HUMAN  POPULATION
          EXPOSURES TO  COKE
          OVEN  ATMOSPHERIC
          EMISSIONS
          Final Report
          October 1978
          (Revised May 1979)
          Prepared for:

          U.S. Environmental Protection Agency
          Office of Research and Development and
          Office of Air Quality Planning and Standards
          Project Officers:

          Alan P. Carlin
          Joseph 0. Cirvello
          Technical Monitor:

          Justice A. Manning
          Contracts 63-01 -4314 and 68-02-2835

          SRI Projects EGU-5794 and CRU-6780


          Prepared by:
          Benjamin E. Suta
          Center for Resource and
          Environmental Systems Studies


          CRESS Report No. 27
333 Ravenswood Ave. « Menio Park, California 94025
(415) 326-6200 • Cable: STANRES, Menlo Park • TWX: 910-373-1246

-------
HUMAN  POPULATION
EXPOSURES  TO COKE
OVEN ATMOSPHERIC
EMISSIONS
Final Report


October 1978
(Revised May 1979)
Prepared for:

U.S. Environmental Protection Agency
Office of Research and Development and
Office of Air Quality Planning and Standards
Project Officers:

Alan P. Carlin
Joseph 0. CIrvello
Technical Monitor:

Justice A. Manning


Contracts 68-01 -4314 and 68-02-2835

SRI Projects EGU-5794 and CRU-6780


Prepared by:

Benjamin E. Suta
Center for Resource and
Environmental Systems Studies


CRESS Report No. 27

-------
                                 NOTICE
     This report has been released by the U.S. Environmental Protection
Agency (EPA) for public review and comment and does not necessarily re-
flect Agency policy.  This report was provided to EPA by SRI International,
Menlo Park, California, in partial fulfillment of Contract Nos. 68-01-
4314 and 68-02-2835.  The contents of this report are reproduced herein
as received by SRI after comments by EPA.  The opinions, findings, and
conclusions expressed are those of the authors and not necessarily those
of EPA.  Mention of company or product names  is not to be considered as
an endorsement by the EPA.

-------
                             ACKNOWLEDGMENT

     It is a pleasure Co acknowledge the cooperation and guidance given
by Alan Carlin and Justice Manning of the U.S. Environmental Protection
Agency in the preparation of this report.  In addition, a number of other
people generously supplied input data.   They include:
     Mark Antell, U.S. Environmental Protection Agency
     Robert Armbrust, New York State Department of Environmental
     Conservation
     Thomas Au, Pennsylvania Department of Environmental Resources
     Bernard Bloom, U.S. Environmental Protection Agency
     Walter Cooney, Maryland State Environmental Health Administration
     Ron Dubin, Pennsylvania Department of Environmental Resources
     Arvid Ek, Allegheny County Health Department
     Clemens Lazenka, Philadelphia Air Quality Division
     Jim Payne, Texas Air Control Board
     C. B. Robison, Jefferson County, Alabama Board of Health
     Terry Sweitzer, Illinois Environmental Protection Agency
     Peter Warner, Wayne County, Michigan Health Department
     Robert Yuhnke, Pennsylvania Department of Environmental Resources.
                                   iii

-------
                                PREFACE
     There is a substantial body of evidence, both direct and indirect,
that the mixture that coke oven emissions represent is carcinogenic and
toxic.  Current U.S. Environmental Protection Agency (EPA) policy states
that there is no zero risk level for carcinogens.  To determine what
regulatory action should be taken by EPA on atmospheric emissions of
coke ovens, three reports have been prepared:  (1) a health effects
assessment, (2) a population exposure assessment, and (3) a risk assess-
ment document based on the data in the first two assessments.  This
document is the human population exposure assessment and presents esti-
mates of the number of people in the general population of the United
States exposed to atmospheric concentrations of coke oven emissions.
Estimates are provided of population exposures to ambient concentrations
of benzo(a)pyrene (BaP) and benzene soluble organics (BSO) material caused
by coke oven emissions.

-------
                               CONTENTS

ACKNOWLEDGMENT 	      iii
PREFACE	        v
LIST OF ILLUSTRATIONS	       ix
LIST OF TABLES	       xi
LIST OF CHEMICAL ABBREVIATIONS	       xv
  I  INTRODUCTION	        1
 II  SUMMARY AND CONCLUSIONS 	        3
     A.   Overview	        3
     B.   At-Risk Population 	        6
     C.   Population Estimation	        6
     D.   Population Exposures to BaP Emitted by Coke
          Ovens	        7
     E.   Population Exposures to 3SO Emitted by Coke
          Ovens	       10
     F.   Considerations in the Use of the Annual Average
          As a Measure of Exposure to Coke-Oven Emissions. .  .       10
     G.   Accuracy of Estimated Exposures	       13
     H.   Other Potential Human Exposure Routes	       16
III  SOURCES OF COKE OVEN EMISSIONS	       21
     A.   The Coking Process	       21
     B.   Environmental Emissions During Coking	       22
     C.   Coke Processing Plants	       24
 IV  A METHOD OF ASSESSING BaP AND BSO CONCENTRATIONS IN THE
     VICINITY OF COKE OVENS	       29
     A.   General	       29
     B.   Categorization of Coke Plants by Emission Control.  .       30
     C.   Background Concentrations	       32
     D.   Evaluation of Ambient Concentration Data for Coke
          Plant Locations	       35

                                  vii

-------
     E.   Relationship Between BaF and BSD Atmospheric
          Concentrations 	       37
     F.   Population Exposure Estimates	       40

APPENDICES

     A    AMBIENT ATMOSPHERIC BaP AND BSO CONCENTRATIONS ...       45

     B    STATISTICAL EVALUATION OF BaP ATMOSPHERIC
          CONCENTRATION DATA RECORDED IN THE VICINITY
          OF COKE PLANTS	 .  .       77

     C    DETAILED ESTIMATES OF POPULATIONS AND BaP
          CONCENTRATIONS FOR INDIVIDUAL COKE FACILITIES. ...      101

BIBLIOGRAPHY 	      107
                                  viii

-------
ILLUSTRATIONS
IV-1
IV-2
IV- 3
B-l

B-2

B-3

B-4

B-5

B-6

B-7

B-8

B-9

B-10

B-ll

B-12

B-13

B-L4

Relationship Between BSD and BaP Atmospheric
Relationship Between BaP and BSO Concentrations
Relationship Between BaP and BSO Airborne
Concentrations for Occupational Coke Oven
Statistical Distribution for Atmospheric BaP

Atmospheric Concentrations

Atmospheric Concentrations
Utah 	
Atmospheric Concentrations

Atmospheric Concentrations

Atmospheric Concentrations
New York 	
Atmospheric Concentrations

Atmospheric Concentrations
Illinois 	
Atmospheric Concentrations

Atmospheric Concentrations
Ohio 	
Atmospheric Concentrations

Atmospheric Concentrations

Atmospheric Concentrations

Atmospheric Concentrations


of

of

of

of

of

of

of

of

of

of

of

of

of


BaP

BaP

BaP

BaP

BaP

BaP

BaP

BaP

BaP

BaP

BaP

BaP

BaP


for

for

for

for

for

for

for

for

for

for

for

for

for


Johnstown,

Geneva,

Wayne

Allegheny

Buffalo,

Birmingham,

Granite City,

Sparrows

Cleveland,

Monessen,

Gadsden,

Duluth,

Philadelphia,

38
39
41
80

84

85

86

87

88

89

90

91

92

93

94

95

96
     IX

-------
                                TABLES
 II-l     Estimated BaP Emissions in the United States
          (1972) 	         4

 II-2     Summarization of Ambient BaP and BSO Data	         5
 II-3     Estimated Number of People Exposed to BaP from Coke-
          Oven Emissions, Assuming Variable Location
          Background Concentrations	         8
 II-4     Estimated Number of People Exposed to BaP from Coke-
          Oven Emissions, Assuming Uniform Background
          Concentrations	'	         9

 II-5     Estimated Number of People Exposed to BaP from Coke-
          Oven Emissions, Including Background 	        11

 II-6     Estimated Number of People Exposed to BSO from Coke-
          Oven Emissions	        12

 II-7     Relative Error in Exposure Estimates Assuming
          Circular Isoconcentration Contours 	        17

 II-8     Benzo(a)Pyrene Concentrations in Foods 	        18

III-l     Correlations Among PAH Compounds in the Air over
          Greater Birmingham, Alabama, 1964 and 1965	        23

III-2     Correlation Coefficients Among Log Concentrations
          of 13 PNA and BSO Samples Taken Within Five Coke
          Plants	        23

III-3     By-Product Coke Plant Locations and Capacities ...        25
III-4     Estimated Size and Productive Capacity of By-Product
          Coke Plants in the United States on December 31,
          1975	        27
III-5     Directory of U.S.  Beehive-Coke Plants	        28
 IV-1     Assumed Emission Weighting Factors for Plant
          Compliance Status	        31

 IV-2     Classification of Coke Plants into Emission
          Categories (1974-1975) 	        31
 IV-3     Estimated Annual Background Concentrations of BaP
          for Coke Plant Locations	        33
  A-l     Monessen Air Study, 24-Hour Sample
          Characteristics	        48

  A-2     Ambient BaP Concentrations for Allegheny County,
          Pennsylvania 	        49


                                   xi

-------
A-3     BaP Data Obtained During First Stage Alerts at
        Liberty Borough—Site 8790	       50
A-4     Atmospheric BaP Concentrations near the Geneva
        Works in Utah	       51
A-5     Ambient BaP Concentrations for Wayne County,
        Michigan	       51
A-6     Ambient BaP Concentrations for Buffalo, New York .  .       52
A-7     Ambient BaP Concentrations for Duluth, Minnesota .  .       53
A-8     Ambient BaP Concentrations for Gadsden, Alabama. .  .       54
A-9     Ambient BaP Concentrations for Birmingham,
        Alabama	       54
A-10    CHAMP Site Ambient Atmospheric BaP Data for the
        Birmingham Area (1975 Data)	       55
A-11    Ambient BaP, BSO, and TSP Concentrations  for
        Johnstown, Pennsylvania	       56
A-12    Ambient BaP, BSO, and TSP Concentrations  for
        Philadelphia, Pennsylvania 	       58
A-13    Ambient BaP Concentrations for Granite City,
        Illinois	       59
A-14    Additional Atmospheric Ambient Data for Granite
        City, Illinois	       59
A-15    Ambient BaP Concentrations for Houston, Texas.  ...       59
A-16    Ambient BaP Concentrations for Cleveland, Ohio  ...       60
A-17    Ambient Atmospheric BaP and BSO Concentrations
        for Sparrows Point, Maryland  	       60
A-18    Ambient BaP and BSO Concentrations for Chattanooga,
        Tennessee	.	       61
A-19    Annual Average Ambient BaP Concentrations at NASN
        Urban Stations	       63
A-20    Seasonal Variations of Benzene Soluble Organic
        Substances	       65
A-21    Summarization of Ambient BaP  and  BSO Data	       67
A-22    Annual BaP Averages for Selected  Cities	       67
A-23    Ambient BaP Concentrations for Pennsylvania, 1976.  .       69
A-24    Distribution of BaP Concentrations in Ambient Air
        at Charleston, South Carolina.	       73
A-25    Ambient Atmospheric BaP and BSO Concentrations  for
        Maryland Locations 	       74
A-26    Atmospheric BaP and BSO Concentrations for CHESS
        and CHAMP Sites (1975 Data)	       76
                                xii

-------
B-l     Statistical Summary for Sampling Data Taken from
        a Number of Locations	        82

B-2     Estimated Parameter Values for Regression
        Approximations to Ambient Data	        98

B-3     Statistical Evaluation of Regression
        Approximations	        99
C-l     Detailed BaP Population Exposures	       104

C-2     BaP Exposures for Persons in Locations Having
        More than One Coke Facility	       106
                                xiii

-------
                        CHEMICAL ABBREVIATIONS
A         anthracene
Ant       anthanthrene
BaA       benz(a)anthracene
BaP       benzo(a)pyrene
BbF       benzo(b)fluoranthene
BcA       benzo(c)acridene
BeP       benzo(e)pyr ene
BghiP     benzo(g,h,i)perylene
BjF       benzo(j)fluoranthene
BkF       benzo(k)fluoranthene
BSO       benzene soluble organics
Chr       chrysene
Cor       coronene
DBahA     dibenz(a,h)anthracene
Flu       fluoranthene
Per       perylene
Pyr       pyrene
TSP       total suspended particulates
                                  xv

-------
                            I  INTRODUCTION
     The primary objective of this study has been  to quantify the environ-
mental atmospheric exposure of the general human population to coke-oven
emissions of benzo(a)pyrene (BaP) and benzene soluble organics (BSO).  To
do so, we have located and characterized coke production plants, estimated
atmospheric environmental concentrations of pollutants resulting from
coke production, and estimated human populations exposed to various  levels
of these pollutant concentrations.

     In this report, we  indicate human exposure to coke-oven emissions in
terms of the annual average concentrations for residential population sub-
groups.  Note that this  study reports exposures that took place before
biological sorption occurred and that the degree of sorption is not  con-
sidered.  In addition, because the results of this study are intended to
serve as input to another study in which health effects are to be assessed,
health effects are not addressed.

     The main findings of this report are provided in tables and figures.
The text describes the methodologies, assumptions, and data sources  used.
All estimates given in this report depend in large part on data relia-
bility and availability, both of which varied widely.  Some discussion of
this variability is provided in Appendices A and B.

-------
                      II  SUMMARY AND CONCLUSIONS
A.   Overview

     There are 65 by-product coke planes in the United States.   (Some
authors list 62, omitting separate operations for three of  the  locations.)
These plants consist of an estimated 231 coke-oven batteries, containing
13,324 ovens that have a theoretical maximum annual productive  capacity
of 74.3 million tons of coke.  The industry generally operates  at about
807. of the theoretical capacity.

     Environmental emissions occur in the coking operation  during charging,
from leaks in the oven doors and the tops of ovens, from  the waste gas
stack, during pushing and quenching, and from by-product  processing.  The
various batteries are characterized by different types of control and
operational procedures which affect the amount of their emissions.   In
general, the measurement of environmental emissions from  coke ovens  has
been limited to some atmospheric sampling of BaP for about  one-third of
the locations.  Atmospheric concentrations of TSP have also been measured
for many of the locations, and the BSO fraction of the TSF  has  been
measured for a few locations.  Atmospheric concentrations of other sub-
stances that may be emitted by coke ovens have generally  not been recorded.
In addition, very little work has been done to characterize detailed
emission factors for coke ovens.  Because of these limitations,  this
report's estimates of nonoccupational exposures to coke-oven emissions
are based on the two substances for which some atmospheric  concentration
data are available--BaP and BSO.  These two substances might be considered
as substitute or surrogate measures of total exposure.  However, much more
monitoring data will be required before we can conclude that concentra-
tions of these two substances always correlate well with  other  emitted
substances that are important from a health viewpoint.

     Atmospheric concentration data recorded during 1964  and 1965 for
Birmingham, Alabama, with several coke plants located in  the vicinity,
showed that the correlation coefficient for BaP with 11 other polynuclear
aromatic compounds ranged from 0.65 to more than 0.99.  For BSO with 11
other substances, it ranged from 0.58 to 0.88 (U.S. EPA,  1975).  In
addition, occupational exposure data recorded by NIOSH (1974) show correla-
tion coefficients between BSO and 13 other polynuclear aromatic compounds
to range from 0.71 to 0.94.  The same study also showed correlation  co-
efficients for BaP with 12 other polynuclear aromatic compounds ranging
from 0.57 to 0.95.  The substances used in these correlation studies are
given in Section III of this report.

     It is difficult to use ambient data to assess exposures to coke-oven
emissions; most communities have other sources of the same  substances,

-------
generally associated with coal and other fossil fuel combustion.  Hence,
any evaluation of population exposures to coke-oven emissions must
separate the background concentration from the coke-oven contribution.
Of course, for health risk assessment, the summation of the two  is  im-
portant.  Table II-l reports a BaP emission inventory made by the Environ-
mental Protection Agency (EPA) for 1972.  Stationary sources account  for
987. of the nationwide estimate.  Estimates of BaP emissions from coke
ovens for 1972 were approximately 170 metric tons (mt) per year, based on
a crude emission factor of 2.5 g of BaP per ton of coal processed.  Coke
production is estimated to account' for approximately 19% of the  nation-
wide BaP emissions.  EEA (1978) estimated the 1975 BaP emissions from
coke production to be 100 mt.  Assuming a 4.27. annual growth in  coke
demand and improved emission controls, they estimate that the 1985  BaP
emission will be 21 mt.
                              Table II-l

          ESTIMATED BaP EMISSIONS IN THE UNITED STATES (1972)
                                                           Emissions
                       Source Type	     (tonne/yr)
   Stationary Sources
     Coal, .hand-stoked residential furnaces                   300
     Coal, intermediate-size units                              7
     Coal, steam power plants                                  <1
     Oil, residential through steam power type                  2
     Gas, residential through steam power type                  2
     Wood, home fireplace                                      25
     Enclosed incineration-apartment through municipal          3
     Vehicle disposal                                          25
     Forest and agriculture                                    11
     Other open burning                                        10
     Open burning, coal refuse                                310
     Petroleum, catalytic cracking                              7
     Asphalt air blowing                                       <1
     Coke production                                          170
   Mobile Sources
     Gasoline-powered automobiles and trucks                   11
     Diesel-powered trucks and buses                           <1
     Tire degradation                                          11
   Source:  U.S. EPA (1974)

-------
     The exposure estimates given in this report are based on ambient
monitoring data recorded in the vicinity of coke-oven locations, generally
from 1974 to 1976.  Production data used are for 1975.  Hence,  the ex-
posure estimates apply to the 1975 time period.  According to AISI (1978)
many new coke-oven emission controls were installed in 1976 and 1977.
These were required for implementation of OSHA's new coke-oven  emissions
standard under the terms of various consent orders.  Comparable 1975 and
1976 BaP monitoring data are available from the National Air Surveillance
Network (NASN) program for 18 cities having coke plants.  These cities had
an average BaP concentration of 1.1 ng/m3 during 1975 and 1.0 ng/m3 for
1976.  These concentrations are not significantly different at  the 0.05
confidence level.  However, NASN data for a few cities for the  first three
quarters of 1977 indicate a possible reduction for that year.   Because of
the latency period associated with cancer development, the 1975 exposure
estimates given in this report are considered more relevant for estimating
future cancer cases over the next decade than are current exposures.

     BaP may also have natural sources, including bituminous coal which
also contains benzo(a)anthracene and other polycyclic organic matter.
Two of three types of asbestos used industrially were found to  contain
oils with BaP.  Mold may constitute another source (U.S. EPA, 1974).

     The NASN routinely monitors suspended particulate levels in urban and
nonurban areas.  This program is described in more detail in Appendix A.
BaP and BSO are monitored for 40 locations that include cities  with and
without coke ovens and rural areas.*  The BaP and BSO concentrations re-
corded for this program are summarized  in Table II-2.  The BaP  concentra-
tions are generally 0.1 ng/m3 for rural locations.  Most urban  locations
without coke ovens have average concentrations of  less than 1 ng/m3
                              Table II-2

               SUMMARIZATION OF AMBIENT BaP AND  BSO DATA
                             Cities With   Cities Without
  Pollutant     Statistic     Coke Ovens     Coke Ovens     Rural Areas

BaP (ng/m3)    Average         1.21           0.38            <0.1
1975 data      Sample size     21             13              3
               Range           0.3-4.7        0.03-0.9        <0.1


 1971-72 data
BSO (iJ,g/m3)    Average         4.21           3.75            0.95
               Sample size     25              12               2

               Range           2.1-7.3         1.9-5.6          0.8-1.1
 BSO monitoring was discontinued  in  1972.

-------
(Che average is 0.38 ng/m3); however, areas with coke ovens generally
have average concentrations in excess of 1 ng/m3 (the average is 1.21
ng/m3).


B.   At-Risk Population

     The at-risk populations to coke-oven emissions are defined as the
resident populations exposed to coke-oven atmospheric emissions.  Exposure
is based on the estimated average annual concentrations occurring at the
place of residence of at-risk population subgroups.  Average daily human
inhalation exposure can be calculated as the product of the average annual
concentrations and human daily ventilation rate.  The Radiological Health
Handbook (1960) gives the daily ventilation rate for a standard man as
15 m3/day.  In a kepone assessment report, EPA (1976b) used a rate of
8.6 m3/day.
C.   Population Estimation

     An evaluation of the concentration data shown in Appendix A indicates
that coking operations may possibly affect atmospheric concentrations out
to a radius of 30 km or more from the operations.  For most cases, the
affected radius is considerably less than 30 km; however, for conservative
analysis, population residing within a 30-km radius from each coke plant
is considered as the maximum potential exposure population.  For the
estimation of populations at-risk to selected concentrations resulting
from coke ovens, the resident populations were calculated in a series of
seven concentric rings about each coke plant.  The spacing of the rings
was based on the shape of the concentration versus distance functions
illustrated in Appendix B.  The distances are 0-0.5, 0.5-1.0, 1.0-3.0,
3.0-7.0, 7.0-15, 15-20, and 20-30 km.

     Geographic coordinates of most of the coke plants were obtained from
the U.S. EPA NEDS data system.  The remainder were obtained from consulting
maps or by telephone conversations.  The population residing in each con-
centric ring about each coke plant was obtained from the Urban Decision
Systems, Inc., Area Scan Report, a computer data system that contains the
1970 census data in the smallest geographic area available (city blocks
and census enumeration districts).  The total population residing in each
of these rings for all the coke plants is as follows:
                 Distance from Coke Plant     Resident
                 	(km)	    Population

                            0-0.5                32,700
                          0.5-1                 in£ nnn
                            1-3
                            3-7
                            7-15
                           15-20
                           20-30
    32,700
   116,000
 1,644,000
 7,226,000
22,200,000
15,283,000
25,583,000

-------
The total population residing within 30 km of the coke plants is ap-
proximately 73 million.  This is an overestimate of the total number of
people exposed* because more than one coke plant is located in a number
of areas.
D.   Population Exposures to BaP Emitted by Coke Ovens

     The annual average BaP atmospheric concentrations were estimated
for each of the seven concentric rings around each of the coke plants.
Recorded ambient data were used for those locations having a sufficient
number of samples and monitoring sites; otherwise, the extrapolative
modeling techniques described in Section IV were used.  These modeling
techniques assume that the annual average atmospheric BaP concentrations
near coke ovens can be expressed as a mathematical function of the amount
of coal processed, emission controls, distance from the coke ovens, and
background concentrations from other sources.  The constants of  the
models were statistically estimated, based on ambient data recorded for
some of the locations.  The models were then extrapolated to other loca-
tions for which no ambient data have been recorded.  Because of  the
difficulty in estimating background concentrations for coke-oven loca-
tions, two different types of models are used to give alternative exposure
estimates.  One model type assumes variable BaP background concentrations,
whereas the other assumes constant background concentrations.  For loca-
tions with several coke plants, a procedure was devised to assess the
combined atmospheric concentrations by summing the contribution  for in-
dividual plants for areas in overlapping geographic rings.

     Atmospheric concentrations resulting from coke-oven emissions were
calculated, as were total concentrations including background plus coke-
oven emissions.  Because of the- uncertainty involved  in estimating ex-
posures, several alternative estimates are given under varied assumptions.
Three ,of these alternatives are given in Table II-3, one assumes that
coke plants, for which no monitoring data are available, can be  assigned
to either well and poorly controlled groups; two other estimates use
available monitoring data and assume all other plants are either well
or poorly controlled.

     Two additional exposure estimates are given  in Table II-4.   Each
assumes a uniform background BaP concentration for all locations (i.e.,
0.4 and 1.0 ng/m-3).  Monitoring data are used when available for the ex-
posure estimates.  For other well and poorly controlled plants,  mathe-
matical functions are used.

     All alternative exposure estimates give annual average population
exposure concentrations  ranging from 0.1 (the  lower value used)  to  100
ng/m-3.  Depending on the assumptions used, the' total  number of  exposed
  Coke-oven emissions  resulted in an increase in the average annual
  atmospheric BaP  concentrations  of 0.1  ng/m3 or more.

-------
                        Table II-3
         ESTIMATED NUMBER OF PEOPLE EXPOSED TO BaP
        FROM COKE-OVEN EMISSIONS, ASSUMING VARIABLE
            LOCATION BACKGROUND CONCENTRATIONS*
   Subgroup
Concentration
    Range t
   (ng/m3)

   90-100
   80-90
   70-80
   60-70
   50-60
   45-50
   40-45
   35-40
   30-35
   25-30
   20-25
   15-20
   10-15
   8-10
   6-8
   5-6
   4-5
   3-4
   2-3
   1-2
   0.5-1
   0.1-0.5

     Total
               Assumption
Stratified
 Grouping

     1,800
 All Well
Controlled?

     1,800
50
100
2,400
1,400
1,500
1,600
480
24,000
38,000
43,000
230,000
580,000
440,000
1,200,000
640,000
1,200,000
3,400,000
8,500,000
8,100,000
20,000,000
— •
—
—
870
1,500
1,600
—
1,000
8,700
41,000
110,000
440,000
450,000
620,000
180,000
490,000
990,000
4,400,000
7,900,000
26,000,000
44,000,000    42,000,000
All Poorly
Controlled''1
     1,800
        10
        60
     4,000
     2,400
     1,400
     4,400
     8,600
       480
    25,000
    40,000
   140,000
   360,000
   620,000
   880,000
 1,700,000
   750,000
 1,100,000
 5,600,000
13,000,000
11,000,000
18.000,000

53,000,000
   — No exposures are estimated for this concentration
      range.
   *
    Excludes background concentrations.

   'Based on annual averages.

    That is, all coke plants for which no monitoring data
    exist.

-------
               Table II-4

ESTIMATED NUMBER OF PEOPLE EXPOSED TO BaP
    FROM COKE-OVEN EMISSIONS, ASSUMING
    UNIFORM BACKGROUND CONCENTRATIONS*
   Subgroup
Concentration
    Range''"          Assumed Background
   (ng/m3)        0.4  ng/m3      1.0 ng/m3

   90-100             1,800         1,800
   80-90
   70-80
   60-70
   50-60
   45-50
   40-45
   35-40
   30-35
   25-30
   20-25
   15-20
   10-15
   8-10
   6-8
   5-6
   4-5
   3-4
   2-3
   1-2
   0.5-1
   0.1-0.5

     Total       50,000,000    44,000,000
50
130
2,400
1,400
1,500
1,600
480
24,000
40,000
43,000
220,000
980,000
610,000
620,000
730,000
1,500,000
3,200,000
11,000,000
6,800,000
24,000,000
50
100
2,400
1,400
1,500
1,600
480
24,000
38,000
39,000
230,000
580,000
440,000
1,100,000
380,000
1,500,000
3,500,000
6,200,000
11,000,000
19,000,000
   — No exposures are estimated for this
      concentration range.
   *
    Excludes background concentrations.

   'Based on annual averages.

-------
people  so estimated  ranges  from 42  to  53 million.  The  "better  estimates"
give  44 to  50 million people  exposed.  In no  case were  exposures  estimated
beyond  30 km from the plants.   Had  this constraint been relaxed,  more
people  exposed would have been  assigned to  the  lower  concentrations.

     The total background plus  coke-oven exposures are  given  in Table
II-5.   Note  that,  because the procedure used  first estimated  total ex-
posure  from which  background  concentrations are  subtracted, total ex-
posures are  the same for the  two uniform background assumption  cases.
Detailed exposure  estimates for the variable  background are given in
Appendix C.
E.   Population Exposures  to BSO Emitted by Coke Ovens

     Sufficient data have  not been collected near coke plants nor have
emission factors been developed for adequately assessing the atmospheric
BSO concentrations resulting from the plants' emissions.  The approach
taken here is to estimate  the BSO concentrations, based on  the estimated
BaP concentrations.  A number of problems are associated with this ap-
proach, however, and the results can, at best, be described as "ballpark
estimates."  Further work  on assessing plant emission factors or measuring
environmental concentrations should help to improve the quality of future
estimates.

     The approach taken here, which is described in Section IV, assumes
that BaP constitutes 17. of the total BSO emitted by coke ovens.  Five
of the alternative BaP exposure estimate techniques are used to provide
five alternative BSO estimates.  The estimated exposures to BSO from
coke plants only are given in Table II-6.  Annual average concentrations
are shown to range as high as 10 ng/nr*.  Because of the lack of sufficient
data, estimates are not given for BSO exposures due to coke ovens plus
background.  Table II-2 shows that the 1971-72 concentrations for cities
without coke ovens averaged 3.75 jig/m^.  Hence, adding this value to the
exposure concentrations given in Table II-6 would give a rough estimate
of total BSO exposure.  This, however, may not be proper because the
constitution of the background BSO may likely differ from that of the
BSO emitted by coke ovens.
F.   Considerations in the Use of the Annual Average as a Measure
     of Exposure to Coke-Oven Emissions

     Exposure estimates in this report are given in terms of the daily
exposure averaged over a year.  Statistically, this measure represents
the expected daily exposure; multiplied by 365,- it gives the total ex-
pected annual exposure.  However, the statistical distribution of con-
centrations for a specific location is not symmetrical; rather, it takes
the form of many relatively small observations and a few relatively
                                   10

-------
                  Table II-5

ESTIMATED NUMBER OF PEOPLE EXPOSED TO BaP FROM
   COKE-OVEN EMISSIONS, INCLUDING BACKGROUND
    Subgroup
 Concentration
     Range
    (ng/m3)
    90-100
    80-90
    70-80
    60-70
    50-60
    45-50
    40-45
    35-40
    30-35
    25-30
    20-25
    15-20
    10-15
    8-10
    6-8
    5-6
    4-5
    3-4
    2-3
    1-2
    0.1-1
    0.1-0.5

      Total
  Background Assumption
 Variable'

     1,800

        50
       100
     3,300
       550
     1,500
     1,600
     1,500
    27,000
    36,000
    43,600
   660,000
   570,000
   600,000
   900,000
   770,000
 1,800,000
 7,900,000
17,000,000
24,000,000
 4.800.000

59,000,000
 Uniform^

     1,800

        50
       130
     3,200
       550
     1,500
     1,600
     1,500
    27,000
    36,000
    43,000
   670,000
   550,000
   600,000
   730,000
   950,000
 1,600,000
 6,800,000
13,000,000
26,000,000
 8.000.000

59,000,000
 — No exposures are estimated for this concen-
    tration range.
 *
  Based on annual averages.

 'Based on the stratified grouping assumptions.
 TBased on the 0.4-ng/m  background assumption.
                      11

-------
                             Table II-6


              ESTIMATED NUMBER OF PEOPLE EXPOSED TO BSO

                      FROM COKE-OVEN EMISSIONS*
Subgroup
Concentration
Range7
(ug/m3)
9-10
8-9
7-8
6-7
5-6
4.5-5.0
4.0-4.5
3.5-4.0
3.0-3.5
2.5-3.0
2.0-2.5
1.5-2.0
1.0-1.5
0.8-1.0
0.6-0.8
0.5-0.6
0.4-0.5
0.3-0.4
0.2-0.3
0.1-0.2
0.05-0.1
0.01-0.05
Estimation Procedure
Variable
BaP
Background
1,800
—
50
100
2,400
1,400
1,500
1,600
480
24,000
38,000
43,000
230,000
580,000
440,000
1,200,000
640,000
1,200,000
3,400,000
8,500,000
8,100,000
20,000,000
BaP Back-
ground of
0.4 ng/m3
1,800
—
50
130
2,400
1,400
1,500
1,600
480
24,000
40,000
43,000
220,000
980,000
610,000
620,000
730,000
1,500,000
3,200,000
11,000,000
6,800,000
24,000,000
BaP Back-
ground of
1 ng/m3
1,800
—
50
100
2,400
1,400
1,500
1,600
480
24,000
38,000
39,000
230,000
580,000
440,000
1,100,000
380,000
1,500,000
3,500,000
6,200,000
11,000,000
19,000,000

All Well
Controlled
1,800
—
—
—
—
870
1,500
1,600
—
1,000
8,700
41,000
110,000
440,000
450,000
620,000
180,000
490,000
990,000
4,400,000
7,900,000
26,000,000

All Poorly
Controlled
1,800
10
60
4,100
2,400
1,400
4,400
8,600
480
25,000
40,000
140,000
360,000
620,000
880,000
1,700,000
750,000
1,100,000
5,600,000
13,000,000
11,000,000
18,000,000
  Total
44,000,000  50,000,000  44,000,000  42,000,000  53,000,000
— No exposures are estimated for this concentration range.

*
 Excludes background concentrations.

 Based on annual averages.
                                 12

-------
larger observations.  Examples of these distributions are given  in
Appendix B.  The averages for these types of distributions are much
larger than the median and, generally, only 207« to 40% of the observa-
tions might be expected to exceed the mean in value.  The geometric
average, rather than the arithmetic average, is a better measure to
characterize the central location of these distributions.  The overall
arithmetic average was found to be 1.8 times as large as the geometric
average (Appendix B).

     Calculations of averages and standard deviations are given  in Ap-
pendix B for BaP concentration data recorded over a number of different
days at a specific location.  For most of these locations, the average
was found to equal the standard deviation.  Thus, concentrations for an
individual worst case day could easily be three times the annual average.


G.   Accuracy of Estimated Exposures

     1.   BaP Estimates

          The overall accuracy of the BaP exposure estimates is  difficult
to assess because many relevant factors are either unknown or violate
statistical randomization principles.  In addition, certain sources of
error cannot be adequately quantified with available information.  Examples
of the major sources of error are described in the following paragraphs.


          a.   Errors in Estimating Background Concentrations.  Monitoring
data contain both background and coke-oven emission contributions to BaP
concentrations.  Hence, background concentrations must be subtracted from
monitoring data to estimate coke-oven contributions.  However, accurately
estimating background concentrations for most locations is difficult be-
cause of insufficient monitoring data or lack of other source emission
analyses.  The rough estimates made of background concentrations may, in
some cases, be quite inaccurate.  Because this is an important source of
error in estimating the population exposures to the lower concentrations,
the effect of the assumed background concentrations on population ex-
posures has been parametrically analyzed by using three alternative back-
ground concentration cases.  These are shown in Tables II-3 and  II-4.
The background assumptions can account for errors in total exposures of
about ±30%.
          b.   Errors from Using Samples from a Limited Number of Days
to Represent Annual Averages.  Data given in Appendix B show  that the
day-to-day variations in BaP concentrations for a monitoring  station
follow a skewed statistical distribution of a log-normal  type.  The
average geometric standard deviation is approximately 2 and the coefficient
of variation is 100, indicating that the one-standard deviation confidence
limits for variations in 24 hour concentrations are  1007.  of the mean.
Appendix B also gives factors to adjust the standard deviation for


                                   13

-------
individual daily means  to  the  standard deviation  for  annual mean.  These
factors are based on  the number of days of available  data.  Use of these
factors shows  that  the  one-standard deviation confidence  limits for  the
annual mean based on  the sample mean would vary from  257.  to 1007..  That
is, the one-standard  deviation confidence limits  would  be 0.8  to  1.25
times the estimated mean for the best case and 0.5  to 2 times  the estimated
mean for the worst  case.
          c.   Errors  in Applying Observed Data  from Coke Plants  to Coke
Plants with No Observed Data.  The potential errors in  estimated  ex-
posure concentrations  were addressed by using  the variable background
model to predict average annual concentrations for 1- and 3-km distances
for a number of coke plants  for which environmental monitoring data are
available.  The discrepancies between the observed and  predicted  concen-
trations then provide  an estimate of the accuracy of the procedure.  The
one-standard deviation on these discrepancies was about 1007..  This
indicates that the one-standard deviation confidence limits on a  predicted
value are one-half to  two times that value.  An  error of at least  this
magnitude would be expected  when the model is  applied to locations for
which no monitoring data have been recorded.   In addition, alternative
exposure estimates are arrived at by assuming  that all  plants for which
no monitoring data are available are either poorly or well controlled.


          d.   Errors  from Assuming Circular Concentration Contours.  The
actual residential population distribution was used for the exposure
estimates given in this report; however, it was  assumed that the  annual
average BaP isoconcentration contours about each coke plant are circular.
This, in effect, assumes that over a year the wind blows equally  from all
directions; more specifically, the wind blows  toward the direction of the
major populated areas  in a manner similar to that if it were blowing
equally toward all directions.  This obviously is not true for all coke
plant locations.  Generally, the annual average  isoconcentration  contours
will be elongated in the directions in which the wind blows more  frequently
and shortened in the directions in which the wind blows less frequently.
Depending on the actual residential locations  for specific coke plants,
this assumption may result in over- or under-estimates  of exposed popula-
tions.  However, this  report presents total estimated population  exposures
for all coke plants.   Because more than 60 coke  plants  are located
throughout the country, the  over- and under-estimates for specific loca-
tions should tend to compensate for one another, resulting in national
exposure estimates that are  approximately correct.

          The circular isoconcentration contour  assumption was analyzed
to determine if it would result in gross over- or under-estimates of
total exposures.  The  analysis compared the ratio of the actual and assumed
wind directional-frequency data (which are summarized on a figure called
a wind rose) with the  actual population distribution.   The wind roses
used in this analysis  show the percent of time,  over a  year, that the
wind blows from each of 16 equally spaced compass directions and  the


                                  14

-------
percent of time that calm conditions exist.  Each of the 16 compass
directions represent a 22.5° sector.  The compass direction having the
highest percentage of winds is sometimes called the direction of pre-
vailing wind.  However, note that to fulfill this definition the wind
need only blow from the prevailing wind direction a small fraction of
the time more than from other directions.

          The uniform wind direction assumption implies that the wind
blows 6.257o of the time from each of the sectors represented by the
16 compass direction wind rose (assuming no calm periods).  The ratio
of the actual percent of time the wind blows to 6.25% for a compass
direction indicates the degree in which the isoconcentration contours
should be elongated or shortened toward the diametrically opposing
22.5° sector.  This ratio, therefore, indicates the degree of over- or
under-exposure estimation for that sector resulting from the circular
isoconcentration assumption.  Mathematically, this may be expressed as:
where

          g. = the ratio of actual to assumed wind frequencies  for
               a sector
          a. = the actual wind frequency (7.) for a sector  taken
               from a wind rose

          U. = 6.2570, the uniform wind frequency.

          As examples, if g = 0.5, population exposure concentrations  for
that sector are overestimated by 50%; if g = 2.0, then population exposure
concentrations for that sector are underestimated by 50%.  The  total
fractional error for the estimated exposures for one location can be
approximated by

                               16     , 16
                              i
                          E =
where

           E = the total fractional error

          g. = the ratio as previously defined

          P. = the residential population  in each 22.5°  sector.

          The SRI data base containing the U.S. population on  1-km grids
was printed for the area about each coke plant.  These areas were each
divided in sixteen 22.5° sectors radiating from each  coke plant.  It was
then determined whether a relatively  significant number  of residents


                                   15

-------
would be exposed  in each  sector and, based on  this determination, corre-
sponding values of 1 or 0 were assigned to P^.  The total  fractional
error (E) was then calculated for each coke plant; these are given  in
Table II-7.  Individual coke plant values were  found  to vary from 0.55
to 1.58 with an average value of 0.96.  The average weighted by  population
residing within 15 km of  the coke plants was 0.97.  Hence, when  calm con-
ditions are excluded, the circular isoconcentration contour assumption
may result in overexposure estimate of about 37. to 4%.  The average calm
conditions occur  at each  plant 47. of the time.  During calm conditions,
adjacent residential populations will be exposed  to very high concentra-
tions, which were not considered in this report.  Thus, the total over-
estimate should be less than 4%.
          e.   Total Error.  Because of  the many uncertainties,  an
accurate assessment of  the  total error associated with the BaP exposure
estimates cannot be made.   The estimates are not expected to be  highly
accurate for any location.  The overall  exposure estimates, which are a
summation of the exposures  about individual plants, are expected to be
more accurate because they  were formulated by using averages of  parameters
that represent a range  of meteorological, geographical, and emission
control conditions.
     2.   BSO Estimates

          The BSO exposure  estimates are based on  the BaP  estimates.
Thus, they contain not only the errors associated with estimating BaP
exposures but also an additional error source caused by  relating BSO  to
BaP.  Occupational data are evaluated in Section IV-D of this report  to
determine the magnitude of  this error source.  To do this, the BaP con-
centrations were used to estimate corresponding BSO concentrations; the
estimated concentrations were then compared with recorded  BSO concentra-
tions.  The one-standard deviation for the differences between the pre-
dicted and observed concentrations was found to be 537, for individual
values and 107. for the mean.
H.   Other Potential Human Exposure Routes

     There are potential human exposure routes for coke-oven emissions
other than inhalation.  These include ingestion of contaminated food and
water and dermal contact.  In addition, family members of occupational
workers might be exposed through particulates brought home on clothing
and other equipment such as  lunch pails and automobiles.  An assessment
of potential human exposures via these routes was excluded from the scope
of this study because they either appear to be .much less significant than
the inhalation route or because of the lack of available data.
                                  16

-------
                                              Table II-7


           RELATIVE ERROR IN EXPOSURE ESTIMATES ASSUMING CIRCULAR ISOCONCENTRATION CONTOURS
Site   No. of   Relative  Site   No. of   Relative  Site   No. of   Relative  Site   No. of   Relative
No.*  Sectors*   Error**  No.*  Sectors*   Error**  No.*  Sectors*   Error**  No.*  Sectors*   Error**
1
3
4
5
6
7
8
10
11
12
13
14
15
16
8
8
10
12
9
11
7
11
11
11
11
6
13
13
0.96
0.81
0.95
1.01
0.96
0.77
0.55
0.92
0.90
0.89
0.89
0.87
0.96
1.05
17
18
19
20
21
22
23
24
25
27
28
29
30
31
7
7
6
7
9
9
12
9
11
4
7
6
8
10
0.73
0.73
0.78
0.73
0.84
0.77
1.09
1.12
1.13
1.58
1.04
1.48
1.41
1.28
32
33
35
36
•37
38
39
40
41
42
43
44
45
46
5
4
8
6
9
13
7
12
15
13
10
16
7
11
0.59
0.75
0.91
0.51
1.34
1.04
1.36
1.04
0.94
1.00
0.95
1.09
0.53
0.94
47
49
50
51
52
53
54
55
56
57
58
60
64

11
10
13
11
10
11
11
8
11
9
4
9
10

0.94
1.19
0.99
1.18
0.67
1.18
0.82
0.59
1.17
0.69
1.19
1.11
0.83

*vV
 Corresponds  to site numbers given in Table III-3.   Some sites  are  excluded  because  of  insignificant
 exposures  or insufficient data available.

 Number  of  the sixteen 22.5° sectors for which significant  population  exposures  are  estimated.
t
 Vdlues  less  than  I  indicate overestimates  and values  more  than 1  indicate underestimates.

-------
     Foods can become contaminated because of atmospheric fallout of
particulates or by way of contaminated water released by the coke plant.
The contamination may be on the surface of plants from fallout or in-
cluded by root-uptake.  Animals can become contaminated by drinking con-
taminated water, eating contaminated foods, or breathing contaminated
air.  Contamination may also result from other man-made or natural sources.
Processed foods may contain additional contaminations from the combustion
of fuels used in smoking, roasting, or broiling.  Foods in general have
been found to contain concentrations of polynuclear aromatic hydrocarbons
such as benzo(a)anthracene, chrysene, and benzo(a)pyrene (Radding et al.,
1976).   Table II-8 lists concentration levels of BaP in some foods.   As
expected, the BaP concentration of certain prepared foods is higher than
for other foods.  At present, insufficient information is available to
access the potential contamination of foods by coke-oven emissions.
                              Table II-8

                BENZO(A)PYRENE CONCENTRATIONS IN FOODS
                                   Concentration
          	Food               (ng/kg)       Reference

          Cereals                    0.3-0.3            A
          Potato peelings              0.36             A
          Potato tubers                0.09             A
          Barley, wheat, rye         0.2-4.1            B
          Cabbage                     24.5              B
          Spinach                      7.4              C
          Lettuce                    2.8-12.8           B
          Tomatoes                     0.22             B
          Fruits                     2.0-8.0            C
          Refined fats and oils      0.9-15             C
          Fresh fish                  <0.1              D
          Broiled meat and fish      0.2-0.6            C
          Smoked fish                1.0-78.0           E
          Smoked meat/sausage       0.02-107.0          C
          Roasted coffee             0.3-0.5            B
          Roasted coffee             0.1-4.0*           C
          Teas                       3.7-3.9*           B
          Whiskey                      0.04*            B
          A—Shabad (1972)
          B—Grummer (1968)
          C —IRAC (1973)
          D--Gorelova (1971)
          E--Andelman (1970)
                                  18

-------
     Polynuclear aromatic hydrocarbons find their way into waterways
already absorbed onto aerosols or bacteria.  Although their solubility
in pure water is essentially zero, they may exist in water in association
with organic matter or colloids (Radding et al., 1975).  The IRAC (1973)
report lists BaP concentrations in drinking water of 0.0001 to 0.023
                                    19

-------
                  Ill  SOURCES OF COKE OVEN EMISSIONS

                       *
A.   The Coking Process

     Coke is a porous cellular residue from the destructive  distillation
or carbonization of coal.  It is used as a fuel and reducing agent  in
blast furnace operations, and in foundries as a cupola  fuel.   Of  the
approximately 60 million tons of coke produced annually in the United
States, 92% is used in blast furnaces, 57. in foundry operations,  and 3%
in other types of industrial plants.  Of the total coke production,
approximately 907. is produced by steel industry plants,  87, by foundry
plants, and 17<> by beehive ovens.

     Two basic processes are used in the production of  coke:   One recovers
vapors and other by-products from the coking process (by-product  ovens),
and one does not (beehive ovens).  The beehive oven, an older design,
that has been steadily replaced by the newer by-product design is excluded
from this analysis.

     A by-product coke battery consists of 10 to  100 ovens made up  of
chambers for heating, coking, and regeneration.   Heating and coking flues
alternate with each other so that there is a heating flue on either side
of a coking flue; the regenerative flues are located underneath.

     The coking cycle begins with the introduction of coal into the coke
oven.  This operation, called "charging," is carried out with a mechanical
"larry car" on rails on the top of the battery.   The larry car receives
a load of coal from the coal bunker at the end of the battery.  The car
moves down the battery to the oven to be charged.  The  lids  on the  oven
charging holes are removed, the larry car is positioned over the  holes,
and the hoppers are emptied.  During the charge,  the oven is aspirated
by steam jets in the standpipes connecting the by-product gas collector
main with the oven.  This operation, called "charging the main" is  designed
to limit the escape of gas from the oven during the charging process.
After charging is completed, the lids are replaced and  the aspiration
system is shut off.

     The "coking time," the time required to produce coke from coal, is
governed by numerous factors, including the condition and design  of the
oven heating system, width of the coking chamber, coal  moisture,  and the
nature of the coals being coked.  The coking time for blast  furnace coke
varies from 16 to 20 hours.  Coking times for foundry coke are longer than
*
 The material contained in this section is summarized  from  the  Federal
 Register (October 22, 1976).


                                   21

-------
for blast furnace coke because coke of different physical characteristics
is required for foundry operations.

     When the coal is coked, the doors on each side of the oven are re-
moved and the coke is pushed out.  A large mechanically operated ram
attached to a pusher machine moves the coke out the opposite side of the
oven called the "coke side," through the "coke-guide" attached to the
door machine and into a railroad car called the "hot car" or "quench car."
The quench car moves down  the battery to a "quench tower" where the hot
coke is cooled with water.  The quenched coke is then dumped onto the
coke wharf, from which it  is conveyed to the screening station for sizing,
then to the blast furnace, or removed for other purposes.  When the doors
on the oven are replaced,  the oven is ready to be charged again.
B.   Environmental Emissions During Coking

     Environmental emissions can occur during charging; during coking
from leaks in the doors and on the top of the oven; from the waste gas
stack; and during pushing and quenching, and from by-product processing.
Coke-oven emissions are described as a complex mixture of particulates,
vapors, and gases (Federal Register. October 22, 1976).

     Because of the effort and complexity that would be required in
characterizing all of the constituents of coke-oven emissions, various
surrogate measures have been used in the past.  These usually are of
three types:  TSP,* BSO, and BaP.  TSF is generally considered not to be
a specific enough measure for assessing total occupational health effects
(Federal Register. October 22, 1976).  The concept of a surrogate measure
would be valid if it could be shown that that measure correlates well
with the presence of other emitted substances known to have adverse
health effects.   Atmospheric concentration data recorded during 1964 and
1965 for Birmingham, which has several coke plants in the surrounding
area, showed that the correlation coefficient for BaP with 11 other sub-
stances ranged from 0.65 to more than 0.99.   For BSO with 11 other sub-
stances the coefficient ranged from 0.53 to 0.88 (U.S. EPA, 1975),
indicating a fairly good association.  These are given in Table III-l.
In an occupational exposure study, the atmospheric concentrations of
13 polynuclear aromatics (PNAs) and the total benzene soluble organics
were recorded.  A correlation study was made of these data using logarith-
mic transformations because the data followed a log-normal distribution
(NIOSH, 1974).  The correlation of the PNAs with BaP and BSO are given
in Table III-2.   Except for one case, all the correlation coefficients
exceeded 0.7, thus indicating a fairly good correlation.  The correlation
of BSO with the 13 PNAs was generally better than the similar correlations
for BaP.
*
 TSP--total suspended particulates.
                                   22

-------
              Table III-l

  CORRELATIONS AMONG PAH COMPOUNDS
 IN THE AIR OVER GREATER BIRMINGHAM,
       ALABAMA, 1964 AND 1965

Compound
Flu
Pyr
BaA
Chr
BeP
BaP
Per
BghiP
A
Cor
TSP
BSO

BaP
0.916
0.935
0.988
0.980
0.998
1.000
0.985
0.966
0.971
0.815
0.789
0.651
Compound
BSO
0.582
0.684
0.597
0.746
0.677
0.651
0.689
0.804
0.672
0.867
0.880
1.000

TSP
0.668
0.730
0.742
0.842
0.823
0.789
0.830
0.839
0.716
0.856
1.000
0.888
 Source:' U.S. EPA (1975)
             Table III-2

 CORRELATION COEFFICIENTS AMONG LOG
  CONCENTRATIONS OF  13  PNA AND BSO
SAMPLES TAKEN WITHIN FIVE COKE PLANTS
Compound
Flu
Pyr
BcA
Chr
BaA
BbF
BjF
BkF
BeP
BaP
DBahA
BghiP
Ant
BSO
BaP
0.797
0.740
0.569
0.857
0.824
0.776
0.768
0.813
0.950
1.000
0.694
0.855
0.892
0.914
                            BSO
    Source:  NIOSH  (1974)

                  23

-------
     The occupational and Che Birmingham correlation studies provide some
justification for using a surrogate measure rather than trying to identify
and control each of the PNA compounds emitted by coke ovens.
C.   Coke Processing Plants

     In 1975, 57.2 million tons of coke were produced in the United States.
By-product ovens produced 98. 77, of the total production, with beehive
ovens accounting for the remaining 1.37..  Approximately 90% of the coke
is used in blast furnace plants, whereas 2% is exported.  The remainder
is primarily used in foundries.  The yield of coke from coal, which
averaged 68.4% in 1975, has remained fairly constant during the past
decade (Sheridan, 1976).

     In the United States, 65 plants produce coke.  (Some authors list
only 62 by combining three pairs of closely collocated plants, where
each pair of plants are owned by the same corporation.)  The 65 plants
are listed in Table III-3 which also lists the coal capacity and the
1974 coal consumption on a plant-by-plant basis.  The plants consist of
an estimated 231 coke-oven batteries containing 13,324 ovens that have a
theoretical maximum annual productive capacity of 74.3 million tons of
coke.  Because of depressed economic activity in 1975, the industry
operated at only 76% of this capacity.  Coke production on a state-by-
state basis is given in Table III-4.

     The Keystone Coal Industries Manual (1975) lists six beehive-coke
plants.  These operate in two states (Pennsylvania and Virginia).  Although
excluded from this analysis, they are listed in Table III-5.
                                  24

-------
                  Table  III-3



BY-PRODUCT COKE PLANT LOCATIONS AND CAPACITIES
State. City
Alabama
1. Tarranc
2. Holt
3. Woodward
i. Gad sen
5. Thomas
6. Birmingham
7. Fatrfield
California
8. Foncana
Colorado
9. Pueblo
Illinois
10. Cranita City
11. Chicago
12. Chicago
13. South Chicago
Indiana
14. Chesterton
IS. Indianapolis
16. Terre Haute
17. Eaac Chicago
18. Ease Chicago
19. Gary
20. Indiana Harbor
Kentucky
11. Ashland
Maryland
22. Sparrows Point
Michigan
23. Detroit
24. Dearborn
25. Zug Island
(Detroit)
Minnesota
26. St. Paul
27. Duluth
Missouri
28. St. Louis
New York
29. Buffalo
30. Lackawaoa
31. Buffalo
Plant N*ne
Tarrant Plant
Holt Plant
Woodward Plant
Cadsdan Plant
Thomas Plant
Birmingham Plant
Fairfield Plant

Fontana Plant
Pueblo Plant

Granite City Steel Div.
Chicago Plant
Wisconsin Steel Works
South Chicago Plane

Burns Harbor Plant
Prospect Street Plant
Terre Haute Plant
Plant No. 2
Plant No. 3
Gary Plant
Indiana Harbor Plant
Semet
Sparrows Point Plant
Semet
Steel Plant
Zug Island Plant

St. Paul Plant
Duluth Plane
St. Louis Plane
Harriet Plant
Lackawana Plane
Dormer -Hanna Plant
Company
Alabama By-Products Co.
Empire Coke Co.
[Coppers Company, Inc.
Republic Steel Corp.
Republic Steel Corp.
U.S. Pipe and Foundry Co.
U.S. Steel Corp.

Kaiser Steel Corp.
C7U Sceel Corp.

National Steel Corp.
Interlace, Inc.
International Harvester Co. ,
Wisconsin Steel Dlv.
Republic Seeel Corp.

Bethlehem Seeel Corp.
Citizens Gas & Coke Utility
Indiana Gas and Chemical Corp.
Inland Steel Co.
Inland Steel Co.
U.S. Seeel Corp.
Yougseovn Sheet and Tube Co.
Solvay Dlv., Allied Chemical Corp.
Bethlehem Steel Corp.
Solvay Dlv., Allied Chemical Corp.
Ford Motor Co.
Great Lakes Seeel Dlv. , Nacional
Seeel Corp.

Koppers Company, Inc.
U.S. Steel Corp.
Greae Lakes Carbon Corp., Missouri
Coke & Chem. Dlv.
Seme c- Solvay Div., Allied Chemical
Corp.
Bethlehem Seeel Corp.
Donner-Hanna Coke Corp.
Annual Coal
Capacity
(tons)
1,200,000
ISO. 000
800,000
820,000
185.000
1,175,000
2.500,000

2,336,000
1.332,000

1,132,000
949,000
991.000
590,000

2,630,000
675,000
204,000
3.102.000
1,642,000
3,700,000
2.100,000
1,600,000
4,820.000
900,000
1.800,000
2,850,000

250,000
850,000
450,000
400,000
4,250,000
1.387,000
1974 Coal
Consumption
(:ons)


1,760,000


900,000
643.000

2,525,000
584,838
193,000
3.096.000
1.258,000
1.750,000

4.100,000




3.385,000
                      25

-------
                                Table III-3 (Concluded)


Scace. City
Ohio
32. Ironcon

33. New Miami
34. Mlddlecown
35. Painesvllle
36. Portsmouth

37. Toledo
33. Cleveland
39. Masailon
40. Warren
41. Youngscovn
42. Lorain
43. Campbell
Pennsylvania
44. Swede land
45. Bethlehem
46. Johnstown
47. Johnstown
4a. Midland

49. Allquippa
50. Pittsburgh
51. Erie
52. Philadelphia
53. Pittsburgh
54. Claircon
55. Fairless Hills
56. Mono s sen
Tennessee
57. Chattanooga
Texas
58. Houston
59. Lone Scar
Utah
60. Provo
West Virginia
61. Weircon

62. Ueircon

63. Fairmont
64. Follonsbee
'J i scons in
63. Milwaukee



Plane Name

Ironcon Plane

Hamilton Plane
Mlddlecown Elanc
Painesvllle Plane
Empire

Toledo Plane
Cleveland Plane
Mass i Ion Plane"
Warren Plane
Youngstown Plane
Lorain Cuyahoga Works
Campbell Plane

Alan Hood Plane
Bethlehem Plane
Rosedale Oiv.
Franklin Dlv.
Alloy & Stainless Sceel
Div.
Aliquippa Plane
Pittsburgh Plane
Erie Plane
Philadelphia Plane
Neville Island Plant
Claireon Plane
Fair less Bills Plane
Wheeling

Chattanooga Plane

Houston Plane
£. 3. Germany Plane

Geneva Works

Weir ton Mainland Plant

Weircon 's Brown's Island
Plane
Fairmont Plane
Ease Seeubenville Plane

Milwaukee Solvay Coke
Co.


Company

Semee-Solvay Div.. Allied Chemical
Corp.
Armco Sceel Corp.
Armco Sceel Corp.
Diamond Shamrock Corp.
Detroit Sceel Div. of Cyclops
Corp.
Ineerlaka Inc.
Republic Scael Corp.
Republic Sceel Corp.
Republic Seeel Corp.
Republic Seeal Corp.
U.S. Sceel Corp.
Youngstown Sbeec and Tube Co.

Alan Wood Sceel Co.
Bethlehem Seeel Corp.
Bethlehem Sceel Corp.
Bethlehem Seeel Corp.
Crucible Inc. , Dlv. Cole
Industries
Jones and Laugnlin Sceel Corp.
Jones and Laughlin Steel Corp.
Koppers Company, Inc.
Philadelpnia Coke Division
Shenango Inc.
U.S. Sceel Corp.
U.S. Sceel Corp.
Pittsburgh Sceel Corp.

Chattanooga Coke and Chemicals Co.

Armco Steel Corp.
Lane Scar Sceel Co.

U.S. Sceel Corp.

Ueircon Sceel Dlv.. National Sceel
Corp.
Weirton Seeel Div. . National Sceel
Corp.
Sharon Sceel Corp.
Wheeling-Pittsburgh Sceel Corp.

A Division of Plcklands Mather
and Co.
Annual Coal
Capacity
(tons)

1.230,000

934,000
748,000
215,000
600,000

438.000
2,220,000
250,000
650,000
1,500,000
2,700,000
2,300,000

303,000
2,210,000
550,000
1,680,000
657,000

2,250.633
2,587,404
290,000
715,400
1,022,000
9,670,000*
1,300,000
750,000

204,400

584,000
498,000

2,000,000

2,500.000

1,325,000

300,000
2,500,000

347,000

1974 Coal
Consumption
( tons)





210,000








1,395,116


2,105,000
545.000
1.645,000
630,000




385,000
323,900







492,000







284,000




Based on a 1973 emission inventory.
Sources: Keystone
Coal Industries Manual (1975)
Varga (1974)
                                          26

-------
                             Table  III-4

         ESTIMATED  SIZE AND PRODUCTIVE  CAPACITY  OF  BY-PRODUCT
        COKE PLANTS IN THE UNITED STATES ON DECEMBER 31, 1975
State
Alabama
California
Colorado
Illinois
Indiana
Kentucky
Maryland
Michigan
Minnesota
Missouri
New York
Ohio
Pennsylvania
Tennessee
Texas
Utah
West Virginia
Wisconsin
Undistributed
Total
Maximum Annual
Theoretical
Number Number Productive
of Number of of Capacity
Plants Batteries Ovens (tons)
7
1
1
4
6 (7)
1
1
3
2
1
3
12
12 (13)
1
2
1
3 (4)
1
__
62 (65)
28
7
4
9
31
2
12
10
5
3
10
35
51
2
3
4
13
2
_ _
231
1,401
315
206
424
2,108
146
758
561
200
93
648
1,795
3,391
44
140
252
742
100
__
13,324
6,961,000
1,547,000
1,261,000
2,523,000
11,925,000
1,050,000
3,857,000
3,774,000
784,000
257,000
4,053,000
9,960,000
18,836,000
216,000
839,000
1,300,000
4,878,000
245,000
_-
74,266,000
Coke
Production
in 1974
(tons)
5,122,000
*
*
1,912,000
9,073,000
*
*
3,259,000
*
*
•*
8,842,000
16,318,000
*
*
*
3,555,000
*
12,656,000
60,737,000
 Included in Undistributed.

Source:  Sheridan (1976)
                                  27

-------
                        Table III-5

           DIRECTORY OF U.S.  BEEHIVE-COKE PLANTS
Name or Location
    of Plant
County
Company
Pennsylvania
1.
2.
3.
Mahoning
Daugherty
Laughead
Armstrong
Fayette
Fayette
Caipentown Coal & Coke Co.
Bortz Coal Company
Ruane Coal & Coke Company
Virginia
5.
6.
Vansant
Esserville
Buchanan
Wise
Jewell Smokeless Coal Corp.
Christie Coal & Coke
Source:  Keystone Coal Industries Manual (1975)
                            28

-------
        IV  A METHOD OF ASSESSING BaP AND BSO CONCENTRATIONS
                    IN THE VICINITY OF COKE OVENS
A.   General

     All available ambient concentration data recorded  for BaP and BSO  in
the vicinity of coke ovens are presented in Appendix A  and analyzed in
Appendix B.  These data (mostly for BaP) have been recorded in 15 loca-
tions, some of which contain several coke plants; as a  result, approxi-
mately one-third of the coke plants are represented.  However, in many
cases, the data were recorded for only a few days and for only a few
sampling stations, thus making exposure estimates based solely upon them
unreliable.  Moreover, it was necessary to devise some  method of pre-
dicting ambient concentrations for coke plant areas in  which no atmo-
spheric data have been recorded.  A procedure for doing this is given
here.  One approach considered was to model the concentrations mathe-
matically, basing it in part on emission factors, amount of coal pro-
cessed, and local meteorology.  When this approach was  tried by the
EFA (Youngblood, 1977), it was concluded that, because  of the uncertain-
ties in characterizing the sources themselves, definitive estimation of
air quality impact of coke ovens by means of dispersion calculations is
impossible at this time.  The EFA is currently working  on developing
better emission factors for coke ovens.  Because these  will not be
available  for some time, however, it was decided to develop a procedure
to extrapolate the available ambient data that have been recorded in
the vicinity of coke plants to other locations for which no data has
been recorded.  When possible and when they seem reliable, the actual
recorded ambient concentration data have been used to estimate popula-
tion exposures.

     The procedure that was devised required the following steps, which
are described in more detail  in subsequent sections of  this report:

     (1)   Information on the  type of environmental controls at coke
           plants is evaluated to determine if facilities can be grouped
           by their degree of  control.

     (2)   The background concentrations estimated for each coke plant
           location are those  that would exist if the batteries were not
           in operation.

     (3)   Existing ambient concentration data are evaluated to determine
           if atmospheric concentrations can be expressed as a function
           of distance from the coke plants.
                                  29

-------
     (4)  These concentration functions are evaluated to determine if
          relationships can be derived from them, based on  the amount of
          coal processed and the degree of environmental controls.

     (5)  The functions are then used to estimate atmospheric concen-
          trations in the vicinity of coke plants, with subsequent esti-
          mation of human population exposures.


B.   Categorization of Coke Plants by Emission Control

     Emission factors are not well-developed for coking operations.  Among
other factors, they are thought to be a function of process equipment,
environmental controls, and operating procedures.  In theory, a different
set of emission factors exists for each battery.  These battery emission
factors would be composed of emission factors for such sources as charging,
door leaks, pushing, topside leaks, by-product processing,  quenching, and
the waste gas combustion stack.

     The most detailed source of information on coke battery pollution
control compliance is based on a survey conducted by FEDCo  during September
1974 to April 1975 (Kuliyian, 1976).  Among other items reported in  this
survey was the compliance status of each plant or battery with regard to
charging, doors, waste gas combustion stacks, pushing, and  quenching.
Compliance or noncompliance provide only a general indication of environ-
mental emissions.  In addition, some of the batteries have  reduced their
emissions since 1975.  However, this time frame is consistent with the
dates when much of the environmental concentration data were recorded.

     Weighting factors were assigned to each compliance status listed in
the PEDCo survey (in, out, at least one battery out, under  a Legal plan,
undetermined).  These weighting factors are based on work performed  by
EFA personnel, who were familiar with coke operations, to roughly estimate
BaP emission factors (Manning, March 18, 1977).  This assignment of
weights assumes that an in-compliance status indicates low  emissions and
that an out-compliance status indicates high emissions.  Because the EFA
work gives emission factors for clean and dirty operations, the clean
factor was assigned to the in-compliance status and the dirty factor was
assigned to the out-compliance status.  Plants having at least one battery
out of compliance and at least one battery in compliance were assigned a
weighting factor half-way between the out and in factors.   These weighting
factors are given in Table IV-1.  Note that the quenching weighting
factors dominate those for all other sources.  Individual weights were
assigned to each compliance status within plants and summed to give  a
total for each plant.  These sums formed the basis for classifying plants
into two groupings.  Plants for which no compliance data are available
are assigned to a separate group.  Plant assignments are shown in Table
IV-2.  This method of assignment can, and obviously has, led to some mis-
classifications.  At best, it should be regarded as a technique to be
used to form strata for statistical sampling.  In theory, stratified
samples usually have increased precision over simple samples.  As will be
                                  30

-------
                             Table IV-1

                 ASSUMED EMISSION WEIGHTING FACTORS
                     FOR PLANT  COMPLIANCE STATUS
Compliance Status

Emission Source
Charging
Doors
Pushing
Topside**
Quenching
Waste gas stacks

In
1.5
16
N*
1.6
175
N

Out
80
130
3
65
350
0.7

Undetermined
80
130
3
65
350
0.7
At Least One
Battery Out
40
73
1.5
33
260
0.4
Under a
Legal Plan
40
73
1.5
33
260
0.4
^--Negligible.

Topside compliance was assumed to be the same as door compliance.


                             Table IV-2

             CLASSIFICATION OF COKE PLANTS INTO EMISSION
                        CATEGORIES  (1974-1975)
Plant
Number
1*
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Classification
K**
F
K
K
F
F
F
F
F
F
F
K
F
F
F
F
K
X
K
K
K
F
                         Plant                      Plant
                        Number   Classification   Number

                          23            F           45
                          24            F           46
                          25            F           47
                          26            F           48
                          27            K           49
                          28            F           50
                          29            K           51
                          30            K           52
                          31            F           53
                          32            K           54
                          33            K           55
                          34            K           56
                          35            F           57
                          36            K           58
                          37            K           59
                          38            K           60
                          39            F           61
                          40            X           62
                          41            F           63
                          42            F           64
                          43            F           65
                          44            F
Classification

       F
       F
       F
       F
       F
       F
       F
       X
       F
       K
       X
       K
       F
       K
       F
       F
       X
       X
       X
       F
       F
Plant numbers correspond to plant names given in-Table III-3.

F indicates clean and K indicates dirty.  The X  indicates that  insuffi-
cient data were available to classify the plant.
                                 31

-------
later shown, atmospheric concentration versus distance from  the coke plant
relationships for  the two  strata, when scaled for plant production, were
different.  This indicates that  the stratification method did, in  this
case, provide increased precision.
C.   Background Concentrations

     Because substances emitted  to the atmosphere by coke ovens can also
be emitted by other  sources, it  is necessary to consider atmospheric con-
centrations as a  sum of background plus coke-oven emissions.  The coke
plants should only be assigned responsibility for their contribution to
the total.

     Background concentrations are difficult to assess because ambient
concentrations are seldom measured in an area when the coke-ovens are not
in operation.  Moreover, upwind  ambient concentrations, recorded near coke
plants, appear to have been influenced by  the coking operations.  In fact,
ambient atmospheric  concentrations of BaP  or BSO have not been measured
at all for many of the coke-oven locations.  It is therefore necessary to
estimate background  concentration by using data recorded at a sufficient
distance from the coke plant or  by using data recorded at "similar" loca-
tions.  "Similar" cities were taken to be  cities in the same general geo-
graphic area for which ambient data were available.  No attempt was made
to select cities based on other  factors such as manufacturing.  Indeed,
this would not have  been possible given the paucity of data.  Either of
these methods has inherent error.  In addition, background concentrations
have been shown to vary from location to location within a city and with
the season (see monitoring data  given in Appendix A).

     The available BaP atmospheric concentration data for cities without
coke plants are given in Appendix A.  They were reviewed to identify a
"similar" noncoke plant location for each  coke plant location.  For
example, the average BaP concentration over Montgomery, Jacksonville, and
Charleston was used  to represent Birmingham.  The assumed annual average
BaP backgrounds are  given in Table IV-3.   They vary from 0.04 ng/m^ for
Houston to 1.6 ng/m^ for Pittsburgh.

     To illustrate the effects of various  assumed background concentrations
on exposure estimates, the estimates of population exposures draw on three
sets of assumptions.  These assumptions are (1) the background concentra-
tions are variable as given in Table IV-3, (2) the background concentra-
tion for all locations is 0.4 ng/m-* (the average for NASN cities without
coke ovens shown  in  Appendix A), and (3) the background concentration
for all locations is 1 ng/nr (the average  for all locations shown in
Appendix A).
                                   32

-------
                Table IV-3

ESTIMATED ANNUAL BACKGROUND CONCENTRATIONS
     OF  BaP FOR COKE PLANT LOCATIONS
plant*
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
BaP
(ng/m3)
0.4
0.4
0.4
0.4
0.4
0.4
0.4
1.2
0.6
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.4
0.8
1.1
1.1
1.1
0.4
0.3
0.3
0.8
0.8
0.8
0.6
0.6
0.6
Remarks**
Montgomery, Jacksonville, Charleston
Montgomery, Jacksonville, Charleston
Montgomery, Jacksonville, Charleston
Montgomery, Jacksonville, Charleston
Montgomery, Jacksonville, Charleston
Montgomery, Jacksonville, Charleston
Montgomery, Jacksonville, Charleston
Average of 5 sites in the Los Angeles area
Spokane
HaiHBiong
Hammond
Hammond
Hammond
Hammond
Hammond
Hammond
Hammond
Hammond
Hammond
Hammond
Norfolk, Charleston
Riviera Beach, Maryland
Site 30 km away
Site about 30 km away
Site about 30 km away
NASH site
NASN site
NASN site
Site about 30 km away
Site about 30 km away
Site about 30 km away
Average of Pennsylvania and Ohio sices
Average of Pennsylvania- and Ohio sites
Average of Pennsylvania and Ohio sites
                    33

-------
                  Table IV-3 (Concluded)
Plant
Number*
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
BaP
0.6
0.6
0.4
0.5
0.6
0.6
0.8
0.6
0.6
0.8
0.8
0.8
0.8
0.8
0.8
1.6
0.8
0.8
1.6
1.6
0.8
0.8
0.4
0.04
0.04
0.5
0.5
0.5
0.5
0.5
0.7
Remarks**
Average of Pennsylvania and Ohio sites
Average of Pennsylvania and Ohio sites
NASN site
Site about 12 km away
Average of Pennsylvania and Ohio sites
Average of Pennsylvania and Ohio sites
Average of several Pennsylvania basins
Average of Pennsylvania and Ohio sites
Average of Pennsylvania and Ohio sites
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Sites about 10 km away
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Sites about 10 km away
Sites about 10 km away
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Montgomery, Jacksonville, Charleston
Austin and Brownwood
Austin and Brownwood
Sices 20 co 30 km away
Charleston
Charleston
Charleston
Charleston
Hammond
Plane numbers correspond co plant names given in Table III-3.
Cities on locations used for reference concentrations.
                           34

-------
D.    Evaluation of Ambient Concentration Data for Coke Plant Locations

     Available ambient data that were recorded in the vicinity of coke
plants have been evaluated to determine if it is possible to represent the
relationship of concentration mathematically as a function of distance
from a coke plant.  An analysis of the results of the dispersion calcula-
tions performed by EPA (Youngblood, 1977) indicate that such a procedure
should be possible.  An analysis of data given in Appendix B shows that
the BaP atmospheric concentration versus distance relationship about coke
plants can be represented by a power function.  The procedure taken here
is to modify the power function formulation to include allowances for
background concentrations and for coke plant capacities.  The function
selected is as follows:


                          Cd = B + V • A • Db                       (1)


where

         C, is the atmospheric BaP at some distance (D) from the
            coke plant.

          B is the location's nominal background concentration.

          V is the amount of coal processed annually by the coke plant.

          A and b are constants determined by regression.

          0 is the distance from the plant.

Least squares techniques were used to fit the available data to this
function to estimate values for A and b.

     To extrapolate these functional representations from areas where data
are available to areas where data are not available, it is first necessary
to determine if the functional parameters (A and b) are consistent within
the emission control grouping given in Table IV-2.  If they are found to
be consistent within groupings, average values can be used to represent
a group.  The parameter designated as A in Equation (1) relates to the
atmospheric concentration resulting from coke-oven emissions at a distance
of 1 km from the plant.  It could be estimated for more plants than the
slope parameter (b) because of the type of available data.

     'In the first case evaluated, the background concentrations are assumed
to be variable.  For five plants representing the better control classi-
fication group, the A parameters had an average value of 2.8 X 10"6,
whereas for seven plants representing the poorer control group, the average
was 1.2 x 10"5.  Data were insufficent to show a difference in the slope
parameter (b) for the two control groupings.  The average value for eleven
locations was found to be approximately -1.0.  This is consistent with
the dispersion modeling data, which gave values of about -0.9 to -1.0 (it
is also consistent with standard assumptions sometimes used in diffusion


                                   35

-------
equations).  Hence, this analysis suggests that Equation  (1) be used with
a value of -1.0 for the parameter b.  The value of  the parameter A depends
on the grouping in which the plant  is placed.  For  the F  grouping, a value
of 2.8 x 10~6 is used, and for the  K. grouping a value of  1.2 x 10"5 is
used.

     A slightly different formulation was derived for the cases in which
the background was assumed to be constant for all coke plant locations.
The formulation selected is as follows:


                             C, = B' + VADb
                              Q

where B' is  the assumed background  concentration.

     The slope parameter b was again assigned a value of  -1.  The A para-
meter had  an average  value of 3.0 x 10"^  for the better control group  and
1.2 x 10~5 for the poorer control group when a background concentration
of 0.4 was assumed.   Its value was  2.7 x  10~^ for the better control group
and 1.1 x  10~5 for the poorer control group when a  background concentra-
tion of 1.0  was assumed.

     Data from only 12 plant locations were used in the development of the
models because:

     (1)   Sufficient  data had to be available to develop  a concentration-
          versus-distance function, resulting in the 16 locations shown
          in Table B-2 of Appendix  B.
     (2)   The data for Duluth and Johnstown were also eliminated from
          development of the functions because they had much higher
          concentrations than other plants in the better  control grouping.
          Inclusion of these two plants would greatly increase the popula-
          tion exposure estimates.

     (3)   The data for Wayne County and Dormer-Hanna were also excluded
          because they could not be related well enough to specific coke
          plants.

Data for the above locations were, however, used to assist in deriving
the location specific exposures.

     If the plants were not divided into clean and dirty  categories,  the
A parameter would have a value of 1.0 x 10"^ for the variable background
formulation.  This value, which includes data for Johnstown and Monessen,
is almost equal to the A parameter value for the dirty plant model.  When
the data for Johnstown and Monessen are excluded, the combined A parameter
value is 2.9 X 10"^,  which is very  close to the A parameter value for the
clean plant model.
                                   36

-------
E.   Relationship Between BaP and BSO Atmospheric Concentrations

     Because so few data are available for BSO atmospheric concentrations
taken in the vicinity of coke production plants, an analysis has been made
to determine if the BaP data can be used to predict BSO atmospheric con-
centrations, that is, to determine if some mathematical relationship
exists between BaP and BSO concentrations.  Some of the potential hindrances
to establishing this type of relationship are that BaP and BSO are emitted
from other sources besides coke ovens and that the precise relationship
of BSO to BaP for coke-battery emissions is unknown.

     The available BSO concentration data (Appendix A) have been plotted
against the BaP data on Figure IV-1 for sampling sites that collected
both types of data.  Average values were used.  Data sources included
the 1972 NASN urban data, data recorded at sampling sites near coke
plants, and Maryland data.  The data from the various sources appear to
form an increasing function with the cities without coke ovens repre-
senting the lower end of the scale and the data recorded near coke plants
representing the upper scale.  Figure IV-2 is a plot of only more recent
data found near the coke plants.

     Statistical regression techniques were used to fit mathematical
functions to various selected combinations of data given in Figures IV-1
and IV-2.  The functional equation used was of  the  type:


                            BSO = a  • BaPb                          (2)
where
           BSO  is  the  atmospheric  BSO  concentration (p-g/m ).
                                                         *j
           BaP  is  the  atmospheric  BaP  concentration (ng/nr).

           a, b are  constants.

      The  values of  the  constants  were found to be as follows:
                                              Parameter
                         Data Set	
                All  data                     3.80   0.19
                Data for noncoke locations   3.82   0.15
                Data for coke locations      2.15   0.53


 The regression coefficients (R2)  were found to be around 0.4,  indicating
 a less than good fit to the data.   The equations fit to all of the data
 appear to underestimate the BSO concentrations for the higher  BaP con-
 centrations near coke ovens.
                                   37

-------
CO
oo
            40
            10
        I
         I
        8
            1.0
        o
        E
        UJ
           0.1
                     J	L
JJ_L
                                             T    I—I   I  I I  I I
       J	L
                                T	1—I  I  I  I I I
                                                                ALL DATA

                                                                DATA FOR NON-COKE OVEN
                                                                CITIES ONLY
              10
                -1
       J	I	I  I  I  I I I
                                T	1—I  I  I  I I I
10'
101
                                                                                      10'
                                            ATMOSPHERIC B»P CONCENTRATION - ng/m*
                                                                                                          I I I I
                                                                                                               10
                     FIGURE  IV-1.  RELATIONSHIP BETWEEN  BSD AND BiP ATMOSPHERIC CONCENTRATIONS

                                  FOR ALL LOCATIONS

-------
   200
   100
    80
                   I   I   T
                              
-------
     Data collected in  the occupational  environment  should provide  a
better estimate of the  relationship between  BaP and  BSO  emitted  by  coke
ovens than do environmental data;  the concentrations are much higher  and
have not yet been diluted or mixed in the environment.   From June 1971 to
January 1972, Schulte et al. (1975) collected  and analyzed 1,440 airborne
samples from within five coke-oven facilities.  Their raw data are  not
presented; however, they state that the  data collected on the high-volume
filters (similar to those used in ambient sampling)  were fairly  consistent.
The weight of the BSO extract was generally  207.-407.  of the weight of  the
entire sample, and the  weight of the BaP was 1% of the BSO extract.

     Smith (1971) presents BaP and BSO data  for 14 airborne samples
collected on the platform of the larry car inside a  coke-oven battery.
These data are plotted  in Figure IV-3.   Ante11 (1977) supplied BaP  and
BSO data on 12 additional occupational airborne samples  collected during
1974 and 1975.  These data are also plotted  on Figure IV-3.  A statistical
t-test shows that, based on the data given in  Figure IV-3, the assumption
that the BaP is 17. of the BSO cannot be  rejected at  the  0.05 significance
level.  No statistically significant difference was  found between the BaP
versus BSO relationships in the Smith data and in the Ante11 data.

     Differences between the actual BSO  concentrations and predicted  BSO
concentrations (assuming that BaP  is 17.  of BSO) were evaluated.  The  one-
standard deviation between the differences in  the predicted and  observed
concentrations was found to be 537. for individual values and 107. for
the mean.

     The indicated procedure for estimating  BSO exposures is first  to
estimate BaP exposures  due solely  to coke ovens.  The BSO exposures are
made assuming that the  BaP constitutes 17. of the BSO.  This procedure
clearly includes all the error present in estimating BaP exposures, plus
an additional source of error in relating BSO  to BaP.


F.   Population Exposure Estimates

     The estimated population exposures  to coke-oven emissions are  given
in the summary section  of this report and are  not repeated here.  How-
ever, a general discussion of the  approach is  included.

     Resident populations were estimated for seven concentric geographic
rings about each plant. The radii of the rings were taken as 0  to  0.5,
0.5 to 1, 1 to 3, 3 to  7, 7 to 15, 15 to 20, and 20  to 30 km.  These
spacings were selected  to correspond to  the  shape of the concentration
versus distance curves  shown in Appendix B.  Resident population for
each of the geographic  rings was obtained from the Urban Decision Systems,
Inc., Area Scan Report, a computer data  system-that  contains the 1970
census data in city blocks and census enumeration districts and  from  the
SRI BESTPOP computer system.
                                  40

-------
   to
"6
1
I   '
8
   0.5
   0.1
               T    T   I   I   I  I  i  i
                         I   I   I  I  I  I
I      I
                                                       DATA SOURCE
                                                      O  SMITH (1971)
                                                      D  ANTELL (1977)
      ]    1   1 I  1  I 1
                            5         10
                           BaP CONCENTRATION-
             GO
100
          TWO OF SMITHS DATA POINTS CANNOT BE SHOWN ON THIS FIGURE.
          BOTH HAVE BaP-4 pg/m3 AND BSO- 0.0 mg/m3.
 FIGURE IV-3.  RELATIONSHIP BETWEEN BaP AND BSO AIRBORNE CONCENTRATIONS
              FOR OCCUPATIONAL COKE OVEN  LOCATIONS
                                  41

-------
     Average annual BaP concentrations for each geographic ring were
estimated by using the empirical models for those coke plants for which
questionable or no monitoring data were available.  The concentration at
0.4 km was used to estimate exposures for the innermost ring.  For the
other rings, the concentration at the center of each ring was selected
as the exposure concentration for all people residing in that ring.  The
center point is the radius which best represents the average population
exposures in the geographic ring when a uniform ring population density
and decay slope of -1 are assumed.  That is, when
                     2rr                R    2rr
                 C                      2

               /     /  RVAR"1fd9dR =  /     /

              •/R.  Jo                •'R   Jo
solving this equation  gives:
                                  R2
where

          V, A are  as previously defined  in  the exposure equations
        R_, R. are  ring  outside and  inside radii respectively

             f is ring population density
            R  is the radius where the exposures inside and
             c outside R  are equal.

The models were used for 45 of the 65 coke plants.  The plant-specific,
best fit equations  given in Appendix B were  used for  locations for which
sufficient monitoring data were available.   On a few  locations, the
monitoring data were used to fix the concentration at a related distance
from the plant and  the empirical model slope of -1.0  was used to estimate
concentrations at other  distances.   In all,  some monitoring data were used
in making exposure  estimates for 20  of the coke plants.  In addition,
monitoring data were used to estimate the concentration of the outermost
population ring about the Fairfield  plant near Birmingham.  Thus, some
monitoring data were used in estimating the  exposures for 21 coke plants.
Available monitoring data were not used for  an additional seven coke-plant
exposures for the following reasons:  The monitoring  stations were
situated too far from the Woodward,  Thomas,  Fairless  Hills, Alan Wood,
and Chattanooga coke plants to provide useful exposure data for nearby
residential populations.   The Philadelphia monitoring data were not used
because they are reported as having  been  upwind of the plant during collec-
tion and are much lower  than indicated by the NASN data or the Pennsyl-
vania monitoring data.   The Houston  data  were not used because they appear
to be questionably  low for the city, even if it had no coke plant.
According to AISI (1978),  "Houston is a large industrial and transporta-
tion center involved in  petrochemical operations — which are important


                                  42

-------
sources of BaP and benzene solubles--in addition to being a major port."
For locations with more than one coke plant, the exposures were estimated
for overlaps of the geographic rings.  To make these estimates, the geo-
graphic rings about the various coke plants in the area were drawn to
scale.  The BaP concentrations for the overlaps was the sum of the con-
centrations for the various coke plants.  The population of the overlaps
was based on the area contained in the overlaps and the ring populations.

     Concentration subgroups were then developed, based on the range of
concentrations for the estimated exposures, and the total number of
residents for each exposure subgroup were calculated.  The population
residing within a subgroup was excluded if its average annual BaP con-
centration due only to coke-oven emissions was less than 0.2 ng/m3.
These subgroupings were made for exposures to coke-oven emissions only
and to coke-oven emissions plus background concentrations.

     Population exposures to BSO were calculated using the procedure
given in Section IV-E.  This procedure estimates BSO exposures based on
estimated BaP exposures.
                                  43

-------
                  Appendix A




AMBIENT ATMOSPHERIC BaP AND BSO CONCENTRATIONS
                      45

-------
                              Appendix A

            AMBIENT ATMOSPHERIC BaP AND BSO CONCENTRATIONS
A.   General
     This appendix presents BaP and BSO atmospheric concentration data
recorded in the vicinities of coke manufacturing plants.  Data are also
presented that give background concentrations for locations that contain
and do not contain coke ovens.  Most of the data given is based on 24-hr
samples; however, some are based on monthly or quarterly composites.   Un-
less otherwise specified,  sample size refers to the number of 24-hr days
of data available.  All data used in this report are based on high-volume
filter samples.  In addition, many of the sampling programs were con-
ducted over a relatively few days within 1 or 2 consecutive months; thus,
they may not be entirely representative of an area's average annual con-
centration.   The implications of this sampling approach in estimating
population exposures is described in further detail in Appendix B.
B.   Atmospheric BaP and BSO Concentration Data Recorded Near Coke
     Manufacturers

     Atmospheric data that have been recorded near coke manufacturers are
described in the following paragraphs.
     1.   Monessen Area Air Quality Study, Pennsylvania

          The Pennsylvania Department of Environmental Resources conducted
an air quality study to determine the distribution and magnitude of total
suspended particulates (TSP), benzene soluble organics (BSO), and benzo-
(a)pyrene (BaP) concentrations in the Monessen area.   The impact and ex-
tent of air pollution due to sources at the Wheeling-Pittsburgh Steel
Corporation, Monessen, were evaluated, with sampling conducted from April
6 to June 21, 1976, at three sites near the steel plant.   Meteorological
and selective sector actuator techniques were included in the sampling
program (DER, 1977A).

          A statistical summary of the data for the three sites is given
in Table A-l.  The average TSP concentrations ranged from 79 to 166 ^g/m^,
average BaP concentrations from 2.7 to 40.8 ng/m^, and the average BSO
from 2.6 to 9.2 p,g/m3.  Selective sector actuator sampling and a
concentration-wind direction frequency weighting technique all confirmed
that the steel plant is the major source of TSP and BaP.   The average
concentrations found in areas in the direction of winds coming from the
                                   47

-------
                               Table A-l

          MONESSEN AIR STUDY, 24-HOUR SAMPLE CHARACTERISTICS
                                                      Geometric
    TSP
      Station 2
      Station 6
      Station 7
    BaP (ng/m3)

      Station 2
      Station 6
      Station 7

    BSO (tig/m3)

      Station 2
      Station 6
      Station 7
                  Sample
                   Size
29
28
31
29
28
31
29
28
31
       Average
166.0
 79.0
113.0
 40.8
  2.7
 22.8
  9.2
  3.3
  2.6
           Range
27.0-360.0
22.0-165.0
26.0-300.0
 0.3-206.4
 0.2-10.8
 0.4-100.3
 1.5-25.4
 0.6-9.1
 0.9-19.3
             Mean
145.0
 71.0
 93.0
 10.0
  1.6
 10.1
  6.5
  2.6
  3.8
        Standard
        Deviation
1.
1.
1.
76
64
91
7.60
2.78
4.57
2.34
2.01
2.36
    Station 2 is 1 km ESE of the coke ovens.
    Station 6 is 2.1 km NW of the coke ovens.
    Station 7 is 1.8 km ENE of the coke ovens,

    Source:  DER (1977A)
plant are between 1. 5 and 3_ times the average concentrations for winds
from all other directions (DER, 1977A).
     2.   Allegheny County. Pennsylvania

          Three coke batteries are located in Allegheny County:  U.S.
Steel Corporation in Clairton, Jones and Laughlin in Hazelwood, and
Shenango, Inc. on Nevell Island.  From April to September 1976, high-
volume particulate samples taken from 11 sites were analyzed for BaP.
The sampling schedule included two 10-week periods of four and two samples
per week, respectively (Ek, 1977).

          Table A-2 shows  the results obtained 'during the sampling.  The
average BaP concentrations for the 11 locations varied between 1.64 and
51.95 ng/m3.  Eight additional samples were collected during first-stage
alerts at Liberty Borough  in April and June 1976.  Four were collected
over 24 hours and four over 8 to 12 hours.  These data which are given
                                  48

-------
\o
                                                      Table A-2
                          AMBIENT BaP CONCENTRATIONS FOR ALLEGHENY COUNTY, PENNSYLVANIA
                                                     (ng/m3)
Site
Number
7102
5702
8601
8704
5802
7570
8602
6903
8790
5602
7004
Site Location*
10.5 km N of USS, 8 km E of J&L, 21.5 km SE of S
18 km NW of USS, 4.5 km N of J&L, 12 km ESE of S
0.5 km SE of USS, 14 km SE of J&L, 28 km SE of S
2.0 km NE of USS, 12 km SE of J&L, 27 km SE of S
18 km NW of USS, 6 km NW of J&L, 10 km SW of S
8.5 km NNE of USS, 9.5 km ESE of J&L, 24 km SE of S
1.5 km NNW of USS, 12 km SE of J&L, 26 km SE of S
12 km NW of USS, 1 km SSE of J&L, 16 km SE of S
2 km NE of USS, 13 km SE of J&L, 27.5 km SE of S
16 km NNW of USS, 5 km NNE of J&L, 15 km ESE of S
12.5 km N of USS, 6 km E of J&L, 18.5 km SE of S
Sample
Size
2
2
6
5
2
5
4
10
20
2
2
Average
1.64
2.62
13.63
15.00
2.29
6.12
28.17
3.95
51.95
3.78
1.66
Range
0.2- 3.1
0.3- 4.9
0.9- 67.7
0.3- 40.4
1.8- 2.8
0.9- 20.1
2.8- 83.5
0.5- 19.0
0.4-310.0
3.1- 4.5
1.4- 1.9
         USS  is  U.S.  Steel,  J&L is  Jones and Laughlin,  and S is Shenango.
         Source:   Ek  (1977)

-------
in Table A-3, show average BaP concentrations about six times higher than
for the regular sampling given in Table A-2.


                               Table A-3

            BaP DATA OBTAINED DURING FIRST STAGE ALERTS AT
                      LIBERTY BOROUGH—SITE 8790
                               (ng/m3)

        Sample
        Number              24-Hour Data         8-12 Hour Data

          1                    427.9                 405.8

          2                    277.8                 458.8

          3                    320.4                 189.8

          4                    171.0                 155.6

        Average                299.3                 302.3


     3.   Geneva Works. Utah

          The data collected during October and November 1976 for BaP
concentrations near the U.S. Steel Geneva Works located near Provo, Utah,
are summarized in Table A-4.  Eight stations within 4 km of the coke
batteries showed average BaP concentrations of 1.47 to 3.81 ng/m^.  Two
background stations 20 to 30 km away showed average BaP concentrations
of 0.12 and 0.83 ng/m^.


     4.   Wayne County. Michigan

          Three companies operating coke batteries are located at Wayne
County, Michigan:  Solvay, Ford, and Great Lakes Steel.  Ambient atmo-
spheric BaP concentration data were reported annually for seven sites in
the general area for 1971 to 1975 and are given in Table A-5.   Annual BaP
concentrations for the various sites varied between 0.34 to 14.72 ng/m^.


     5.   Buffalo. New York

          Three companies operate coke batteries near Buffalo, New York:
Semet-Solvay, Bethlehem Steel, and Donner-Hanna.-  Atmospheric BaP con-
centration data were recorded from 1973 to 1974 on 13 sites, in addition
to data recorded at the National Air Surveillance Network (NASN) site.
These data, which are given in Table A-6, indicate the average BaP con-
centrations ranged from 0.45 to 27.10 ng/nP.
                                  50

-------
                               Table A-4

               ATMOSPHERIC BaP CONCENTRATIONS NEAR THE
                         GENEVA WORKS IN UTAH
                               (ng/m3)
Scacion
Number
1
2
3
4
5
6
7
8
9
10
Location In Relation
to Battery
2.0 km NW
2.7 km NW
2. it km NW
1.8 km N
1.3 km NE
2.4 km SE
4.0 km NW
2.6 km S
30.0 km S
20.0 km N
Sample
Size
9
6
9
11
11
11
3
11
11
9

Average
2.08
3.81
3.15
2.41
3.13
1.63
2.10
1.47
0.12
0.83

Range
0.40-4.42
2.52-5.27
0.97-6.30
0.44-5.85
0.54-6.29
0.46-3.44
0.87-3.53
0.38-3.35
0.01-0.32
0.05-2.77
                               Table A-5

              AMBIENT BaP CONCENTRATIONS FOR WAYNE COUNTY,
                               MICHIGAN
                                (ng/m3)
Sice
Number 1971







Mo.
No.
No.
No.
No.
02
04
05
06
08
11
NASN
2 is
4 is
5 is
6 is
8 is
3
2
9
3
2
1
1
14 km NE
.00
.97
.32
.62
.39
.30
.40
of
1972
2.44
3.14
5.95
2.62
2.56
1.32
1.90
Solvay, 14
7.2 km NNE of Solvay,
1.6 km N
15.3 km
10.5 km
of
NNW
Solvay, 4
of Solvay,
SW of Solvay,
1973
3.02
4.16
11.78
3.12
2.70
2.00
1.00
.5 km NE
9.3 km NE
km NNE of
16.5 km
1974
1.
1.
10.
0.
0.
0.

of G.
of G
G.L.
46
70
83
52
44
34

L.*;
.L.,
, and
NNW of G.L
8.5 km SW of G
.L. ,
1975
3.
4.
14.
1.
2.
0.
1.
and
and
4.4
43
85
72
47
54
73
00
18
Average
2.
3.
10.
2.
2.
1.
1.
km ENE
9 km NE of
km
. , and
and
9.3
67
36
52
27
13
14
33
of Ford
Ford.
E of Ford.
11.7 km
km SSW
NNW of Ford
of Ford.
No. 11 is 30 km SW of Solvay, 29 km SW of G.L., and 30 km  SW  of  Ford.
 G.L. - Great Lakes Steel.
                                   51

-------
                                                   Table A-6


                                AMBIENT BaP CONCENTRATIONS FOR BUFFALO, NEW YORK
Oi
ro
Site
Number
1
2
3
tt
5
6
7
8
9
10
11
12
13
NASN
Site Location
3.1 km E of Beth, and 3.4 km SE of D-H*
1.9 km ESE of Beth, and 3.8 km S of D-H
3.8 km NE of Beth, and 1.5 km SE of D-H
1 km ENE of Beth, and 2.8 km S of D-H
3 km N of Beth, and 1.4 km WNW of D-H
4.3 km NE of Beth, and 1.1 km ENE of D-H
5.6 km NE of Beth, and 2.1 km NNE of D-H
0.6 km W of Allied
1.2 km ESE of Allied
2.4 km NE of Allied
30 km SW of Beth, and 34 km SU of D-H
3.2 km NW of Allied
4 km SW of Beth, and 6.2 km S of D-H
8.8 km NNW of Beth, and 5.6 km N of D-H
Sample
Size
37
81
48
7
78
65
41
73
76
44
44
28
7
Average
5.99
8.99
11.38
27.10
2.78
9.10
7.29
1.29
3.80
3.74
0.82
0.45
1.29
0.70**
Range
0.27-30.5
0.26-48.7
0.06-68.4
2.76-48.8
0.05-23.8
0.20-65.6
0.07-46.2
0.05-21.2
0.13-81.4
0.01- 9.3
0.04-24.5
0.06- 3.1
0.40- 3.7
     **
 Beth.  Is Bethlehem Steel; D-H Is Donner-llanna.
*
 Two-year composite

-------
     6.   Duluth. Minnesota

          Thirty-eighc samples for ambient BaP concentrations were ob-
tained during the period of January 1974 to November 1975 from two sites
within 3 km of the U.S. Steel coke batteries in Duluth, Minnesota.  These
data are summarized in Table A-7.  Average BaP concentrations of 0.22 and
1.45 ng/m3 were found for the two sites.  When most of the samples were
collected, the wind was blowing  in the general direction of the collec-
tion sites from the plant.


                               Table A-7

           AMBIENT BaP CONCENTRATIONS FOR DULUTH, MINNESOTA
                                 (ng/m3)


   Site                                  Sample
  Number     Distance from Coke Ovens     Size   Average      Range

    1            2.1 km SW                 18     1.45      BDM-7.02*

    2            2.7 km N                  20     0.22      BDM-1.25
BDM - below detectable minimum.

Source:  Jungers (1977A).




     7.   Gadsden, Alabama

          The Republic Steel Corporation operates coke ovens in Gadsden,
Alabama.  Atmospheric BaP concentrations were sampled at two sites within
1.6 km of the coke ovens during 1974, 1975, and 1976.  The data from this
sampling which are summarized in Table A-8, indicate the annual atmo-
spheric BaP concentrations varied from 0.44 to 5.06 ng/m^.


     8.   Birmingham, Alabama, Area

          Five coke battery facilities, which are within about 20 km of
Birmingham, are located at Tarrant, Woodward, Thomas, Birmingham, and
Fairfield.  Atmospheric BaP concentrations were sampled at Tarrant and
Fairfield during 1976, and NASN data are available for Birmingham.  These
data are given in Table A-9.  The average BaP concentrations ranged from
2.5 to 4.5 ng/m^.  BaP data were also recorded for five CHAMP sites in
the Birmingham area.  These data are given in Table A-10.
                                   53

-------
                               Table A-8
   Sice
  Number

    1

    2

  NASN
            AMBIENT BaP CONCENTRATIONS FOR GADSDEN, ALABAMA
                                (ng/m3)
                                              Average Concentrations*
Distance from Coke Ovens
   1974
 1975  1976   3-Year
     1.6 km E

     1.1 km SW

  Same as Station 1
5.06(0.44)**  0.75  0.58 2.13(0.60f*

   0.97      0.44  1.89    1.10

   0.50      0.60   —     0.55
 *
  Sample size for each year for Sites 1 and 2 was 5 days.
**                                           3
  Excludes one high observation of 23.55 ng/m .

 Source:  Jungers (1977B).
                                 Table A-9

            AMBIENT BaP CONCENTRATIONS FOR BIRMINGHAM, ALABAMA
                                 (ng/m3)
     Site
    Number

      1

      2

     NASN*
Distance from Coke Plants

Tarrant (0.5 km NW)

Fairfield (0.5 km ESE)
   Sample
    Size

     2

     3
Average
Range
 4.46    0.06-8.86

 2.79    1.10-5.31

 2.50
    1974 sample composite.

   Source:  Jungers  (1977B).
                                   54

-------
                              Table A-10

           CHAMP SITE AMBIENT ATMOSPHERIC BaP DATA  FOR  THE
                     BIRMINGHAM AREA  (1975 Data)
Site
Number
304
305
306
307
323
331
Distance 1
Fair field
11.4
25.8
10.3
4.9
16.4
13.4
                            3.8

                            5.0

                           10.8

                           18.4

                            2.4

                            3.2
 2.4

13.0

13.4

21.4

 1.9

 5.4
Sample
 Size*

   6

  12

  12

   6

  12

  12
                                                          BaP  (ng/ni)
Average
4.2
1.8
1.5
2.4
2.9
3.5
Range
0.7-9.2
0.6-3.7
0.4-4.0
0.9-4.3
1.2-6.6
1.4-5.5
   Number of months  for which data are available;  for
   individual months data were  generally  collected for
   25  to 31 days.
     9.   Johnstown Air Basin. Pennsylvania

          Two coke plants are located near Johnstown, Pennsylvania
(Bethlehem Steel's Franklin and Rosedale Divisions).  An air quality
study was conducted from August through November 1975 to determine the
distribution and magnitude of TSP, BSO, and BaP concentrations in the
Johnstown area.  Concentration data were obtained for eight sampling
sites 0.6 to 7.8 km from the Franklin Works (Table A-ll).  BaP concen-
trations ranged from 85.3 ng/m3 for the site nearest the Franklin plant
to 3.6 ng/m3 for the site farthest from the plant.

          Wind-actuated sampling was also conducted for TSP, BSO, and
BaP.  For all three, in-sector sample concentrations were almost double
the out-sector concentrations.
     10.  Philadelphia. Pennsylvania

          One coke facility is situated in Philadelphia (Philadelphia
Coke Division),  and another two coke facilities are within 12 km of the
city at Alan Wood and Fairless Hills.   Air quality data were collected
at four different times from November 1976 to January 1977 to determine
                                   55

-------
Ul
                                                    Table A-ll


                     AMBIENT  BaP,  BSO,  AND TSP  CONCENTRATIONS FOR JOHNSTOWN,  PENNSYLVANIA
Site
Number
1
2
3
4
5
6
7
8
Distance from Franklin
Coke Ovens
7.8 km WSW
3.8 km W
2.9 km SU
1.0 km NNE
4.6 km SSW
3.4 km SSW
0.6 km ESE
1.9 km SE
Number of
Samples
30
32
33
32
28
31
34
31
BaP
Average
3.6
13.8
7.7
23.4
6.0
6.8
85.3
19.9
(ng/raj)
Range
0.5- 15.4
2.0-110.9
0.9- 41.8
3.6-246.6
1.5- 11.0
1.4- 24.5
1.5-575.9
1.2-102.9
Average
TSP*
32
70
71
142
55
58
179
70
(pg/mj)
BSO
2.2
5.5
5.4
9.7
3.9
4.1
14.1
5.6
      *
       Geometric  mean.
        Source:   DER (1977B),

-------
the distribution and magnitude of TSP, BSO, and BaP in Philadelphia.
Concentration data were obtained for 13 sampling stations about 2 to 14
km from Philadelphia Coke Division.  These data are summarized in Table
A-12.  The average BaP concentrations for the 13 sampling sites ranged
from 0.97 to 4.70 ng/tn^.  BSO average concentrations ranged from 3.05 to
8.56
     11.  Granite City. Illinois

          BaP was measured at eight sampling sites during March to May
1975 between 0.5 to 3.5 km from the National Steel coke ovens in Granite
City, Illinois.  The data obtained during this sampling, which are sum-
marized in Table A-13, indicate that average atmospheric BaP concentra-
tions for the stations ranged from 2.6 to 12.2 ng/m^.

          More recently TSP,  BSO, and BaP data have been obtained for 3
days on two sites within 0.8 km of the coke batteries (Table A-14).  BaP
measurements from individual observations ranged from 1.6 to 278 ng/m^.
     12.  Houston, Texas

          Atmospheric BaP samples were obtained from seven sites located
up to 5.5 km from the Armco Steel coke ovens situated in Houston, Texas.
Samples were recorded at various times from 1973 to 1976.  The data sum-
marized in Table A-15 show that average concentrations by site varied
from 0.03 to 0.28 ng/ra^.  These concentrations are much lower than those
recorded at similar distances from other coke-oven locations, perhaps
indicating any of the following factors:

          •  Good emission control.

          •  Faulty measurement techniques.

          •  All samples recorded upwind.

          •  Ovens not in operation when measurements were recorded.
     13.  Cleveland. Ohio

          King et al. (1976) have reported on atmospheric BaP concentra-
tions for a number of sites in Cleveland, Ohio.   These data are summarized
in Table A-16.  The geometric means are given rather than the arithmetic
means.
     14.  Sparrows Point, Maryland

          The Maryland State Division of Air Quality Control measures
ambient BaP and BSO concentrations for many sites within the state, four
of which are located within approximately 12 km of the coke batteries.
Data for 1976 are given in Table A-17.
                                   57

-------
                                                    Table A-12
                     AMBIENT BaP, BSO, AMD TSP CONCENTRATIONS FOR PHILADELPHIA, PENNSYLVANIA
Ul
00
 Site     Location from Philadelphia     Number of
Number    	Coke Ovens	      Samples
   1             14.1 km SW                  4
   2             3.5 km SW                   3
   3             19 km SW                    4
   4             12.5 km WSW                 4
   5             13.7 km WNW                 4
   6             9.3 km NNE                  4
   7             5.8 km W                    4
   8             10 km SW                    4
   9             2 km WNW                    4
  10             8.8 km SW                   4
  11             13 km SSW                   3
  12             10 km SW                    3
  13             5.2 km SW                   4
 NASN
BaP
Average
3.82
1.61
2.27
0.97
1.34
2.54
4.24
1.96
2.78
4.42
3.68
4.70
4.10
2.10
(ng/m3)
Range
2.09-8.81
0.82-2.26
1.09-5.35
0.21-1.81
0.35-2.31
1.38-4.53
1.90-6.29
1.38-2.31
1.26-4.46
2.28-7.15
1.55-6.70
2.44-8.06
1.57-9.90
_
Average
TSP
76.5
130.5
102.5
44.8
36.3
48.0
85.5
60.3
60.0
109.7
133.3
95.0
102.3
_
(Mg/m3)
BSO
5.44
A. 00
5.75
3.05
3.11
4.15
7.22
4.77
6.42
7.76
8.05
8.56
6.59
4.66
       Source:   Lazenka (1977).

-------
                                 Table A-13
            AMBIENT BaP CONCENTRATIONS FOR GRANITE CITY, ILLINOIS
                                   (ng/m3)
  Sice Number   Distance from Coke Ovens   Sample Size   Averag
                                              Range
      NW2
      008
      006
      007
      009
      010
      Oil
 0.7 km N
 0.6 km SSW
 1.1 km NE
 2.4 km WNW
 1.8 km NW
 1.5 km W
 3.5 km WNW
 2.9 km WSW
3
3
2
2
2
1
2
2
• 
-------
                              Table A-16
            AMBIENT BaP CONCENTRATIONS FOR CLEVELAND, OHIO
                              Sample Size
                                  21
                                  37
                                  23
                                  28
                                  22
                                  38
                                  28
                                  30
                                  33
                                  32
                                  23
                                  22
                                  21
                                  32
                                  19
                                  22
Sice Number
1
3
4
5
6
7
8
9
10
12
13
14
15
17
20
21
Location from
Coke Battery*
0.8 km N
4.8 km SW
4.4 km NE
4.4 km SE
12.0 km NE
7.2 km W
6.8 km SW
1.0 km SE
6.0 km NNE
13.2 km ESE
4.4 km S
9.6 km SE
3.2 km W
6.4 km NE
16.8 km NE
4.0 km NNW
BaP
Geometric Mean**
1.40
0.62
0.64
0.58
0.71
0.46
0.44
3.60
0.74
0.43
0.85
0.47
0.51
0.91
0.50
1.10
Maximum
41.0
3.1
15.0
3.3
3.0
2.1
2.3
130.0
7.2
2.0
14.0
3.7
3.5
49.0
6.9
17.0
  Locations are only approximate.
**
  Arithmetic mean not reported.
Source:  King ec al. (1976)
                              Table A-17
            AMBIENT ATMOSPHERIC BaP AND BSO CONCENTRATIONS
                     FOR SPARROWS POINT, MARYLAND
                                         BaP
                                                             BSO
 Distance from
Coke Batteries*
12 km N
7 km NNW
3 km W
4 km SSW
Sample Size
2
12
10
10
Average
1.4
1.4
1.9
2.4
Range
1.1-1.7
0.2-4.4
0.1-2.6
0.4-5.4
Average
6.5
5.4
--
4.8
Range
6.3-6.8
3.0-8.0
--
2.6-8.7
  Locations are only approximate.
  Number of months for which data are available.
                                  60

-------
     15.  Chattanooga. Tennessee

          As part of the CHESS and CHAMP programs, BaP samples were col-
lected for nine sites in the Chattanooga area.  These data for 1975 are
summarized in Table A-18.
                              Table A-18

                  AMBIENT BaP AND BSO CONCENTRATIONS
                      FOR CHATTANOOGA, TENNESSEE
Site
Number
621
622
631
632
633
634
635
641
642
Distance from
Coke Ovens
7.6 km
8.9 km
20.2 km
15.2 km
16.4 km
23.8 km
14.3 km
13.0 km
15.1 km
Number of
Samples*
12
12
12
12
12
12
12
10
12
BaP
(ns/m3)
Mean
3.83
3.49
1.63
1.85
1.55
0.82
1.23
2.35
2.66
Range
1.0-8.4
0.4-8.5
0.2-3.6
0.2-5.9
0.1-4.2
0.0-2.7
0.1-3.0
0.2-8.6
0.2-5.6
BSO
dig/m3)
Mean
3.69
4.51
3.04
2.93
2.33
1.73
2.66
3.26
3.60
Range
2.2-6.5
2.6-9.1
1.4-5.2
1.1-6.4
0.6-4.6
0.3-2.7
1.5-4.1
1.8-5.3
1.5-7.0
  Number of months for which data are available; for individual months
  data were generally collected for 20 to 31 days.
C.   Ambient Background BaP and BSO Concentration Data

     Because coke ovens are not the only sources of BaP and BSO concen-
trations in the atmosphere, the coke-oven contributions must be placed in
perspective with each area's nominal background concentrations.  Data are
presented here for ambient background concentrations measured in cities
in which coke ovens are located, cities without coke ovens, and remote
rural areas.
     1.   NASN Air Quality System Data

          NASN routinely monitors suspended particulate concentration
levels in urban and nonurban areas,  generally reporting them as quarterly


                                   61

-------
composites for stations  in  the network.  The composite, which pools all
samples collected during the quarter, assists in generating sufficient
material for laboratory analysis.

          Before 1971, BaP  analysis was made for more than 120 sites per
year.  For 1971 and subsequent years, the sites were limited to 40 be-
cause of time and resource  restrictions.  These 40 sites were selected
to update BaP concentrations in cities with and without coke ovens.  Three
sites were selected in National Parks to provide nonurban background
readings (U.S. EPA, 1974).

          Annual average BaP concentrations for 1967 to 1976 are given in
Table A-19 for the 40 NASN  sites.  Table A-20 gives BSO data recorded at
these sites for 1971 and 1972.  The BaP and BSO concentrations are sum-
marized in Table A-21.  The BaP concentrations are generally less than
0.1 ng/m-* for rural locations.  Most urban locations without coke ovens
have average concentrations of less than 1 ng/m^ (the average is 0.38
ng/m3); however, areas with coke ovens generally have average concentra-
tions in excess of 1 ng/rn-^  with Ashland's 4.7 ng/rn-^ the highest and
Dearborn's 3.1 ng/m^ the next highest.  Coke ovens are located in both
Ashland and Dearborn.  The  overall average for cities with coke ovens is
1.21 ng/m-3.  The concentrations for coke-oven cities are significantly
higher than for the noncoke-oven cities at the 0.01 significance level.

          The BSO concentrations were generally less than 5 p,g/nr*.  The
average concentrations of most urban locations range from 1 to 4 ng/m^.
Ashland, Chattanooga, Pittsburgh, Buffalo, and Hammond have concentra-
tions exceeding 5 [ig/m^.  The concentrations for the coke-oven and
noncoke-oven cities are not significantly different at the 0.05 level.

          Table A-22 shows  the change from 1966 to 1976 in BaP concentra-
tions in the atmosphere for cities with and without coke ovens.  Both
classes of cities have shown a reduction; however, the average difference
between the two types of cities has been fairly constant since 1968.  The
decrease in concentrations  is statistically significant at the 0.01 level
for both coke-oven and noncoke-oven cities.  Both coke- and noncoke-oven
cities have shown an average BaP reduction of 9% per year over the last
9 years.
     2.   Pennsylvania Air Quality System

          The Pennsylvania Division of Technical Services and Monitoring,
Bureau of Air Quality and Noise Control has systematically surveyed air
quality since 1970.  As part of this program, the division monitors sus-
pended and settleable particulates at 91 locations.  Suspended particu-
lates are collected on a glass fiber filter with a high-volume air sampler.
Each sample represents the particulate matter filtered from approximately
2000 m-3 of air over 24 hours.  Samples are taken from midnight to mid-
night every 6 days (DER, 1977).
                                   62

-------
                           Table A-19

ANNUAL AVERAGE AMBIENT BaP CONCENTRATIONS AT NASN URBAN STATIONS
                             (ng/m3)
Location
Montgomery, AL
Chicago, IL
Detroit, MN
New York, NY
Toledo, OH
Philadelphia, PA
Pittsburgh, PA
Shenandoah Park, VA
Charleston, WV
Grand Canyon, AZ
i
Gadsden, AL
Gary, IN
Indianapolis, IN
Baltimore, MD
Trenton, NJ
St. Louis, MO
Youngs town, OH
Chattanooga, TN
Spokane, WA
1967
2
3
5
3
1
5
7
0
-
0

-
-
5
3
-
2
8
22
-
.3
.0
.4
.9
.9
.9
.0
.3
-
.2

-
-
.7
.8
-
.3
.2
.9
-
1968
2.9
3.1
5.1
--
1.8
2.9
6.3
0.3
4.6
0.2

2.4
--
4.1
2.3
1.4
--
5.6
7.4
--
1969
2.0
3.9
3.9
3.6
1.5
4.0
13.8
0.3
2.6
0.2

1.8
--
5.2
2.8
1.6
3.3
9.9
4.2
--
1970
1.3
2.0
2.6
3.0
1.4
2.4
5.9
0.2
2.1
0.1

2.5
--
2.3
2.1
0.8
--
7.1
5.5
--
1971
0.5
2.5
1.4
2.3
0.8
2.3
6.1
--
0.9
--

1.2
1.6
0.9
2.8
0.7
0.8
3.7
--
1.7
1972
0.5
1.3
1.9
1.8
0.4
0.9
10.6
0.1
0.7
--

1.2
1.2
4.9
1.3
0.5
0.6
3.2
9.9
1.5
1973
0.3
0.4
1.0
0.7
0.4
0.6
--
0.1
0.2
>0.1

0.8
0.3
0.4
0.4
0.1
0.2
1.1
--
0.4
1974
0.4
--
--
0.9
0.2
0.8
1.3
--
0.5
>0.1

0.5
0.5
--
0.5
--
0.3
1.9
--
--
1975
0.3
1.0
1.0
0.9
0.4
0.6
2.1
--
0.5
—

0.6
2.2
--
0.6
--
0.3
2.1
0.8
0.6
1976
0.3
0.5
1.1
0.5
0.5
1.0
2.0
O.I
0.4
0.05

0.6
0.7
0.6
0.5
0.3
0.3
1.5
0.7
1.1

-------
Table A-19 (Concluded)
Location
Milwaukee, WI
Birmingham, AL
Jacksonville, FL
Honolulu, HI
Terre Haute, IN
Ashland, KY
Baton Rouge, LA
New Orleans, LA
Dearborn, MI
Duluth, MN
Buffalo, NY
Cleveland, OH
Bethlehem, PA
Erie, PA
Houston, TX
Newport News, RI
Norfolk, VA
Seattle, WA
St. Paul, MN
Arcadia National Park,
ME
Hammond, IN
1967
--
--
--
0.5
3.7
--
--
1.8
--
--
--
2.9
2.9
--
--
--
3.5
1.8
2.3
	
2.5
1968
4.7
--
2.9
0.6
--
9.3
--
1.6
--
2.7
--
3.0
2.1
--
--
--
4.9
2.0
1.8
0.3
2.1
1969
4.0
--
2.3
0.6
4.0
10.9
--
1.5
--
2.1
--
3.8
2.0
--
—
—
3.9
1.6
1.8
0.1
3.3
1970
2.5
--
1.4
0.2
2.8
6.7
--
1.1
--
l.l
--
2.8
2.7
--
1.2
--
1.8
1.5
1.0
0.2
1.7
1971
1.8
4.0
2.2
0.2
--
9.0
0.4
0.9
--
4.8
—
--
0.9
1.5
0.5
0.4
1.2
0.5
0.5
— _
3.8
1972
3.6
2.3
0.4
0.1
1.1
8.5
0.2
0.4
0.6
19.1
1.5
1.3
0.8
2.4
0.4
0.3
0.6
0.5
0.5
0.3
1.4
1973
0.6
1.5
0.2
0.2
--
2.9
0.1
0.3
1.0
0.3
0.6
--
0.5
0.7
0.4
0.2
0.4
0.3
0.1
_ —
0.2
1974
-
2
-
0
0
-
0
0
1
0
0
-
0
0
0
-
0
-
0
0
0
-
.5
-
.4
.3
-
.1
.3
.7
.2
.8
-
.1
.6
.2
-
.2
-
.5
.1
.4
1975
1.1
—
0.4
0.03
0.6
4.7
0.1
0.2
3.1
0.3
0.5
--
--
0.4
0.2
--
0.2
0.4
0.4
0.1
0.7
1976
0.3
1.6
0.3
0.02
0.6
4.7
0.3
0.2
--
0.2
--
0.6
0.3
0.2
--
--
0.5
0.7
0.5
0.09
0.5

-------
                                                  Table A-20
o>
Ul
Birmingham, AL
Gadsden, AL
Montgomery, AL
Grand Canyon, AZ
Jacksonville, FL
Honolulu, HI
Chicago, IL
Gary, IN
Hammond, IN
Indlanopolls, IN
Terre Haute, IN
Ashland, KY
Batan Rouge, LA
New Orleans, LA
Baltimore, MD
Dearborn, MI
Detroit, MI
Trenton, NJ
Duluth, MM
M VARIATIONS
OF BENZENE SOLUBLE ORGANIC SUBSTANCES (pg/ra3)
1971
1
3.1
2.9
3.4
1.2
4.3
2.3
4.3
4.7
3.8
2.6
4.1
6.8
2 £
4.0
7.3
-
2.6
1.7
1.8
2
6.7
3.6
4.2
0.9
3.0
0.1
5.7
2.7
4.7
3.1

7.4
1.9
3.5
4.5
3.2
3.0
3.0
2.5
3
-
2.1
2.4
-
2.1
1.2
-
-
6.0
3.2
3.6
4.0
-
3.1
-
3.1
2.4
2.6
2.1
4
4.8
4.5
3.4
-
2.2
1.4
4.5
5.7
7.0
3.7

8.3
3.4
3.5
4.3
-
3.9
2.8
3.6
1
3.6
2.7
3.3
-
2.3
1.4
3.4
2.7
2.1
2.9
2.5
7.8
3.2
3.7
5.0
3.6
3.2
1.8
2.0
1972
2
7.5
4.2
2.9
-
5.4
2.3
2.5
4.1
9.4
4.9
5.7
7.2
4.1
4.9
3.6
7.3
3.3
1.7
5.9
3
4.0
2.4
2.2
-
4.4
3.3
2.7
3.0
6.3
3.0
4.0
7.9
3.5
5.5
-
4.6
3.4
2.0
4.5
4
5.2
2.3
2.6
-
6.0
3.0
3.9
2.5
5.0
-
6.3
9.2
5.3
4.2
4.5
4.5
3.0
1.5
12.5
Average
4.99
3.09
3.05
1.05
3.71
1.88
3.86
3.63
5.54
3.34
4.37
7.33
3.43
4.05
4.87
4.38
3.10
2.14
4.36

-------
                                             Table A-20 (Concluded)
                                          1971
                                                                     1972
en
St. Paul, m
St. Louis, MO
Buffalo, NY
New York, NY
Cleveland, OH
Toledo, OH
Youngstown, OH
Bethlehem, PA
Erie, PA
Philadelphia, PA
Chattanooga, TN
Houston, TX
Newport News, VA
Norfolk, VA
Shennandoah, VA
Pittsburgh, PA
Seattle, WA
Spokane, WA
Charleston, WV
Milwaukee. UI
1
-
5.5
-
5.5
3.6
2.1
2.9
3.8
2.6
6.0
4.8
3.5
2.7
4.9
-
3.8
5.6
3.5
-
-
2
2.8
3.1
-
6.2
3.6
2.4
4.9
3.8
2.5
4.0
5.1
3.2
2.9
4.0
0.7
4.4
4.1
4.4
5.0
3.8
3
2.2
2.0
2.9
-
-
-
3.5
3.6
-
3.8
-
3.8
3.1
3.5
0.9
-
5.3
3.5
2.0
3.8
4
3.0
-
3.6
4.5
-
3.1
6.4
3.8
4.2
7.4
-
5.7
4.5
4.2
-
6.9
5.4
3.6
3.6
4.8
1
2.9
2.9
3.0
4.5
3.1
1.8
4.1
2.9
1.3
4.7
4.1
4.5
1.3
2.5
1.0
6.1
1.6
3.8
2.8
3.0
2
7.9
3.7
9.3
5.3
6.5
-
3.9
5.2
6.8
3.8
11.0
5.9
3.3
3.8
0.9
4.9
4.2
3.1
2.7
6.5
3
5.6
2.3
3.4
4.7
-
2.9
4.2
4.2
5.8
4.3
3.7
5.9
4.4
3.8
0.8
4.7
3.9
4.6
3.6
3.2
4
4.6
2.6
7.8
3.9
4.6
2.7
4.6
4.6
-
3.3
-
4.9
3.7
3.2
0.6
6.6
5.4
4.0
2.8
3.4
Average
4.14
3.16
5.00
4.94
4.28
2.50
4.31
3.99
3.87
4.66
5.74
4.68
3.24
3.74
0.81
5.34
4.44
3.81
3.21
4.07

-------
                             Table A-21
              SUMMARIZATION OF AMBIENT BaP AND BSO DATA
Pollutant
BaP (ng/m3) 1975 data

BSO (ng/ra3) 1971-72 data

Statistic
Average
*
Sample size
Range
Average
*
Sample size
Range
Cities
With
Coke Ovens
1.21
21
0.3-4.7
4.21
25
2.1-7.3
Cities
Without
Coke Ovens
0.38
13
0.03-0.9
3.75
12
1.9-5.6
Rural
Areas
<0.10
3
<0. 10
0.95
2
0.8-1.1
Number of cities with annual average concentration data.
                             Table A-22
               ANNUAL BaP AVERAGES FOR SELECTED CITIES
                               (ng/m3)
                                      Cities Without Coke Ovens
                                              2.76 (7)
                                              2.29 (8)
                                              2.64 (8)
                                              2.14 (11)
                                              1.41 (11)
                                              1.22 (8)
                                              0.64 (11)
                                              0.38 (13)
                                              0.41 (13)
       Number of cities included in average.
      Source of 1966-1972 data:  U.S.  EPA (1974)
Year

1966
1967
1968
1969
1970
1971
1972
1975
1976
Cities With Coke Ovens
*
4.74 (15)
5.34 (15)
3.75 (18)
4.41 (23)
3.02 (21)
2.18 (11)
2.14 (19)
1.21 (21)
0.93 (20)
                                  67

-------
          During 1976, samples taken by this surveillance system were
also analyzed for BaP concentrations.  The yearly average for these data,
based on one day sampled per month, are given in Table A-23 by sampling
location within the air basin.  The highest average annual concentration
was 56.35 ng/tn^ for Montessen and the next highest was 17.10 ng/m^ for
Johnstown.  Both locations have coking operations.  The lowest average
concentration was 0.40 ng/ra^ for Hanover Green.
     3.   Charleston, South Carolina

          BaP was analyzed for three collection sites in Charleston, South
Carolina, which has no coke ovens.  The data are summarized in Table A-24.
The average concentration for the city was 0.69 ng/m .


     4.   Maryland Atmospheric Data

          The Maryland State Division of Air Quality Control reports
monthly composite BaP and BSO concentrations for many sites throughout
the state.  Data, primarily for 1976, are summarized in Table A-25.   The
average annual BaP concentrations ranged from 0.43 ng/m^ for Harwood to
6 ng/ra^ for Catonsville.
     5.   CHESS and CHAMP Site Data

          Atmospheric BaP and BSO data have been recorded for a number of
CHESS and CHAMP sites throughout the country.  These data are summarized
in Table A-26.  Average annual concentrations ranged from 0.63 ng/nH for
Thousand Oaks, California to 4.2 ng/m^ for one site in Birmingham,  Alabama.
                                   68

-------
                       Table A-23
     AMBIENT BaP CONCENTRATIONS FOR PENNSYLVANIA, 1976
                        (ng/m3)
Allentown-Eastern Air Basin
  Allentown
  Tatamy
  Bethlehem
  Easton
  Bethlehem East
  Emmaus
  Allen Twp.
  Northampton
       (Basin average)
Beaver Valley Air Basin
  New Castle
  Bessemer
  Koppel
  Beaver Falls
  Vanport
  Rochester
  Ambridge
  Baden
  Midland
  Brighton
       (Basin average)
Erie Air Basin
  Millcreek Twp.
  Erie Central
  Erie South
  Erie East
'early
iverage*
0.71
0.80
1.11
1.86
1.46
1.29
0.55
0.76
1.08
3.06
1.41
9.43
5.03
2.27
4.19
6.18
9.00
3.13
2.42
4.73
0.45
2.04
1.16
1.62
Monthly
Range
0.09- 2.30
0.11- 2.74
0.22- 4.15
0.39- 9.28
0.24- 6.34
0.10- 7.62
0.06- 2.24
0.10- 3.43
~
0.13-11.36
0.46- 2.21
0.30-78.08
0.42-12.65
0.16- 5.44
0.35-13.96
0.75-31.96
0.40-43.48
0.31- 8.60
0.34- 9.74
~
0.12- 0.87
0.26- 7.13
0.21- 3.77
0.23- 6.33
 Based on one sample per
Source:  Dubin (1977)
                         for 12 months
                          69

-------
                   Table A-23  (Continued)
                                  Yearly        Monthly
                                  Average        Range
  Harborcreek Twp.                 0.60       0.13- 3.40
       (Basin average)             1.20           —
Harrisburg Air Basin
  Middletown                       0.83       0.12-2.10
  Swatara Twp.                     0.65       0.18- 1.42
  Steelton                         1.03       0.32- 2.98
  Lemoyne                          0.92       0.28- 2.38
  Susquehanna Twp.                 0.90       0.13- 2.55
  Harrisburg                       0.81       0.15-2.00
  Sumnerdale                       0.61       0.14- 1.60
       (Basin average)             0.82
Johnstown Air Basin
  Westmont                         1.00       0.15- 5.05
  Johnstown North                 17.14       0.31-75.54
  Johnstown Central                4.41       0.24-10.69
  E. Conemaugh                    16.30       1.21-50.74
  Johnstown South                  4.78       0.32-23.01
  Hornerstown                      3.17       0.13- 8.16
       (Basin average)             7.51           —
Lancaster Air Basin
  Lancaster Twp.                   0.54       0.15- 1.77
  Lancaster General                1.01       0.21- 2.74
  Lancaster East                  10.82       0.19-122.7
  Lancaster North                  0.72       0.27- 2.53
  Lancaster West                   0.91       0.25- 3.10
  Neffsville                       0.68       0.11- 1.81
  Manheim Twp.                     0.73       0.12- 2.75
       (Basin average)             2.28
                            70

-------
                   Table A-23 (Continued)
                                  Yearly        Monthly
                                  Average        Range
Monessen Valley Air Basin
  New Eagle                        2.78       0.31-7.51
  Monessen                        56.38       1.05-206.3
  Lover                            2.61       0.61- 9.66
  Elco                             0.96       0.12- 3.94
  Brownsville                      9.05       0.59-57.00
  Charleroi                        2.47       0.13- 6.99
       (Basin average)            12.69           —
Reading Air Basin
  Leesport                         0.56       0.07- 1.60
  Reading South                    0.94       0.18-3.20
  Shillington                      1.02       0.10-4.09
  Sinking Spring                   0.73       0.05- 2.26
  Reading Central                  0.33       0.17-2.67
  Temple                           0.90       0.15- 3.73
  Laureldale                       0.94       0.20-3.32
       (Basin average)             0.85           —
Scranton-Wilkes-Barre Air Basin
  Hanover Green                    0.40       0.09- 1.04
  Dickson City                     1.35       0.18-3.32
  Jessup                           2.00       0.15-13.70
  Pittston                         1.49       0.14-3.60
  Swoyersville                     1.67       0.42- 3.67
  Nanficoke                        0.94       0.11-3.26
  Wilkes-Barre                     1.82       0.19-9.00
  Scranton                         2.06       0.28- 4.25
  Dupont                           1.30       0.27- 2.41
  Avoca                            0.44       0.11- 0.97
  West Nanticoke                   0.79       0.14- 2.05
       (Basin average)             1.32
                            71

-------
                  Table A-23 (Continued)
                                  Yearly       Monthly
                                  Average       Range
Southeast Pennsylvania Air Basin
  Pottstown                        1.06       0.36- 3.09
  Bristol                          0.91       0.20- 3.03
  Willow Grove                     1.05       0.32-3.93
  Dowingtown                       0.69       0.19- 2.48
  Doylestown                       0.76       0.12- 3.21
  Media                            1.00       0.24- 3.11
  Chester                          0.56       0.14- 1.78
  Perkasie                         0.73       0.16- 2.53
  Quakertown                       0.48       0.08- 1.84
  West Chester                     0.81       0.11-2.62
  Lansdale                         1.36       0.18-4.57
  Conshohocken                     2.06       0.40- 3.24
  Phoenixville                     0.80       0.12- 2.47
  Morrisville                      0.79       0.07-2.55
  Coatsville                       0.64       0.07-1.42
       (Basin average)             0.92
York Air Basin
  York East                        0.98       0.17- 2.59
  York Central                     0.96       0.17- 3.13
  West Manchester Twp.             0.78       0.12- 3.38
  Manchester Twp.                  0.41       0.07- 1.12
  West York                        0.77       0.19- 2.16
  Springettsbury                   1.15       0.11- 7.49
       (Basin average)             0.84           —
Altoona Area
  Altoona Central                  3.49       0.29-17.10
  Altoona East                     5.80       0.31-22.20
       (Area average)              4.64           —
                            72

-------
                       Table A-23 (Concluded)
     Farrell-Sharon Area
       Farrell
       Sharon
            (Area average)
     Williamsport Area
       Williamsport Central
       Williamsport East
            (Area average)
                                       Yearly
                                       Average
    2.46
    2.45
    2.46

    1.02
    1.28
    1.15
                  Monthly
                   Range
        0.44- 8.54
        0.24- 9.22
        0.23- 4.14
        0.15- 8.56
                             Table A-24
        DISTRIBUTION OF BaP CONCENTRATIONS IN AMBIENT AIR
                  AT CHARLESTON, SOUTH CAROLINA*
                              (ng/m3)
Site
Number
1
2
3
Location
Radio Station WTMA
Queen St. Fire Station
Mt. Pleasant, Post Office
Sample
Size
22
22
22
Average
0.5711
0.7441
0.7448
Range
0.0028-1.2409
0.1693-1.6787
0.1995-1.9767
            Total
66
0.6866
0.0028-1.9767
 There are no coke ovens in Charleston.
Source:  Spangler and de Nevers (1975).
                                73

-------
                             Table A-2S




AMBIENT ATMOSPHERIC BaP AND BSD CONCENTRATIONS FOR MARYLAND LOCATIONS
Location
Cumberland
Hagerstovn
Adams town
Frederick
Myersville
Buckeyscown
Glen Burnie
Harmons
Harwood
Linthicum
Odenton
Riviera Beach
Annapolis
Baltimore
Lexington and Gay
Sun Avenue
1900 Argonne
5700 Reisterstown
5700 Eastern
Fon thill St.
Cockeysville Ind. Pk.
Cockeysville Police Station
Police Barracks
3001 Eastern Blvd.
Catonsville
Dundalk (8801 Wise Ave.)
Edgemere
Essex
Fort Howard
Towson
Kiddle River
Dundalk (Kavanaugh Rd. )
Dundalk (Reg. Voc. Training)
Westminster
Sample
Size*
12
12
7
12
7
8
11
12
12
12
12
12
11

12
12
12
12
12
12
11
8
12
11
12
10
10
12
10
10
2
11
10
12
(n
Average
4.48
1.40
0.83
1.29
0.17
0.80
1.03
0.54
0.43
0.96
0.71
0.80
0.75

1.95
2.03
1.21
1.37
1.92
1.66
0.80
0.50
0.78
1.29
5.98
1.10
1.85
1.37
2.39
0.75
1.43
1.61
3.30
0.49
BaP
g/m3>
Range
0.40-20.22
0.20-3.87
0.31-2.12
0.18-3.55
0.05-0.55
0.16-2.52
0.15-2.32
0.09-2.10
0.04-1.13
0.15-2.18
0.08-1.75
0.10-2.79
0.10-1.83

0.41-4.46
0.50-4.43
0.27-3.51
0.42-4.76
0.38-5.36
0.37-4.80
0.09-4.83
0.12-1.76
0.10-2.06
0.18-3.97
0.09-2.08
0.56-2.42
0.07-2.60
0.19-4.36
0.38-5.41
0.09-2.95
1.14-1.71
0.53-4.60
0.31-10.35
0.09-1.49
Sample
Size*
12
12
--
12
--
—
12
—
12
12
12
12
12

12
12
12
12
12
12
--
3
12
11
12
—
--
12
10
3
2
—
~
12
BSO
(iig/m3)
Average
8.44
4.74
—
4.80
—
—
4.84
--
2.93
3.99
3.69
4.25
3.76

6.27
6.88
4.21
5.26
6.38
5.24
--
4.11
4.79
4.89
4.02
—
—
5.35
4.79
5.48
6.52
—
—
3.07
Range
6.03-18.14
3.34-7.40
—
3.48-6.77
—
--
2.67-7.76
--
1.52-4.51
2.21-7.49
1.93-5.74
2.50-6.14
2.29-4.93

4.34-8.64
3.35-9.50
2.51-7.24
3.25-8.11
4.25-10.06
3.01-3.63
—
2.15-9.99
2.35-7.40
3.27-6.60
2.53-6.22
--
—
3.00-7.98
2.64-8.66
4.16-6.41
6.25-6.79
--
--
1.95-4.78

-------
                                Table A-25  (Concluded)
Location
Gaithersburg
Silver Spring (1901 Randolph)
Kensington
Poolesville
Silver Spring (Rock Creek Forest)
Rockville
Bechesda
Accokeek
Cheverly
Largo
Laurel
Orme
Oxo n Hill
LapLata
Elkton
Cambridge
Salisbury
Sample
Size*
10
10
11
11
4
12
4
2
11
1
10
10
8
10
12
11
9
BaP
(n«{/m3)
Average
0.62
1.14
0.59
0.46
1.83
0.96
1.25
1.06
0.65
1.30
0.52
0.49
0.70
0.34
1.02
0.62
0.58
Range
0.09-2.13
0.09-5.80
0.09-2.07
0.06-1.40
0.58-3.93
0.07-4.57
0.71-1.68
0.73-1.38
0.17-1.50
--
0.09-1.37
0.08-1.90
0.07-1.50
0.17-1.68
0.18-2.73
0.18-1.58
0.10-1.59
Sample
Size*
10
11
11
12
—
12
--
--
11
--
10
10
8
10
12
11
10
BSO
(L1B/-33)
Average
3.60
4.83
4.26
3.36
--
4.51
--
--
4.57
--
3.56
3.56
4.03
3.27
4.60
4.05
4.46
' Range
2.00-5.17
2.51-11.69
3.05-7.56
1.89-5.18
«
2.56-8.89
--
--
3.45-6.24
--
1.90-5.31
2.51-5.33
3.08-5.80
2.38-4.95
3.55-6.76
2.55-5.80
3.03-5.50
Number of months of data used in calculating  the  average  and  range.
                                           75

-------
                               Table A-26

               ATMOSPHERIC BaP AND BSO CONCENTRATIONS FOR
                   CHESS AND  CHAMP SITES  (1975 DATA)
Location *
Charlotte, NC (1)
Charlotte, NC (2)
Riverhead, NY
Queens , NY
Brooklyn, NY
Bronx, NY
Ogden, Ut
Salt Lake City, Ut
Kearns, Ut
Magna, Ut
Vista, CA
Santa Monica, CA
Thousand Oaks, CA
Garden Grove, CA
Glendora, CA
West Covina, CA
Anaheim, CA
Sample
Size**
6
6
12
12
12
12
12
12
12
12
12
6
12
12
12
12
12
BaP
Mean
1.44
2.36
0.66
1.07
1.57
2.11
2.05
2.37
1.20
1.09
1.03
1.46
0.63
2.42
0.91
1.98
2.36
(ng/m3)
Range
0.3-2.7
0.5-4.4
0.0-3.6
0.1-3.1
0.3-4.0
0.2-4.3
0. 0-7. 2
0.2-5.0
0.1-3.6
0.1-2.9
0.1-4.9
0.2-3.5
0.1-1.4
0.3-7.5
0.1-2.2
0.2-5.0
0.4-7.1
BSO
Mean
1.79
2.91
1.29
1.99
3.70
3.24
2.41
3.26
1.43
1.48
2.07
3.91
2.31
3.86
4.13
5.85
4.77
(yg/m3)
Range
1.1-2.3
2.2-3.8
0.6-2.0
0.9-2.8
1.1-7.6
2.0-3.6
0.7-8.8
1.6-7.7
0.7-3.2
0.5-3.4
0.8-6.7
1.1-6.1
1.1-4.8
0.8-11.9
0.5-6.5
2.6-9.5
1.6-11.2
  Data for Birmingham and Chattanooga are given with the city
  coke oven data in Tables A-8 and A-15, respectively.
**
  Number of months for which data are available; sample size for
  individual months generally ranged from 20 to 31 days.
                                 76

-------
                      Appendix B

STATISTICAL EVALUATION OF BaP ATMOSPHERIC CONCENTRATION
     DATA RECORDED IN THE VICINITY OF COKE PLANTS
                           77

-------
                              Appendix B
        STATISTICAL EVALUATION OF BaP ATMOSPHERIC CONCENTRATION
             DATA RECORDED IN THE VICINITY OF COKE PLANTS
A.   General

     This appendix presents a statistical evaluation of the BaP atmospheric
concentration data recorded in the vicinity of coke plants.  Factors ad-
dressed here include the following:

     •  What is the statistical distribution for atmospheric BaP con-
        centrations over time at a given location?
     •  What errors are introduced by using estimated annual atmospheric
        concentrations, based on a small sample size?

     •  Can the average BaP concentrations around a coke plant be
        described as a mathematical function relating average concen-
        tration to distance from the plant?


B.   Statistical Distribution of 24-Hour BaP Atmospheric Concentrations

     Because of changes in meteorological conditions and other factors,
the atmospheric BaP concentration at a specified location in the vicinity
of a coke plant will vary from day to day.  The day-to-day variations in
the recorded 24-hour concentrations form a statistical distribution.  The
long-term concentration for a specified location is generally characterized
by some central parameter for the distribution like the arithmetic or
geometric mean or the median.  Obviously, the atmospheric concentration
data have been found to follow a distribution having relatively many small
values, with a few observations ranging to fairly high values.  These are
called skewed distributions, as contrasted with symmetrical distributions.
They are sometimes found to follow what is known as two- or three-
parameter lognormal distributions.

     Figure B-l illustrates the cumulative statistical distribution for
BaP atmospheric concentrations from some sampling locations.  Because the
plotted points approximate a straight line, the statistical distributions
might be approximated by a lognormal distribution.  The central measure
that best characterizes this type of distribution is the geometric rather
than the arithmetic average.  The geometric average for these types of
distributions is smaller than the arithmetic average.

     The properties of the lognormal distribution should be used when
describing the probability that a particular BaP atmospheric concentration

                                    79

-------
   1000
             I      I
i     r   i    i    i    i     i    i    i      r
n
 i


o
p
Ul

u
    100
e
UJ


i
o
              0  JOHNSTOWN STATION No. 7



              H  MONESSEN STATION No. 2
             I	I
     I    I     111     III
                   10      20        40       60



                             CUMULATIVE PERCENTAGE
                           80
90    95    98%
          FIGURE B-1.  STATISTICAL DISTRIBUTION  FOR ATMOSPHERIC

                       BaP CONCENTRATIONS
                                    SO

-------
will occur at a specified location.  However, the arithmetic average
should be used when estimating expected population exposures.  That is,
the arithmetic average concentration when used with daily human ventila-
tion rates gives the expected daily inhalation exposure.  This expected
daily inhalation exposure multiplied by 365 gives the estimated total
annual exposure.  The point here is that the arithmetic average should be
used in estimating expected population exposures, and the properties of
the lognormal distribution should be used in estimating the probability
of a specified exposure.

     Table B-l summarizes the arithmetic and geometric average and standard
deviations for samples recorded at a number of stations.  It is of in-
terest and potentially useful that the coefficient of variation (standard
deviation divided by the average) has a value near 1 (i.e., the standard
deviation generally equals the average).
C.   Precision of Estimates Based Upon Small Sample Sizes

     Most of the ambient sampling data available for this study are based
on 24-hour samples collected during limited sampling days.  The ambient
concentrations recorded for these dates, for a given location, are averaged
and used as an estimate of the annual average concentration for that loca-
tion.  It is, therefore, desirable to know how well an estimated average
annual concentration approximates the actual annual concentration.

     From a statistical viewpoint, it is first necessary to know if the
sampling dates or period of dates were selected at random.  In fact,
sampling was probably conducted when people get around to it or are forced
to do it and not because of any particular meteorological or seasonal
reasons.  If this is the case, it might be assumed that the sampling
period was selected in a quasirandom manner.

     The next point has to do with weighting the samples for individual
dates by the fraction of time the meteorological condition on that date
occurs during the year.  This generally is not done because in some cases
the meteorological conditions at the time of sampling are not reported or
because representative sampling is not available for a range of probable
meteorological conditions.  If it can be assumed that the sampling period
is taken at random and that no weighting of the samples is to be made,
the estimation reduces to a simple statistical random sampling problem.
In this case, the average of the available data becomes an unbiased esti-
mate of the average concentration over the year.  However, the number of
dates for which data are available greatly affects the precision of the
estimated annual average.

     The precision of an estimated value is measured by its standard de-
viation.  For a simple random sampling problem, the standard deviation
for an estimated annual average is given by:
                              P = Sf

                                   81

-------
                                                    Table B-l

                     STATISTICAL SUMMARY FOR SAMPLING  DATA TAKEN  FROM A NUMBER OF LOCATIONS
oo
  Sampling
  Location

Monessen, 2
Monessen, 6
Monessen, 7
Johnstown, 1
Johnstown, 2
Johnstown, 3
Johnstown, 4
Johnstown, 5
Johnstown, 6
Johnstown, 7
Johnstown, 8
Utah, 1
Utah, 2
Utah, 3
Utah, 4
Utah, 5
Utah, 6
Utah, 7
Utah, 8
Utah, 9
Utah, 10
Gadsden, 1
Gadsden, 2
Duluth, 1
Duluth, 2
Sample
 Size

  29
  28
  31
  30
  32
  33
  32
  28
  31
  34
  31
   9
   6
   9
  11
  11
  11
   3
  11
  11
   9
   5
   5
  18
  20
Arithmetic

Average
40.8
2.7
22.8
3.6
13.8
7.7
23.4
6.0
6.8
85.3
19.7
2.1
3.8
3.2
2.4
3.1
1.6
2.1
1.5
0.1
0.8
0.6
1.9
1.5
0.2
Standard
Deviation
58.9
2.8
26.3
3.3
19.8
7.5
43.2
3.0
5.0
112.0
28.0
1.3
0.9
2.0
1.6
1.9
1.0
1.3
1.0
0.1
0.8
0.6
1.4
2.0
0.3
Coefficient
of Variation*
1.4
1.0
1.2
0.9
1.4
1.0
1.8
0.5
0.7
1.3
1.4
0.6
0.2
0.6
0.7
0.6
0.6
0.6
0.7
1.0
1.0
1.0
0.7
1.3
1.5
Geometric

Average
10.0
1.0
10.1
2.6
8.6
5.7
13.2
5.2
5.6
44.5
8.3
1.7
3.7
2.6
2.0
2.5
1.4
1.8
1.2
0.1
0.5
0.3
1.6
0.3
0.1
Standard
Deviation
7.6
2.8
4.6
2.3
2.5
2.2
2.5
1.8
1.9
3.6
3.9
2.2
1.3
1.9
2.0
2.2
1.9
2.0
2.2
2.9
3.4
1.8
1.9
13.2
17.7
                 The standard deviation divided by the average.

-------
where

          P = the standard deviation of the estimated annual average.

          S = the calculated standard deviation for the sampling data.

          ,    7365  -  n
          f = V  365 n '
          n = the sample size.

The factor labeled as f is called the finite sample correction factor,
some values of which follow:
                                     Finite Sample
                  Sample Size      Correction Factor

                        1                0.999
                        5                0.444
                       10                0.319
                       20                0.217
                       30                0.175
                       50                0.131
                      100                0.085
                      200                0.048
                      365                0.000
Note that the finite sample correction factor reduces in size rapidly with
additional sampling when the sample size is small.  Depending on the
standard deviation for the sampling data, one might reasonably want sample
sizes in excess of 30.   Estimates based on sample sizes of less than 10
might be suspected of being quite imprecise.
D.   Evaluation of Ambient Concentration Data as a Function of
     Distance from Coke Plant Locations

     Available ambient data that have been recorded in the vicinity of
coke plants (Appendix A) are evaluated to determine if it is mathematically
possible to represent the relationship of BaP concentration as a function
of distance from the coke plant.  To investigate the feasibility of an
approach using recorded ambient concentrations,  the average atmospheric
concentrations have been plotted as a function of distance from the coke
plants (Figures B-2 through B-14).  As might be expected, the atmospheric
concentrations decrease with increasing distance from the coke plants,
thus indicating that the coke plants are a possible source of BaP.   The
moderate amount of scatter in these relationships is probably due to such
factors as the location of the sampling site with respect to the coke
plant, local meteorological conditions, and local geography.  In addition,
because many of the areas have several coke plants, it is difficult to
characterize the contribution to the environment for an individual  plant.


                                   83

-------
   100

<
C
8     h
u
E
tu
z
                        DISTANCE FROM COKE PLANT - km
FIGURE B-2. ATMOSPHERIC CONCENTRATIONS OF BaP FOR JOHNSTOWN. PENNSYLVANIA
                                   84

-------
oo

01
               10
           1
            I
I

ui
o


O
U

o

E

              1.0
              0.1
                0.1
                          I     I    I   I  I  I  I I
                          I     I    I   I  I  I  I I
                                            I     I   I  I   I  II I
                                                           O
                                            I
I    I   I  I  I  I l
                                   1.0                            10



                                     DISTANCE FROM COKE PLANT - km
                         I     I    I   I  I  I I  L
O

1   L  L  1 1 11
                                             100
                               FIGURE B 3. ATMOSPHERIC CONCENTRATIONS OF BaP FOR GENEVA. UTAH

-------
oo
Ot
I    I   I  I  I  I I
                                                                                            I    I   I  I  I
                                                                               O RELATION TO SOLVAY


                                                                               D RELATION TO GREAT LAKES


                                                                                  RELATION TO FORD
                          I     I   I   I   I  I I I
                         I         I   I   I  I I  I
                                               1.0                            10


                                                 DISTANCE FROM COKE PLANT - km
                                                                            100
                          FIGURE B4.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR WAYNE COUNTY. MICHIGAN

-------
oo
             100
           n
            E
o


-------
100
                                       O  RELATED TO BETHLEHEM



                                       O  RELATED TO OONNER-HANNA



                                       A  RELATED TO ALLIED
                  I    I   I   I   I  I
                        DISTANCE FROM COKE PLANT - km
 FIGURE B-6.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR BUFFALO, NEW YORK
                                  88

-------
CO
               10
            I
            IE


            BI.O
            z
8
            X
            111
              0.1
                0.1
                      O RELATIVE TO TARRENT


                      Q RELATIVE TO FAIRFIELO


                         RELATIVE TO BIRMINGHAM
                                                           0
                                                         I     I    I   I  I   I I  I
                                                              I   I   I  I  I I  I
                                                                            I     I    I   I  I  I I L
                                                                                 I    I   I  I  I I  I
                                   1.0
10
100
                                                  DISTANCE FROM COKE PLANT - km
                             FIGURE B 7.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR BIRMINGHAM, ALABAMA

-------
   50
   10
                     I    I   I   I  I  I I
                                  CD
                      T
I    I   I   I  I  I  I
                                        O SAMPLING PERIOD 2

                                        O SAMPLING PERIOD 1
 I
 O
 in
 8
 O
 £
 tu
   1.0
                                                       0
   0.1
                              0
                                      0
    0.1
            1.0

DISTANCE PROM COKE PLANT - km
                10
FIGURE B-8.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR GRANITE CITY, ILLINOIS
                                     90

-------
10
          1     I    I  I   I  I I  I
I     I    I   I  I  I  I I
I     I    I   I  I  I  I L
                                                                          I    I   I  I  i  i I
                                 DISTANCE FROM COKE PLANT - km
          FIGURE B-9.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR SPARROWS POINT. MARYLAND

-------
   100
    10
            ©
                               I    I   I   I  I  I  i
                                             Q ARITHMETIC AVERAGES BASED
                                                ON GEOMETRIC AVERAGES AND
                                                STANDARD DEVIATIONS
c
C

1U
u
C
ui
   1.0
     0
   0.1
               1.0
                                   G

                               0         ©3
                    0
                     0  O
                  10
DISTANCE FROM COKE PLANT - km
                                                                   I    I  I  I  I  I
100
       FIGURE B-10.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR CLEVELAND, OHIO
                                       92

-------
    100
     10

  o

  H


  DC



  ui
  O
  u

  £
  in
  <  1.0
     0.1
       0.1
                 i     I    I   1  I  I  I  I
                            I    I   I  TTT
                                                                 1
                            L   I   I  I  I  I  I
            1.0



DISTANCE FROM COKE PLANT - km
10
FIGURE B-11. ATMOSPHERIC CONCENTRATIONS OF BaP FOR MONESSEN. PENNSYLVANIA
                                     93

-------
    10
 I


o
   1.0
u

o

£
iu
   0.1
     0.1
               I      I    I   I   I  I  I I
                       I      I    I   I  I  I   I I
             1.0



DISTANCE FROM COKE PLANT - km
10
 FIGURE B-12. ATMOSPHERIC CONCENTRATIONS OF BaP FOR GADSDEN, ALABAMA
                                     94

-------
    10.
I
o

H

OC
a   1.0
z
I
UJ
Z
    0.1
     0.1
                     I    I   I  I   I I  I
                       ©
                                III  I  I J-
             1.0


DISTANCE FROM COKE PLANT - km
10
FIGURE B-13.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR DULUTH, MINNESOTA
                                    95

-------
VO
              10
                         1     1    1   1  1  1  1 1
            i
R ATI ON
CE
b
            u

            £
            Ul



            8
              0.1
                     ALL UPWIND STATIONS
1     1    1   1   1  1 1 1
               0.1
                               1     1    1   1  1  1  1  1
                                                               O
1     1    1   1   1  1 1 1
                               1     1    I   I   1  1  |
                                                                                 0
1     1   1  1   1  1  1 1
                     1.0                            10


                       DISTANCE FROM COKE PLANT - km
                                                   100
                        FIGURE B 14.  ATMOSPHERIC CONCENTRATIONS OF BaP FOR PHILADELPHIA, PENNSYLVANIA

-------
If ambient data are to be used to characterize human exposure, it would
be desirable to have data from many monitoring sites located at different
directions from the coke plant and to have data recorded for each moni-
toring site over a large number of days.  Much of the recorded data do
not meet these requirements; the number of sampling stations by plant
ranged from 1 to 16.

     Where elevated concentrations exist, the relationship of average
atmospheric concentration to distance does appear to follow a mathematical
function of the type:


                             C = aDb
where

          C = the average concentration at some distance from the coke
              plant.

        a,b = model parameters fit by regression techniques.

          D • the distance from the plant.

Least squares regression techniques have been used to fit the data to
this mathematical function for each coke plant for which data are avail-
able.  The results of this evaluation are given in Table B-2.  The re-
gression coefficients given in Table B-2 indicate how well the data fit
the function.  For most cases, the regression coefficients ranged from
0.5 to 1.0, suggesting fairly good approximations.  The coke plants with
only two monitoring stations do not have sufficient data to perform
statistical confidence tests.  The model parameters based on actual am-
bient data can be compared to similar fits to the atmospheric dispersion
modeling data conducted by Youngblood (1977).  Two conditions are given
in Table B-2 for comparison (one for a dirty plant and one for a clean
plant).

     The magnitude of the model parameter a relates to the amount of BaP
emitted from the source, and the model parameter b relates to decay in
the concentration versus distance function.  Note that for all coke loca-
tions with more than two sampling stations on Table B-2 the b parameter
varies between -0.32 to -1.42 with an average of -0.9.  When locations
that have more than one coke plant are also excluded, the parameter has
an average value of -1.0.  This compares favorably with the modeling
data, which give a value of b of about -0.95.

     Table B-3 gives statistical tests used to determine if the decay
parameter  (b) is significantly less than zero.  The table shows that in
none of the 13 cases is the parameter significantly less than zero at
the  0.05 confidence level.  For the four cases where  significance was not
found,  the data either were highly variable or there  were other coke
plants  in  the area.

                                   97

-------
                            Table B-2


            ESTIMATED  PARAMETER VALUES  FOR REGRESSION

                 APPROXIMATIONS TO AMBIENT DATA
Model
Number of Parameters
Location Stations
Johns town
Gadsden
Duluth
Monessen
Utah
Wayne County
Buffalo--Beth.
Buffalo— D.H.
Buffalo— Allied
Cleveland
Pittsburgh— USS
Pittsburgh— J&L
Tarrant
Granite City
Sparrows Point
Fairfield
Dirty Plant Model
**
Clean Plant Model
8
2
2
3
10
6
5
7
4
16
5
6
4
10
4
2
5
4
a
35.24
1.28
379.39
49.99
4.70
13.09
15.96
8.40
1.96
12.42
22.00
4.32
3.98
11.50
3.07
2.67
135.84
60.66
b Cc
-1.13
-1.62
-7.50
-2.92
-0.84
-0.69
-0.99
-0.75
-0.33
-1.42
-0.32
-0.37
-0.33
-0.57
-0.33
-0.07
-0.96
-0.95
Regression
jefficient (R)2
0.96
*
*
0.64
0.76
0.92
0.79
0.60
0.06
0.72
0.16
0.52
0.75
0.10
0.61
*
0.99
0.98
Data  available  for only two  monitoring  stations.
t
Uses  only data  for distances of  1  km and  greater.
                               98

-------
                          Table B-3
     STATISTICAL EVALUATION OF REGRESSION APPROXIMATIONS
                                           Significance Level
                                                  0.01
                                                   NS
                                                  0.01
                                                  0.01
                                                  0.05
                                                  0.01
                                                   NS
                                                  0.01
                                                   NS
                                                  0.01
                                                  0.01
                                                   NS
                                                  0.01
S   —Standard error of the estimate in terms of natural
      logarithms.
Sfc--Standard error of the regression slope parameter (b).
t(b < o)--Students-t test value for testing if the slope
          parameter (b) is less than zero.
NS--Not statistically significant at the 0.05 level.
Location
Johnstown
Monessen
Utah
Wayne County
Buffalo-Beth.
Buffalo--D.H.
Buffalo— Allied
Cleveland
Pittsburgh--USS
Pittsburgh — J&L
Tarrant
Granite City
Sparrows Point
S
0.12
1.50
0.29
0.04
0.62
0.63
1.47
0.14
0.74
0.09
0.08
1.56
0.04
Sb
0.05
2.71
0.09
0.02
0.24
0.22
1.13
0.04
0.36
0.05
0.03
0.75
0.04
t(b < o)
-21.7
-1.08
-9.03
-41.2
-4.06
-3.46
-0.29
-33.02
-0.88
-7.12
-10.00
-0.76
-8.68
                             99

-------
                       Appendix C

DETAILED ESTIMATES OF POPULATIONS AND BaP CONCENTRATIONS
             FOR INDIVIDUAL COKE FACILITIES
                           101

-------
                              Appendix C

       DETAILED ESTIMATES OF POPULATIONS AND BaP CONCENTRATIONS
                    FOR INDIVIDUAL COKE FACILITIES
     This appendix includes the detailed population and BaP concentration
estimates for each defined geographic population ring (i.e., 0 to 0.5,
0.5 to 1, 1 to 3, 3 to 7, and 7 to 15 km) about each coke facility.  These
estimates are given in Tables C-l and C-2.  The concentrations include the
summation of atmospheric concentrations from both the coke ovens and back-
ground.  The population within a geographic ring was considered not to be
excessively exposed to coke-oven emissions if its -estimated average annual
BaP concentration was less than 0.1 ng/nH.  For some locations, several
separate coke facilities are located within 15 km of each other.  In these
cases, it was necessary to estimate geographic population ring overlaps
and total ring BaP concentrations.

     The exposure estimates given here use the model that assumes variable
background concentrations (see Section IV-D).
                                   103

-------
            Table C-l

DETAILED BdP POPULATION EXPOSURES
 (Coke Emissions  plus Background)
             (ng/ipj)
    Distance from Cuke Facility  (km)
0-0.5
SUe
No.
1
2
3
4
5
6
7
a
9
10
11
12
11
14
IS
16
17-18b
19
20
21
21
23
24
25
26
27
28
29
JO
31
12
33
34
Popu-
lalJon
3UU
0
0
478
1.656
975
0
0
0
0
82 7
0
2,618
0
57
512
0
0
99
552
11
0
0
0
991
0
3,368
0
0
I.LB4
0
6
0
Concen-
l rut ion
*
1.5
24.4
6.0
1.7
*
17.9
17.5
9.9
2.0
7.3
30.4
4.8
19.1
5.4
2.1
143.0
112.0
63.7
48.4
4.3
*
*
*
2.2
4.2
3.5
2.7
*
*
37.5
28.6
23.0
0.
Popu-
lation
1,756
0
0
916
532
5.279
0
1.416
0
2.244
7.307
0
2,494
0
8.176
3.059
53
33
482
0
3,416
3.008
2,197
51
3.901
0
2.450
202
3,113
2,537
2,373
2.608
0
5-1
Cuncen-
t rat Ion
*
1.0
13.2
2.0
1.1
*
9.7
9.9
5.6
1.4
4.2
16.6
2.9
JO. 5
3.2
1.5
76.6
60.0
34.3
26.0
3.5
*
*
*
1.3
2.0
2.0
2.1
*
*
20.3
15.5
12.6
1
Popu-
laLlou
13,880
0
21.495
19,693
36.497
28.195
22.105
9.493
00
30.475
58.244
00
27.666
00
71.661
36.083
21,279
26.829
25.740
15.511
15.948
37.283
46,224
33.878
61,720
4,219
30.734
7.874
17.140
58,527
11,701
8,567
455
-3
Concen-
l rat Ion
*
0.6
5.2
0.4
0.7
*
3.9
4.5
2.5
0.7
2.0
6.6
1.5
4.4
1.6
1.0
29.3
22.9
13.3
10. 0
2.5
*
*
*
0.8
0.8
0.9
1.6
*
*
8.0
6.2
5.1
3-7
Popu-
lation
114,873
5,843
80,440
36,345
158,506
120.414
120,356
51,629
0
46,890
248,247
2,828
187,310
16,253
207,269
42,851
51,533
122,882
70,004
35,072
97,933
289,066
322,403
172,788
255,071
7,036
224,719
111,218
138,521
257,202
55,346
57,955
749
Concen-
tratlon
*
0.5
2.3
0.4
0.5
*
1.8
2.5
1.3
0.7
1.2
3.1
1.0
2.2
1.1
0.8
12.1
9.6
5.7
4.2
1.9
*
*
*
0.5
0.3
0.6
1.2
*
*
3.6
2.8
2.4
7-15
Popu-
lation
278.354
7.802
196.316
21,099
283,180
297,002
256.857
161.178
52
704,796
1,153,057
14.649
1.223.577
73.329
444.465
24,392
520,919
265,439
637,625
90.904
481.929
1.190.455
1.274.124
892,126
727,327
76.894
714.782
601.056
531,748
533,320
37,066
83,131
17.189
Concen-
L rat Ion
*
0.4
1.3
0.4
0.4
*
*
1.8
0.9
0.7
0.9
1.8
0.9
1.4
0.9
0.8
5.9
4.7
3.0
2.1
0.8
*
X
*
0.5
0.3
0.4
—
*
*
1.9
1.6
1.4
15-20
Popu-
lation
110,032
23.498
133.219
9.035
88.482
97,271
113.551
254.702
123
558.193
859.055
9.847
787,330
165,955
88.930
14.169
626,382
191.054
740,842
55.563
592.595
862.111
872,932
841.553
1,007
63.748
388,813
230,660
257.603
151.272
34,939
124,988
22,212
Concen-
craclon
*
0.4
0.9
0.4
0.4
*
*
1.6
0.8
0.7
0.9
1.4
0.8
1.1
0.8
0.7
4.0
3.2
2.1
1.5
0.8
*
*
*
0.4
0.3
0.4
*
*
*
1.4
1.2
1.1
20-30
Popu-
lation
132,291
81.950
150.334
33.966
65,967
97.546
95,069
428,754
1.661
741,093
1,581.162
100,951
1,840,707
324.094
122,293
41.807
1,521,807
785,467
1.456,265
57,907
617.561
1.220,338
1,139.598
1.389,647
3,605
33,818
716,207
226,767
194.061
160.244
95,537
430,148
92.077
Coucen-
LralJon
*
0.4
0.8
0.4
0.4
*
*
1.5
0.7
0.7
0.8
1.2
0.8
1.0
0.8
0.7
3.0
2.5
1.7
1.2
0.8
*
*
*
0.4
0.3
0.4
*
*
*
1.2
1.0
1.0

-------
                                                        Table (>J (Concluded)
                                                        Distance from Coke facility (km)
0-0.5
Sice
No.
35
36
37
38
39
40
41
42
43
44
45
46-47l>
48
4!)
50
51
52
5)
54
55
56
57
58
59
60
61-62b
63
64
65
Popu-
1 at ion
0
0
66 J
1,530
0
0
0
13
0
0
3,960
1,804
0
0
0
2,628
2,433
0
1,024
2,374
0
1.185
6
0
19
0
0
J
854
Concen-
tration
2.1
18.6
13.5
42.0
2.4
20.1
11.3
19.5
16.7
6.4
16.3
100.0
5.4
16.5
*
2.8
22. J
*
*
54.8
50.0
1.8
17.6
3.5
10.0
130.3
9.5
18.0
3.2
0.
Popu-
lation
0
0
1.597
4,228
71
0
1,986
2,858
4,074
0
1.593
870
0
0
10.170
10,041
4.526
75
1,365
1.757
1,559
5,483
0
0
0
0
0
0
3,900
5-1
Conuen-
t ration
1.4
10.2
7.4
25.0
1.5
11.0
6.4
10.7
9.2
3.8
9.1
50.0
3.2
9.2
*
1.9
12.2
*
*
29.6
38.0
1.2
9.4
1.9
5.8
69.7
5.3
9.8
2.0
1-3
Popu-
lation
10,963
5,352
22,961
7J.580
17,663
20,260
45.234
72,578
33,723
26,142
42.961
35,460
573
19,922
115.372
56,243
54,405
31,991
16.819
30,671
22,763
16.595
4.866
1.046
3,044
0
0
20,971
73.254
Concen-
l rat. Ion
0.9
4.2
3.0
4.5
1.0
4.5
2.9
4.4
3.8
1.9
3.9
17.0
1.7
4.0
*
1.2
5.1
*
*
11.6
10.0
0.7
3.5
0.7
2.6
26.5
2.3
4.0
1.2
3-7
Popu-
IdClOll
30,503
34,067
155.445
399,565
40,897
82,028
147.771
378,615
115,698
110,829
64.472
51,502
12.276
61,371
396,226
87,797
486,896
122,412
83,779
117,606
42,498
81,104
99.968
2,219
28,887
3.570
0
39,901
328,749
Concen-
tration
0.7
2.0
1.5
2.3
0.7
2.2
1.6
2.1
1.9
1.2
2.0
5.2
1.2
2.1
*
1.0
2.5
*
*
5.1
2.0
0.5
1.4
0.3
1.2
10.9
1.2
1.9
0.9
7-15
Popu-
lation
50.851
24.213
258.780
859,264
152.877
96,129
133,429
860,216
164,425
734,938
229,293
38,180
91,115
116,901
748,696
40,701
1,685.267
632.088
407.475
354.575
80.359
117.888
350,402
8.902
72,123
8,598
1,410
65,031
60 i, 106
Concen-
Irallon
0.7
1.3
0.9
0.5
0.7
1.3
1.2
1.3
1.2
1.0
1.4
2.2
1.0
1.4
*
0.9
1.6
*
*
2.8
0.9
0.5
0.7
0.2
0.6
5.2
0.8
1.1
0.8
15-20
Popu-
lation
50,962
15,450
99,927
283.361
92,928
131.626
87,533
313,924
95,320
1.270.355
125.342
23,466
128,565
86,055
207.185
10.856
670.578
416.361
594,186
234.800
91.526
57,125
420,504
4.884
2.800
30,057
2,094
28,883
127.088
Concen-
t rat Ion
0.6
1.0
0.7
0.5
0.6
1.0
1.0
1.0
1.0
0.9
1.2
1.5
0.9
1.2
*
0.8
1.3
*
*
2.0
0.8
0.4
0.4
0.2
0.5
3.5
0.7
0.9
0.8
20-30
Popu-
lation
177,825
24,710
97,007
304,663
197,213
260,433
253,967
298,122
252,616
1,587,922
142,000
63,508
179,905
471,960
365,233
32,420
1,026,074
576,450
654,955
833,846
698,784
49,386
685,288
14,240
41,916
165,797
10,004
83,458
176.532
Concen-
t rat Ion
0.6
0.9
0.6
0.5
0.6
0.9
1.0
0.9
0.9
0.9
1.0
1.2
0.9
1.1
*
0.8
1.1
*
*
1.7
0.8
0.4
0.3
0.1
0.5
2.6
0.6
0.8
0.7
dSite numberb correspond to coke facilities listed in Table 111-3.
 Indicates tliut die two facilities were treated as though there were collated.
 Indicates that one or more coke facilities are located within 15 km of that facility.  Estimated concentrations are given
 In Table C-2.

-------
                            Table C-2

             BaP EXPOSURES FOR PERSONS IN LOCATIONS
               HAVING MORE THAN ONE COKE FACILITY
                                                       Exposure
                                                    Concentration
	Location	     Exposed Population        (ng/m3)

Birmingham, Alabama                    975               8.2
                                       388               5.8
                                    14,025               5.6
                                     7,035               4.5
                                    28,054               3.0
                                   110,000               2.6
                                   106,951               1.8
                                   135,893               1.6
                                   108,302               1.4

Detroit, Michigan                       51              12.0
                                     5,000               8.1
                                     2,197               7.8
                                    41,000               6.7
                                     3,008               4.5
                                    19,900               4.0
                                   330,000               3.6
                                    14,913               3.4
                                 1,274,124               2.1
                                   869,325               1.9
                                 1,166,511               1.5

Buffalo, New York                    3,113              22.0
                                     1,184              17.0
                                    19,000              12.0
                                     2,537              10.0
                                    11,300               8.0
                                    39,000               5.0
                                   213,178               4.2
                                   193,691               3.2
                                   533,000               1.6

Pittsburgh, Pennsylvania             1,024              30.0
                                     1,365              24,0
                                    83,800              13.0
                                   407,500              10.0
                                   405,911 -             8.8
                                   532,215               6.6
                                    10,170               4.8
                                   147,363               3.0
                                   396,226               2.0
                                    16,819               1.8

                               106

-------
                             BIBLIOGRAPHY
American Iron and Steel Institute, "Comments of the American Iron and
     Steel Institute on the External Review Drafts of Documents Assessing
     the Public Health Consequences of Exposures to Coke Oven Emissions"
     (June 30, 1978).

Andelman, J. B., and M. J. Suess, "Polynuclear Aromatic Hydrocarbons in
     the Water Environment," World Health Organization. 43_:479-508 (1970).

Antel, M., personal communication (May 1977).

Department of Environmental Resources, "Monessen Area Air Quality Study,
     Monessen, Pennsylvania," Commonwealth of Pennsylvania (1977A).

Department of Environmental Resources, "Johnstown Air Basin Air Quality
     Study, Johnstown, Pennsylvania," Commonwealth of Pennsylvania (1977B).

Department of Environmental Resources, "1976 Pennsylvania Air Quality
     Surveillance System Air Quality Report," Commonwealth of Pennsylvania
     (April 1977C).

Dubin, R., personal communication, Pennsylvania Department of Environ-
     mental Resources (May 1977).

Ek, A., private correspondence with B. E. Suta, Allegheny County Health
     Department (May 2, 1977).

Energy and Environmental Analysis, Inc., "Preliminary Assessment of the
     Sources, Control, and Population Exposure to Airborne Polycyclic
     Organic Matter (POM) as Indicated by Benzo(a)pyrene (BaP)" (Draft
     May 1978).

Federal Register, "Exposure to Coke Oven Emissions," Occupational Safety
     and Health Administration, 4J,(206): 46742-46790 (October 22, 1976).

Gorelova, N. D., P. P. Dikun, L. D. Kostenko, 0. P. Gretskaya, and A. V.
     Emshanova, "Detection of the Possible Presence of 3,4-Benzopyrene
     in Fresh Fish," Novosti Onkol. 8-12 (1971).

Grimmer, G., "Carcinogenic Hydrocarbons in the Human Environment,"
     Deut. Apoth.-Ztg. 108(16) =529-533 (1968).-

Jungers, R. H., "BaP Analysis of Duluth Samples," memorandum to J. Manning,
     U.S. Environmental Protection Agency (February 2, 1977A).
                                  107

-------
Jungers, R. H.,  "BaP Analysis, Alabama," memorandum Co J. Manning, U.S.
     Environmental Protection Agency (February 4, 1977B).

King, R. B.,  J.  S. Fordyce, A. C. Antoine, H. F. Leibecki, H. E. Neustadter,
     and S. M. Sidik, "Extensive One-Year Survey of Trace Elements and
     Compounds in the Airborne Suspended Particulate Matter  in Cleveland,
     Ohio," Lewis Research Center (1976).

Keystone Coal Industry Manual (McGraw-Hill Publications,. New York, New
     York, 1975).

Kulujian, N.  J., "By-Product Coke Battery Compliance Evaluation," PEDCo-
     Environmental Specialists, Inc. (1975).

Lazenka, C.,  private communications to Ben Suta, City of Philadelphia,
     Department of Public Health (May 9, 1977).

Manning, J.,  "Coke Oven Dispersion Analysis  for BaP and BSD," memorandum,
     U.S. Environmental Protection Agency (March 18, 1977).

Manning, J.,  personal communication (August  1977).

National Academy of Sciences, "Particulate Polycyclic Organic Matter--
     Report on Biological Effects of Atmospheric Pollutants," Report No.
     PB-212 940/1 (1972).

NIOSH, "Correlations Among Log Concentrations of 13 Polynuclear Aromatic
     Compounds (FNAs) and the Log Concentrations of the Benzene Soluble
     Fraction," memorandum (December 6, 1974).

Radding, S. B.,  T. Mill, C. W. Gould, D. H.  Liu, H. L. Johnson, D. C.
     Bomberger,  and C. V. Fojo, "The Environmental Fate of Selected Poly-
     nuclear Aromatic Hydrocarbons," prepared by Stanford Research
     Institute for the Environmental Protection Agency, Report No. EPA
     560/2-75-009 (1976).

Schulte, K. A.,  D. J. Larson, R. W. Hornung, and J. V. Crable, "Analytical
     Methods Used in a Study of Coke Oven Effluent," Am. Ind. Hyg.
     Assoc. J. 36(2) .-131-139 (February 1975).

Shabad, L. M., and Y. L. Cohan, "Contents of Benzo(a)pyrene  in Some
     Crops," Arch. Geschwulstforsch. 40(3):237-243 (1972).

Sheridan, E.  T., "Supply and Demand for United States Coking Coals and
     Metallurgical Coke," U.S. Department of the Interior (1976).

Smith, W. M. , "Evaluation of Coke Oven Emissions," J. Occ. Med. 13:69-74
     (1971).
                                  108

-------
Spangler, C., and N. deNevers, "Benzo(a)pyrene and Trace Metals in
     Charleston, South Carolina," U.S. Environmental Protection Agency,
     Report No. EPA-450/2-75-004 (1975).

U.S. Environmental Protection Agency, "Preferred Standards Path Report
     for Polycyclic Organic Matter" (1974).

U.S. Environmental Protection Agency, "Scientific and Technical Assess-
     ment Report on Particulate Polycyclic Organic Matter (PPOM),"
     Report No. EPA-600/6-75-001 (1975).

U.S. Environmental Protection Agency, "Compilation of Air Pollutant Emis-
     sion Factors," Report No. AP-42 (1976A).

U.S. Environmental Protection Agency, "The Carcinogen Assessment Group
     Analysis of Kepone," Draft Report (1976B).

U.S. Public Health Service," Radiological Health Handbook," U.S. Depart-
     ment of Health, Education, and Welfare (1960).

Varga, J., Jr., "Screening Study on By-Product Coke Oven Plants," Draft
     Report, Battelle Research Institute (1974).

Youngblood, P. L., "Rough Estimation of Ambient Impact of Organic Emissions
     from Coke Ovens," Draft Memorandum, U.S. Environmental Protection
     Agency (May 27, 1977).
                                  109

-------