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:
Aian P. Carlin
Joseph D. Cirvello
Technical Monitor:
Justice A. Manning
Contracts 58-01 -4314 and 83-02-2835
SRI Projects SGU-57S4 and CRU-6780
Prepared by:
Benjamin £. Suta
Center 'or Resource and
Environmental Systems Studies
CRESS Report No. 2'
333 Pavenswood Ave. • Menio Park. California 9<*C25
(415) 326-6200 • Cable: STANRES, V.enlo Parx • TVVX: 910-373-12^8
-------
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:
AlanP.Carlin
Joseph D.Cirvello
Technical Monitor:
Justice A. Manning
Contracts 68-01 -431 4 and 68-02-2835
SRI Projects EGU-5794 and CRU-6730
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 (SPA) 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 to 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.
-------
PREFACE
There is a substantial body of evidence, both direct and indirect,
chat the mixture chat coke oven emissions represent is carcinogenic and
coxic. Current U.S. Environmental Protection Agency (SPA) policy states
chat 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-
siates 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 co ambient concentrations
of benzo(a)pyrene (BaP) and benzene soluble organics (3SO) material caused
by coke oven emissions.
v
-------
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. Ac-Risk Population 6
C. Population Estimation 6
D. Population Exposures to 3aP 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
3. 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 3aP and BSO Atmospheric
Concentrations 37
F. Population Exposure Estimates. . 40
APPENDICES
A AMBIENT ATMOSPHERIC BaP AND BSO CONCENTRATIONS ... 45
B STATISTICAL EVALUATION OE 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 L07
viii
-------
ILLUSTRATIONS
IV -1
IV-2
IV-3
3-1
3-2
3-3
B-4
B-5
3-6
3-7
3-8
3-9
3-10
B-ll
B-12
3-13
3-14
Relationship Between BSO and BaP Atmospheric
Concentrations for ail Locations '
Relationship Between BaP and BSO Concentrations
for Locations near Coke Ovens
Relationship Between BaP and BSO Airborne
Concentrations for Occupational Coke Oven
Locations
Statistical Distribution for Atmospheric BaP
Concentrations
Atmospheric Concentrations
Pennsylvania
Atmospheric Concentrations
Utah
Atmospheric Concentrations
County, Michigan
Atmospheric Concentrations
County, Pennsvlvania . . .
Atmospheric Concentrations
New York
Atmospheric Concentrations
Alabama
Atmospheric Concentrations
Illinois
Atmospheric Concentrations
Point, Maryland
Atmospheric Concentrations
Ohio
Atmospheric Concentrations
Pennsylvania
Atmospheric Concentrations
Alabama
Atmospheric Concentrations
Minnesota
Atmospheric Concentrations
Pennsylvania
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
3aP
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
SO
84
35
86
37
38
89
90
91
92
93
94
95
96
-------
TABLES
II-l Estimated BaP Emissions in Che United Scares
(1972)
II-2 Summarization of Ambienc 3aP and 3SO Data.
II-3 Estimated Number of People Exposed co BaP from Coke-
Oven Emissions, Assuming Variable Location
Background Concentrations 3
II-4 Estimated Number of People Exposed Co BaP from Coke-
Oven Emissions, Assuming Uniform Background
Concentrations ' 9
II-5 Estimated Number of People Exposed to 3a? from Coke-
Oven Emissions, Including Background II
II-6 Estimated Number of People Exposed to 3SO from Coke-
Oven Emissions 12
11-7 Relative Error in Exposure Estimates Assuming
Circular Isoconcentracion Contours 17
II-8 Benzo(a)Pyrene Concentrations in Foods 13
III-l Correlations Among PAH Compounds in che Air over
Greater Birmingham, Alabama, 1964 and 1965 23
III-2 Correlation Coefficients Among Log Concentrations
of 13 PNA and 3SO Samples Taken Within Five Coke
Planes 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 23
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 Obcained 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 Ba? 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, BSD, and TSP Concentrations for
Johnstown, Pennsylvania 56
A-12 Ambient BaP, BSD, 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 Daca Taken from
a Number of Locations 32
3-2 Estimated Parameter Values for Regression
Approximations to Ambient Data 98
3-3 Statistical Evaluation of Regression
Approximations 99
C-L Detailed 3aP Population Exposures 104
C-2 3aP Exposures for Persons, in Locations Having
More than One Coke Facility 106
xiii
-------
CHEMICAL ABBREVIATIONS
A anthracene
Ant anthanthrene
3aA benz(a)anthracene
BaP benzo(a)pyrene
BbF benzo(b)fluoranthene
BcA benzo(c)acridene
BeP benzo(e)pyrene
BghiP benzo(g,h, Dperylene
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
?yr pyrene
TSP total suspended particulates
xv
-------
I INTRODUCTION
The primary objective of this study has been co quantify che environ-
mental atmospheric exposure of the general human population to coke-oven
emissions of benzo(a)pyrene (3aP) and benzene soluble organics (3SO). 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
cerms 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 cables 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 3.
-------
II SUMMARY AND CONCLUSIONS
A. Overview
There are 65 by-produce coke plants in the Unitsd 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 •
S07<> 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 TSP 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 wich 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-1 reports a BaP emission inventory made by the Environ-
mental Protection Agency (EPA) for 1972. Stationary sources account for
987o 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 197, of the nation-
wide BaP emissions. ESA (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 IS THE UNITED STATES (1972)
Emissions
Source Type C.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 bloving <1
Coke production 170
Mobile Sources
Gasoline-powered automobiles and trucks 11
Diesel-powered trucks and buses <1
Tire degradation 11
Source: U.S. E?A (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/m^ 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 L
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 (^g/m3) Average 4.21 3.75 0.95
Sample size 25 12 2
Range 2.1-7.3 1.9-5.6 0.3-1.1
BSO monitoring was discontinued in 1972.
-------
(the average is 0.38 ng/m^); however, areas with coke ovens generally
have average concentrations in excess of 1 ng/ni^ (the average is 1.21
ng/m-3).
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 m^/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
0.5-1 116,000
1-3 L. 644,000
3-7 7',226',000
7-15 22,200,000
15-20 15,283,000
20-30 25,583,000
6
-------
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 bv 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/nr). 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^, 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/tiH or more.
-------
Table II-3
ESTIMATED NUMBER OF ?EOPL£ EXPOSED TO 3aP
FROM COKE-OVEN EMISSIONS, ASSUMING VARIABLE
LOCATION BACKGROUND CONCENTRATIONS*
Subgroup
Concentration
Range ^
(ng/a3)
90-100
80-90
70-30
60-70
50-60
45-50
40-45
35-40
30-35
25-30
20-25
15-20
10-15
3-10
6-3
5-6
4-5
3-4
2-3
1-2
0.5-1
0.1-0.5
Total
Assuznotion
All Well j
Controlled'
1,300
Stratified
Grouping
1,300
50
100
2,400
1,400
1,500
1,600
430
24,000
38,000
43,000
230,000
580,000
440,000
1,200,000
640,000
1,200,000
3,400,000
3,500,000
3,100,000
20,000.000
44,000,000 42,000,000
All Poorly.
Controlled?
1,300
10
60
4,000
4
7
26
1
1
1
8
41
110
440
450
620
130
490
990
,400
,900
,000
—
870
,500
,600
—
,000
,700
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
1,
1,
5,
13,
11,
13,
2
1
4
3
25
40
140
360
620
380
700
750
100
600
000
000
000
,400
,400
,400
,600
430
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
7000
53,000,000
— No exposures are estimated for chis concentracion
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
Rangei" Assumed Background
(ng/m3) 0.4 ng/m3 1.0 ng/m3
90-100 1,300 1,800
30-90
70-80
60-70
50-60
45-50
40-45
35-40
30-35
25-30
20-25
15-20
10-15
3-10
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
6,300,000
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 tan from the plants. Had this constraint been relaxed, more
people exposed would have been assigned to the lover 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.
£. Population Exposures to 350 Emitted by Coke Ovens
Sufficient data have not been collected near coke plants nor have
emission factors been developed for adequately assessing the atmospheric
3SO concentrations resulting from the plants' emissions. The approach
taken here is to estimate the 3SO concentrations, based on the estimated
3aP 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 3a? constitutes 17. of the total 3SO emitted by coke ovens. Five
of the alternative 3aP exposure estimate techniques are used to provide
five alternative 3SO estimates. The estimated exposures to 3SO from
coke plants only are given in Table II-6. Annual average concentrations
are shown to range as high as 10 u,g/m^. Because of the lack of sufficient
data, estimates are not given for 3SO exposures due to coke ovens plus
background. Table II-2 shows that the 1971-72 concentrations for cities
without coke ovens averaged 3.75 p.g/m^. Hence, adding this value to the
exposure concentrations given in Table II-6 would give a rough estimate
of total BSD exposure. This, however, may not be proper because the
constitution of the background BSO may likely differ from that of the
3SO emitted by coke ovens.
F. 'Considerations in the Use of the Annual Average as a Measure
of Exposure co Coke-Qven Emissions
Exposure estimates in this report are given in cerms 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
che 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"7 Uniform?
1,800
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
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
59,000,000
— No exposures are estimated for this concen-
tration range.
*
Based on annual averages.
'Based on the stratified grouping assumptions.
"Based on the 0.4-ng/m background assumption.
11
-------
Table II-6
ESTIMATED NUMBER OF PEOPLE EXPOSED TO 3SO
FROM COKE-OVEN EMISSIONS*
Subgroup
Concentration
Range"
(u2/tn3)
9-10
3-9
7-3
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
0.3-1.0
0.6-0.3
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
3a?
Background
1,800
—
50
100
2,400
1,400
1,500
1,600
430
24,000
38,000
43,000
230,000
580,000
440,000
1,200,000
640,000
1,200,000
3,400,000
3,500,000
3,100,000
20,000,000
3aP 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
6,300,000
24,000,000
3aP Back-
ground of
1 ng/m3
1,300
—
50
100
2,400
1,400
1,500
1,600
430
24,000
38,000
39,000
230,000
530,000
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
1,300
-—
—
—
—
870
1,500
1,600
—
1,000
3,700
41,000
110,000
440,000
450,000
620,000
130,000
490,000
990,000
4,400,000
7,900,000
26,000,000
All Poorly
Controlled
1,300
10
60
4,100
2,400
1,400
4,400
8,600
430
25,000
40,000
140,000
360,000
620,000
380,000
1,700,000
750,000
1,100,000
5,600,000
13,000,000
11,000,000
13,000,000
Total
44,000,000 50,000,000 44,000,000 42,000,000 53,000,000
— Mo 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 2070 to 4070 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
co 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 Concentracions. 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 Davs
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 10070 of the mean.
Appendix 3 also gives factors to adjust the standard deviation for
13
-------
individual daily means Co che 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 Apolving Observed Data from Coke Plants to Coke
Plants with Mo 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 ail plants for which
ao 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 Ba? 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 L6 equally spaced compass directions and che
14
-------
percent of time chat 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 6.2570 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 6.257. 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, chis may be expressed as:
i • (V"J
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 507«; if g = 2.0, then population exposure
concentrations for that sector are underestimated by 5070. The total
fractional error for the estimated exposures for one location can be
approximated by
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 che 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 chis determination, corre-
sponding values of 1 or 0 were assigned co P^. The cocal 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 47.. 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 47..
e. Total Error. Because of the many uncertainties, an
accurate assessment of the total error associated with the 3aP exposure
estimates cannot be made. The estimates are not expected co 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
chat represent a range of meteorological, geographical, and emission
control conditions.
2. 3SO Estimates
The 3SO exposure estimates are based on the 3a? estimates.
Thus, they contain not only the errors associated with estimating 3aP
exposures but also an additional error source caused by relating 350 to
3aP. Occupational data are evaluated in Section IV-D of chis report to
determine the magnitude of this error source. To do this, the 3aP con-
centrations were used to estimate corresponding 3SO concentrations; the
estimated concentrations were then compared with recorded 350 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 10% for the mean.
3. 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 chey either appear co be much less significant chan
che 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.*
1
3
4
5
6
7
8
10
11
12
13
14
15
16
No. of
Sectors^
8
8
10
12
9
11
7
11
11
11
11
6
13
1 J
Relative
Error**
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
Site
No.*
17
18
19
20
21
22
23
24
25
27
28
29
30
31
No. of
Sectors!
7
7
6
7
9
9
12
9
11
4
7
6
8
10
Relative
Error**
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
SLte
No.*
32
33
35
36
37
38
39
40
41
42
43
44
45
46
No. of
Sectors^
5
4
8
6
9
13
7
12
15
13
10
16
7
11
Relative
Error**
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
Site
No.*
47
49
50
51
52
53
54
55
56
57
58
60
64
No. of
Sectors^
11
10
13
11
10
11
11
8
11
9
4
9
10
Relative
Error**
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
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.
•!
Values less than 1 indicate overestimates and values more than I indicate underestimates.
-------
Foods can become contaminated because of atmospheric fallout of
parciculates 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-3 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
3ENZO(A)PYHENE CONCENTRATIONS IN FOODS
Concentration
Food ('j.g/kg) Reference
Cereals 0.3-0.8 A
Potato peelings 0.36 A
Potato tubers 0.09 A
Barley, wheat, rye 0.2-4.1 3
Cabbage 24.5 3
Spinach 7.4 C
Lettuce 2.8-12.8 3
Tomatoes 0.22 3
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 3
Roasted coffee 0.1-4.0* C
Teas 3.7-3.9* B
Whiskey 0.04* B
A—Shabad (1972)
3—Grunaner (1968)
C —IRAC (1973)
D— Gorelova (1971)
E--Andelman (1970)
L3
-------
Polynuclear aromatic hydrocarbons find cheir 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, 927, is used in blast furnaces, 57, in foundry operations, and 37,
in other types of industrial plants. Of the total coke production,
approximately 907, is produced by steel industry plants, 87, by foundry
plants, and 170 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 rr.ade 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, L976).
21
-------
for blase 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.
3. 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, L976).
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. TS? 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 0.58 to 0.38 (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 (PJtAs) 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 ?NAs 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
0.330
0.339
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
Comoound
BaP
BSO
Flu
Pyr
BcA
Chr
BaA
BbF
BjF
BkF
BeP
BaP
DBahA
BghiP
Ant
BSO
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
0.914
0.862
0.713
0.936
0.909
0.884
0.894
0.915
0.922
0.914
0. 725
0.875
0.905
1.000
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 907. of the coke
is used in blast furnace plants, whereas 27. is exported. The remainder
is primarily used in foundries. The yield of coke from coal, which
averaged 68.47. 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 stata-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.
26.
-------
Table III-3
BY-PRODUCT COKE PLANT LOCATIONS AND CAPACITIES
State. City
Alabama
1. Tarrane
2. Hole
3. Woodward
-. Gadsen
5. Thomas
6. Birmingham
7. Falrfteld
California
8. Fontana
Colorado
9. Pueblo
Illinois
10. Granite City
11. Chicago
12. Chicago
13. South Chicago
Indiana
14. Chesterton
15. Indianapolis
16. Terra Haute
17. Caac Chicago
18. East Chicago
19. Gary
20. Indiana Harbor
Kentucky
11. Ashland
Maryland
22. Sparrows Point
Michigan
23. Detroic
24. Dearborn
25. Zug Island
(Detroit)
Minnesota
26. SC. Paul
27. Duluch
Missouri
28. St. Louts
Nev York
29. Buffalo
30. Lackawana
31. Buffalo
Plant Name
Tarrant Plant
Hole Plant
Woodvard Plant
Gadsden Plant
Thomas Plant
Birmingham Plant
Fairfleld Plant
Fontana Plant
Pueblo Plant
Granite City Steel Div.
Chicago Plane
Wisconsin Steel Works
South Chicago Plane
Burns Harbor Plant
Prospect Street Plant
Tarre Haute Plane
Plant No. 2
Plant No. 3
Gary Plane
Indiana Harbor Plane
Sem.c
Sparrows Poinc Plane
Semec
Seeel Plane
Zug Island Plane
St. Paul Plane
Duluch Plane
St. Louis Plane
Harrlae Plane
Lackawana Plant
Donner-Hanna Plane
Company
Alabama Byproducts Co.
Empire Coke Co.
Hoppers Company, lac.
Republic Seeel Corp.
Republic Steel Corp.
U.S. Pipe and Foundry Co.
U.S. Steel Corp.
Kaiser Steel Corp.
CFil Steel Corp.
National Seeel Corp.
laterlake. Inc.
International Harvester Co. ,
Wisconsin Seeel Div.
Republic Steel Corp.
BechlehcB Steel Corp.
Cicizens Gas & Coke Utilicy
Indiana Gas and Chemical Corp.
Inland Steel Ca.
Inland Seeel Co.
U.S. Seeel Corp.
Yougacown Sheet and Tube Co.
Solvay Div. , Allied Chemical Corp.
Bethlehem Seeel Corp.
Solvay Div., Allied Chemical Corp.
Ford Motor Co.
Great Lakes Seeel Div. , National
Steel Corp.
Koppers Company, Inc.
U.S. Seeel Corp.
Greac Lakes Carbon Corp. , Missouri
Coke &> Chen. Div.
Semae-Solvay Div., Allied Chemical
Corp.
Bethlehem Steel Corp.
Donner*Hanna Cokai Corp.
Annual Coal
Capacity
(cons)
1.200,000
150,000
800,000
320,000
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
4,320,000
900,000
1,300,000
2,350,000
250,000
850,000
450,000
400,000
4,250,000
1.387,000
1974 Coal
Consumption
Cans)
1,760,000
900.000
643,000
2,525,000
584,338
193,000
3,096,000
1,258,000
1,750,000
4,100,000
3,385,000
25
-------
Table IZI-J (Concluded)
Stats. City
Ohio
32. Ironton
33. New Miami
34. Mlddleeown
35. Painesville
36. Portsmouth
37. Toledo
33. Cleveland
39. Maaailoa
40. Warrea
41. Youngscovn
42. Loraln
43. Campbell
Pennsylvania
44. Swede land
45. iethlanan
-6. Johnstown
47. Johnstown
43. Midland
49. Aliquippa
50. Pittsburgh
51. Erie
52. Philadelphia
53. Pittsburgh
54. Claireoa
55. Fairlese ill Is
56. Monessea
Tennessee
57. Chattanooga
Texas
53. Houston
59. Lone Scar
Utah
60. Provo
Vase Virginia
51. Weir-on
52. Weirtoa
63. Fairmont
64. Folloasbee
Wisconsin
63. Milwaukee
leaed on a 1973 m
Varga ( 19
Plane Sane
Ironton Plane
Hamilton Plane
Middletown Plane
Painesville Plane
HJnplre
Toledo Plane
Cleveland Plane
Maasiloa Plane
Warren Plane
Youngs eowa Plane
Loraln Cuyahoga Works
Campbell Plane
Alan Wood Plane
Bethlehem Plane
loaadale Olv.
Franklin Olv.
Alloy & Stainless Stael
Olv.
Aliquippa Plane
Pittsburgh Plane
Erie Plane
Philadelphia Plane
Seville Island Plane
Claire on Plane
F airless -dills Plane
Wheeling
Chattanooga Plane
Houston Plane
2. 3. Germany Plane
Geneva Works
Weirton Mainland Plane
Weirtoa1 s 3 rown's Island
Plane
Fairmont Plane
Saac Seeubeaville Plane
Milwaukee Solvay Coke
Co.
isaloa Inventory.
74)
Coooany
Scaee-Solvay Oiv. , Allied Chemical
Corp.
Armco Seeel Corp.
Armco Stael Corp.
Diamond Shamrock Corp.
Detroit Seeel Oiv. ot Cyclope
Corp. •
Incerlaka Inc.
Republic Sceel Corp.
Republic Sceel Corp.
Republic Stael Corp.
Republic Sceel Corp.
U.S. Seeel Corp.
Youngs town Sheet and Tube Co.
Alan Wood Seeel Co.
Bethlehem Sceel Corp.
aechiehem Sceel Corp.
Sachiehem Stael Corp.
Crucible lac., Olv. Colt
Industries
Jones and Laughlin Sceel Corp.
Jones and Laughiln Sceel Corp.
•Coopers Company, Inc.
Philadelphia Coke Division
Shaaango Inc.
U.S. Stael Corp.
'J.5. Steal Corp.
PLttsburgn Stael Corp.
Chattanooga Coke and Chemicals Co.
Armco Seeel Corp.
Lone Sear Sceel Co.
U.S. Sceel Corp.
Weireoa Sceel Dlv. , National Seael
Corp.
Weirtoa Seeel Olv. , National Sceel
Corp.
Sharon Steel Corp.
Wheeling-Pittsburgh Sceei Corp.
A Division of Plcklaads Mather
and Co.
3)
934,000
744,000
212,000
600,000
433,000
2,220,000
220,000
620,000
1,200,000
2,700,000
2,300,000
303,000
2,210,000
550,000
1,530.000
637,000
:,220,633
2,537,404
290,000
712.400
1,022,000
9,570,000*
1.300,000
750,000
, -oo
534,000
-93,300
2.000,000
:,joo,ooo
1,323.000
300,000
2.JOO.OOO
347,000
210,000
1,393,116
2,105,000
545,000
1,545,000
630,000
335,000
323,900
492,000
234,000
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
Mew 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
*
*
*v
3,842,000
16,313,000
*
*
*
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 PLATES
Name or Location
of Plane
Pennsylvania
1. Mahoning
2. Daugherty
3. Laughead
Virginia
5. Vansanc
6. Esserville
Countv
Armstrong
Fayecte
Fayecte
Buchanan
Wise
Company
Caipencown Coal & Coke Co.
Bortz Coal Company
Ruane Coal & Coke Company
Jewell Smokeless Coal Corp.
Christie Coal & Coke
Source: Keystone Coal Industries Manual (1975)
23
-------
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
EPA (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 EPA 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
-------
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.
3. Categorization of Coke Plants bv 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 PEDCo during September
1974 to April 1975 (Xuliyian, 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
EPA 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 EPA
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 seme ais-
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-I
ASSUMED EMISSION WEIGHTING FACTORS
FOR PLANT COMPLIANCE STATUS
Coraoliance Status
Emission Source
Charging
Doors
Pushing
Topside**
Quenching
Waste gas stacks
In
L.5
16
N*
1.6
175
N
Oue
80
130
3
65
350
0.7
Undetermined
80
130
3
65
350
0.7
Ac Lease One
Bacterv Ouc
40
73
1.5
33
260
0.4
Under a
Lezal Plan
40
73
1.5
33
260
0.4
If—Negligible.
Topside compliance was assumed co be Che same as door compliance.
Table IV-2
CLASSIFICATION OF COKE PLANTS INTO EMISSION
CATEGORIES (1974-1975)
Classification
Plant Plane
Number Classification Number
23 F 45
24 F 46
25 F 47
26 F 48
27 X 49
28 F 50
29 X 51
30 K 52
31 F 53
32 K 54
33 X 55
34 X 56
35 F 57
36 X 53
37 K 59
38 K 60
39 F 61
40 X 62
41 F 63
42 F 64
43 F 65
44 F
Plant numbers correspond co plant names given in Table III-3.
F indicat.es clean and K indicates dirty. The X indicates that insuffi-
cient data were available co classify the plant.
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
X
X
F
F
F
F
F
F
F
X
7
f
F
F
K
X
X
X
X
F
r
F
X
F
X
X
K
F
x
F
c
X
X
X
3L
-------
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 3SO 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. Mo 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/a-3 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 MASN cities without
coke ovens shown in Appendix A), and (3) the background concentration
for all locations is 1 ng/rn^ (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
3
9
10
11
12
13
14
15
16
17
13
19
20
21
22
23
24
25
26
27
23
29
30
31
32
33
34
3aF
(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
0.3
0.3
0.3
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 or 5 sites in the Los Angeles area
Spokane
rrdffiDonQ
Hammond
HBnnnond
Hammond
Uamnond
Hammond
Hammond
Hammond
Haannond
H&xQnond
Hammond
Norfolk, Charleston
Riviera Beach, Maryland
Site 30 km away
Site about 30 km away
Site about 30 km away
NASN 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 sites
Average of Pennsylvania and Ohio sites
Average of Pennsylvania and Ohio sites
33
-------
Table IV-3 (Concluded)
Plane
^hl3lbe^*
35
36
37
38
39
40
41
42
43
44
45
46
47
43
49
50
51
52
53
54
55
56
57
58
59
60
61
62
62
64
65
3aP
fag/ a-3)
0.6
0.6
0.4
0.5
0.6
0.6
0.3
0.6
0.6
0.3
0.3
0.3
0.3
0.3
0.3
1.6
0.3
0.3
1.6
1.6
0.3
0.3
0.4
0.04
0.04
0.5
0.5
0.5
0.5
0.5
0.7
Xemarks**
Average of Pennsylvania and Ohio sices
Average of Pennsylvania and Ohio aices
HASH sice
Sice about 12 ka away
Average of Pennsylvania and Ohio sices
Average of Pennsylvania and Ohio sices
Average of several Pennsylvania basins
Average of Pennsylvania and Ohio sices
Average of Pennsylvania and Ohio sices
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Average of several Pennsylvania basias
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Average of several Pennsylvania basins
Sices abouc 10 ioa away
Average of several Pennsylvania basins
Average of several Pennsylvania basias
Sices abouc 10 '
-------
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 Co 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.
D 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" ,
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 lO"6 is used, and for the X grouping a value of 1.2 x 10"^ 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' + VAD°
a
where 3' is the assumed background concentration.
The slope parameter b was again assigned a value of -1. The A para-
aeter 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"-3 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 3-2 of Appendix. 3.
(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 Donner-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 • 3aPb (2)
where
BSO is the atmospheric BSO concentration (ng/m )•
BaP is the atmospheric BaP concentration (ng/nr3).
a, b are constants.
The values of the constants were found to be as follows:
Parameter
Data Set
All data
Data for noncoke locations
Data for coke locations
The regression coefficients (R^) 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
-------
oo
40
to
*
I
S
8
t.o
u
c
UJ
O.I
T 1 I I I I II
A .* • i_
!!• •• I. ^-~=-=Z
to "
J I I Mill
1—I I I I II
J I I Mill
1—I I I I II
ALL DATA
DATA FOR NON-COKE OVEN
CITIES ONLY
J I I Mill
1 1 I I I II
10'
10'
ATMOSPHERIC B»f CONCENTRATION - ng/m1
J I 1 Mill
10'
FIGURE IV 1. RELATIONSHIP BETWEEN BSO AND B.P ATMOSPHERIC CONCENTRATIONS
FOR ALL LOCATIONS
-------
300
100
80
00
40
t
§
U 10
8 •
1 1—I I I I 111
I I I i I I i
i i i i i i i
I
T 1—I I I I
I I I I
8 10 30 40 60 80 100
BaP CONCENTRATION - ng/m3
300
400 600 800 1000
FIGURE IV-2. RELATIONSHIP BETWEEN BaP AND BSO CONCENTRATIONS FOR
LOCATIONS NEAR COKE OVENS
-------
Oaca collected in the occupational environment should provide a
better estimate of the relationship between Bap and 3SO 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 17. of the. BSO extract.
Smith (1971) presents 3a? 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. Antell (1977) supplied-BaP and
BSO data on 12 additional occupational airborne samples collected during
1974 and 1973. These data are also plotted on Figure IV-3. A statistical
c-cest 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 3a?
versus BSO relationships in the Smith data and in the Antell 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 3a? exposures due solely to coke ovens. The 3SO 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 3SO 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 plane. 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 3, 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 3ESTPOP computer system.
40
-------
10
f
E
01
U
z
0-5
0.1
TTTf
DATA SOURCE
O SMITH (1971)
Q ANTEU. (1977)
I I
5 10
8aP CONCENTRATlON-yg/m3
SO
100
TWO OF SMITHS DATA POINTS CANNOT BE SHOWN ON THIS FIGURE.
BOTH HAVE BaP-4 f/g/m3 AND BSO- OJ) 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 planes 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
- .
RVAR fdGdR = / RVAR fdQdR
solving this equation gives:
|R
Ro=^
where
V, A are as previously defined in the exposure equations
R., H_ 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 3 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 MASN 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 3aP and benzene solubles--in addition Co being a major pore."
For locations with more than one coke plane, che exposures were estimated
for overlaps of che 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/m^.
These subgroupings were made for exposures to coke-oven emissions only
and to coke-oven emissions plus background concentrations.
Population exposures Co 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 BSD 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 Studv, 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 u.g/nH,
average BaP concentrations from 2.7 to 40.8 ng/ra^, and the average BSO
from 2.6 to 9.2 ug/m^. 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
Sample
Size
Average
Ranee
Mean
TS?
Station 2
Station 6
Station 7
3aP (ng/nH)
Station 2
Station 6
Station 7
BSO (lig/ni3)
Station 2
Station 6
Station 7
29
28
31
29
28
31
29
28
31
166.0
79.0
113.0
27.0-360.0
22.0-165.0
26.0-300.0
145.0
71.0
93.0
40.8
2.7
22.8
9.2
3.3
2.6
0.3-206.4
0.2-10.8
0.4-100.3
1.5-25.4
0.6-9.1
0.9-19.3
10.0
1.6
10.1
6.5
2.6
3.3
Standard
Deviation
1.76
1.64
1.91
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 3aP.
The sampling schedule included two 10-week periods of four and rwo samples
per week, respectively (Ek, 1977).
Table A-2 shows "he results obtained during che sampling. The
average 3a? concentrations for the 11 locations varied between 1.64 and
51.95 ng/m . 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 3 to 12 hours. These data which are given
48
-------
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 Laughlln, 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-gour 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 kin of the coke
batteries showed average BaP concentrations of 1.47 to 3.31 ng/in^. Two
background stations 20 to 30 km away showed average BaP concentrations
of 0.12 and 0.83 ng/m2.
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 197& on 13 sites, in addition
Co data recorded at the National Air Surveillance Network (NASN) site.
These data, which are given in Table A-6, indicate the average 3a? con-
centrations ranged from 0.45 to 27.10 ng/m^.
50
-------
Table A-4
ATMOSPHERIC BaP CONCENTRATIONS NEAR THE
GENEVA WORKS IN UTAH
(ng/m^)
Station
Number
1
2
3
4
5
6
7
3
9
10
Location in Relation
to Battery
2.0 km NW
2.7 tan NW
2. it km NW
1.3 taa N
1.3 km ME
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
Ran?e
0.40-4.42
2.52-5.27
0.97-6.30
0.44-5.35
0.54-6.29
0.46-3.44
0.37-3.53
0.38-3.35
0.01-0.32
0.05-2.77
Table A-5
AMBIENT BaP CONCENTRATIONS FOR WAYNE COUNTY,
MICHIGAN
(ng/m3)
Site
Number 1971
No.
No.
No.
No.
No.
No.
02
04
05
06
08
11
NASN
2 is
4 is
5 is
6 is
3 is
11 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
1973
3.02
4.16
11.78
3.12
2.70
2.00
1.00
Solvay, 14.5 km NE
7.2 km NNE of Solvay,
1.6 km N
of
15.3 km NNW
10.5 km
Solvay, 4
of Solvay
SW of Solvay,
30 km SW of
Solvay,
9.3 km NE
km MNE of
, 16.5 km
8.5 km SW
1974
1
1
10
0
0
0
of G
of
G.L
NNW
of
29 km SW of G.
.46
.70
.83
.52
.44
.34
•"
.L.
G.L
• |
of
G.L
L. ,
1975
3.
4.
14.
1.
2.
0.
1.
*', and
. , 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
G.L. , and
. , and
and 30
9.3
km
67
36
52
27
13
14
33
of Ford
Ford.
E of Ford.
11.7 km
km SSW
SW of
NNW of Ford
of Ford.
Ford.
G.L. - Great Lakes Steel.
51
-------
Table A-6
AMBIENT BaP CONCENTRATIONS FOR BUFFALO, NEW YORK
Ol
K>
Site
Number
I
2
3
4
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-ll
3.8 km ME of Beth, and 1.5 km SE of D-ll
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-ll
4.3 km ME of Beth, and 1.1 km ENE of D-H
5.6 km NE of Beth, and 2.1 km NNE of D-ll
0.6 km U 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 SU of Beth, and 6.2 km S of D-ll
8.8 km NNU 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**
Kan^e
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
—
**
Ueth. ib Bethlehem Steel; D-ll Is Donner-llunita.
k
Two-year coiupoulte
-------
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/m-3 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
flDM - 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
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/m3. 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 3aP CONCENTRATIONS FOR GADSDEN, ALABAMA
Cng/m3)
Average Concentrations*
Distance from Coke Ovens
1974
1975 1976 3-Year
1.6 km £
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
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 from Batteries (km) Sample
Fairfield Birmingham Tarranc Size*
BaP (ng/m )
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
6
12
12
6
12
12
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, BSD, 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/rn-^ for the site nearest the Franklin plant
to 3.6 ng/m^ 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
-------
Table A-11
AMBIENT BaP, BSO, AND TSP CONCENT HAT IONS FOR JOHNSTOWN, PENNSYLVANIA
Sice
Number
1
2
3
4
5
6
7
8
Distance from Franklin
Coke Ovens
7.8 km USU
3.8 km W
2.9 km SW
1.0 km NNE
4.6 km SSU
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/m )
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
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/m^. BSO average concentrations ranged from 3.05 to
8. 56 (j,g/m3.
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 site's 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/m^. 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 UaP, DSO, AND TSP CONCENTRATIONS FOR PHILADELPHIA, PENNSYLVANIA
In
oo
Site Location from Philadelphia Number of
Number Coke Ovens Samples
1
2
3
4
5
6
7
8
9
10
11
12
13
NASN
14.1 km SU
3.5 km SU
19 km SU
12.5 km USU
13.7 km UNU
9.3 km NNE
5.8 km U
10 km SU
2 km UNU
8.8 km SU
13 km SSU
10 km SU
5.2 km SU
—
4
3
4
4
4
4
4
4
4
4
3
3
4
-
BaP (ng/m3)
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
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 I
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
—
(ug/m3)
BSO
5.44
4.00
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/aP)
Site Number
NW2
008
006
007
009
010
Oil
Distance from Coke Ovens Sample Size
0.7 km N 3
0.6 km SSV -3
1.1 km ME 2
2.4 km WNW 2
1.8 km NW 2
1.5 km W 1
3.5 km WNW 2
2.9 km WSW 2
Table A-14
ADDITIONAL ATMOSPHERIC AMBIENT DATA FOR GRANITE
Station Distance from Coke Ovens Pollutant 1
1
™
Site Number
256034
256015
233006
256017
236019
256028
256005
0.8 km S TSP (ng/m3) 113
3SO (iig/m3) 4,9
BaP (ng/m3) 2.1
0.5 km S TSP (ng/«3) 193
3SO (ng/m3) 17
BaP (ng/m^) 124
Table A-15
AMBIENT BaP CONCENTRATIONS FOR HOUSTON,
(ng/m3)
Distance from Coke Ovens Sample Size
1.0 km NW 6
0.9 km NNW 5<
2.2 km NE 4
0.8 km W 6
2.2 km WSW 4
0.8 km SSE 7
5.4 km SW 1
Average Range
8.
4.
2.
8.
5.
3.
12.
7.
CITY,
Dav
2
344
18
278
113
4.2
2.6
TEXAS
Avera
0.17
0.15
0.05
0.03
0.33
0.28
0.16
60 1.1-16.5
83 0.6-9.5
65 1.3-3.5
15 8.0-8.3
20 3.9-6.5
50 . «
15 1.8-22.5
15 3.9-10.4
ILLINOIS
3 Averagi
263 238
14 12
202 161
33 130
0.5 7
1.6 43
se Range
0.05-0.62
0.07-0.35
0.02-0.11
0.02-0.05
0.04-1.00
0.05-0.34
—
59
-------
Table A-16
AMBIENT Ba? CONCENTRATIONS FOR CLEVELAND, OHIO
3aP
Sice Sumber
1
3
4 .
5
6
7
3
9
10
12
13
14
15
17
20
21
Location from
Coke 3atcerr*
0.8 ka S
4.8 ka SB
4.4 ka NE
4.4 ka SE
12.0 km N£
7.2 ka W
6.3 km S«
1.0 ka SE
6.0 ka SUE
13.2 ka ISE
4.4 km S
9.6 ka SE
3.2 ka W
6.4 ka XE
16.3 ka NE
4.0 ka SOT
Samel e Size
21
37
23
28
22
38
23
30
33
32
23
22
21
32
19
22
(ag/nJ^
Geometric Mean1™
1.40
0.62
0.64
0.58
0.71
O.i6
0.44
3.60
0.74
0.43
0.35
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 ara only approximate.
•Mto
Arithmetic mean not reported.
Source: King at al. (1976)
Table A-17
AMBIENT ATMOSPHERIC 3aP AND 3SO CONCSNTSATIONS
FOR SPARROWS PODJT, MARYLAND
3aP
Distance from
3SO
Coke Batteries*
12 ka X
7 km NOT
3 kB '4
4 kB SSW
Same la 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
0.4-3.4
Average
6.5
5.4
—
4.3
Range
6.3-6.3
3.0-8.0
—
2.5-3.7
Locations are only approxiaate.
dumber of aonths 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
(t
Mean
3.83
3.49
1.63
1.85
1.55
0.82
1.23
2.35
2.66
BaP
ig/m3)
Ran^e
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
(ug/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/nH (the average is 0.38
ng/m^); however, areas with coke ovens generally have average concentra-
tions in excess of 1 ng/m^ with Ashland's 4.7 ng/m^ 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/nP. 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 lig/nr. The
average concentrations of most urban locations range from 1 to 4 y.g/mj.
Ashland, Chattanooga, Pittsburgh, Buffalo, and Hammond have concentra-
tions exceeding 5 yig/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 1963. 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 97, 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 che 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
Gadsden, AL
Gary, IN
Indianapolis, IN
Baltimore, MD
Trenton, NJ
St. Louis, MO
Youngstown, Oil
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. I
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. I
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. I
0. 5
0.5
1.0
2.0
0. 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, UI
Birmingham, AL
Jacksonville, PL
Honolulu, III
Terre Haute, IN
Ashland, KY
Baton Houge, LA
New Orleans, LA
Dearborn, MI
Uuluth, UN
Buffalo, NY
Cleveland, Oil
Bethlehem, PA
Erie, PA
Houston, TX
Newport News, RI
Norfolk, VA
Seattle, UA
St. Paul, MN
Arcadia National Park,
ML:
Haiiunond, IN
1967 1968
4
--
2
0.5 0
3.7
9
__
1.8 1
__
2
__
2.9 3
2.9 2
__
__
__
3.5 4
1.8 2
2.3 1
0
2.5 2
.7
-
.9
.6
-
.3
-
.6
-
.7
-
.0
.1
-
-
-
.9
.0
.8
.3
.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
-
I
0
2
6
-
1
-
1
-
2
2
-
1
-
1
I
I
0
I
.5
-
.4
.2
.8
.7
-
.1
-
.1
-
.8
.7
-
.2
-
.8
.5
.0
.2
.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
O.I
I. I
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
O.I
0.3
1.0
0.3
0.6
--
0.5
0.7
0.4
0.2
0.4
0.3
O.I
— —
0.2
1974
--
2.5
--
0.4
0.3
--
0.1
0.3
1.7
0.2
0.8
--
O.I
0.6
0.2
--
0.2
--
0.5
O.I
0.4
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
Birmingham, AL
Gadsden, AL
Montgomery, AL
Grand Canyon, AZ
Jacksonville, FL
Honolulu, HI
Chicago, IL
01 Gary, IN
Hammond, IN
Indlanopolls, IN
Terre Haute, IN
Ashland, KY
Batan Rouge, LA
New Orleans, LA
Baltimore, MD
Dearborn, Ml
Detroit, MI
Trenton, NJ
Duluth, MN
AL VARIATIONS OF BENZENE SOLUBLE ORGANIC SUBSTANCES (pg/m3)
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 Jb
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
O>
St. Paul, MN
St. Louis. MO
Buffalo, NY
New York, NY
Cleveland. OH
Toledo. Oil
Youngstown, Oil
Bethlehem. FA
Erie. PA
Philadelphia. PA
Chattanooga, TN
Houston, TX
Newport News, VA
Norfolk. VA
Sltennaudoah, VA
Pittsburgh, PA
Seattle. WA
Spokane. WA
Charleston. WV
Milwaukee. Wl •
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
j.O
-
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
?.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
3aP (ng/m3) 1975 data
BSO (^g/m3) 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/m^)
Year Cities With Coke Ovens Cities Without Coke Ovens
1966
1967
1968
1969
1970
1971
1972
1975
1976
*
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)
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)
67
-------
During 1976, samples Caken by this surveillance system were
also analyzed for 3aP 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/m^ for Montessen and the next highest was 17.10 ng/m-3 for
Johnstown. Both locations have coking operations. The lowest average
concentration was 0.40 ng/rn^ for Hanover Green.
3. Charleston, South Carolina
3aP 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 3aP and 3SO concentrations for many sites throughout
the state. Data, primarily for 1976, are summarized in Table A-25. The
average annual 3aP concentrations ranged from 0.43 ng/m^ for Harwood to
6 ng/nP for Catonsville.
5. CHESS and CHAMP Site Data
Atmospheric SaP and 3SO 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/m^ for
Thousand Oaks, California to 4.2 ng/nP for one site in Birmingham, Alabama.
63
-------
Table A-23
AMBIENT BaP CONCENTRATIONS FOR PENNSYLVANIA, 1976
(ng/m3)
Yearly Monthly
Average* Range
Allentown-Eastern Air Basin
Allentown 0.71 0.09-2.30
Tatamy 0.80 0.11- 2.74
Bethlehem 1.11 0.22- 4.15
Easton 1.86 0.39- 9.28
Bethlehem East 1.46 0.24- 6.34
Emmaus 1.29 0.10- 7.62
Allen Twp. 0.55 0.06- 2.24
Northampton 0.76 0.10- 3.43
(Basin average) 1.08
Beaver Valley Air Sasin
New Castle 3.06 0.13-11.36
Bessemer 1.41 0.46- 2.21
Koppel 9.43 0.30-78.08
Beaver Falls 5.03 0.42-12.65
Vanport 2.27 0.16-5.44
Rochester 4.19 0.35-13.96
Ambridge 6.18 0.75-31.96
Baden 9.00 0.40-43.48
Midland 3.13 0.31-8.60
Brighton 2.42 0.34-9.74
(Basin average) 4.73 —
Erie Air Basin
Millcreek Twp. 0.45 0.12-0.87
Erie Central 2.04 0.26-7.13
Erie South 1.16 0.21- 3.77
Erie East 1.62 0.23- 6.33
Based on one sample per mnnt-h ?<*>T 12 mnnfha.
Source: Dubin (1977)
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
Summerdale 0.61 0.14- 1.60
(Basin average) 0.32 —
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. Conenaugh 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 0.11-1.31
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 3ange
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
Dovingtown 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 0.13-4.57
Conshohocken 2.06 0.40- 3.24
Phoenixville 0.30 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 —
Alcoona 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)
Yearly Monthly
Average Range
Farrell-Sharon Area
Farrell 2.46 0.44- 8.54
Sharon 2.45 0.24- 9.22
(Area average) 2.46 —
Williamsport Area
Williamsport Central 1.02 0.23- 4.14
Williamsport East 1.28 0.15- 8.56
(Area average) 1.15
Table A-24
DISTRIBUTION OF BaP CONCENTRATIONS IN AMBIENT AIR
AT CHARLESTON, SOUTH CAROLINA*
(ng/m3)
Site Sample
Number Location Size
1 Radio Station WTMA
2 Queen St. Fire Station
3 Mt. Pleasant, Post Office
Total
22
22
22
66
Average
0.5711
0.7441
0.7448
0.6866
Range
0.0028-1.2409
0.1693-1.6787
0.1995-1.9767
0.0028-1.9767
*
There are no coke ovens in Charleston.
Source: Spangler and de Nevers (1975).
73
-------
Table A-25
AMBISOT ATMOSPHERIC BaP AND 3SO CONCENTRATIONS FOR MARYLAND LOCATIONS
Location
Cumberland
3agerstovn
Ad ana town
Frederick
>*jrersville
Buckeystovn
Glen Burnie
Ha roons
Harvood
Lintnicum
Odenton
Riviera Beach
Annapolis
Baltimore
Lexington and Gay
Sun Avenue
1900 Argonne
5700 Rei3 cars town
5700 Eastern
ronthill St.
Cockeysville Ind. ?Jc.
Cockeysville Police Station
Police Barracks
3001 Eastern Blvd.
Catonsville
Dundalk (3801 Vise Ave. )
Edgemere
Essex
Port Howard
Tooeon
Kiddie River
Dundalk (Savanaugh Rd. )
Dundalk (Reg. ?oc. Training)
Wes coins ter
Samole
Size*
12
12
7
12
7
3
11
12
12
12
12
12
11
12
12.
12
12
12
12
11
3
12
11
12
10
10
12
10
10
2
11
10
12
(
Average
4.48
1.40
0.33
1.29
0.17
0.30
1.03
0.54
0.43
0.96
0.71
0.30
0.75
1.95
2.03
1.21
1.37
1.92
1.66
0.30
0.50
0.78
1.29
5.98
1.10
1.35
1.37
2.39
0.75
1.43
1.61
3.30
0.49
BaP
ag/»3>
Range
0.40-20.22
0.20-3.37
0.31-2.12
0.13-3.55
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
0.10-1.33
0.41-4.46
0.50-4.43
0.27-3.51
0.42-4.76
0.38-5.36
0.37-4.30
0. 09-4. 33
0.12-1.76
0.10-2.06
0.13-3.97
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
Samole
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
(y
Average
3.44
4.74
—
4.30
«
--
4.34
—
2.93
3.99
3.69
4.25
3.76
6.27
5.38
4.21
5.26
6.38
5.24
—
4.11
4.79
4.39
4.02
—
"
5.35
4.79
5.48
6.52
—
--
3.07
3SO
.d/a^)
Range
6.03-13.14
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
4.34-3.64
3.35-9.50
2. 51-7. -24
3.25-8.11
4.25-10.06
3.01-4.63
--
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
74
-------
Table A-25 (Concluded)
Location
Gaithersburg
Silver Spring (1901 Randolph)
Kensington
Poolesville
Silver Spring (Rock Creek Forest)
Rockville
Bethesda
Accokeek
Cheverly
Largo
Laurel
Orae
Oxon Hiil
Laplata
Elkton
Cambridge
Salisbury
Number of aonths of data used in calculating the average and range.
i tuple
ize*
10
10
11
11
4
12
4
2
11
1
10
10
8
10
12
11
9
(r
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
BaP
ig/ta3>
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
0.18-1.53
0.10-1.59
Sample
Size*
10
11
11
12
--
12
—
—
11
—
10
10
3
10
12
11
10
1
Averagi
3.60
4.33
4.26
3.36
--
4.51
—
--
4.57
--
3.56
3.56
4.03
3.27
4.60
4.05
4.46
BSO
! ' 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
3.08-5.30
2.38-4.95
3.55-6.76
2.55-5.30
3.03-5.50
75
-------
Table A-26
ATMOSPHERIC BaP AND BSO CONCENTRATIONS FOR
CHESS AND CHAMP SITES (1975 DATA)
Sample BaP (ng/m3) BSO
Location *
Charlotte, NC (1)
Charlotte, NC (2)
Riverhead, NY
Queens, NY
Brooklyn, NY
Bronx, NY
Ogden, Ut
Sale 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
Size**
6
6
12
12
12
12
12
12
12
12
12
6
12
12
12
12
12
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
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
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
Range
1.1-2.3
2.2-3.8
0.6-2.0
0.9-2.3
1.1-7.6
2.0-3.6
0.7-8.8
1.6-7.7
0.7-3.2
0.5-3. 4
0.3-6.7
1.1-6.1
1.1-4.3
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.
**
Mumber of months for which data are available; sample size ror
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
•a 100
z
o
o
u
0.
o
cc
Ul
a
10
0 JOHNSTOWN STATION No. 7
Q MONESSEN STATION No. 2
I
j i
10 20 40 60
CUMULATIVE PERCENTAGE
30
90
98%
FIGURE 3-t. STATISTICAL DISTRIBUTION FOR ATMOSPHERIC
BaP CONCENTRATIONS
80
-------
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
-------
CO
Tahle U-l
STATISTICAL SUMMARY FOH SAMPLING DATA TAKEN FKOM A NUMIiKK OF LOCATIONS
Arithmetic Geomctrii
Sampling
Location
Monessen, 2
Monessen, 6
Monessen, 7
Johnstown, I
Johnstown. 2
Johnstown, 3
Johnstown, 4
Johnstown, 5
Johnstown, 6
Johnstown, 7
Johnstown, 8
Utah, I
Utah, 2
Utah, 3
Utah, 4
Utah, 5
Utah, 6
Utah, 7
Utah, 8
Utah, 9
Utah, 10
Gadsden, 1
Gadsden, 2
Dulutli, 1
Duluth, 2
Sample
Size
29
28
31
30
32
33
32
28
31
34
31
9
6
9
I
I
I
3
1
1
9
5
5
18
20
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. I
1.6
2. I
1.5
0. I
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
I. 5
f 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.
f = V .. " n .
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
n
!
a
£
c
K
Z
IU
-------
oo
10
n
O
H
UJ
u
o
o
u
-------
100
00
I
I
a 10
§
o
-------
100
00
r»
z
o
UJ
o
z
o
o
o
5
UJ
a.
1.0
0.1
T 1 I I I I I I
O
J I I I i I I I
1 I T
0
0
O
i.o
Q El
I I I I I I I I
1 I 1—I I I I I_l
O RELATION TO U.S. STEEL
Q RELATION TO J&L
10
J I I I I I M
100
OlSTANCe FROM COKE PLANT - km
FIGURE B-6. ATMOSPHERIC CONCENTRATIONS OF BaP FOR ALLEGHENY COUNTY. PENNSYLVANIA
-------
100
10
I I I I I I
I I I I I I i I
O RELATED TO IETHL£HEM
Q RELATED TO DONNER-HANNA
£ RELATED TO ALUED
c
I
O
Z
UJ
CJ
a
UJ
z
o
< 1.0
r-
0.1
B
0.1
1.0
DISTANCE PROM COKE PLANT - bn
10
FIGURE B-6. ATMOSPHERIC CONCENTRATIONS OF BaP FOR BUFFALO, NEW YORK
-------
oo
10
I
o
p
1
uj
o
u
E
UJ
I
8
0.1
0.1
O RELATIVE TO TARRENT
Q RELATIVE TO FAIRFIELO
RELATIVE TO BIRMINGHAM
I I I I i I I
T 1 1—I I I I I
I I I I I I I
I i I I I I 1
1.0
10
100
DISTANCE FROM COKE PLANT - km
FIGURE B 7. ATMOSPHERIC CONCENTRATIONS OF BaP FOR BIRMINGHAM. ALABAMA
-------
50
10
1 T
O SAMPUNG PWIOO 2
Q SAMPUNG P€RIOO 1
O
z
UJ
u
O
u
s
UJ
|
O
G
0
0.1
0
0
I I I
0.1
1.0
DISTANCE FROM COKE PLANT - km
10
FIGURE 8-8. ATMOSPHERIC CONCENTRATIONS OF BaP FOR GRANITE CITY. ILUNOIS
90
-------
u
o
u
o
o
I I 1 I I I I L
I I I I I I I I
I I I I I i
I I i l i i i i
i.o 10
DISTANCE FROM COKE PLANT - km
100
FIGURE B 9. ATMOSPHERIC CONCENTRATIONS OF BaP FOR SPARROWS POINT. MARYLAND
-------
100
-MM
10
1 I
I ! i
Q ARITHMETIC AVERAGES tASED
ON GEOMETRIC AVERAGES AND
STANDARD DEVIATIONS
n
c
i
Ul
U
O
u
x
in
Q
1.0
O
0.1
1.0
0
0
0
0 0
10
DISTANCE FROM COKE PLANT - Ion
100
FIGURE 8-10. ATMOSPHERIC CONCENTRATIONS OF BaP FOR CLEVELAND, OHIO
92
-------
100
10
I
o
P
oc
H
UJ
u
O
(J
c
HI
i
1.0
0.1
I I I I I I
I I I I I U
I I I I I
0.1
1.0
DISTANCE FROM COKE PLANT - km
10
FIGURE 8-11. ATMOSPHERIC CONCENTRATIONS OF BaP FOR MONESSEN, PENNSYLVANIA
93
-------
4
o
H
x
8 i.o
o
u
u
£
0.1
0.1
1.0
DISTANCE FROM COKE PLANT - km
10
FIGURE 3-12. ATMOSPHERIC CONCENTRATIONS OF B*P FOR GAOSOEN, ALABAMA
94
-------
z
o
cc
z
111
(J
o
o
u
£
-------
VO
to
n
i
CONCENTRATION -
b
o
ATMOSPHERI
0.1
0
r- 1 1 1 1 1 1 1 1
—
—
I
—
—
ALL UPWIND STATIONS
1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1
-— °G 0(
0 ' TD
(
O
1 1 1 1 1 1 1 I
1 1 1 1 1 1 1 I
GO
r^^-
o
0 ~
—
—
—
1 1 ! 1 1 1 1 1
1 1.0 10 10
DISTANCE FROM COKE PLANT - km
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 = aD ,
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 SaP
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
Johnstown
Gadsden
Duluth
Monessen
Utah
Wayne County
Buffalo — Beth.
Buffalo— D.H.
Buffalo— Allied
Cleveland
Pittsburgh— USS
Pittsburgh— J&L
T arrant
Granite City
Sparrows Point
Fairfield
Dirty Plant Model
**
Clean Plane 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
3.40
1.96
12.42
22.00
4.32
3.98
11.50
3.07
2.67
135.34
60.66
b C<
-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
sefficient (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.
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
y • x
3 ' logarithms.
Stj--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--J6eL
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
L01
-------
Appendix C
DETAILED ESTIMATES OF POPULATIONS AND Ba? 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/m^. 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
-------
Till) I
C-l
DETAIl.lil) Uul* POPULATION KXI'USUKKS
(Cuke Eiuittiiloiiij (i lua background)
(•IB/I"3)
Distance froia Coke Fuel Illy (km)
Situ
..No._
,
2
3
4
5
6
7
8
9
10
11
12
11
14
15
16
1 7- 1 81*
19
20
21
22
23
24
25
26
27
28
2'J
30
31
32
31
34
0-
Popu-
lat Ion
388
0
0
478
1.656
975
0
0
0
0
827
0
2.618
0
57
' 512
0
0
•J9
552
il
0
0
0
991
0
3,368
0
0
1,184
0
6
0
0. 5
Concen-
t rat Ion
A
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.
Poun-
latlon
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
Concen-
t rat Ion
*
1.0
13.2
2.0
1.1
A
9.7
9.9
5.6
1.4
4.2
16.6
2.9
10.5
3.2
1.5
76.6
60.0
34 . 3
26.0
3.5
Ik
A
*
1.3
2.0
2.0
2.1
A
*
20.3
15.5
12.6
1-3
Popu-
latlon
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
Concen-
t 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
*
A
*
0.8
0.8
0.9
1.6
A
A
8.0
6.2
5.1
3-7
Popu-
lut loii_
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
A
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
A
A
A
0.5
0.3
0.6
1.2
A
A
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-
tration
A
0.4
1.3
0.4
0.4
A
A
1.8
0.9
0.7
0.9
l.B
0.9
1.4
0.9
0.8
5.9
4.7
3.0
2.1
0.8
A
*
A
0.5
0.3
0.4
—
A
A
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-
t rat Ion
A
0.4
0.9
0.4
0.4
A
A
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
A
*
A
0.4
0.3
0.4
*
*
A
1.4
1.2
1.1
V
1.
1
1
4
3
1.5
1
l.«
3
1
1.5
7
l.<
£
1.2
1.1
1.3
3
2
1
1
4
20-30
Popu- Concen-
|atloa__ tratIon
132.291 *
81.950 0.4
150.334 0.8
33.966 0.4
65.967 0.4
97.546 *
95.069 *
428.754 1.5
1,661 0.7
741,093 0.7
1.581.162 0.8
100.951 1.2
1.840.707 0.8
324.094 1.0
122.293 0.8
41.807 0.7
1.521.807 3.0
785,467 2.5
1.456,265 1.7
57.907 1.2
617.561 0.8
20.338 *
39.598 *
89,647 *
3.605 0.4
33.818 0.3
716.207 0.4
226,767 *
194,061 *
160,244 *
95,537 1.2
430,148 1.0
92,077 1.0
-------
Table C-l (Conclude.!)
Uiulance trou Coke
0-0.5
Site
Mo.
15
16
17
IB
19
40
41
42
41
44
45
46-4 7»»
48
44
SO
51
52
51
54
55
56
57
58
59
60
6l-62b
61
64
65
Popu-
1 UL Inn
0
0
66)
1.510
0
0
0
n
0
0
1.960
1.804
0
0
0
2.62B
2.411
0
1.024
2,1/4
0
1.185
6
0
19
0
0
1
U54
Cone cu-
ll ul lull
2.1
1B.6
11.5
42.0
2.4
20.1
11.1
19.5
16.7
6.4
16.1
100.0
5.4
16.5
*
2. B
22.1
*
*
54. B
50.0
1.8
17.6
1.5
10.0
110.1
9.5
IB.O
1.2
0.
Popu-
lation
0
0
1.597
4.226
71
0
1.9B6
2.858
4.074
0
1.591
870
0
0
10.170
10.041
4.526
75
1.165
1.757
1.559
5.481
0
0
0
0
0
0
1.900
5-1
Conceu-
llul lun
1.4
10.2
7.4
25.0
1.5
li.O
6.4
10.7
9.2
1.8
9.1
50.0
1.2
9.2
*
1.9
12.2
*
*
29.6
ia.o
1.2
9.4
1.9
5.8
69.7
5.1
9. a
2.0
1-1
Popu-
lation
10.961
5.152
22.961
71.580
17.661
20,260
45.214
72.578
11.721
26.142
42.961
15.460
571
19.922
115.172
56.241
54.405
11.991
16,819
10.671
22.761
16.595
4.866
1.046
1.044
0
0
20.971
71.254
Coucen-
[ rut Ion
0.9
4.2
1.0
4.5
1.0
4.5
2.9
4.4
i.a
1.9
1.9
17.0
1.7
4.0
*
1.2
5.1
*
*
11.6
10.0
0.7
1.5
0.7
2.6
26.5
2.1
4.0
1.2
1-
Popu-
lation
10.501
14.067
155.445
199.565
40.897
82,028
147.771
178.615
115,698
110.829
64,472
51.502
12.276
61.171
196.226
87.797
486.896
122,412
81,779
117,606
42.498
81.104
99.968
2.219
28,887
1.570
0
19.901
128.749
7
Conccii-
trallou
0.7
2.0
l.S
2.1
0.7
2.2
1.6
2.1
1.9
1.2
2.0
5.2
1.2
2.1
*
1.0
2.5
A
*
5.1
2.0
0.5
1.4
0.1
1.2
10.9
1.2
1.9
0.9
Facility (km)
7-15
Popu-
lat lOII
50.851
24.211
258.780
859,264
152.877
96.129
111.429
860.216
164.425
714.918
229.291
18.180
91.115
116,901
748,696
40,701
1.685.267
612.088
407.475
154.575
80,159
117.888
150.402
8.902
72.121
8.598
1,410
65.011
601.106
Conceii-
t rat Inn
0.7
1.1
0.9
0.5
0.7
1.1
1.2
.1
.2
.0
.4
.2
.0
.4
*
0.9
1.6
A
*
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
281.161
92,928
111.626
87.511
111,924
95,120
1.270,155
125.142
21.466
128.565
86.055
207.185
10,856
670.578
416.161
594,166
214,800
91.526
57,125
420.504
4.884
2.800
10.057
2.094
28,881
127.088
Couceik-
1 ration
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
1.1
*
*
2.0
0.8
0.4
0.4
0.2
0.5
1.5
0.7
0.9
0.8
20-10
Popu-
lul Ion
177.825
24.710
97,007
104.661
197.211
260,411
251.967
298.122
252.616
1.587.922
142.000
61.508
179.905
471.960
165.211
12,420
1,026,074
576,450
654.955
811.846
698.784
49.186
685,288
14,240
41,916
165.797
10.1)04
81,458
176.512
Coitcen-
t rut Itiu
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
0.1
0.1
0.5
2.6
0.6
0.8
0.7
Situ Humbert! correspond Co coke facilities lifted In Table 111-1.
Inillualed tliul I lie two facilities were treated ad though lliure were collated.
*
luUlcaleu lliul. one or more coke taclllllea arc located wlllilu 15 km of that facility. Kut liuuleil concentrations are given
In T.ible C-2.
-------
Table C-2
BaP EXPOSURES FOR PERSONS IN LOCATIONS
HAVING MORE THAN ONE COKE FACILITY
Exposure
Concentration
Location Exposed Population (ng/m-Q
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 Scael 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 3. 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 3enzo(a)pyrene (3a?)" (Draft
May 1978).
Federal Register, "Exposure to Coke Oven Emissions," Occupational Safety
and Health Administration, 41/206):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. Aooth.-Ztg. 108(16):529-533 (1968).
Jungers, R. H., "3aP Analysis of Duluth Samples," memorandum to J. Manning,
U.S. Environmental Protection Agency (February 2, L977A).
107
-------
Jungers, R. H., "BaP Analysis, Alabama," memorandum to J. Manning, U.S.
Environmental Protection Agency (February 4, 1977B).
Xing, 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 BSO," 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 (PNAs) 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. Hv°.
Assoc. J. 36(2):131-139 (February 1975).
Shabad, L. M., and Y. L. Cohan, "Contents of 3enzo(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 Polycyciic Organic Matter" (1974).
U.S. Environmental Protection Agency, "Scientific and Technical Assess-
ment Report on Particulate Polycyciic 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
------- |