450585002
vvEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/5-85-002
April 1985
          Air
Inorganic Arsenic
Risk Assessment
For  Primary and
Secondary  Lead
Smelters, Primary
Zinc Smelters,
Oxide Plants,
Cotton Gins, and
Arsenic Chemical
Plants

-------
9

-------
                        NOTICE
Thisdocument has not been formally released by EPA and should not now be construed to represent Agency
policy. It is being circulated for comment on its technical accuracy and policy implications.
                                   EPA-450/5-85-002
  Inorganic Arsenic Risk Assessment for
  Primary and Secondary Lead Smelters,
Primary Zinc Smelters, Zinc Oxide Plants,
Cotton Gins, and Arsenic Chemical Plants
                 Strategies and Air Standards Division
               U.S. ENVIRONMENTAL PROTECTION AGENCY
                    Office of Air and Radiation
               Office of Air Quality Planning and Standards
               Research Triangle Park, North Carolina 27711

                        April 1985

-------
This report has been reviewed by the Strategies and Air Standards Division of the Office of Air Quality
Planning and Standards, EPA, and approved for publication. Mention of trade names or commercial products
is not intended to constitute endorsement or recommendation for use. Copies of this report are available
through the Library Services Office (MD-35), U.S. Environmental Protection Agency, Research Triangle
Park,  N.C. 27711, or from National Technical Information Services, 5285 Port Royal Road  Sprinqfield
Virginia 22161.                                                                    '      a

-------
                             TABLE OF CONTENTS

Title                                                                 Page

  1  INTRODUCTION 	  1

     1.1      Overview	1
     1.2      The Relationship of Exposure to Cancer Risk	1
     1.3      Public Exposure	4
     1.4      Public Cancer Risks  	  5

  2  THE UNIT RISK ESTIMATE FOR INORGANIC ARSENIC	    6

     2.1      The Linear No-Threshold Model for Estimation of
               Unit Risk Based on Human Data (General)	6
     2.2      Unit Risk Estimates Derived from Epidemiologic  Studies  .  9

  3  QUANTITATIVE EXPRESSIONS OF PUBLIC EXPOSURE TO INORGANIC ARSENIC
     EMISSIONS	13

     3.1      EPA's Human Exposure Model (HEM) (General)  	 13
     3.1.1    Pollutant Concentrations Near A Source 	 13
     3.1.2    Expansion of Analysis Area	14
     3.2      Methodology for Reviewing Pollutant Concentrations  ... 15
     3.2.1    Use of Ambient Data	19
     3.2.2    The People Living Near A Source	19
     3.2.3    Exposure	20
     3.3      ASARCO-East Helena  	 22
     3.3.1    Public Exposure to Inorganic Arsenic Emissions  from
               Primary Lead Smelters	24
     3.3.1.1  Source Data	24
     3.3.1.2  Exposure Data	24
     3.4      Murph Metals-Dallas and Quemetco-Seattle 	 29
     3.4.1    Public Exposure to Inorganic Arsenic Emissions  from
               Secondary Lead Smelters	32
     3.4.1.1  Source Data	32
     3.4.1.2  Exposure Data	32
     3.5      Public Exposure to Inorganic Arsenic Emissions  from
               Primary Zinc Smelters	39
     3.5.1    Source Data	39
     3.5.2    Exposure Data	39
     3.6      Public Exposure to Inorganic Arsenic Emissions  from
               Zinc Oxide Plants	45
     3.6.1    Source Data	45
     3.6.2    Exposure Data	45
     3.7      Methodology for Reviewing Pollutant Concentrations  -
               Cotton Gins	51
     3.7.1    Public Exposure to Inorganic Arsenic Emissions  from
               Cotton Gins	53
     3.7.1.1  Source Data	53
     3.7.1.2  Exposure Data	53

                                     iii

-------
Title

     3.8      Public Exposure to Inorganic Arsenic  Emissions  from
               Arsenic Plants .....  	   68
     3.8.1    Source Data	   68
     3.8.2    Exposure Data	   68

  4  QUANTITATIVE EXPRESSIONS OF PUBLIC  CANCER  RISKS  FROM INORGANIC
     ARSENIC EMISSIONS 	   74

     4.1      Methodology (General)  	   74
     4.1.1    The Two Basic Types of Risk	   74
     4.1.2    The Calculation of Aggregate Risk	   74
     4.1.3    The Calculation of Individual  Risk	   76
     4.2      Risks Calculated for Emissions of Inorganic Arsenic ...   76

  5  ANALYTICAL UNCERTAINTIES APPLICABLE TO THE CALCULATION  OF  PUBLIC
     HEALTH RISKS CONTAINED IN THIS  DOCUMENT 	  85

     5.2      Public Exposure	86
     5.2.1    General	86
     5.2.2    The Public	87
     5.2.3    The Ambient Air Concentrations	88

  6  REFERENCES	90

-------
                               LIST OF  TABLES
Table
  1       Summary of Quantitative Risk  Analyses  	  10
  2       Combined Unit Risk Estimates  for Absolute Risk  Linear Models.  .  12
  3       Arsenic Concentrations Near ASARCO-East Helena  Primary
           Lead Smelter	16
  4       Identification of Primary Lead Smelters 	  25
  5       Input Data to Exposure Model  Primary Lead Smelting Industry
            (Assuming Baseline Controls) 	  26
  6       Total Exposure and Number of People  Exposed Primary Lead
           Smelting Industry  	  27
  7       Public Exposure for Primary Lead Smelting Industry as
           Produced by the Human Exposure Model  (Assuming Baseline
           Controls)	28
  8       Arsenic Concentrations Near Select Secondary Lead Smelters  .  .  35
  9       Identification of Secondary Lead Smelters  	  34
  10       Secondary Lead Industry  Inputs to HEM Model
           (Assuming  Baseline Controls)  	  35
  11       Total  Exposure and Number of  People Exposed Secondary Lead
           Smelting Industry  	  37
  12       Public Exposure for Secondary  Lead Smelters as Produced
           by  the Human Exposure Model  (Assuming Baseline Controls)  ...  38
  13       Arsenic Concentrations Near  Select Primary Zinc Smelters   ...  40
  14       Identification of Primary Zinc Smelters   	  41
  15       Input  Data  to Exposure Model  Primary Zinc  Smelting  Industry
            (Assuming  Baseline Controls)  	 42
  16       Total  Exposure and  Number of  People Exposed Primary
           Zinc  Smelter	43
  17       Public Exposure  for Primary  Zinc  Smelters  as Produced by  the
           Human Exposure  Model (Assuming  Baseline Controls)   	 44
  18       Arsenic  Concentrations  Near  Select  Zinc  Oxide  Plants   	 46
  19        Identification  of Zinc Oxide Plants	47
                                       v

-------
Table                                                                    Page

 20       Input Data to Exposure  Model  Zinc Oxide Plants
           (Assuming Baseline  Controls)  	  48

 21       Total Exposure and Number  of  People Exposed  (Zinc Oxide
           Plants)	49

 22       Public Exposure for  Zinc Oxide  Plants as Produced by the
           Human Exposure Model  (Assuming Baseline Controls)  	   50

 23       Arsenic Concentrations  Near Two Texas Cotton Gins   ......   52

 24       Identification of  Model Cotton  Gins   	 .....   54

 25       Input Data to Exposure  Model  Cotton Gins (Assuming  Baseline
           Controls)	  .   55

 26       Public Exposure for  4 Bales/Hour Model Cotton Gin (Hutto.TX)
           as Produced by the  Human  Exposure Model (Assuming  Baseline
           Controls)	56

 27       Public Exposure for  7 Bales/Hour Model Cotton Gin (Hutto.TX)
           as Produced by the  Human  Exposure Model (Assuming  Baseline
           Controls)	57

 28       Public Exposure for  12  Bales/Hour Model Cotton Gin  (Hutto.TX)
           as Produced by the  Human  Exposure Model (Assuming  Baseline
           Controls)	58

 29       Public Exposure for  20  Bales/Hour Model Cotton Gin  (Hutto.TX)
           as Produced by the  Human  Exposure Model (Assuming  Baseline
           Controls)	59

 30       Public Exposure for  4 Bales/Hour Model Cotton Gin (Buckholts,
           TX) as Produced by  the Human Exposure Model (Assuming
           Baseline Controls)  ....  	   60

 31       Public Exposure for  7 Bales/Hour Model Cotton Gin (Buckholts,
           TX) as Produced by  the Human Exposure Model (Assuming
           Baseline Controls)  	   61


 32       Public Exposure for  12  Bales/Hour Model Cotton Gin  (Buckholts,
           TX) as Produced by  the Human Exposure Model (Assuming
           Baseline Controls)  	   62

 33       Public Exposure for  20  Bales/Hour Model Cotton Gin  (Buckholts,
           TX) as Produced by  the Human Exposure Model (Assuming
           Baseline Controls)  	   63

                                     vi

-------
34       Public  Exposure  for  4 Bales/Hour Model Cotton Gin  (Itasca.TX)
          as  Produced  by  the  Human  Exposure Model  (Assuming Baseline
          Controls)  ..........................  64

35       Public  Exposure  for  7 Bales/Hour Model Cotton Gin  (Itasca.TX)
          as  Produced  by  the  Human  Exposure Model  (Assuming Baseline
          Controls)  ..........................  65

36       Public  Exposure  for  12 Bales/Hour Model Cotton Gin (Itasca.TX)
          as  Produced  by  the  Human  Exposure Model  (Assuming Baseline
          Controls)  ..........................  66

37       Public  Exposure  for  20 Bales/Hour Model Cotton Gin (Itasca.TX)
          as  Produced  by  the  Human  Exposure Model  (Assuming Baseline
          Controls)  ..........................  67

38       Arsenic Concentrations Near Select Arsenic  Chemical  Plants  .  .  69

39       Identification of Arsenic  Chemical Plants ..........   70

40       Input Data  to Exposure Model Arsenic  Chemical  Plants
          (Assuming  Baseline  Controls) ................   71

41       Total Exposure and Number  of People Exposed (Arsenic
          Chemical Plants) ......................   72

42       Public Exposure  for  Arsenic Chemical  Plants as  Produced  by
          the Human  Exposure  Model  (Assuming Baseline Controls)   ...   73

43       Maximum Lifetime Risk and  Cancer  Incidence  for  Primary Lead
          Smelters (Assuming  Baseline Controls) ............   78

44       Maximum Lifetime Risk and  Cancer  Incidence  for Secondary
          Lead Smelters  (Assuming Baseline  Controls) .........   79

45       Maximum Lifetime Risk and  Cancer  Incidence  for  Primary
          Zinc Smelters  (Assuming Baseline  Controls) .........   80

46       Maximum Lifetime Risk and  Cancer  Incidence  for Zinc  Oxide
          Plants  (Assuming Baseline Controls)  ............   81

47       Maximum Lifetime Risk and Cancer  Incidence  for Model Cotton
          Gins  (Assuming Baseline Controls) .............   82
 48       Lifetime Risk for Two Texas Cotton Gins (Assuming Baseline
          Controls)  ..........................  83

 49       Maximum Lifetime Risk and Cancer Incidence for Arsenic
          Chemical Plants (Assuming Baseline Controls) ........   84

                                    vii

-------
                              LIST OF FIGURES

Figure                                                                   Page

  1       Group 2 BG/ED Interpolation 	   17

  2       Predicted Versus  Measured Inorganic Arsenic  Ambient
           Concentrations  (ASARCO-East Helena, MT)   	   23

  3       Predicted Versus  Measured Inorganic Arsenic  Ambient
           Concentrations  (Murph Metals-Dallas, TX)  	   30

  4       Predicted Versus  Measured Inorganic Arsenic  Ambient
           Concentrations  (Quemetco,  Seattle, WA)  	   31
                                     vm

-------
INORGANIC ARSENIC RISK ASSESSMENT FOR PRIMARY AND SECONDARY  LEAD BELTERS,
PRIMARY ZINC SMELTERS AND ZINC OXIDE  PLANTS,  COTTON GINS AND ARSENIC  CHEMICAL
PLANTS

1   INTRODUCTION

1.1  Overview

     The quantitative expressions of public cancer risks presented in this
document are based on (1) a dose-response model  that numerically relates
the degree of exposure to airborne inorganic  arsenic to the  risk of getting
lung cancer, and (2) numerical expressions of public exposure to ambient
air concentrations of inorganic arsenic estimated to be caused by emissions
from stationary sources.   Each of these factors  is discussed briefly  below
and details are provided in the following sections of this document.

1.2  The Relationship of Exposure to Cancer Risk

     The relationship of exposure to the risk of contracting lung cancer  is
derived from epidemiological studies in occupational settings rather  than
from studies of excess cancer incidence among the public.  The epidemiological
methods that have successfully revealed associations between occupational
exposure and cancer for substances such as asbestos, benzene, vinyl chloride,
and ionizing radiation, as well as for inorganic arsenic, are not readily
applied to the public sector, with its increased number of confounding
variables, much more diverse and mobile exposed  population,  lack of consoli-
dated medical records, and almost total absence  of historical exposure
data.   Given such uncertainties, EPA considers it improbable that any
association, short of very large increases in cancer, can be verified in
the general population with any reasonable certainty by an epidemiological
study.  Furthermore, as noted by the National Academy of Sciences (NAS)1,
"...when there is exposure to a material, we  are not starting at an origin

-------
of zero cancers.  Nor are we starting at an origin  of zero  carcinogenic
agents in our environment.  Thus, it is likely that any  carcinogenic  agent
added to the environment will  act by a particular mechanism on  a  particular
cell population that is already being acted on by the same  mechanism  to
induce cancers."  In discussing experimental  dose-response  curves,  the NAS
observed that most information on carcinogenesis  is derived from  studies  of
ionizing radiation with experimental animals  and  with humans which  indicate
a linear no-threshold dose-response relationship  at low  doses.  They  added
that although some evidence exists for thresholds "in some animal  tissues,
by and large, thresholds have  not been established  for most tissues.  NAS
concluded that establishing such low-dose thresholds "...would  require
massive, expensive, and impractical  experiments  ..."  and recognized that
the U.S. population "...is a large, diverse,  and  genetically heterogeneous
group exposed to a large variety of toxic agents."   This fact,  coupled with
the known genetic variability  to carcinogenesis and the  predisposition of
some individuals to some form  of cancer, makes it extremely difficult, if
not impossible, to identify a  threshold.

     For these reasons, EPA has taken the position, shared  by other Federal
regulatory agencies, that in the absence of sound scientific evidence to
the contrary, carcinogens should be considered to pose some cancer  risk
at any exposure level.   This no-threshold presumption is based  on the view
that as little as one molecule of a carcinogenic  substance  may  be sufficient
to transform a normal  cell into a cancer cell.  Evidence is available from
both the human and animal  health literature that  cancers may arise  from a
single transformed cell.  Mutation research with  ionizing radiation in cell
cultures indicates that such a transformation can occur  as  the  result of
interaction with as little as  a single cluster of ion pairs.   In  reviewing
the available data regarding carcinogenicity, EPA found  no  compelling
scientific reason to abandon the no-threshold presumption for inorganic
arsenic.

-------
     In developing the exposure-risk relationship for inorganic  arsenic,  EPA
has assumed that a linear no-threshold relationship  exists at and  below  the
levels of exposure reported in the epidemiological  studies of occupational
exposure.   This means that any exposure to inorganic arsenic  is  assumed
to pose some risk of lung cancer and that the linear relationship  between
cancer risks and levels of public exposure is the same as that between cancer
risks and levels of occupational  exposure.   EPA believes that this assumption
is reasonable for public health protection in light  of presently available
information.   However, it should be recognized that  the case  for the  linear
no-threshold dose-response relationship model  for inorganic arsenic is not
quite as strong as that for carcinogens which interact directly  or in
metabolic form with DMA.   Nevertheless, there is no  adequate  basis for
dismissing the linear no-threshold model  for inorganic arsenic.  Assuming
that exposure has been accurately quantified, it is  the Agency's belief
that the exposure-risk relationship used by EPA at low concentrations
represents only a plausible upper-limit risk estimate in the  sense that  the
risk is probably not higher than the calculated level  and could  be much
lower.

     The numerical constant that defines the exposure-risk relationship
used by EPA in its analysis of carcinogens is called the unit risk estimate.
The unit risk estimate for an air pollutant is defined as the lifetime cancer
risk occurring in a hypothetical population in which all individuals  are
exposed throughout their lifetimes (about 70 years)  to an average  concentration
of 1 ug/m3 of the agent in the air which they breathe.   Unit  risk  estimates
are used for two purposes:  11) to compare the carcinogenic potency of several
agents with each other, and (2) to give a crude indication of the  public
health risk which might be associated with estimated air exposure  to  these
agents.

-------
     The unit risk estimate for inorganic arsenic that is used in this
appendix was prepared by combining the five different exposure-risk numerical
constants developed from four occupational  studies.2 The methodology used
to develop the unit risk estimate from the  four studies is described in
Section 2 below.

1. 3  Public Exposure

     The unit risk estimate is only one of  the factors needed to  produce
quantitative expressions of public health risks.   Another factor  needed
is a numerical expression of public exposure,  i.e.,  the numbers of
people exposed to the various concentrations of inorganic arsenic.   The
difficulty of defining public exposure was  noted by  the National  Task
Force on Environmental Cancer and Health and Lung Disease in  their 5th
Annual  Report to Congress,  in 1982. 3  They  reported  that "...a large
proportion of the American  population works some distance away from their
homes and experience different types of pollution in their homes,  on the
way to and from work, and in the workplace.   Also, the American population
is quite mobile,  and many people move every few years."  They also noted the
necessity and difficulty of dealing with long-term exposures  because of
"...the long latent period  required for the development and expression
of neoplasia [cancer]..."  The reader should note that the unit risk estimate
has been changed from that  value used in the inorganic NESHAP proposal  as a
result of EPA's analysis of several  occupational  epidemiological  studies that
have recently been completed.

     EPA's numerical expression of public exposure is based on two estimates.
The first is an estimate of the magnitude and  location of long-term average
ambient air concentrations  of inorganic arsenic in the vicinity of emitting
sources based on dispersion modeling using  long-term estimates of source
emissions and meteorological conditions.  The  second is an estimate of the
number and distribution of  people living in the vicinity of emitting sources
based on 1980 Bureau of Census data which "locates"  people by population

-------
centrolds in census tract areas.   The people and  concentrations are  combined
to produce numerical  expressions of public exposure by an  approximating
technique contained in a computerized model.   The methodology  is described
in Section 3 below.

1.4  Public Cancer Risks

     By combining numerical  expressions of public exposure with the  unit
risk estimate, two types of numerical expressions of public cancer risks are
produced.  The first, called individual risk, relates to the person  or
persons estimated to live in the area of highest  concentration as estimated
by the computer model.  Individual risk is expressed as "maximum lifetime
risk."  As used here, the work "maximum" does not mean the greatest  possible
risk of cancer to the public.   It is based only on the maximum annual  average
exposure estimated by the procedure used.   The second, called  aggregate  risk,
is a summation of all the risks to people estimated to be living within  the
vicinity (usually within 50 kilometers) of a source and is customarily summed
for all  the sources in a particular category.  The aggregate risk is expressed
as incidences of cancer among all of the exposed  population after 70 years of
exposure; for convenience, it is often divided by 70 and expressed as cancer
incidences per year.   These calculations are described in more detail  in
Section 4 below.

     There are also risks of nonfatal cancer and  other potential health  effects,
depending on which organs receive the exposure.  No numerical  expressions
of  such  risks have been developed.

-------
 2.   THE UNIT RISK ESTIMATE FOR INORGANIC ARSENIC?

      The following discussion is summarized from a more detailed description
 of the Agency's derivation of the inorganic arsenic unit risk  estimate as
 found in EPA's "Health Assessment Document for Inorganic Arsenic"  (EPA-600/
 8-83-021F).

 2<1  The Linear No-Threshold Model  for Estimation  of Unit Risk Based  on
      Human  Data (General)

      The methodologies used to  arrive  at  quantitative estimates  of risk
 must  be  capable of being implemented using  the data  available in existing
 epidemic!ogle  studies  of exposure  to airborne  arsenic.  This requires
 extrapolation  from the  exposure  levels  and  temporal  exposure patterns  in
 these studies  to  those  for  which risk  estimates are  required.  It is assumed
 that  the  age-specific mortality  rate of respiratory  cancer per year per
 100,000 persons for a particular 5-year age interval, i, can be
 represented using  the  following linear  absolute or additive risk model:

                          aj(D) = ai + lOO.OOOa'D                  ( 1)

With  this model, a^ is  the age-specific mortality rate per year of
 respiratory cancer in a control  population not exposed to arsenic,  a1  is
 a  parameter representing the potential  of airborne arsenic to cause
 respiratory cancer, and D is some measure of the exposure to arsenic  up
 to the ith age  interval.  For example,  D might be the cumulative  dose
 in years-pg/m3, the cumulative dose neglecting exposure  during  the  last
 10 years prior  to the ith age interval, or the average dose  in  ug/rn3
 over  some time period prior to the ith  age interval.   The forms  to  be  used
 for D are constrained by the manner in  which dose was treated in  each
 individual epidemiologic study.   At low exposures the extra  lifetime
probability of  respiratory cancer mortality will  very correspondingly
( e.g.  , linearly).

-------
     The dose-response data available in the epidemiologic  studies  for  esti-
mating the parameters in these models consists  primarily  of a  dose  measure
Di for the jth exposure group, the person-years of  observation Yj,  the  observed
 J
number of respiratory cancer deaths Oj,  and the number Ej of these  deaths
expected in a control population with the same  sex  and age  distribution as
the exposure group.  The expected number Ej is  calculated as
                          j = £ Yjjai/100,000                        (2)

 here YJ-J is the number of person-years of observation in the ith age cate-
gory and the jth exposure group (Yj = 2 Y ji).  This is actually a simplified

representation, because the calculation also takes account of the change in
the age-specific incidence rates with absolute time.  The expected number
of respiratory cancer deaths for the ith exposure group is
                 E(0j) = Z Yji (aj + 100,OOOa'Dj)/100,000

                 =Ej+a'YjDj                                      (3)

 under  the  linear absolute risk model.  Consequently, E(0j) can be expressed
 in  terms of quantities typically available from the published epidemiologic
 s tu d i es.

     Making the reasonable assumption that Oj has a Poisson distribution,
 the parameter  a' can  be estimated from the above equation using the method
 of  maximum likelihood.  Once  this parameter  is estimated, the age-specific
 mortality  rates for  respiratory cancer can be estimated for any desired ex-
 posure pattern.

-------
     To estimate the corresponding additional  lifetime  probability  of  res-
piratory cancer mortality, let bi,...,bi8 be the  mortality  rates,  in the
absence of exposure, for all  cases per year per 100,000 persons  for the age
intervals 0-4, 5-9,..., 80-84, and 85+, respectively;  let ai	aig represent
the corresponding rates for malignant neoplasms of  the  respiratory  system.
The probability of survival to the beginning of the ith 5-year age  interval
is estimated as
                           i-1
                            n [1 - 5bj/100,000]                     (4)
Given survival to the beginning of age interval  i,  the probability of  dying
of respiratory cancer during this 5-year interval  is  estimated as

                           Sai/lOO.OOO                              (5)

     The probability of dying of respiratory cancer given  survival  to  age
85 is estimated as ai8/&i8«  Therefore,  the probability of dying of respir-
atory cancer in the absence of exposure  to arsenic  can be  estimated as:

                      17                i-1
                 PQ = Z  [5a-j 7100,000)   n  (l-5bj/100,000)]
                      i=l               j=l
                                   17
                      +(ai8/bis)    n (1 - 5bj/100,000)
                                   j=l
Here the mortality rates a-j apply to the target  population for which risk
estimates are desired, and consequently  will be  different  from those in

-------
(l)-(5), which applied to the epidemiologic study cohort.   If the 1976  U.S.
mortality rates (male, female, white, and non-white combined) are used  in
this expression, then Pg = 0.0451.
     To estimate the probability PEP of respiratory cancer mortality when
exposed to a particular exposure pattern EP,  the formula 16)  is again used,
but a-,- and b-j are replaced by a-j(D-j) and b-j(Df), where D-j is  the exposure
measure calculated for the ith age interval from the exposure pattern EP.
For example, if the dose measure used in (1)  is cumulative dose to the  be-
ginning of the ith age interval in ug/m3-years, and the exposure pattern
EP is a lifetime exposure to a constant level  of 10 ug/m , then D^ =
( i-l)(5)( 10), where the 5 accounts for the fact that each age interval  has
a width of 5 years.  The additional  risk of respiratory cancer mortality is
estimated as

                              PEP - PO                             (7)

If the exposure pattern EP is constant exposure to 1 Mg/m , then PEP -  PQ is
called the "unit risk."

This approach can easily be modified to estimate the extra probability  of
respiratory cancer mortality by a particular  age due to any specified
exposure pattern.

2.2  Unit Risk Estimates Derived from Epidemiologic Studies
     Prospective studies of the relationship  between mortality and exposure
to airborne arsenic have been conducted for the Anaconda, Montana smelter
and the Tacoma, Washington smelter.   Table 1  summarizes the fit of the
absolute linear model to dose-response data from 4 different studies at the
two smelters.  (See the "Health Assessment Document for Inorganic Arsenic",

-------
                                            10
                                          Table 1
                            Summary of Quantitative Risk Analyses
Exposed
Population
Anaconda
smel ter
workers


ASARCO
smel ter
workers
Study and
Data Source Model
Lee-Feldstein absolute risk
( heavy exposure
omitted)
Higgins et al. absolute risk
Brown & Chu absolute risk
Enterline 4 Marsh absolute risk
( zero lag)
Results of Goodness-of-Fit Test
x2(d. f. ) p-value "unit" risk3
12.7(5) 0.025 2.80(-3)&
1.2(3) 0.75 4.901-3)
7.01(7) 0.41 1.25(-3)
5.5(4) 0.24 6.8K-3)
               Enterline 4 Marsh     absolute risk
               (10-year lag)
7.0(4)       0.14      7.60(-3)
 Additional lifetime risk of respiratory cancer mortality  from Hfetfnie environmental exposure"
 to 1 ugm-5 arsenic.
b2.80 (-3) means 2.80 x 10'3

-------
                                     11
Chapter 7, EPA-600/8-83-021F for detailed description  of  occupational studies.)
Table 1 also displays the carcinogenic  potencies  a*.   It  should be noted
that the potencies estimated from different models  are in different units,
and are therefore not comparable.

     The estimated unit risk is presented for each  fit for which  the chi-
square goodness-of-fit p-value is greater than 0.01.   The unit risks derived
from linear models—8 in all—range from 0.0013 to  0.0136. The  largest of
these is from the Ott et al. study, which probably  is  the least  reliable
for developing quantitative estimates,  and which  also  involved exposures to
pentavalent arsenic, whereas the other  studies involved trivalent arsenic.
The unit risks derived from the linear  absolute-risk models are considered
to be the most reliable; although derived from 5  sets  of  data involving 4
sets of investigators and 2 distinct exposed populations, these estimates
are quite consistent, ranging from 0.0013 to 0.0076.

     To establish a single point estimate, the geometric  mean for data sets
is obtained within distinct exposed populations,  and the  final estimate is
taken to be the geometric mean of those values.  This  process is  illustrated
in Table 2.

-------
                                     12
                                  Table 2
        Combined Unit Risk  Estimates  for Absolute-Risk Linear Models
                                                    Geometric      Final
                                                    Mean Unit    Estimated
Exposure Source      Study            Unit Risk         Risk      Unit Risk


Anaconda smelter   Brown  &  Chu       1.25 x 10~3
                   Lee-Feldstein     2.80 x 10~3    2.56 x 10~3
                   Higgins  et al.    4.90 x 10~3                  4.29 x 10~3

ASARCO smelter     Enterline &
                    Marsh           6.81 x 10'3
                                    7.60 x lO'3    7.19 x 10-3

-------
                                     13
3.    QUANTITATIVE EXPRESSIONS OF PUBLIC EXPOSURE  TO  INORGANIC  ARSENIC
     EMISSIONS

3.1  EPA's Human Exposure Model  (HEM) (General)

     EPA's Human Exposure Model  is a general  model capable of  producing
quantitative expressions of public exposure to ambient air concentrations
of pollutants emitted from stationary sources.   HEM  contains ( 1)  an  atmospheric
dispersion model, with included meteorological  data, and (2) a population
distribution estimate based on Bureau of Census data.   The input data  needed
to operate this model are source data, e.g.,  plant location, height  of the
emission release point, and volumetric rate of release temperature of  the
off-gases.  Based on the source data, the model  estimates the  magnitude  and
distribution of ambient air concentrations of the pollutant in the vicinity
of the source.  The model is programmed to estimate  these concentrations
for a specific set of points within a radial  distance of 50 kilometers from
the source.   If the user wishes to use a dispersion  model other than the
one contained in HEM to estimate ambient air concentrations in the vicinity
of a source, HEM can accept the concentrations if they are put into  an
appropriate format.

     Based on the radial distance specified, HEM numerically combines the
distributions of pollutant concentrations and people to produce quantitative
expressions of public exposure  to the pollutant.

3.1.1  Pollutant Concentrations Near a Source

     The  HEM  dispersion model is a climatological model which is a sector-
averaged  gaussian dispersion algorithm that has been simplified to improve
computational efficiency.^

-------
Stability array (STAR) summaries are the principal  meteorological  input to
the HEM dispersion model.   STAR data are standard climatological  frequency-
of-occurence summaries formulated for use in EPA models and available for
major U.S. meteorological  monitoring sites from the National  Climatic Center,
Asheville, N.C.  A STAR summary is a joint frequency-of-occurence  of wind
speed, atmospheric stability, and wind direction, classified according to
Pasquill's categories.  The STAR summaries in HEM usually reflect  five years
of meteorological  data for each of 314 sites nationwide.   The model  produces
polar coordinate receptor  grid points consisting of 10 downwind distances
located along each of 16 radials which represent wind directions.   Concen-
trations are estimated by  the dispersion model  for  each of the 160 receptors
located on this grid.  The radials are separated by 22.5-degree intervals
beginning with 0.0 degrees and proceeding clockwise to 337.5  degrees.  The
10 downwind distances for  each radial  are 0.2,  0.5, 1.0,  2.0, 5.0, 10.0,
20.0, 30.0, 40.0,  and 50.0 kilometers.  The center  of the receptor grid for
each plant is assumed to be the plant center.  Concentrations at other
points were calculated by  using a log-linear scheme as illustrated in
Figure 1.

3.1.2  Expansion of Analysis Area

     At proposal,  exposure and risk were estimated  for people residing
within 20 kilometers of the smelter.  Some comrnenters pointed out  that
since people beyond 20 kilometers are exposed to some level  of arsenic due
to a source's emissions, EPA's proposal  analysis underestimates the  total
exposure and risk.  EPA agreed with the commenters  and expanded its  analysis
out to 50 kilometers.  When applying air dispersion models,  the EPA's
modeling guidelines recommend that, because of  the  increasing uncertainty
of estimates with  distance from the modeled source  and because of  the
paucity of validation studies at larger distances,  the impact may  extend

-------
                                     15
out to 50 kilometers but the analysis should generally be limited to this
distance from the source.^    Such site-specific factors as terrain features
(complex or flat), the objectives of the modeling exercise, and distance to
which the model has been validated will determine the appropriate distance
(whether greater than or less than the guideline distance) for which the
Agency should apply the model.

3.2  Methodology for Reviewing Pollutant Concentrations

     Before making HEM computer runs, EPA reviewed small-scale U.S. Geological
Survey topographical  maps (scale 1:24000) to verify locational data for each
arsenic source.  Plants were given accurate latitude and longitude values which
were then incorporated into the HEM program.
     After completing the HEM runs, nearby monitoring sites with ambient
air quality data were identified by a computer search of EPA's National
Aerometric Data Bank (NADB) (Table 3).  At some sites, data collected over
several years along with annual  averages (based on different numbers of
sample sizes for the years monitored) for each year were available.  In
these instances, weighted multi-year averages were calculated to provide an
overall mean for each monitoring site.  For purposes of annual mean calculations,
values measured below mimimum detection limits were considered by EPA to be
equal  to one-half the detection  limit.  These ambient arsenic data were
                                    »
then compared to HEM predicted values in order to gauge the accuracy of the
air dispersion model's estimates.  As noted above, HEM predicted values
were based on concentrations at  160 polar coordinate receptor grid points
consisting of 10 downwind distances located along each of 16 radials which
represented wind directions.  Because the actual  monitoring site locations
identified in the NADB retrieval  usually did not  correspond to exact grid
point  locations, a  log-linear interpolation scheme (Figure 1) was used to
calculate an estimated concentration at the site.

-------
16
Table 3
Arsenic Concentrations Near ASARCO-East Helena
Primary Lead Smelter
Plant

ASARCO-East
Helena








Company Data





f Obs.


27
41
137
25
31
36
81
23
20
1460
1460
1460
638
1460
274
Distance1
(km)

.5
.7
.8
.9
1.4
1.5
3.9
4.7
7.2
1.1
1.3
1.3
2.1
6.1
7.2
Bearing

119.6
11.5
20.4
343.9'
45.3
156.9
176.5
270.4
273.4
275
5
145
92
275
162
Predicted2
lug/m-5)

0.230
0.078
0.056
0.050
0.076
0.047
0.0159
0.005
0.003
0.024
0.050
0.077
0.071
0.0037
0.0084
Measured3
(Ug/m^)

0.108
0.151
0.242
0.161
0.078
0.109
0.031
0.025
0.030
0.059
0.24
0.078
0.074
0.024
0.028
MDL4
(ygTm3")

0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.0055
—
—
~
—
~
—
Percent! le5


30< * <50
<10
<10

-------
                                    17
Figure  1     Group 2 BG/ED  Interpolation
                                                      — A
                               R           R2
Given:
A   -  The angle in radians subtended clockwise about  the  source  from  due
       south to the BG/ED centroid;
Al  -  The angle from due south to the radial  line  immediately  counter-
       clockwise of A, or passing through A if  there  is  an exact  match;
A2  -  The angle from due south to the radial  line  immediately  clockwise of
       Al (A2 is 0 if 1t is due south);         ;   -
R   -  The distance in km from the source to the BG/ED centroid;
Rl  -  The distance from the source to the largest  circular arc of  radius
       less than R;
R2  -  The distance from the source to the smallest circular arc  of
       radius greater than or equal  to R;
Cl  -  The natural logarithm of the concentration value  at (Al, Rl);
C2  -  The natural logarithm of the concentration value  at (Al, R2);

-------
                                      18

C3  -  The natural logarithm of the concentration value  at  (A2,  Rl);
C4  -  The natural logarithm of the concentration value  at  (A2,  R2);
then:
RTEMP - ln(R/Rl)/ln(R2/Rl);
ATEMP - (A-A1)/(A2-A1);
CA1   - exp(Cl + (C2-Cl)xRTEMP);
CA2   - exp(C3 + (C4-C3)xRTEMP);  and
CX    - CA1 + (CA2-CAl)xATEMP,
where CX is the interpolated concentration  at the BG/ED  centroid.

-------
                                     19
3.2.1  Use of Ambient Data

       Certain criteria were considered in review of ambient levels.  Mean
concentration values derived from sample sizes of less than 25 data points
were disregarded.  When reviewing the available monitoring data, it appeared
that monitors situated at distances greater than 15 km from the arsenic
source were considered too far from the source to gauge air dispersion
results without interference from other arsenic sources.  Furthermore, at
distances greater than 15 km from the source, plant impacts were often
predicted to be significantly lower than minimum detection limits.  These
data were not incorporated in the analyses.  A third consideration in
reviewing ambient data concerned the percentage of monitored data which
fell below minimum detection limits.  Although some monitoring sites
registered data with over 90 percent of the values above minimum detection
levels, many had about half the data points or more below such levels.
Instances where more than 50 percent of the data were below MDL were dis-
regarded.  It should be noted that the various tables in subsequent sections
display, in addition to company-collected data, all ambient monitoring data
that were collected at sites within 15 kilometers of the source as identified
by EPA's computer search although not all the data were used in the final
analysis.

3.2.2  The People Living Near A Source

       To estimate the number and distribution of people residing within 50
kilometers of the source, the HEM model uses the 1980 Master Area Reference
File (MARF) from the U.S. Bureau of Census.  This data base consists of
enumeration district/block group (ED/BG) values.  MARF contains the population
centroid coordinates (latitude and longitude) and the 1980 population of each
ED/BG  (approximately 300,000) in the United States (50 states plus the District
of Columbia).  HEM identifies the population around each plant, by using the

-------
                                     20
 geographical coordinates of the plant, and identifies, selects, and stores
 for  later use those ED/BGs with coordinates falling within 50 kilometers of
 plant center.

 3.2.3  Exposure^

       The Human Exposure Model (HEM) uses the estimated ground level
 concentrations of a pollutant together with population data to calculate
 public exposure.  For each of 160 receptors located around a plant, the
 concentration of the pollutant and the number of people estimated by the
 HEM to be exposed to that particular concentration are identified.   The HEM
 multiplies these two numbers to produce exposure estimates and sums these
 products for each plant.

       A two-level  scheme has been adopted in order to pair concentrations
 and populations prior to the computation of exposure.   The two level approach
 is used because the concentrations are defined on a radius-azimuth  (polar)
 grid pattern with non-uniform spacing.   At small  radii,  the grid cells  are
 usually smaller than ED/BG's; at large radii,  the grid cells are usually
 larger than ED/BG's.   The area surrounding the source  is divided into two
 regions,  and each ED/BG is classified by the region in which its centroid
 lies.  Population exposure is calculated differently for the ED/BG's located
witliin each region.   For ED/BG centroids located between 0.2 and 3.5 km
 from the  emission source, populations are divided between neighboring
concentration grid  points.   There are 64 (4 x  16)  polar grid points within
this range.   Each ED/BG can be paired with one or many concentration points.
The population associated with the ED/BG centroid is then divided among all
concentration grid  points assigned to it.   The land area within  each polar
sector is considered in the apportionment.

-------
                                     21
     For population centroids between 3.5 and 50 km from the source,  a
concentration grid cell, the area approximating a rectangular shape bounded
by four receptors, is much larger than the area of a typical  EO/BG.   Since
there is an approximate linear relationship between the logarithm of
concentration and the logarithm of distance for receptors more than 2 km
from the source, the entire population of the EO/BG is assumed to be  exposed
to the concentration that is logarithmically interpolated radially and
arithmetically interpolated azimuthally from the four receptors bounding
the grid cell.  Concentration estimates for 96 (6 x 16) grid cell receptors
at 5.0, 10.0, 20.0, 30.0, 40.0, and 50.0 km from the source along each  of
16 wind directions are used as reference points for this interpolation.

     In summary, two approaches are used to arrive at coincident concentration/
population data points.  For the 64 concentration points within 3.5 km  of the
source, the pairing occurs at the polar grid points using an apportionment
of EO/BG population by land area.  For the remaining portions of the grid,
pairing occurs at the ED/BG centroids themselves through the use of log-log
and linear interpolation.  (For a more detailed discussion of the model  used
to estimate exposure, see Reference 5.)

-------
                                    22
 3.3  ASARCO-East Helena

     Predicted (HEM) versus measured data were plotted (Figure 2)  and a
 least squares weighted linear regression analysis  was  run  based on thirteen
 data points  (see Table 3).  The least squares  regression  line  (solid line)
 was determined on the basis of a conparison of National Aerometric Data  Bank
 monitoring data (circumscribed dots) and ASARCO monitoring data (circumscribed
 Xs) with ambient concentrations predicted by the Human Exposure Model.

     The reader should note that a perfect fit for the least squares regression
 analysis results in a line running through the origin  at a 45° angle (dotted
 line on Figure 2).  This means that if the HEM model predicts  the  measured
 data perfectly, then the data points would fall  on the dotted  line.   In  cases
where the HEM model underpredicts  concentrations,  data points  will  be located
above the 45° perfect fit line.  Likewise, when  the HEM model  overpredicts
concentrations, data points will be located below  the  perfect  fit  line.
The regression line resulting from our conparison  of predicted and monitored
data runs nearly parallel to the perfect fit line  but  intersects the ordinate
axis at a value of approximately 0.05 ug/m3.  This result  is consistent  with
the expectation that air dispersion modeling would underpredict ambient  con-
centrations.   The air dispersion modeling did  not  consider other local
sources of arsenic such  as naturally-occurring arsenic in  the  windblown
dust and reentrained arsenic particulate matter  that had settled to  the
earth from past smelter  emissions.

     A study  to determine source apportionment for particulate  lead  and  total
suspended particulates  (TSP) in East Helena  was  completed  in 1982.   High
volume TSP, low volume TSP, and dichotomous  samplers were  co-collected
 (same time period and same site) to permit differences in  sample collection
mass and chemistry to be understood.  Analysis of  hi-vol samples was  carried
out by the State  of Montana and lo-vol  and dichotomous samples  were  analyzed

-------
                                  23
                  FIGURE  2   Predicted  Versus Measured


                  Inorganic Arsenic Ambient Concentrations


                  (ASARCO - East  Helena,  MT)
   0,
J,

c
o
fC
OJ

-------
                                      24
 by NEA, Inc.  In addition  to participate  lead and TSP, sanples were also
 measured in some cases  for arsenic.6

      At six locations where arsenic concentrations were measured using both
 lo-vol  and  hi-vol  samplers, the ratio of  lo-vol to hi-vol in percent ranged
 from 104 to 133  with a  mean of 118%.  This loss of arsenic compounds could
 have occurred  in two areas:  (1)  the volatilization of the arsenic compounds
 from the hi-vol  filter  itself during sampling, and (2) the loss of volatile
 arsenic  compounds  during digestion and storage of sanples prior to analysis.
 However,  based on  the data from the study, EPA concluded that the loss  of
 arsenic  on  hi-vol  filters was relatively minor in nature and within the over-
 all  accuracy goal  of +  15-20% considered adequate for most ambient air  quality
 measurements.

 3*3a   Public Exposure to Inorganic Arsenic Emissions from Primary Lead
       Smelters

 3.3.1.1  Source Data

     Five primary  lead smelters  are included  in  the  analysis.   Table  4
 lists the names and addresses  of  the plants considered,  and  Table  5  lists
the plant data used as  input to  the Human  Exposure Model  (HEM).

3.3.1.2  Exposure Data

     Table 6 lists, on  a plant-by-plant  basis, the total  number of people
encompassed  by the exposure analysis and the  total exposure.  Total exposure
is the sum of the products  of number of  people times  the  ambient air concentration
to which they are exposed,  as calculated by HEM.  Table  7  sums, for the
entire source category  (5 plants),  the numbers of people  exposed to various
ambient concentrations,  as  calculated by HEM.  (Source-by-source exposure
results are  provided in  the EPA docket numbered A-83-23.)

-------
                                    25









                                TABLE 4





                  IDENTIFICATION OF PRIMARY LEAD SHELTERS
Plant Number Code                    Plant Name and Address
                                    ASARCO East Helena. MT
                                    ASARCO El Paso, TX
                                    St. Joe Herculaneum, MD
                                    ASARCO Glover. MO
                                    Amax Boss, MO

-------
                                 26
Table 5  Input  Data  to  Exposure Model  Primary Lead Smelting Industry
                    (Assuming Baseline Controls)
Plant
(Furnace)
ASARCO-East
Helena, MT


ASARCO-
El Paso, TX

Latitude
(Degrees
Minutes
Seconds )
46-34-52


31-47-06

St. Joe- 38-15-47
Herculaneum, MO
ASARCO-
Glover, MO
Amax-
Boss, MO
36-29-46
37-38-31
Longitude
(Degrees
Minutes
Seconds )
111-55-12


106-37-23

90-22-59
90-41-28
91-11-35
Emission
Rate
(Kg/yr)
14700
1680
1680
5040
1680
1680
5040
2772
84
1680
42
1680
50.4
Emission
Point
Elevation
(Meters)
128
130
122
0
186
91
0
107
0
24
0
61
•' o
Emission
Point
Diameter
(Meters)
2.7
3.0
3.4
~ ~
4.3
4.9
~ ~
6.1
1.5
4.6
Emission*
Point
Cross
Sectional
Area (m^)
100
100
100
10000
100
100
10000
100
10000
100
10000
100
10000
Emission
Point Gas
Exit
Velocity
m/sec
16.5
19.4
11.7
--
17.0
3.0
--
13.5
18.3
10.0
Emission
Point Gas
Temp.
(°K)
352
330
375
293
345
330
293
353
293
294
293
355
293
Emission
Point
Type
Stack
Stack
Stack
Fugitive
Stack
Stack
Fugitive
Stack
Fugitive
Stack
Fugitive
Stack
Fugitive

-------
                                 27
         TABLE 6  TOTAL EXPOSURE  AND  NUMBER  OF  PEOPLE  EXPOSED
                      PRIMARY LEAD SMELTING  INDUSTRY*
                            Total                          Total
                          Number of                      Exposure
    Plant              People Exposed               (People  -  ug/m3)
1
2
3
4
5
48,600
497,000
1,510,000
97,300
42,700
215
715
186
27
7
* A 50-kilometer radius was used for the analysis of primary  lead
  smelting industry.

-------
                                            28
                                        TABLE  7

                    PUBLIC EXPOSURE  FOR  PRIMARY  LEAD  SMELTING  INDUSTRY
                         AS PRODUCED BY  THE  HUMAN EXPOSURE MODEL
                               (ASSUMING BASELINE CONTROLS)
Concentration
Level (ug/m^)
0.437
U.25
:).!
U.05
0.025
0.01
0.005
0.0025
0.001
0.0005
U. 00025
0.0001
0 .00005
0.0000269
Population
Exposed
(Persons)*
<1
1
40
441
1240
7700
15900
71700
340000
545000
657000
1470000
2080000
2190000
Exposure
(Persons - ug/m^)**
0
0
6
33
62
144
199
398
801
945
983
1100
1150
1150
 *Column 2 displays  the  computed  value,  rounded to the nearest whole number, of the
  cumulative number  of people  exposed to the matching and higher concentration levels
  found in column  1. For  example,  0.5 people would be rounded to 0 and 0.51 people
  would be rounded to 1.

**Column 3 displays  the  computed  value of the cumulative exposure to the matching
  and higher concentation  levels  found in column  1.

-------
                                     29
3.4  Murph Metals-Dallas  and Quemetco-Seattle

     Predicted (HEM) versus measured data for  Murph  Metals-Dallas  and
Quemetco-Seattle were plotted (Figures  3 and 4)  and  a  least  squares weighted
linear regression analysis was run based on a  number of  data points.   The
least squares regression  line (solid line) was determined on the basis of  a
comparison of National Aerometric Data  Bank monitoring data  (circumscribed
dots) and State agency monitoring data  (circumscribed  Xs) with  ambient con-
centrations predicted by  the Human Exposure Model.

     The reader should note that a perfect fit for the least squares
regression analysis results in a line running  through  the origin at  a  45°
angle (dotted lines in Figures 3 and 4).  This means that if the HEM model
predicts perfectly, then  the data points would fall  on the 45°  line.   In
cases where the HEM model underpredicts concentrations,  data points will be
located above the 45° perfect fit line.  Likewise, when the  HEM model
overpredicts concentrations, data points will  be located below  the perfect
fit  line.  The regression  line resulting from our comparison of predicted
and  monitored data  lies above the perfect fit  line,  intersecting  the
ordinate axis at values of approximately 0.011 ug/m3 and 0.026  ug/m3 for Murph
Metals and Quemetco respectively.  This result is consistent with  the
expectation that air  dispersion modeling would underpredict  ambient  con-
centrations.  The air dispersion modeling did not consider other  local sources
of arsenic such as  naturally-occurring arsenic in the windblown dust and  re-
entrained  arsenic particulate matter that had settled to the earth from past
smelter emissions.

-------
                 30
FIGURE 3  Predicted Versus Measured
Inorganic Arsenic Ambient Concentrations
(Murph Metals - Dallas, TX)
      !  MODEL
      I  UNDERPREDICTION
I  OVERPREDICTION  |j|g
r                  rS tfft
                                  Perfect Fit
                                  Linear Regression
                                  Municipal Data
     0.02                0.04

    Predicted Concentration
        0.06

-------
                                                 Measured  Concentration
 -5
 n>
 Q.
 _j.
 O
 O
 O
 O
 fD
 -5
 CD
 O
 3
03



 OO
                                                                                                                                            o
o
o
                                                                                                                                            oo
                                                                                                                                            fD
                                                                                                                                            fD
                                                                                                                                                 3 i—i
                                                                                                                                                 O O
                                                                                                                                                 -s c:
                                                                                                                                                 U3 73
                                                                                                                                                 fa m
     O
     -s -s

     fD Q.
     3 ->•
     -"• O
     n <-+
        fD
     CT <
     -•• fD
     fD -S
     3 t/1

        l/J
     O
     O 3
     3 fD
     O O<
     fD Wl
     3 C
     <-»• -s
     -S fD
     Q) f^>
     rt-

     O
     3
     cn

-------
                                     32
3-4.1  Public Exposure to Inorganic Arsenic  Emissions  from Secondary  Lead
       Smelters

3.4.1.1  Source Data

         Thirty-five secondary lead smelters  are  included  in  the  analysis.
Table 8 lists arsenic concentrations near  select  secondary  lead smelters.
Table 9 lists the names and addresses  of the  plants considered, and Table 10
lists the plant data used as input  to  the  Human Exposure Model (HEM).

3.4.1.2  Exposure Data

         Table 11 lists, on a plant-by-plant  basis, the total number of
people encompassed by the exposure  analysis and the total exposure.  Total
exposure is the sum of the products of number of  people times the ambient
air concentration to which they  are exposed,  as calculated by HEM.  Table
12 sums, for the entire source category (35 plants), the number of people
exposed to various ambient concentrations, as calculated by HEM.  (Source-
by-source exposure results are provided in the EPA docket numbered A-83-9.)

-------
                                            33

                                          Table 8

                             Arsenic Concentrations Near Select
                                  Secondary Lead Smelters
Plant

General Battery,
Reading, PA
Murph Metals-
Dallas, TX


Murph Metals-Dallas
Texas Air Control
Board Data
Quemetco-City of
Industry, CA







I Obs.

29

86
21
93
57
28
31
31
29
25
81
85
27
47
30
121
29
Distance^ Bearing
(km)
5.1

3.6
3.7
7.6
9.0
0.2
0.2
0.5
17.8
22.8
23.6
24.0
31.2
32.2
35.1
36.1
38.1

189.5

181.6
181.5
311.6
256.2
0
337.5
157.5
314.0
281.9
164.2
275.2
218.2
161.0
300.0
84.8
235.8
Predicted2
(ng/m3)
0.00104

0.0024
0.0024
0.00095
0.0004
0.062
0.042
0.014
0.00037
0.000133
0.00018
0.00023
0.000083
0.000112
0.000106
0.000113
0.000066
Measured3 MDL4
(wg/m3)
0.009

0.028
0.010
0.029
0.025
0.085
0.077
0.025
0.005
0.003
0.003
0.006
0.003
0.003
0.003
0.005
0.004
(ug/m3)
0.0055

0.05
0.0055
0.05
0.05
— —
..
--
0.0055
0.0055
0.0055
0.0055
0.0055
0.0055
0.0055
0.0055
0.0055
Percentile5

30< 5

90< !
30< !
90, 5

— —
__
--
70< 5

90< 5
70< 5

70< 5

70< 5
70< 5

I <50

C <95*
I <50*
6 <95*
>99*



6 <90*
>99*
I <95*
I <90*
>99*
6 <90*
>99*
6 <90*
i <90*
Quemetco-
 Indianapol 1s, IN
64
12.5
78.2
0.00040
0.005   0.0055  50< % <70*
Quemetco-Seattle,
WA

Quemetco- Seattle
Washington State Dept.
of Ecology Data
80
60
72
60
60

1.9
3.2
13.5
0.2
1.4

150.8
30.3
2.6
157.5
180

0.0036
0.00183
0.00047
0.031
0.0075

0.041
0.038
0.020
0.09
0.03

0.0055
0.0055 30< !
0.0055 30< !
..
—

<10
6 <50
I <50



*Indicates data point was disregardfed - see Section 3.3.1.1.
^Distance from source to monitor (km).
Concentration predicted by Human Exposure Model  (HEM).   See Section  3.1.
3The measured values are weighted averages.  When the sampled  arsenic concentrations  were
 below the MDL, a value of 1/2 MDL was assumed for purposes  of calculating the annual  averages.
^Minimum Detection Limit.
5Percentile indicates percentage of data falling  below minimum detectable  levels.

-------
                    34





                 Table  9



Identification of Secondary  Lead Smelters
Plant Number Code

2

4
0
6
1



10

12

14
Ib
16
I/
18
19
20
21
22
23 x
24
25
26
27
28
29
30
31
32
33
34
35
Plant Name and Address
Alco Pacific Gardena, CA
Bergsoe St Helens, OR
Chloride Metals Columbus, (iA
Chloride Metals Tampa, FL
Dixie Metals Dallas, TX
East Penn Lyons Station
Federated Metals San Fran, CA
General Battery Reading, PA
General Smelting college Grov, TN
Gopher Eugene, Minn
Gould Frisco, TX
Gould Vernon, CA
Gulf Coast Tampa, FL
Hyman Viener Richmond, VA
Interstate Lead Leeds, AL
Lancaster Lancaster, PA
Master Metals Cleveland, OH
Murph Metals Dallas, TX
National Smelting Atlanta, GA
National Smelting Pedricktown, NJ
Quemetco City of Industry, CA
Quemetco Indianapolis, IN
Quemetco Seattle, WA
Refined Metals Beach Grove, IN
Refined Metals Memphis, TN
Revere Wall , NY
Ross Metals Rossville, TN
Sanders Lead Troy, Al
Schuylkill Baton Rouge, LA
Schuylklll Forest City, MO
Standard San Antonio, TX
Taracorp Atlanta, GA
Taracorp Granite City, IL
Tonolli Nesquehoning, PA
USS Lead E. Chicago, IN

-------
35
Table 10
Secondary Lead Industry Inputs to HEM Model
(Assuming Baseline Controls)
Process

As
Emission
Latitude /Longitude
Plant (Deqrees-Minutes-Seconds)
Alco Pacific 33-50-20/118-18-07
Bergsoe 45-50-58-122-49-3
Chloride 32-26-00/84-56-00
Metal s/GA
Chloride 27-54-5/82-24-12
Metal s/FL
Dixie Metals 32-44-49/96-46-37
East Penn 40-28-19/75-58-23
Federated 37-44-/122-23
Metals/CA
General Battery 40-22-45/75-54-50
General Smelting 35-48-00/86-40-05
Gopher 40-50-/93-7-30
Gould/TX 33-08-38/96-49-44
Gould/CA 34-00-14/118-13-45
Gulf Coast 27-57-44/82-22-53
Hyman Viener 37-31-10/77-24-54
Interstate Lead 33-31-58/86-32-00
Lancaster 40-03-11/76-19-52
Master Metals 41-28-52/81-40-48
Murph Metals 32-46-40/96-52-21
National 33-47-31/84-24-18
Smelting/GA
National 39-45-30/75-25-30
Smelt ing/NJ
Rate
kg/yr
7.6
86.6
34.9

34.9

52.3
52.3
14.5

91.5
17.4
43.6
87.2
53.4
36.3
11.6
69.7
0.4
69.7
174.3
43.6

174.3


Stack
Ht.
m
31
31
31

31

31
31
31

26
31
31
31
26
31
31
31
31
31
26
31

26


Stack
Diam.
m
0.62
0.92
0.92

0.92

0.92
0.92
0.62

1.2
0.62
0.92
0.92
1.2
0.92
0.62
0.92
0.62
0.92
1.2
0.92

1.2


Stack
Vel.
m/sec
29
25.9
25.9

25.9

25.9
25.9
29

16.6
29
25.9
25.9
16.6
25.9
29
25.9
29
25.9
16.6
25.9

16.6


Exit
Tenp.
°K
331
400
400

400

400
400
400

400
400
400
400
331
400
400
400
400
400
400
400

400

Process


Emission Stack
Rt.
kg/yr
.25
2.6
0.5

0.5

0.75
0.75
0.21

2.98
0.25
0.62
1.24 -
3.83
0.52
0.17
0.99
0.01
0.99
2.48
0.62

2.48

Ht.
m
15.4
15.4
15.4

15.4

15.4
15.4
15.4

15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4

15.4



Fugitive

Stack
Diam.
m
1.2
1.2
1.2

1.2

1.2
1.2
1.2

1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2

1.2


Stack
Vel.
m/sec
10.4
13.3
13.3

13.3

13.3
13.3
10.4

10.6
10.4
13.3
13.3
10.6
13.3
10.4
13.3
10.4
13.3
10.6
13.3

10.6



Area
As

Exit Emission
Temp . Rt .
°K k/yr
311 4.0
311 0
311 17.8

311 13.8

311 168.9
311 20.0
311 4.0

311 253.8
311 7.1
311 46.2
311 13.8
311 0
311 16.9
311 7.1
311 67.6
311 7.1
311 1.3
311 692.1
311 18.7

311 115.6

Area
m
2,945
5,862
5,862

5,862

5,862
5,862
2,945

8,788
2,945
5,862
5,862
8,788
5,862
2,945
5,862
2,945
5,862
8,788
5,862

8,788


-------
                                                             36
       Plant
 Quemetco/CA
 Quemetco/IN
 Quemetco/WA
 Refined
  Metals/IN
 Refined
  Metals/TN
 Revere.NY
 Ross Metals
 Sanders Lead
 Schuylkill/LA
 Schuylkill/MO
 Standard
 Taracorp/GA
 Taracorp/IL
Tonolli
USS Lead
    Lat/Lonq	

 34-01-30/117-58-58
 39-45-14/86-17-59
 47-34-44/122-21-04
 39-42-36/86-03-5 4"

 35-05-13/90-04-10

 41-27-37/74-21-35
 35-02-42/89-34-30
 31-47-28/85-58-16
 30-58-08/91-14-40
 40-01-59/95-13-59
29-20-00/98-29-38
33-47/84-22
38-42-05/90-08-37
40-51-03/75-52-46
41-36-58/87-27-47
                                                    Table 10 (Continued)

                                        Secondary Lead Industry Inputs to HEM Model
                                              (Assuming Baseline Controls)

As
Emission
Rate
kg/yr
61.0
139.5
69.7
52.3
87.2
139.5
27.9
116.2
232.4
104.6
23.5
87.2
244.1
57.2
58.1
P
Stack
Ht.
m
26
26
31
31
31
26
31
31
26
31
31
31
26
26
31
roc ess
Stack
Oiam.
m
1.2
1.2
0.92
0.92
0.92
-1.2
0.62
0.92
1.2
0.92
0.62
0.92
1.2
1.2
0.92

Stack
Vel.
m/sec
16.6
16.6
25.9
25.9
25.9
16.6
29
25.9
16.6
25.9
29
25.9
16.6
16.6
25.9

Exit
Terrp.
°K
331
400
400
400
400
400
400
400
400
400
400
400
400
331
400


Emission Stack
Rt. Ht.
kg/yr m
1.99
1.99
0.99
0.75
1.24
1.99
0.39
1.66
3.31
1.49
0.34
1.24
3.48
1.86
0.83
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
15.4
• 15.4
15.4
15.4
15.4

fugitive
Stack Stack
Diam. Vel.
m m/sec
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
10.6
10.6
13.3
13.3
13.3
10.6
10.4
13.3
10.6
13.3
10.4
13.3
10.6
10.6
13.3
Exit
Tenp.
°K
311
311
311
311
311
311
311
311
311
311
311
311
311
311
311
Area
As
Emission
Rt. Area
692.1
692.1
461.4
76.9
76.9
692.1
0
29.3
20.9
13.8
7.6
13.8
23.1
69.3
61.3
8,788
8,788
5,862
5,862
5,862
8,788
2,945
5,862
8,788
5,862
2,945
5,862
8,788
8,788
5,862

-------
                                     37
       Plant
                                  Table 11

                Total Exposure and Number of People Exposed
                      Secondary Lead Smelting Industry*
     Total
   Number of
People Exposed
      Total
     Exposure
(People - ug/m3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
8,450,000
1,120,000
314,000
1,670,000
2,350,000
1 , 310 ,000
3,370,000
1,260,000
679,000
58,200
1,800,000
8,900,000
1,690,000
766,000
844,000
1,160,000
2,530,000
2,560,000
1,920,000
4,210,000
8,860,000
1,150,000
2,060,000
1,180,000
927 ,000
948,000
902,000
92,000
143,000
149,000
1,050,000
1,920,000
2,190,000
934,000
5,280,000
32
15
17
19
228
16
22
250
3
1
17
189
32
15
41
6
59
668
51
197
2300
460
576
67
115
249
4
11
4
2
24
83
159
21
103
* A 50-kilometer radius was used for the analysis  of  secondary  lead
  smelters.

-------
                                      38
                                   Table 12
                 Public Exposure for Secondary Lead Smelters
                   as Produced by the Human Exposure Model
                          (Assuming Baseline Controls)
Concentration
Level (ug/m3)
0.101
0.1
0.05
0.025
0.01
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
Population
Exposed
(Persons)*
<1
<1
256
2880
16000
53100
152000
7 43000
1940000
4510000
12800000
21700000
31100000
44000000
55700000
64100000
70800000
73800000
7 4700000
Exposure
	 (Persons - ug/rn^)**
0
0
16
104
300
543
878
1770
2590
3480
4750
5390
5720
5930
6010
6050
6060
6060
6060
   Column 2 displays  the  computed  value,  rounded to the nearest whole number
   of the cumulative  number of  people  exposed to the matching and higher
   concentration levels found  in column  1.  For exarple, 0.5 people would be
   rounded to 0 and 0.51  people would  be  rounded to 1.

** Column 3 displays  the  conputed  value  of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                     39
3.5  Public Exposure to  Inorganic Arsenic Emissions  from Primary Zinc
     Smelters

3. 5.1  Source Data

         Five primary zinc  smelters are  included  in  the analysis.  Table 13
lists ambient arsenic concentrations near select  primary zinc  smelters.
Table 14 lists the names and addresses of the  plants considered, and Table
15 lists the plant data  used as input to the Human Exposure Model (HEM).

3.5.2  Exposure Data

     Table 16 lists, on  a pi ant-by-plant basis, the  total number of people
encompassed by the exposure analysis and the total exposure.   Total exposure
is the sum of the products of number of  people times the ambient air
concentration to which they are exposed, as calculated by HEM.  Table  17
sums, for the entire source category (5  plants),  the numbers of people
exposed to various ambient concentrations,  as  calculated by HEM.  I Source-
by-source exposure results are provided  in  the EPA docket numbered A-83-23. )

-------
Plant
St . Joe-
Monaca , PA
ASARCO-
Corpus Christi ,
TX





Amax-
Sauget, IL
Jersey Miniere
Zinc Co-
Clarksville, TN
National Zinc-
Bartlesville, OK
Distance-^
# Obs (km)
40
Table 13
Arsenic Concentrations Near
Select Primary Zinc Smelters
Predicted^
Bearing (ug/m3)
No data within Ib km
36 3.0
299 3.8
33 5.5
83 6.2
319 6.9
26 14.9
190 14.9
27 4.1
No data within

No data within
262.8 0.000024
252.1 0.0000149
299.4 0.0000156
289.7 0.0000125
173.0 0.0000067
187.7 0.0000029
187.7 0.0000029
314.4 0.000042
15 km

15 km
Measured3 MDL4
(ug/m3) (|jg/m3) Percentile 5

0.025 0.05 100*
0.026 0.05 95< % <99*
0.025 0.05 100*
0.029 0.05 95< % <99*
0.026 0.05 95< % <99*
0.008 0.0055 30< % <50*
0.028 0.05 90< % <95*
0.007 0.0055 50< % <70*



* Indicates data point was  disregarded; see Section 3.1.
* Distance from source to monitor  (km).
2 Concentration predicted by  Human Exposure Model (HEM).
  The measured values are weighted averages.  When the sampled arsenic concentrations were below the MDL, a
  value of 1/2 MDL was assumed  for purposes of calculating the annual averages.
^Minimum detention limit.
5 Percentile indicates percentage of data falling below minimum detectable levels.

-------
                                41

                             Table 14
              Identification of Primary Zinc Smelters
Plant Number Code                      Plant Name and Address

       1                    St. Joe - Monaca, PA
       2                    ASARCO - Corpus Christi, TX
       3                    Amax - Sauget, IL
       4                    Jersey Miniere Zinc Co - Clarksville, TN
       5                     National Zinc - Bartlesville,  OK

-------
                                                               42
                                                           Table 15

                                                  Input Data to Exposure Model
                                                Primary Zinc Smelting Industry
                                                  (Assuming Baseline Controls)
Plant Lati
(Deg
Min
Sec
tude Longitude
rees (Degrees
utes Minutes
onds) Seconds)
St. Joe - 40-40-12 80-20-10
Monaca, PA




ASARCO - 27-48-00 97-23-46
Corpus Christi,
TX
Amax - 38-36-07 90-10-16
Sauget, IL
Jersey 36-30-54 87-24-14
Miniere
Zinc Coup .
Clarksville, TN
Emission
Rate
(Kg/yr)
109
8.4
67.7
8.4
8.4
. 23.5
25.2
Emission
Point
Elevation
(Meters)
61
31
37
32
92
46
61
Emission
Point
Diameter
(Meters)
3.4
1.8
2.1
2.0
2.0
1.5
1.8
Emission
Point
Cross
Sectional
Area (m^)
100
100
100
100
100
100
100
Emission
Point Gas
Exit
Velocity
m/sec
7.0
1.8
10.7
1.1
2.7
9.7
6.3
Emission
Point Gas
Temp .
325
325
336
325
389
373
334
Emission
Point
Type
Stack
Stack
Stack
Stack
Stack
Stack
Stack
National  Zinc  36-44-24    95-58-59
 Bartelsvi 1 le,
 OK
11.3
31
1.3
100
5.7
325
                                                        Stack

-------
                            43
                         Table 16
        Total  Exposure and  Number of People Exposed
                    Primary Zinc Smelter*
                        Total                           Total
                      Number  of                        Exposure
Plant
1
2
3
4
5
People Exposed
2,000,000
336,000
2,200,000
235,000
120,000
V people - ug/m~j
47
2
16
3
2
* A 50-kilometer radius was used for the analysis of primary
  zinc smelters.

-------
                                      44
                                  Table 17

                  Public  Exposure  for Primary Zinc Smelters
                   as  Produced  by  the Human Exposure Model
                         (Assuming Baseline Controls)
Concentration
Level (ug/m3)
0.00182
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
0.0000001
0.00000005
Population
Exposed
(Persons)*
7
109
2350
19400
91700
212000
450000
1870000
3170000
3960000
4650000
4800000
4850000
4870000
4890000
Exposure
(Persons - ug/m^**
0
0
2
8
18
27
35
55
65
68
69
69
69
69
69
 * Column 2 displays  the  computed  value,  rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration levels found  in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be  rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                    45
3.6   Public Exposure  to  Inorganic Arsenic Emissions from Zinc Oxide
     Plants

3.6.1  Source Data

       Two zinc oxide plants are  included In  the analysis.  Table 18 lists
ambient arsenic concentrations  near  select zinc oxide plants.  Table 19
lists the names and addresses of  the plants   considered, and Table 20 lists
the plant data used as input to the  Human Exposure Model (HEM).

3.6. 2  Exposure Data

       Table 21 lists on a pi ant-by-plant basis, the total  number of people
encompassed by the exposure analysis and the  total exposure.  Total
exposure is tHe sum of tHe products  of number of people times the ambient
air concentration to which they are  exposed,  as calculated  by HEM.
Table 22 sums, for the entire source category (2 plants), the numbers of
people exposed to various ambient concentrations, as calculated by HEM.
I Source-by-source exposure results are provided in the EPA  docket numbered
A-83-11.)

-------
                                          Table 18

                                 Arsenic  Concentrations  Near
                                  Select  Zinc  Oxide  Plants

Plant
ASARCO-
Columbus, OH

# Obs
127

Distance1
(km)
3.8


Bearing
206.0

Predicted2
(uq/m3)
0.0000124

Measured3
(ug/m3)
0.006

MDL4
(ug/m3)
0.0055



Percentile^
70< 51

', <90*

New Jersey Zinc-         No data  within  15  km
 Palmerton, PA
* Indicates data point was  disregarded;  see  Section  3.5.1.1.
* Distance from source to monitor (km).
2 Concentration predicted by Human Exposure  Model  (HEM).
J The measured values  are weighted averages.  When the sanpled arsenic concentrations
  were below the MDL,  a value of  1/2  MDL was assumed for purposes of calculating the
  annual averages.
4 Minimum detection limit.
5 Percentila indicates percentage of  data falling below minimum detectable levels.

-------
                                     47


                                  TABLE 19

                    Identification of Zinc Oxide Plants

Plant Number Code	Plant Name and Address

     1                                                ASARCO-Columbus,  OH
                                                     New Jersey Zinc
                                                     Palmerton, PA.

-------
                      48
                   Table  20

Input Data to Exposure  Model  Zinc Oxide  Plants
         (Assuming Baseline Controls)
Plant
Latitude
(Degrees
Minutes
Seconds)
Longtitude
(Degrees
Minutes
Seconds )
Emission
Rate
(Kg/yr)
Emission
Point
Elevation
(Meters)
Emission
Point
Diameter
(Meters)
Emission*
Point
Cross
Sectional
Area (m^)
Emission
Point Gas
Exit
Velocity
m/sec
Emission
Point Gas
Temp .
ASARCO-
Columbus, OH 39-59-53 82-58-48 11.3 61 1.2 100 7.5 333
Emission
Point
Type
Stack
New Jersey 3155 24 5 100 7.3 411 Stack
Zinc- 40-49-41 75-35-22 2656 18 1.8 100 7.9 364 Stack
Palmer-ton, PA 2754 9 1.2 100 17.4 466 Stack

-------
                                     49





                                  Table 21



                Total  Exposure and Number of People Exposed



                            (Zinc  Oxide Plants)*
Plant
1
2
Total Number of
People Exposed
1,210,000
907,000
Total Exposure
(People - ug/m^)
8
1260
* A 50 kilometer radius was used for the analysis of zinc oxide plants,

-------
                                        50
                                   Table 22
                    Public Exposure for Zinc Oxide Plants
                   as Produced by the Human Exposure Model
                         (Assuming Baseline Controls)

 Concentration                  Population Exposed                Exposure
 Level  (ug/mj)	;	(Persons)*
**
0.269
0.25
0.1
0.05
0.025
0.01
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
2
2
138
1160
3990
11700
21300
54000
392000
732000
883000
908000
913000
976000
1160000
1360000
1610000
1840000
1960000
2110000
o
(J
17
-L /
79
180
300
366
474
921
-s I— J.
1200
1250
1260
1260
1260
1260
1260
1270
1270
1270
1270
 * Column 2 displays  the  computed  value,  rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration levels found in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be  rounded to 1.

** Column 3 displays  the  conputed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                     51

3.7  Methodology for Reviewing Pollutant Concentrations - Cotton Gins

     A total  of 320 cotton gins were  identified  as processors of arsenic
desiccated cotton.  Due to the large  number  of gins, EPA determined that  it
was impractical to obtain the location  data  necessary for arsenic  risk
assessment.  Based on information regarding  the  range of processing rates
possible, four model plants operating at  4,  7, 12 and 20 bales/hour were
designed that are representative of the operations and  emissions of the gin
population.7 These were located at each of  three sites  typical  of  the areas
in which the gins are located.  Of the 320  gins, it  was assumed that 32
processed 4 bales per hour, 96 processed  7  bales per hour,  160  processed  12
bales per hour, and 32 processed 20 bales per hour.  The Human  Exposure
Model was run for each scenario to establish a  range of exposure and risk
estimates for individual sources.  To provide data for  validating  the model
plant exposure estimates, two operating gins in  south central Texas were
chosen for test sites over a one year period.  Monitors were arranged  in  a
fan-like array of sites positioned at distances  of  100, 200 and 400 meters
downwind of the gin.  Upwind sites were placed at  400m  (one gin only) and
100m.  This configuration provided a total  of 13 sanpling  sites.   The study
was conducted over  a period of one year with intense sampling (4 hour
intervals) for  15 days during  the short ginning season  followed by 6 day
interval sampling for the  remainder of the year.

      Data  from these two  gins  were compared to Human Exposure Model
calculated values  (Table  23).  The comparison was hampered somewhat by
the  large  number  of monitored  values which fell below minimum detection
 limits  --  only 298  measurements  out of 708 were above the MDL of 0.05
ug/m3.   To circumvent  this problem, a  range  of  mean measured values was
 developed. At  one end,  all  values  below MDL  were considered as zero
 values, and at the  other  end,  all  such values were considered equal to the
 MDL of  0.05 ug/m3.

-------
                                   52


                                Table 23

           Arsenic Concentrations Near Two Texas Cotton Gins
Plant
A ( =9 bales/hr)



B ( =12 bales/hr)


Distance ( km)1
0.1
0.2
0.2
0.1
0.1
0.2
0.1
Predicted (HEM)
( uq/m3)
._
0.011
0.011
—
— _
0.011
—
Measured2
( ug/m3)
0.083-0.088
••_
0.051-0.060
0.12 -0.12
0.015-0.024
0.013-0.022
0.013-0.022
Distance from source to monitor.

Weighted mean concentrations for one calendar year.   Lesser value
represents weighted mean concentration calculated with  values less
than minimum detection limit set equal  to zero.   Greater  value
represents weighted mean concentration calculated with  values less
than minimum detection limit set equal  to MDL (0.0065  ug/m3).

-------
                                     53
     When conparing the measured arsenic  values  to  the predicted  con-
centrations from the appropriate model  gin exposure analysis,  EPA found
that the predicted values were reasonably close  to  concentrations measured
very near the gins.  The monitoring study data also showed  that the
arsenic concentrations fell  off very rapidly with distance  from the  gins.
This result suggests that people living at some  distance  from  the gins
are not being significantly  exposed to  the gins' emissions. Such a
result, coupled with the observation that many gins are  in  rural  areas
supports the Agency's conclusion that the aggregate risks for  this source
category are low.

3.7.1  Public Exposure to Inorganic Arsenic Emissions from  Cotton Gins

3.7.1.1  Source Data

     Four model cotton gins  at each of  three geographic  locations are
included in the analysis.  Table 24 lists the names and  addresses of the
plants considered, and Table 25 lists the plant  data used as  input to  the
Human Exposure Model (HEM).

3.7.1.2  Exposure Data

     Tables 26 - 37 sum, for the entire source category  (12 plants), the
numbers of people exposed to various ambient concentrations, as calculated
by HEM.  (Model plant-by-model plant exposure results  are provided in  the
EPA docket numbered A-83-10.)

-------
                                 54
                              Table 24
                Identification of Model Cotton Gins
Model  Plant Location
Model Plant Production
   Hutto, TX
   4 Bales/Hour
   7 Bales/Hour
  12 Bales/Hour
  20 Bales/Hour
   Buckholtz,  TX
   4 Rales/Hour
   7 Bales/Hour
  12 Bales/Hour
  20 Bales/Hour
   Itasca,  TX
   4 Bales/Hour
   7 Bales/Hour
  12 Bales/Hour
  20 Bales/Hour

-------



55




Table 25
Input Data to Exposure Model Cotton Gins

Plant Latitude

(Degrees
Minutes
Seconds )
Hutto, TX 30-33-00
4 Bales/Hour

7 Bales/Hour

12 Bales/Hour

20 Bales /Hour

Buckholts, TX 30-52-00
4 Bales/Hour

7 Bales /Hour

12 Bales /Hour

20 Bales/Hour

Itasca, TX 32-10-00
4 Bales /Hour

7 Bales /Hour

12 Bales /Hour

20 Bales/Hour


Longitude Emission
Rate
.(Degrees
Minutes
Seconds) (Kg/yr)
97-33-00
1.0
1.0
2.7
2.7
4.6
4.6
10.2
10.2
97-OB-OO
1.0
1.0
2.7
2.7
4.6
4.6
10.2
10.2
97-09-00
1.0
1.0
2.7
2.7
4.6
4.6
10.2
10.2
(Assuming
Emission
Point
Elevation
(Meters)

9
5
9
5
9
5
10
5

9
5
9
5
9
5
10
5

9
5
9
5
9
5
10
5
Baseline Controls)
Emission
Point
Diameter
(Meters)

0.3
—
0.4
—
0.4
—
0.4
—

0.3
—
0.4
—
0.4
—
0.4
—

0.3
—
0.4
—
0.4
—
0.4
—
Emission
Point
Cross
Sectional
Area (itr)

25
12
25
12
25
27
25
27

25
12
25
12
25
27
25
27

25
12
25
12
25
27
25
27
Emission
Point Gas
Exit
Velocity
m/sec

20.4
—
20.4
—
20.4
—
20.4
—

20.4
—
20.4
—
20.4
—
20.4
—

20.4
—
20.4
—
20.4
—
20.4
—
Emission
Point Gas
Temp.
(°K)

298
298
298
298
298
298
298
298

298
298
298
298
298
298
298
298

298
298
298
298
298
298
298
298
Emission
Point
Type


Stack
Fugitive
Stack
Fugitive
Stack
Fugitive
Stack
Fugitive

Stack
Fugitive
Stack
Fugitive
Stack
Fugitive
Stack
Fugitive

Stack
Fugitive
Stack
Fugitive
Stack
Fugitive
Stack
Fugitive

-------
                                     56
                                  Table 26

             Public Exposure for 4 Bales/Hour Model  Cotton  Gin
             (Hutto.TX) as Produced by the Human Exposure Model
                        (Assuming Baseline Controls)

Concentration                 Population Exposed                Exposure
Level  (ug/m3)	(Persons)*	(Persons-ug/nr3)**
0.00263
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
1
1
1
6
23
112
177
433
1810
1810
3390
46800
285000
506000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
 * Column  2  displays  the  computed  value, rounded to the nearest whole number,
   of the  cumulative  number  of people exposed to the matching and higher
   concentration levels found  in column  1.  For example, 0.5 people would he
   rounded to 0 and 0.51  people would be rounded to 1.

** Column  3 displays  the  computed  value  of  the cumulative exposure to the
   matching and higher concentration  levels found in column 1.

-------
                                     57
                                  Table 27
             Public Exposure for 7 Bales/Hour Model Cotton Gin
             (Hutto.TX) as Produced by the Human Exposure Model
                        (Assuming Baseline Controls)

Concentration                 Population Exposed                Exposure
                                  (Persons)*	(Persons-ug/m-*)**
O.OU682
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.000000528
1
1
1
8
28
112
282
799
1810
2420
7220
55400
448000
506000
0
0
0
0
0
0
0
0
0
0
0
0
1
1
 * Column 2 displays  the  computed  value,  rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration  levels found  in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be  rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration  levels found in column 1.

-------
                                      58
                                   Table 28
              Public Exposure for 12  Bales/Hour  Model Cotton Gin
              (Hutto.TX) as Produced  by  the  Human  Exposure Model
                         (Assuming Baseline  Controls)

 Concentration                 Population  Exposed                Exposure
 Level (ug/mj)	  (Persons)*	(Persons-ug/m3)**
0.011
0.01
0.005
0.0025
0.001
0.0005
0.00025
0.0001
O.OOOU5
0.000025
0.00001
0.000005
0.0000025
0.000001
1
1
1
5
25
102
161
523
1810
1810
3820
39500
232000
506000
0
0
0
0
0
0
0
0
0
0
0
1
1
2
 * Column 2 displays  the  computed  value, rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration levels found  in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                     59
                                  Table 29

             Public Exposure for 20 Bales/Hour Model Cotton Gin
             (Hutto.TX) as Produced by the Human Exposure Model
                        (Assuming Baseline Controls)

Concentration                 Population Exposed                Exposure
                                  (Persons)*	(Persons-ug/m-3)**
0.0234
0.01
0.005
0.0025
0.001
0 .0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
1
1
5
23
102
169
433
1810
1810
3390
46800
300000
506000
0
0
0
0
0
0
0
1
1
1
1
3
4
 * Column 2  displays  the computed  value, rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration  levels found in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51 people would be rounded to 1.

** Column 3  displays  the computed  value of the cumulative exposure to the
   matching  and higher concentration  levels found  in column 1.

-------
                                      60
                                   Table 30
              Public Exposure for 4 Bales/Hour Model  Cotton Gin
             (Buckholts,TX) as Produced by the Human Exposure Model
                         (Assuming Baseline Controls)

 Concentration                 Population Exposed                Exposure
 Level  (u9/m3)	(Persons)*                 (Persons-ua
**
0.00263
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
0.000000196
<1
<1
<1
2
10
49
77
190
1050
1050
4020
10600
86700
129000
131000
n
\J
o
o
o
o
o
o
o
o
o
o
o
n
\J
o
0
 * Column 2 displays  the  computed  value,  rounded to the nearest whole number
   of the cumulative  number  of people exposed to the matching and higher
   concentration  levels found in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be  rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                     61
                                  Table  31
             Public  Exposure for 7  Bales/Hour  Model  Cotton Gin
           (Buckholts.TX)  as Produced  by  the Human Exposure Model
                        (Assuming Baseline  Controls)

Concentration                 Population  Exposed                 Exposure
                                  (Persons)*	(Persons-ug/nr3)**
0 .00682
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.000000528
<1
<1
<1
3
12
49
124
269
1050
1050
6020
15500
121000
131000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
 *  Column  2 displays the computed value, rounded to the nearest whole number,
   of  the  cumulative number of people exposed to the matching and higher
   concentration  levels found in column 1.  For example, 0.5 people would be
   rounded to  0 and 0.51 people would be rounded to 1.

**  Column  3 displays the computed value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                       62
                                    Table  32

                          x     f°r  1Z  Bales/Hour  Model Cotton Gin
                       .TX)  as  Produced  by  the  Human Exposure Model
                          (Assuming Baseline Controls)
  Concentration
  Level  (uq/m3)

  0.011
  0.01
  0.005
  0.0025
  0.001
  0.0005
  0.00025
  0.0001
  0.00005
  0.000025
  0.00001
  0.000005
  0.0000025
  0.000001
Population Exposed
    (Persons)*
            2
           11
           45
           71
          230
         1050
         1050
         6020
        10100
        81500
       131000
   Exposure
(Persons-ug/rn3)**
       0
       0
       0
       0
       0
       0
       0
       0
       0
       0
       0
       0
       0
       1
   Co.lumn 2 delays the corrputed value,  rounded  to  the  nearest whole  number,
   of the cumulative number of people  exposed  to  the matching  and  higher
   concentration levels found in column  1.   For exanple,  0.5 people would  be
   rounded to 0 and 0.51 people would  be  rounded  to  1.

** Column 3 displays the corrputed value of  the cumulative exposure to  the
   matching and higher concentration levels  found in column 1.

-------
                                     63
                                  Table 33

              Public Exposure for 20 Bales/Hour Model Cotton Gin
            (Buckholts.TX) as Produced by the Human Exposure Model
                         (Assuming Baseline Controls)

 Concentration                Population Exposed                Exposure
 Level  (ug/m3)	(Persons)*	(Persons-ug/m3)**
0.0234
0.01
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000002
<1
<1
2
10
45
74
190
1050
1050
4020
11300
89500
129000
131000
0
0
0
0
0
0
0
0
0
0
1
1
1
1
 * Column 2 displays  the  computed  value, rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration levels found  in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                      64
                                   Table 34
              Public Exposure for 4 Bales/Hour Model  Cotton Gin
              (Itasca.TX) as Produced by the Human Exposure  Model
                         (Assuming Baseline Controls)

 Concentration                 Population Exposed                Exposure
 Level  (ug/m3)	(Persons )*                 (Persons-no
0.0011
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
0.0000001
0.0000000634
1
1
5
19
57
153
489
1280
2140
2660
6520
38900
107000
156000
162000
	 yrci auiia ~ \J\}/ HI ) —
o
n
\J
o
o
\J
0
o
\J
Q
\J
o
o
o
o
0
o
o
0
 * Column 2 displays  the  computed  value, rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration  levels found in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                      65
                                  Table 35

              Public  Exposure for 7 Bales/Hour Model Cotton Gin
             (Itasca,TX) as  Produced by the Human Exposure Model
                         (Assuming Baseline Controls)

 Concentration                Population Exposed                Exposure
 Level  (ug/m^)      	(Persons)*	(Persons-ug/nr)**
0.00285
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.00000025
0.000000171
1
1
7
23
70
167
587
1280
2140
3870
6520
65200
120000
159000
162000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
 * Column 2 displays  the  computed  value,  rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration levels found in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be  rounded to 1.

** Column .3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration levels found in column 1.

-------
                                      66
                                   Table 36
              Public Exposure for 12 Bales/Hour Model  Cotton Gin
              (Itasca.TX) as Produced by the Human Exposure Model
                         (Assuming Baseline Controls)

 Concentration                 Population Exposed                Exposure
 Level  (pg/m-3)	(Persons)*                 fPersons-uo
0.00461
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0.0000005
0.000000293
1
4
22
42
118
489
948
1980
2660
4970
20300
114000
153000
162000
	 yrci auiia ~n^)/ in /
n

-------
                                     67
                                  Table 37
             Public Exposure for 20 Bales/Hour Model Cotton Gin
             (Itasca.TX) as Produced by the Human Exposure Model
                        (Assuming Baseline Controls)

 Concentration                 Population Exposed                Exposure
 ,—T  /.._/_3\                     (Persons)*	 (Persons-ug/rrr)**
0.0097
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0.000005
0.0000025
0.000001
0 .0000006 49
1
4
14
46
146
489
1280
2140
2660
6520
40100
109000
156000
162000
0
0
0
0
0
0
0
0
0
1
1
1
1
1
 * Column 2 displays  the  computed  value,  rounded to the nearest whole number,
   of the cumulative  number  of people exposed to the matching and higher
   concentration levels found  in column 1.  For example, 0.5 people would be
   rounded to 0 and 0.51  people would be  rounded to 1.

** Column 3 displays  the  computed  value of the cumulative exposure to the
   matching and higher concentration  levels found  in column 1.

-------
                                     63
3.8  Public Exposure to Inorganic Arsenic Emissions  from Arsenic
     Plants

3.8.1  Source Data

       Eight arsenic chemical  plants are included in the analysis.   Table 38
lists ambient arsenic concentrations near select  arsenic chemical  plants.
Table 39 lists the names and addresses of the plants considered, and Table
40 lists the plant data used as input to the Human Exposure  Model  (HEM).

3.8.2  Exposure Data

       Table 41 lists, on a plant-by-plant basis, the total  number  of people
encompassed by the exposure analysis and the total  exposure.   Total  exposure
is the sum of the products of numbers of people  times the ambient air con-
centration to which they are exposed, as calculated  by HEM.   Table  42 sums,
for the entire source category (8 plants), the numbers of people exposed  to
various ambient concentrations, as calculated by  HEM.  (Source-by-source
exposure results are provided in the EPA docket  numbered A-83-23.)

-------
69
Table 38
Arsenic Concentrations Near Select
Arsenic Chemical Plants
Distance1 Predicted2 Measured3
Plant
Diamond Shamrock-
Greens Bayou, TX






Koppers Co.-
Conley, GA
Koppers Co.-
Valparaiso, IN
Mineral R&D-
Concord, NC
Osmose Wood
Preserving Co.-
Memphis, TN
Pennwalt Inc.-
Bryan, TX
Vineland Chemical-
Vineland, NJ
# Obs
302
20
37
143
261
45
26
30
76


32
72

107

(km)
1.
5.
5.
6.
8.
8.
11.
12.
13.
No
No
5.
9.

0.
No
6
5
9
2
3
7
5
0
0
data
data
9
3

1
data
Bearing
291.1
229.7
186.3
197.3
70.3
62.2
195.9
143.0
336.1
within 15
within 15
27.7
20.8

296.4
within 15

0
0
0
0
0
0
0
0
0
km
km
0
0


km
(yg/m3)
.00000042
.000000042
.000000051
.000000029
.0000000118
.0000000105
.0000000121
.0000000090
.0000000079


.000099
.000054



(ug/m3)
0
0
0
0
0
0
0
0
0


0
0

0

.027
.026
.008
.027
.025
.026
.025
.025
.008


.005
.004

.026

MOL4
(ug/m3)
0.05
0.05
0.0055
0.05
0.05
0.05
0.05
0.05
0.0055


0.0055
0.0055

0.05



Percent! le^
90< %

50< %
95< %

95< %
100
100
70< %


70< °k
50< 5

95< 51

<95*
>99*
<70*
<99*
>99*
*
*
*
<90*


; <9o*
; <7o*

i <99*

 Group-Bonham.TX


* Indicates data point was  disregarded;  see  Section  3.7.1.1.          ~      ~~~
* Distance from source to monitor (km).
2 Concentration predicted by Human Exposure  Model  (HEM).

3 The measured values  are weighted averages.  When the  sampled  arsenic concentrations were below the  MDL,  a
  value of 1/2 MDi was assumed  for purposes  of calculating  the  annual averages.
4 Minimum detection limit.

  levels*116 1ndicates Percentage of  data  falling below minimum detectable

-------
                                    70

                                 Table 39
                 Identification of Arsenic Chemical Plants
Plant Number Code                          Plant Name and Address

       1                         Diamond Shamrock - Greens Bayou, TX
       2                         Koppers Co.  - Conley, GA
       3                         Koppers Co., - Valparaiso, IN
       4                         Mineral Research & Development Co.  -
                                   Concord, NC
       5                         Osmose Wood Preserving Co., -Memphis,  TN
       6                         Pennwalt Inc. - Bryan, TX
       7                         Vineland Chemical - Vine!and, NJ
       8                         Voluntary Purchasing Group - Bonham, TX

-------
                         71
                      Table  40

Input Data to Exposure Model  Arsenic  Chemical  Plants
             (Assuming Baseline  Controls)
Plant
Latitude
(Degrees
Minutes
Seconds )
Longitude
(Degrees
Minutes
Seconds)
Diamond Shamrock- 29-45-58 95-12-22
Greens Bayou ,TX
Koppers Co.-
Conley, GA
Koppers Co.-
Valparaiso.IN
Mineral R&D Co
Concord, NC
Osmose Wood
Preserving Co
Memphis.TN
Pennwalt Inc.-
Bryan, TX
Vineland
Chemical-
Vineland, NJ
Voluntary
Purchasing
Group-
Bonham.TX
33-38-42 84-19-34
41-28-34 87-04-40
.- 35-24-29 80-34-44
35-05-13 90-04-19
>""
30-40-30 96-22-12
39-55-59 74-44-53
33-34-41 96-10-41
Emission
Rate
(Kg/yr)
0.030
0.027
0.054
0.022
51.3
0.019
0.001
0.019
Emission
Point
Elevation
(Meters)
13
31
9
11
5
10
13
10
Emission
Point
Diameter
(Meters)
0.38
0.61
0.76
0.50
0.36
0.53
0.38
0.53
Emission
Point
Cross
Sectional
Area (m2)
3000
3000
3000
3000,
3000
3000
3000
3000
Emission
Point Gas
Exit
Velocity
m/sec
11.9
0.1
1.6
8.8
14.3
8.9
0.1
8.9
Emission
Point Gas
Temp .
(°K)
298
298
298
298
298
298
298
298
Emission
Point
Type
Stack
Stack
Stack
Stack
Stack
Stack
Stack
Stack

-------



Plant
1
2
3
4
5
6
7
8
72
Table 41
Total Exposure and Number of People
(Arsenic Chemical Plants)*
Total Number of
People Exposed
2,680,000
1,900,000
1,190,000
813,000
927,000
138,000
4,230,000
152,000


Exposed
Total Exposure
(People - ug/m3)
0
0
0
0
68
0
0
0
* A 50-kilometer radius was used for the analysis of arsenic chemical
  olants.

-------
                                     73
Concentration
                  Table 42

Public Exposure for Arsenic  Chemical  Plants
  as Produced by the Human Exposure  Model
        (Assuming Baseline Controls)

              Population Exposed
                  (Persons)*
   Exposure
(Persons-gg/m3)**
0.0541
0.05
0.025
0.01
0.005
0.0025
0.001
0.0005
0.00025
0.0001
0.00005
0.000025
0.00001
0 .000005
0.0000025
0.000001
0.0000005
0.00000025
0.0000001
0.00000005
0.000000025
0.00000001
0.000000005
0.0000000025
0.000000001
0.0000000005
0.00000000025
0.0000000001
0.00000000005
0.0000000000395
31
31
62
472
1440
2720
8130
16700
30900
105000
210000
374000
653000
852000
908000
935000
952000
984000
1040000
1100000
1320000
2230000
3670000
5880000
7750000
7890000
8320000
9910000
11900000
12000000
2
2
3
10
15
19
28
34
38
49
56
61
66
68
68
68
68
68
68
68
68
68
68
68
68
68
68
68
68
68
 *Column 2 displays  the  computed  value,  rounded to the nearest whole number,
  of the cumulative  number  of people  exposed to the matching and higher
  concentration levels found  in column  1.   For example, 0.5 people would be
  rounded to 0 and 0.51  people would  be  rounded to 1.

**Column 3 displays  the  computed  value  of  the cumulative exposure to the
  matching and higher concentration  levels  found  in column 1.

-------
                                     74
4  QUANTITATIVE EXPRESSIONS OF PUBLIC CANCER RISKS FROM  INORGANIC ARSENIC
   EMISSIONS

4.1  Methodology (General)

4.1.1  The Two Basic Types of Risk

     Two basic types of risk are dealt with in  the analysis.   "Aggregate
risk" applies to all of the people encompassed  by  the  particular  analysis.
Aggregate risk can be related to a single source,  to all  of the  sources  in
a source category, or to all  of the source categories  analyzed.   Aggregate
risk is expressed as incidences of cancer among all of the  people included
in the analysis, after 70 years of exposure.  For  statistical  convenience,
it is often divided by 70 and expressed as cancer  incidences per  year.
"Individual  risk" applies to the person or persons estimated to live  in the
area of the highest ambient air concentrations  and it  applies  to  the  single
source associated with this estimate as estimated  by the  dispersion model.
Individual  risk is expressed as "maximum lifetime  risk" and  reflects  the
probability of getting cancer if one were continuously exposed to the
estimated maximum ambient air concentration for 70 years.

4.1.2  The Calculation of Aggregate Risk

     Aggregate risk is calculated by multiplying the total  exposure produced
by HEM (for a single source,  a category of sources, or all  categories of
sources) by the unit risk estimate.   The product is cancer  incidences among
the included population after 70 years of exposure.  The  total exposure,
as calculated by HEM, is illustrated by the following  equation:

                                      N
                     Total  Exposure =  Z    (P-jCj)

-------
                                     75
where
     £  = summation over all  grid points where exposure  is  calculated
     Pi = population associated with grid point i,
     Cj = long-term average inorganic arsenic  concentration at grid  point  i,
     N  = number of grid points to 2.8 kilometers and  number of ED/BG
          centroids between 2.8 and 50 kilometers of each  source.
To more clearly represent the concept of calculating aggregate risk, a
simplified example illustrating the concept follows:
                                  EX/MPLE
     This example uses assumptions rather than actual  data  and uses  only
three  levels of exposure rather than the large number  produced by  HEM.  The
assumed unit risk estimate is 4.29 x 10'3 at 1 gg/m3 and the assumed
exposures are:
            ambient air                      number of people exposed
          concentrations                      to given concentration

          2    ug/m3                                   1,000

          1    ug/m3                                  10,000

          0.5  ug/m3                                 100,000
The  probability of getting cancer if continuously exposed to the assumed
concentrations for 70 years is given by:
   concentration                unit risk               probability  of cancer

     2    pg/m3         x   4.29 x  10-3 (ug/rn3)'1   =         9 x 10~3

     1    ug/m3         x   4.29 x  10'3     "        =         4 x 10'3

     0.5  ug/m3         x   4.29 x  10'3     "        =         2 x 10'3

-------
                                     76
 The 70  year  cancer  incidence among the people exposed to these concentrations
 is given  by:
     probability  of  cancer          number of people at        after 70 years
     at  each  exposure level         each exposure level          of exposure
9 x
4 x
2 x
10-3
10~3
10-3
x
x
X
1
10
100
,000
,000
,000
                                                                     9
                                                                    40
                                                                   200
                                                           TOTAL =  249
The aggregate risk, or total cancer incidence, is 249 and, expressed
as cancer incidence per year, is 249 * 70, or 3.6 cancers per year.   The
total cancer incidence and cancers per year apply to the total  of 111,000
people assumed to be exposed to the given concentrations.
4.1. 3  The Calculation of Individual Risk
     Individual  risk, expressed as "maximum lifetime risk,"  is calculated
by multiplying the highest concentration to which the public is exposed,  as
reported by HEM, by the unit risk estimate.   The product,  a  probability of
getting cancer,  applies to the number of people  which HEM  reports as being
exposed to the highest listed concentration.   The concept  involved is a
simple proportioning from the 1 ug/m3 on which the unit risk estimate is
based to the highest listed concentration.   In other words:
       maximum lifetime risk          the unit risk  estimate
     highest concentration to                1 ug/m3
     which people are exposed
4.2  Risks Calculated for Emissions of Inorganic Arsenic
     The explained methodologies for calculating maximum lifetime risk  and
cancer incidences were applied to each  plant,  assuming  a baseline level of
emissions.   A baseline level  of emissions means  the  level  of emissions  after

-------
                                     77
the application of controls either currently in place or required to be in
place to comply with current state or Federal  regulations but before application
of controls that would be required by a NESHAP.

     Tables 43-49 summarize the calculated risks for each source category.
To understand the relevance of these numbers,  one should refer to the
analytical uncertainties discussed in section  5 below.  Note that the annual
incidence is not calculated for cotton gins.  As mentioned earlier in this
document, it was impractical to identify and locate all  the gins handling
arsenic-acid-desiccated cotton ( = 300 gins).   The Agency does not have enough
available data to provide an estimate of annual cancer incidence that would
be comparable in accuracy to the other source  category estimates.  As outlined
in Section 3.7, three model gins operating at  each of four production rates
were used to establish a range of exposure and risk estimates for individual
sources.  Likewise, two operating gins in south central  Texas were chosen
for ambient air monitoring in order to validate the model plant exposure
estimates.  Maximum lifetime risk estimates were calculated for each of the
three model plants (Table  47) and for the two  operating gins (Table 48).

-------
                                78
                             Table 43
Maximum Lifetime Risk  and Cancer Incidence for Primary Lead Smelters
                   (Assuming Baseline Controls)

                         Maximum
                         Lifetime              Cancer Incidences
 P1ant  	Risk     	Per Year

   1                    2 x 10-3                    0>Q13

   2                    4 x 10-5                    0>Q44

   3                    2 x 10-5                    0
-------
                                  79
                              Table 44
Maximum Lifetime Risk and  Cancer  Incidence for Secondary Lead Smelters
                     (Assuming Baseline Controls)
Plant
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Maximum
Lifetime
Risk
5 x 10-6
4 x ID'6
8 x 10-6
8 x 10~6
8 x ID'5
1. 1 x 10-5
5 x ID'6
1. 1 x 10-4
2 x 10~7
1.6 x 10~5
9 x 10~6
4 x Kr6
1.1 x 10"5
4 x ID'6
6 x lO-5
9 x lO"6
4 x 10-6
3 x 10-4
8 x 10-6
4 x ID'5
4 x 10-4
2 x 10-4
4 x 10 ~4
2 x ID'5
7 x ID'5
3 x lO'4
2 x 10-6
1.2 x ID'5
1.7 x 10-6
1. 1 x 10~6
1.5 x 10-5
6 x ID"6
1.5 x lO-5
3 x lO'5
2 x lO-5
Cancer Incidences
Per Year
0.0019
0.0009
0.0010
0.0011
0.014
0.0010
0.0013
0.015
0.0002
<0.0001
0.0010
0.011
0.0019
0. 0009
0.0024
0.0004
0.0035
0.040
0.0031
0.012
0.14
0.028
0.035
0. 0040
0.0069
0.015
0.0002
0.0007
0.0002
0.0001
0.0015
0.0050
0. 0095
0.0013
0.0062

-------
                                 80
                              Table 45
Maximum Lifetime Risk and Cancer  Incidence for Primary Zinc Smelters
                    (Assuming  Baseline Controls)
          Plant
   Maximum
  Lifetime
    Risk
Cancer Incidences
     Per Year
           1
           2
           3
           4
           5
  8 x lO-6
1.9 x ID'7
1.1 x ID'6
  9 x 10-7
  3 x ID'6
     0.0029
     0.0001
     0.0010
     0.0002
     0.0001

-------
                               81
                            Table 46
Maximum Lifetime Risk and Cancer Incidence for Zinc Oxide Plants
                  (Assuming Baseline Controls)

                      Maximum
                     Lifetime          Cancer Incidences
          Plant	Risk	Per Year

            1         4 x 10-7               0.005

            2       1.2 x 10'-3               0.077

-------
                              82
                           Table 47
Maximum Lifetime Risk  and Cancer Incidence for Model  Cotton Gins
                  (Assuming Baseline Controls)

Model
Plant
Hutto, TX
4 Bales/Hr
7 Bales/Hr
12 Bales/Hr
20 Bales/Hr
Buckholts.TX
4 Bales/Hr
7 Bales/Hr
12 Bales/Hr
20 Bales/Hr
Itasca,TX
4 Bales/Hr
7 Bales/Hr
12 Bales/Hr
20 Bales/Hr
Maximum
Lifetime
Risk

1.1 x lO-5
3 x ID'5
5 x ID'5
1.0 x ID'4

1.1 x ID'5
3 x 10-5
5 x ID'5
1.0 x 10-4

5 x 10-6
1.2 x lO-5
2 x ID'5
4 x 10-5

Cancer Incidences
Per Year

O.OOOl
0.0001
0.0001
0.0002

<0.0001
<0.0001
<0.0001
0.0001

<0.0001
<0.0001
<0.0001
0.0001

-------
                   83
                Table 48

Lifetime Risk for Two Texas  Cotton  Gins
      (Assuming Baseline Controls)
 Plant	   Maximum Lifetime  Risk

   A                  5   x HT4*

   R                  1.0 x 10-4
  * Represents  final  risk  estimate as incorporated
    by EPA.

-------
                                  84
                               Table  49
Maximum Lifetime Risk and  Cancer  Incidence for Arsenic Chemical Plants
                     (Assuming  Baseline Controls)
Plant
1
2
3
4
5
6
7
8
Maximum
Lifetime
Risk
4 x lO'8
7 x ID'9
3 x 10-8
3 x 10-8
2 x 1CT4
3 x ID'8
9 x 1CT10
3 x lO-8
Cancer Incidences
Per Year
<0.0001
<0.0001
<0.0001
<0.0001
0.0042
<0.0001
<0.0001
<0.0001

-------
                                     85
5  ANALYTICAL UNCERTAINTIES  APPLICABLE  TO  THE  CALCULATIONS OF  PURLIC
   HEALTH RISKS CONTAINED IN THIS DOCUMENT

5.1  The Unit Risk Estimate

     The procedure used to develop the  unit risk  estimate is  described  in
referenced.  The model used and its application  to epidemiological data
have been the subjects of substantial comment  by  health  scientists.   The
uncertainties are too complex to be summarized sensibly  in  this  appendix.
Readers who wish to go beyond the information  presented  in  the reference
should see the following Federal Register  notices:   (1)  OSHA's "Supplemental
Statement of Reasons for the Final Rule",  48 FR 1864 (January  14,  1983);
and (2) EPA's "Water Quality Documents  Availability" 45  FR  79318 (November
28, 1980).

     The unit risk estimate used in this analysis applies only to  lung
cancer.  Other health effects are possible; these include skin cancer,
hyperkeratosis, peripheral neuropathy,  growth  retardation and brain
dysfunction among children, and increase in adverse birth outcomes.   No
numerical expressions of risks relevant to these health  effects  is included
in this analysis.

     Although the estimates derived from the various studies  are quite
consistent, there are a number of uncertainties associated with them.  The
estimates were made from occupational studies  that involved exposures only
after  employment age was reached.   In estimating risks from environmental
exposures throughout life,  it was assumed through the absolute-risk model
that the  increase in the age-specific mortality rates of lung cancer  was  a
function  only of cumulative exposures, irrespective of how the exposure was
accumulated.  Although this assunption provides an adequate description of

-------
                                     86
 all  of  the  data,  it may be in error when applied to exposures that begin
 very early  in  life.  Similarly, the linear models possibly are inaccurate
 at  low  exposures, even though they provide reasonable descriptions of the
 experimental data.

      The  risk  assessment methods employed were severely constrained by the
 fact that they were based only upon the analyses performed and reported by
 the  original authors—analyses that had been performed for purposes other
 than quantitative risk assessment.  For example, although  other measures of
 exposure might be more appropriate, the analyses were necessarily based
 upon  cumulative dose, since that was the only usable measure reported.  Given
 greater access to the data from these studies, other dose  measures, as well
 as models other than the sinple absolute-risk model, could be studied.  It
 is possible that such wide analyses would indicate that other approaches
 are  more appropriate than the ones applied here.

 5.2   Public Exposure

 5.2.1  General

      The basic  assumptions implicit in the methodology are that all  exposure
occurs at people's residences, that people stay  at the same location  for 70
years, that the ambient air concentrations and the emissions  which  cause
these concentrations persist  for 70 years, and that the concentrations  are
the same inside and outside the residences.  From this it  can be  seen  that
public exposure is based on a hypothetical  premise.  It is  not  known whether
this  results in an over-estimation or  an underestimation of public  exposure.

-------
                                     87
5.2.2  The Public
     The following are relevant to the public  as  dealt  with  in  this  analysis:

     1.  Studies show that all  people are not  equally susceptible  to cancer.
There is no numerical recognition of the "most susceptible"  subset of the
population exposed.

     2.  Studies indicate that  whether or not  exposure  to a  particular
carcinogen results in cancer may be affected by the person's exposure to
other substances.  The public's exposure to other substances is not
numerically considered.

     3.  Some members of the public included in this analysis are  likely  to
be exposed to inorganic arsenic in the air in  the workplace, and workplace
air concentrations of a pollutant are customarily much  higher than the
concentrations found in the ambient, or public air.  Workplace  exposures
are not numerically approximated.

     4.  Studies show that there  is normally a long latent period  between
exposure and the onset of lung cancer.  This has not been numerically
recognized.

     5.  The people dealt with in the analysis are not  located by  actual
residences.  As explained previously, people are grouped by census districts
and these  groups are located at single points called the population  centroids
The effect is that the actual  locations of residences with respect to the
estimated  ambient air concentrations are not known and that the relative
locations  used  in the exposure model may have changed since the 1980 census.
However, for the population sectors estimated to be at highest risk, U.S.

-------
                                     88
 Geological  Survey  topographical  maps were checked to verify that people did
 live or could live in  locations  near the sources as modeled predictions
 estimated.   Maps  in  certain  instances were old and the possibility could
 not  be  excluded  that additional  areas near sources have been developed
 since publication  of the maps.

     6.   Many people dealt with  in this analysis are subject to exposure to
 ambient  air  concentrations of inorganic arsenic where they travel and shop
 (as  in  downtown areas and suburban shopping centers), where they congregate
 (as  in public parks, sports stadiums, and schoolyards), and where they work
 outside  (as  mailmen, milkmen, and construction workers).  These types of
 exposures are not  numerically dealt with.

 5.2.3.   The  Ambient  Air Concentrations

     The  following are relevant to the  estimated ambient air concentrations
 of inorganic arsenic used in this analysis:

     1.   Flat terrain was assumed in  the dispersion  model.   Concentrations
 much higher  than those estimated would  result  if  emissions  impact on  elevated
 terrain or tall buildings near a plant.

     2.  The estimated concentrations do not account  for the  additive impact
 of emissions from plants  located close  to one  another.

     3.  The increase in  concentrations  that could result from  re-entrainment
of arsenic-bearing dust from,  e.g., city streets, dirt  roads, and vacant
 lots, is not considered.

-------
                                     89
     4.  Meteorological  data specific  to plant sites  are  not used  in  the
dispersion model.  As explained,  HEM uses the meteorological data  from  the
STAR station nearest the plant site.  Site-specific meteorological  data
could result in significantly different estimates,  e.g.,  the estimated
location of the highest  concentrations.

5.  tn some cases, the arsenic emission rates are estimates that are  based
on assumptions rather than on measured data.

-------
                                     90
 6  REFERENCES

 1.   National Academy of Sciences, "Arsenic," Committee on Medical and
     Biological  Effects of Environmental Pollutants, Washington, D.C., 1977.
     Docket  Number  (OAQPS 79-8) II-A-3.

 2.   Health  Assessment Document for Inorganic Arsenic - Final  Report EPA-600/
     8-83-021F March 1984, OAQPS Docket Number OAQPS 79-8, II-A-13.

 3.   U.S. EPA, et.al., "Environmental  Cancer and Heart and Lung Disease,"
     Fifth Annual Report to Congress  by the Task Force on  Environmental  Cancer
     and Health and Lung Disease,  August,  1982.

 4.   OAQPS Guideline Series,  "Guidelines on Air  Quality Models".   Publication
     Number EPA-450/2-78-027, (OAQPS  Guideline No.  1.2-080).

 5.   Systems Application,  Inc.,  "Human Exposure  to  Atmospheric  Concentrations
     of Selected Chemicals."   (Prepared for the  U.S.  Environmental Protection
    Agency, Research Triangle  Park, North  Carolina).   Volume  I, Publication
    Number EPA-2/250-1, and  Volume II, Publication  Number EPA-1/250-2.

6.  NEA, Inc.,  "East Helena  Source Apportionment Study    Particulate  Source
    Apportionment Analysis Using  the  Chemical Mass Balance Receptor Model."
     (Prepared for the Department  of Health  and  Environmental Sciences,  State
    of Montana.) Volume I, September,  1982.

7.  RADIAN  Corporation, "Preliminary  Study  of Sources of  Inorganic Arsenic."
    (Prepared for the U.S. Environmental Protection Agency, Research Triangle
    Park,  North  Carolina.)   Publication Number  EPA-450/5-82-005, August 1982.

-------

-------
  United States                              Office of Air and Radiation
  Environmental Protection                   Office of Air Quality Planning and Standards
  A9encV                                    Research Triangle Park, NC 27711
Official Business
Penalty for Private Use
$300
                                                                          If your address is incorrect, please change on the above label;
                                                                          tear off; and return to the above address.
                                                                          If you do not desire to continue receiving this technical report
                                                                          series, CHECK HERE D; tear off label; and return it to the
                                                                          above address.

-------