
        SOURCE:   EEA, Inc.

-------
                         -18-
C.   Assumptions and Limitations

     An interactive diffusion model, a detailed point and area
source emissions inventory, and a relatively sophisticated
approach for translating the estimated NO  to NO- concentrations
                                         Ji      4b
have been used in this analysis.  Nevertheless, the prediction
models have several limitations, and simplifying assumptions  had-
to be made.  The assumptions and the limitations they place on
the results are described below.

     •    The meteorological conditions are assumed to
          hold for a period of one hour and to be uni-
          form over the whole region.  The latter is
          not likely to be true because of intra-regional
          variations in topography.  In addition, center
          city meteorological conditions are considerably
          altered by the built environment, but a micro-
          scale analysis seemed unwarranted in view of the
          objective of the study and resource limitations.
          The NO- concentrations may have been underesti-
          mated at downtown sites.
          Mobile source emissions are uniformly  dis-
          tributed throughout each grid cell.  This
          results in a loss of spatial resolution which
          may reduce the impact of these emissions on
          air quality.  Ambient levels of mobile source
          pollutants (HC, CO, NO ) tend to peak  next
                                a
          to the roadways and normally diffuse to low

-------
               -19-
values within relatively short distances,
inferring considerable spatial variation in
their concentrations.  Therefore, NG>2 levels,
even though considerably smoothed due to
the slow conversion of NO to N02, may still
be significantly higher in the vicinity of
roadways.  Some version of a highway model
for estimating the impact of mobile sources
would be needed to more accurately assess
this effect.

Ambient levels of NO  due to point source
                    J\
emissions were translated into NO- concen-
trations by the use  a time-dependent, plume-
specific approach incorporated into RAM.
Volume I of this study contains an explana-
tion of the approach.  Area sources were
treated differently.  A constant NO-/NO
                                   £   Jk
ratio of 0.5 was derived from hourly N02
and NO  data for the summer morning period
      a
at the continuous monitoring stations in
Chicago.   (The N02/NOX ratios ranged  from 0.4
to 0.6.)  These monitors are likely to be
influenced by mobile sources at these times.
The factor of 0.5 has been assumed to apply
to all area sources.  NO  emissions from
                        X
area sources within relatively short dis-
tances of their discharge are characterized

-------
                           -20-
           by rapid atmospheric mixing.  It is there-
           fore reasonable to assume that the ratio of
           NO- to NO  concentrations from area sources
             £»      •**
           would be about the same, especially where
           the sources are fairly evenly distributed in
           space.  The narrow range of NO-/NO  ratios
                                         fc   a
           observed at the continuous monitoring sites
           further supports -this deduction.  However, a
           conversion ratio of 0.5 must be considered
           little more than a rough,. initial approxima-
           tion at this time.

           The model is incapable of predicting
           ground level concentrations from point
           sources under certain unusual meteorolog-
           cal conditions such as morning inversion-
           breakup (fumigation), though other conditions
           can be used as surrogates.  Secondly, the
           model assumes no impact on air quality from
           sources with effective plume heights above
           the mixing layer.   For these reasons, NO-
           concentrations from large point sources may
           have been somewhat underestimated.
D.   Relationship of Estimated Area Source One Hour Concentra-
     tions to Observed Annual NO., Levels

     One of the objectives of this case study was to empiri-
cally assess the nature of the area source problem and, if
possible, develop a relationship between the estimated peak

-------
                           -21-
one-hour  levels  and  the  observed  annual  average  N02  concen-
trations.  By  examining  this  relationship  under  meteorolog-
ical  conditions  which  favor contributions  to  ambient levels
from  point sources,  area source additions  to  the modelled
point source levels  could be  derived  for use  in  the  nation-
wide  study.  The results of this  investigation were  incor-
porated in the nationwide study as  reported in Volume II.

      Referring back  to Table  3, one can  see that the average
area  source contributions to  ambient  levels in each  of the
•three cases is very  close to  the  area source  contribution
at  the receptor  with the highest  total concentrations.  That
is,  the  spatial  variation in  the  area source  component of  peak
hourly N02 concentration is low.   However, some  loss of spatial
detail in the  area source concentrations could have  resulted
from  limitations in  the  way RAM models area sources, as noted
earlier.   Regardless,  limited empirical  evidence on  hourly N02
levels observed  supports the  above deduction. The use of  a
single value to  account  for the area  source contribution at  all
receptors in an  AQCR is  therefore reasonable.

      To  quantify the area source  background level  for the  two
sets  of  meteorological conditions used in the national study,
the average area source  contributions to the  peak  one-hour
levels determined conditions  in the Chicago case study
were related  to  the highest observed annual average  N02 con-
centration in  the Chicago AQCR.   Table 4 shows  a comparison
of  the estimated region-wide  average  one-hour area source  con-
tribution and  the highest observed annual average  concentra-
tion  in  the Chicago  AQCR.  For the first case, maximizing  the
point source impact, the average  estimated one-hour  NC^ level
from  area sources alone  is  almost equal  to the highest observed

-------
                         -22-
                         TABLE 4

      COMPARISON OF ESTIMATED AVERAGE ONE-HOUR N02
CONCENTRATION DUE TO AREA SOURCES ALONE  (FOR ALL MODELED
  RECEPTORS) WITH THE HIGHEST OBSERVED ANNUAL AVERAGE
               IN  1975  IN THE  CHICAGO AQCR
                         (yg/m3)

Case
Worst Case
Point
Source
Intermediate
Case
Estimated
One-Hour
Average NO-


111

174
Highest
Observed
Annual NO-


109

109

Ratio


1.0

1.5

-------
                           -23-
annual average value at any monitor; for the intermediate
case, the estimated area source background is higher by
about 50 percent.

     Assuming Chicago to be a reasonable representation of area
source problems in other AQCR's in the country, the highest
observed annual NO- average in an AQCR can be used to represent
the area source contribution to the short-term NO- levels for
meteorological conditions that maximize the point source impact.
For intermediate conditions that result in the overall highest
NO- concentrations, the area source impacts are best reflected by
values 1.5 times the highest observed annual NO- average in an
AQCR.

     For meteorological conditions that maximize the area source
impacts, the relationship between the one-hour and the annual
average can also be derived.  These are the conditions which
are likely to produce high readings at existing monitors since
in most AQCR's  (including Chicago), the monitors are located
to reflect area source influences.*  B.ased on the analysis of the
Chicago data, the peak one-hour ambient value from area sources
alone seem to be four to six times the observed annual average
concentrations.  Therefore, the 6:1 ratio used in the nationwide
area source analysis appears to be conservative (i.e., on the
high side).

E.  Control Options and Cost Analysis

     Control strategy development and cost analysis was con-
ducted only for a proposed one-hour NO- standard of 250 yg/m  .
The objective here was to compare the control costs estimated
in this detailed analysis with those obtained for Chicago
in the nationwide analysis, and to assess the degree of over
or underestimation in the nationwide control costs.
*See Volume  I  for further discussion of  this  issue.

-------
                           -24-
     The point source control costs and effectiveness data
used in this analysis are the same as described in Volume II of
the nationwide study.  No additional area source controls were
considered here.  As in the nationwide analysis, a 20 percent
reduction in the area source emissions was introduced to account
for a one percent'annual increase in the VMT and a 25 percent
decrease in the mobile source emissions by 1982 due to mandatory
emission reduction requirements.
                                             «
     In the nationwide analysis, point source controls to
achieve a proposed short-term N02 standard at the point of
maximum impact of each source were optimized on a plant basis.
That is, all processes within a plant were evaluated individually
in terms of the cost of emission control per unit improvement in
air quality.  Those with the most cost-effective control options
were controlled preferentially so that the desired reduction in
ambient NO- levels at each plant's receptor was achieved in a
least cost manner.

     In the Chicago case study, the least cost optimization
routine simultaneously considered all combustors and processes
in Chicago that had a significant impact  (>1 ug/m ) at any of
the receptors.*  Control cost estimates thus obtained are
theoretically the lowest costs in tKS~shCire AQCR to meet a
specified ambient level at all receptors.  The conventional
control strategies (where individual source categories are
required to control to specified levels regardless of their
contribution to the problem) could result in costs several
orders of magnitude higher.  This is illustrated by the results
in Table 5 which compares the control costs obtained by the
least-cost and the conventional rollback approaches.**

*Appendix A contains details of the least-cost optimization
   model developed for this analysis.
** For a discussion of using economic incentives in N0x
   control strategies, see Reference 6.

-------
                                         TABLE  5

               POINT SOURCE CONTROL COSTS TO MEET A ONE-HOUR NO2 STANDARD OF

                                 250 vig/m3 IN CHICAGO AQCR

                                   (CHICAGO CASE STUDY)
Implementation
   Strategy
     Number of
Sources Contributing
 Sources
Controlled
Capital Cost

  $106/year
Annual Cost

 $106/year
  Least-Cost
  Rollback
        794
        794
     94
    794
     131
    1651
     21
    254
               i
               to
               Ln
               I
 SOURCE:   EEA,  Inc.

-------
                            -26-
     Table "6 shows a comparison between the point source con-
trol costs for the Chicago AQCR obtained in the two studies.
The costs are for the 1982 area source growth case:  a 20 per-
cent reduction in the area source emissions due to mandatory
mobile source emission standards, a one percent growth in the
VMT, and a ratio of mobile to  stationary area  source  emissions
during the peak hours of 3:1.  Point source emissions were
assumed to be the same in 1982 as in 1975.

     The cost estimates in the two analyses are remarkably
similar though there are significant differences in the analytical
approaches and the data bases  used, as summarized in Table 7 .   if
these results can be extrapolated to other AQCR's,  then  the
results of  the nationwide study  are probably not  seriously biased
toward either systematically high or low estimates.

G.   Summary and Conclusions

     •    NO  concentrations from either point or area
            Ji
          sources can result in high short-term NO-
          concentrations.  In an urban area, point
          sources with short-multiple stacks and
          area sources (mobile and dispersed area)
          seem to be the dominant cause of high short-
          term NO2 concentrations.

     •    Two distinct groups  of point sources can
          be identified in terms of their response
          (dilution and N02 formation rate) to dif-
          ferent meteorological conditions:  (1)
          plants with high effective plume heights
          such as utilities,  and (2)  plants with a
          large number of short stacks such as steel
          mills and refineries.  The diffusion charac-
          teristics of the second point source  group
          seem to be similar to those  of the area
          sources.

-------
                           -27-
                        TABLE 6
  COMPARISON OF THE CHICAGO AQCR RESULTS OBTAINED IN THE

         NATIONAL AND CHICAGO CASE STUDY ANALYSIS

                 (250yg/m3) N02 STANDARD)
                                        Chicago Case
          National Anavsisa'           Studv Analysis '

          Capital  Annual              Capital  Annual
           Cost     Cost                Cost     Cost

           (106 $)   (106 $)              (106 $)   (106 $)

           123      34                  131      21
 ' Case 4 with a 20 percent reduction in area source
   emissions.
   Intermediate meteorological conditions with a 20 percent
   reduction in area source emissions.

SOURCE:  EEA, Inc.

-------
                                             TABLE  7

                         MAJOR DIFFERENCES BETWEEN THE NATIONWIDE ANALYSIS
                                    AND  THE CHICAGO CASE STUDY
1.  Treatment of Point Sources
Nationwide Analysis

Each plant was considered
individually.

The maximum ambient contri-
bution from all sources
within the plant were
summed, irrespective of where
the points of maximum impact
were located.
Chicago Case Study

Interaction of sources
within and among plants
was considered by estima-
ting the ambient contribu-
tion of each at each recep-
tor in a specified network
                                                                                                              to
                                                                                                              oo
2.  Treatment of Area Sources
3.  Data Sources
4.  Dispersion Model
Ambient contributions were
estimated from the regionwide
highest observed annual average
NO , and incorporated as a
background value.

NEDS Point Source File was used.
PTMAX, a single source model,
was used with no mixing
height limit.
Emissions from area source
grid cells and their impact
on ambient levels were
modelled explicitly
Combination of NEDS and  Illinois
EPA files were used, with  the
latter updated and allocated on
an hourly basis by Radian  Corp.

RAM, a multiple area and point
source model, was used with a
mixing height limit.
SOURCE;  EEA, Inc.

-------
              -29-
The meteorological conditions that maxi-
mize the impact of sources with high
effective stack heights are at an opposite
extreme to the conditions that result in
high concentrations from both area sources
or point sources with short effective
stack heights.

An intermediate set of meteorological con-
ditions, closer to the area source maximizing
end on the spectrum of diffusion conditions
seems to result in the highest short-
term NO- concentrations in an urban area.

Little  spatial variation is estimated in  the
short-term peak NG>2 concentrations due  to
area sources.

Under meteorological conditions that
maximize the point source impact, the
area source contribution to the short-
term N02 concentrations in Chicago appear
to be equal to the highest observed
annual  average N02 level in the AQCR.

For the intermediate conditions which
result  in the highest short-term N02
concentrations overall, the area source
component can be approximated as one and
a half  times the highest observed annual
average N02 level.
For conditions which maximize the area
source  impact, the estimated peak hourly  con-
centrations are from 4  to 6 times the recorded
annual  average at the same sites.

-------
                 -30-
The least cost NO  control options  for
                 X
the Chicago AQCR estimated in  (a) the
detailed Chicago Case Study and  (b)  the
nationwide study are well within ten
percent.  There is evidence that the nation-
wide study is probably not seriously biased
toward either high or low estimates.

-------
                           -31-
                       REFERENCES
                       VOLUME III
1.   Epright, B.  R.,  et al,  Impact of Point Source Control
    Strategies on N02 Levels,  Draft Report, Radian Corporation,
    February 10, 1978.

2.   Aerotherm Division, Control Techniques for Nitrogen Oxides
    Emissions from Stationary Sources, Second Edition,
    Corporation, December 1977.

3.   California Air Resources Board, Control of Oxides of Nitrogen
    from Stationary Sources in the Soutn coast Air Basin,ARE
    2-1471, California Air Resources Board, September 1974.

4.   Environmental Protection Agency Compilation of Air Pollutant
    Emissions Factors,  (AP-42) Second Edition, August 1977.

5.   Memo The Cause of High Short-Term NOg Levels from Robert E.
    Neligan, Director of MDAD, USEPA, September 12, 1978.

6.   R.  J. Anderson,  Jr., R. 0. Reid, and E. P. Seskin, An Analysis
    of Alternative Policies for Attaining and Maintaining a Short-
    Term NOg Standard, prepared for the President's Council an
    Environmental Quality,  Mathtech, Inc., Princeton, N.J.,
    November 14, 1978.

-------
                          APPENDIX A

             THE EEA LEAST-COST OPTIMIZATION MODEL
     In order to calculate the minimum cost of applying NO  con-
trol technologies necessary to meet a given ambient air quality
standard for NO-, a heuristic integer program was developed.
                                                            I/
The underlying model is a variation of the Knapsack Problem.

     In describing the programming problem, the following nota-
tion will be used:

     i         =  an emissions source.
     N         =  total number of sources.
      s
     t.        =  control technology applied to source i.
     k         =  technology level.
     Klast.     =  "highest level" of control technology applied
                  to source i.
     j         =  receptor.
     N         =  total number of receptors.
     Hr.        =  hourly rate of emissions at source i (10  Btu/
                  hour).
     Hours.     =  total hours of operation per year of source i.
     d..        =  contribution at receptor j due to source i.
     S         =  ambient air quality standard.
     B.        =  contribution at receptor j due to background
                  and area sources.
     See Senju and Toyoda (1968),  Toyoda (1975),  and  Zanakis  (1977)
                             -32-

-------
                              -33-
      C   ,      =  increase  in capital cost rate  associated with
        i
                  moving  from technology k-1  to  technology k  at
                  source  i.
      OM ,      =  increase  in operating and maintenance cost
         i
                  rate associated with moving from  technology
                  k-1 to  technology k at source  i.
     P  ,       =  increase in percent of total emissions con-
       1          trotled associated with moving  from technolo
                  k-1 to technology k at source i.

A.   Control Costs

     The annualized cost (AC) to source i of moving  from tech-
nology k-1 to technology k is calculated as follows:
          AC, .   =   (.16 C. ,HR. )  +   (OM  .Hours. Hr.)       (1)
            tiK           tilc  1         t±K      i  i
Note that the calculation assumes a capital recovery rate of  16
percent per year.  Furthermore, it should be noted that capital
costs are a function of Btu's per hour, while operating and
maintenance costs are a function of Btu's per year.

     Given equation  (1), the total annualized cost  (TAG) for
source i to achieve the "highest level" of control technology
can be expressed as follows:
                           klas^
               TAG.   =      >         AC. .                  (2)
                  1         f  j           U » JV
                            k=0
This is simply the sum- of the increased annualized costs of
applying all control technologies up to and including  the

-------
                              -34-
"highest level" actually employed.  Thus, if klasti = 0, source
i is uncontrolled and the costs of control are assumed to be
zero .  '

B.   Air Quality

     The impact (Impact) of source i on the air quality of the
region can be expressed as a vector of dimension Nr, the number
of receptors.  First, we calculate the impact of source i at
receptor j :
                                        klasti
          impact         di        x  -              k       (3)
This states that the impact is represented by the contribution
at receptor j due to source i multiplied by the percentage of
total emissions that remain uncontrolled.

     Second, we calculate the ambient air quality (Result) at
receptor j as follows:
          Result.   =    B.   +       y      Impact^.        (4)
                                     i = 1
The ambient air quality at receptor j is equal to the background
and area source contributions plus the sum of the impacts of the
individual emissions sources.

     The change in air quality (Delta) at receptor j due to
source i associated with moving from technology k-1 to technolo-
gy k at source i can be expressed as:
 '    Note that klast^ is an integer variable.  In the programming
     problem, klast^ can take on any integer value between zero
     ("no control") and five (the assumed "highest level" of
     control.

-------
            -35-
          Deltaij  -  dijPt.k      3  =  1	Nr            (5)

C.   The Problem

     With the above information, the problem can be stated as
minimizing the following objective function:

                       .  Ns
                                   TAC^                      (6)
                        i = 1

This is the sum over all sources of the total annualized costs
for controlling each source.

     The constraint on the minimization problem can be written
as:

          Result.   -    S  1 0    j  =  l,...,Nr            (7)

This states that the ambient air quality at each receptor must
be less than, or equal to, the ambient standard.

D.   The Solution

     To begin the program, klast. is set equal to zero for all
sources.  This corresponds to no control on all sources.  The
Impact and Result vectors are then given by:

          Impact^.  o  di-         j  =  l,...,Nr            (8)

                                Ns
          Result.   -  B.  +    y>    d±.    j = l,...,Nr    (9)
.

               i = 1

-------
                              -36-
Thus, the result vector is used to represent the present level
of air quality at each receptor.  We then define a limit vector
(Limit) that contains the individual receptor standards.  In our
problem, we are assuming one uniform standard, hence:

          Limit.  =  S             j  =  l,...,Nr            (10)

Next, we define a vector that represents the difference between
the Result vector and the Limit vector at each receptor and
designate this vector as Extra.  This vector represents the
amount by ehich each receptor is in violation of the standard;
it contains no negative entries:
      Extra.
                  Result. - Limit., if Result. > Limit.
                  0,                if Result. < Limit.
(11)
     The program then proceeds to pick and apply control techno-
logies to each source in an effort to bring all receptors into
compliance with.the ambient standard.  This occurs when the Re-
sult vector lies within the feasible region (see Figure A-l
which depicts a two^jreceptor case) .

     Specifically, at each source, the next highest level of
technology is considered.  First, the projection  (Projection) of
the Delta vector on the Extra vector is calculated:

                            Nr         Delta. Extra.
          Projection.  =    \ ^        	
                    1       ^->            Extra.

This gives the total amount that the Extra vector would be
"shortened" if the next "highest level" of technology were
applied to source i  (see Figure A-2 ).  Since Extra, is a common

-------
                                -37-
                            FIGURE A-l
        Diagrammatic Representation of Programming Problem
                       (Two-Receptor Case)
                                             Impact 4
Limit-
              Feasible Region
                            FIGURE A-2
          Projection of the Delta Vector on the Extra Vector
                      in Programming Problem
 Limit.
                                                      Limit
                                                           1

-------
                             -38-
tenn in all the projection calculations , it can be removed.
This leaves simply the dot product (Dot) of the Delta vector and
the Extra vector:

                      Nr
          Dot.   =                  Delta. Extra.             (13)
             i        f ^               3      D
                      j - 1

     Finally, we define the gradient vector  (G) as:

                       Gi   -   ^t.k/^i                   (14)

This expression represents the relative cost of shortening the
Extra vector by a unit when the next "highest level" of control
technology is applied to source i.  Once a gradient is calculated
for each source, the lowest one is chosen and applied.  This pro-
cess is repeated until one of two possible outcomes occurs:

     •    The Extra vector equals zero; or

     •    The "highest level" of control technologies
          has been applied to all sources.

If the first occurs, a least-cost solution has been reached.  If
the second occurs, the problem is infeasible and no solution can
be found.  That is, the standard cannot be met at all receptors
given the available control technologies.

-------
                           -39-
                  REFERENCES TO APPENDIX A
S. Senju, and Y. Toyoda, "An Approach to Linear Programming
  with 0-1 Variables," Management Science 15, pp. 196-207
  (1968).


Y. Toyoda, "A Simplified Algorithm for Obtaining Approximate
  Solutions to 0-1 Programming Problems," Management Science 21,
  pp. 1417-1427 (1975).

S.H. Zanakis, "Heuristic 0-1 Linear Programming: An Experimental
  Comparison of Three Methods, "Management Science 24, no. 1,
  pp. 91-104 (September 1977).

-------