Risk and Technology Review - 

Analysis of Socio-Economic Factors for Populations Living Near Wool
Fiberglass Facilities

Prepared by:

EC/R Incorporated

501 Eastowne Drive, Suite 250

Chapel Hill, NC  27514

EPA Contract No. EP-D-06-119

Work Assignment No. 4-01

Prepared for:

Regina Chappell, Work Assignment Manager

Community and Tribal Programs Group

Office of Air Quality Planning and Standards

U.S. Environmental Protection Agency

Research Triangle Park, North Carolina 27711

August 17, 2011	Disclaimer

Although the analysis described in this document has been funded wholly
or in part by the United States Environmental Protection Agency contract
EP-D-06-119 to EC/R Incorporated, it has not been subject to the
Agency's review and therefore does not necessarily reflect the views of
the Agency, and no official endorsement should be inferred.Contents

  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc301356644"  1.	Introduction
  PAGEREF _Toc301356644 \h  1  

  HYPERLINK \l "_Toc301356645"  2.	Census Data	  PAGEREF _Toc301356645
\h  2  

  HYPERLINK \l "_Toc301356646"  3.	Calculation Methods	  PAGEREF
_Toc301356646 \h  4  

  HYPERLINK \l "_Toc301356647"  3.1	Racial, Ethnic and Age Categories
and the Total Population	  PAGEREF _Toc301356647 \h  4  

  HYPERLINK \l "_Toc301356648"  3.2	Level of Education	  PAGEREF
_Toc301356648 \h  5  

  HYPERLINK \l "_Toc301356649"  3.3	Household Income	  PAGEREF
_Toc301356649 \h  6  

  HYPERLINK \l "_Toc301356650"  3.4	Poverty Level	  PAGEREF
_Toc301356650 \h  7  

  HYPERLINK \l "_Toc301356651"  3.5	Disability	  PAGEREF _Toc301356651
\h  7  

  HYPERLINK \l "_Toc301356652"  3.6	Linguistic Isolation	  PAGEREF
_Toc301356652 \h  8  

  HYPERLINK \l "_Toc301356653"  4.	Results	  PAGEREF _Toc301356653 \h  8
 

  HYPERLINK \l "_Toc301356654"  5.	Uncertainty Discussion	  PAGEREF
_Toc301356654 \h  10  

  HYPERLINK \l "_Toc301356655"  Appendix A	A-  PAGEREF _Toc301356655 \h 
1  

 



Introduction

	This document describes the approach used to evaluate the potential
cancer risks associated with inhalation and air-related exposures to
hazardous air pollutants (HAP) in different social, demographic, and
economic groups within the population living near wool fiberglass
manufacturing facilities in the United States.  This work was carried
out in support of the U.S. Environmental Protection Agency’s Residual
Risk and Technology Review (RTR) for wool fiberglass manufacturing
emissions subject to Maximum Available Control Technology (MACT)
requirements under 40 CFR 63 Subpart NNN.  

	In the RTR analysis, the Human Exposure Model, Version 3 (HEM-3), was
used to estimate cancer risks due to the inhalation of HAP for the
populations residing within 50 kilometers of each wool fiberglass
facility in the country.  HEM-3 estimates cancer risks at the level of
census blocks using the AERMOD state-of-the-art air dispersion model
developed under the direction of the American Meteorological Society
(AMS) / EPA Regulatory Model Improvement Committee (AERMIC).  Each
census block typically includes about 50 people.  Additional information
on the risk analysis is available in the docket for the proposed
National Emission Standards for Hazardous Air Pollutant Emissions: Wool
Fiberglass Manufacturing Operations rulemaking where a report is
provided, covering the inputs and specific assumptions, and addressing
uncertainties.

	In the current analysis, cancer risk estimates from the wool fiberglass
HEM-3 modeling effort were linked to detailed census data in order to
evaluate the distribution of risks for different demographic groups
(including racial, ethnic, age, economic, educational, disabled and
linguistically isolated population categories).  The following
population categories were included in this analysis:

Total population

White

Minority

African American (or Black)

Native Americans

Other races and multiracial

Hispanic or Latino

Children 18 years of age and under

Adults 19 to 64 years of age

Adults 65 years of age and over 

Adults without a high school diploma

Households earning under the national median income

People living below the poverty line

Working aged people (16-64) with a disability

Linguistically isolated people

The HEM-3 results for a particular census block reflect the estimated
level of cancer risk that would be experienced by an individual residing
in the block for 70 years. In this analysis, the demographic composition
of the population estimated to experience a risk greater than 1 in 1
million as a result of wool fiberglass manufacturing emissions is
compared to the demographic composition of the overall nationwide
population.

	The census data used in this analysis is described in Section 2.  The
algorithms used to compute the distributions of risk and exposure are
presented in Section 3.  The results of this analysis are presented in
Section 4.  

Census Data

	Table 1 summarizes the census data used in this analysis, showing the
source of each dataset and the level of geographic resolution.  All of
the data are from the 2000 Decennial census.  Race, ethnicity and age
data were obtained at the census block level.  Distributions regarding
educational status, poverty status, household income, disabilities and
linguistic isolation were obtained at the block group level.  A census
block contains about 50 people on average; and a block group contains
about 26 blocks on average, or about 1,350 people.  (For comparison, a
census tract is larger than a block group, with each tract containing an
average of 3 block groups, or about 4,300 people.)

Table 1.  Summary of Census Data used to Analyze Risks for Different
Socio-economic Groups

Type of population category	Source of data	Level of geographic
resolution

Racial and ethnic categories	Landview®	Census block

Age groups	SF1 Table P12	Census block

Level of education - adults without a high school diploma	SF3 Table 37
Block group

Households earning below the national median income	SF3 Table 52	Block
group

People living below the poverty line	SF3 Table P87	Block group

Working age people (16-64) with a

      disability	SF3 Table P42	Block group

Linguistically isolated people	SF3 Table P20	Block group



	Data on race and ethnicity were obtained primarily from the Landview®
database compiled by the Census Department.  Landview® gives a
breakdown for the population of each census block among different racial
classifications, including: White, African American or Black, American
Indian or Native Alaskan, Asian, Native Hawaiian or other South Pacific
Islander, other race, and two or more races.  In the current analysis,
the Asian, Native Hawaiian or other South Pacific Islander, and other
race categories were combined into a single category.  The Landview®
database also indicates the number of people in each tract that are of
Hispanic or Latino ethnicity.  Landview® covers the 50 states, the
District of Columbia, and Puerto Rico, but does not cover the Virgin
Islands.  Race and ethnicity data on the Virgin Islands were obtained
from the Virgin Islands Summary File.

	Data on age distributions in the U.S. and Puerto Rico for each census
block were obtained from the 2000 Census of Population and Housing
Summary File 1 (SF1) Short Form, Table P12.  Data on poverty status,
household income, education level, disabilities and linguistic isolation
for each block group in the U.S. and Puerto Rico were obtained from the
2000 Census of Population and Housing Summary File 3 (SF3) Long Form. 
For the U.S. this file was accessed on a DVD version prepared by
GeoLytics.  SF3 data for Puerto Rico were obtained from the Census
Department website,  and data for the Virgin Islands were retrieved from
similar tables in the Virgin Islands Summary File.4 

	The SF3 data set consists of over 800 separate tables, each providing
information on a different subject.  For the current analysis, data were
obtained from Tables P20, P37, P42, P52, and P87.  Table P37 analyzes
the level of education attained by men and women over 25 years of age
(e.g. some high school but no high school diploma, high school graduate,
some college, etc.) in each block group.  Table P52 gives information on
household income in 1999.  Table P87 estimates the number of people
living below the poverty line in each block group.  Table P42 estimates
the number of people in different age categories with disabilities in
each block group.  And Table P20 estimates the fraction of households in
each block group that are considered linguistically isolated.

Calculation Methods

	The HEM-3 models the cancer and noncancer risk at a point near the
geographic center of each census block.  For the current analysis, this
risk estimate was assumed to apply to all individuals residing in the
block.  We used block identification codes to link the HEM-3 modeling
results for each block to the appropriate census statistics.  This
allowed us to estimate the numbers of people falling into different
population categories within each block.  We then analyzed the
distribution of estimated inhalation risks within each population
category, given the numbers of people within the category that are
exposed to different risk levels.  Each distribution involved a
tabulation of all the census blocks modeled for the wool fiberglass
source category.  We also computed the average risk for all individuals
in each population category.  

	Distributions of risk and average risks were computed for the raw HEM-3
model results for wool fiberglass operations.  For comparison, the
nationwide demographic composition (i.e., population percentage in each
demographic group for the country as a whole, based on the 2000 Census)
is also provided in the results table.  

	Section 3.1 describes the calculation method used for categories where
block-level data were available – racial, ethnic and age categories
and the total population.  Sections 3.2 through 3.6 describe
calculation methods for education status, household income, poverty
status, disability and linguistic isolation, respectively.  

Racial, Ethnic and Age Categories and the Total Population

	Since race, ethnicity and age data are available at the census block
level, the calculation of risk distributions for these categories
involved a simple block-by-block accumulation of the people in each
category.  We began by identifying a set of bins reflecting the level of
risk.  The population of each block was then assigned to the appropriate
risk bin based on the modeled risk level in the block.  The numbers of
people in each risk bin were then added together for all of the blocks
modeled for the wool fiberglass source category:

∑i(Ra≤Ri<Rb) [N(s,i)]	(1)

where: 

	H(Rab,s) =	the population count for risk bin Rab, which is between Ra
and Rb for population subgroup “s”

	Ri =	the modeled risk level in block “í” (estimated lifetime cases
of cancer per million population)

	∑i(Ra≤Ri<Rb)	refers to the summation over all blocks i where Ri
falls in bin Rab, between Ra and Rb 

	N(s,i) =	the number of people within population subcategory s, in block
i

The same approach was used for the total population.  The average risk
for a given
populati湯挠瑡来牯⁹牯映牯琠敨琠瑯污瀠灯汵瑡潩⁮慷
⁳桴湥挠污畣慬整⁤獵湩⁧桴⁥潦汬睯湩⁧煥慵楴湯ഺ

		A(S) = ∑i [N(s,i) × Ri] ∕ ∑i [N(s,i)]	(2)

where:

	A(s) =	the average risk for population subgroup “s” (estimated
lifetime cases of cancer per million population)

	∑i	refers to the summation over all blocks “í” modeled for the
emission source category

	N(s,i) and Ri 	were defined above

Level of Education

Table P37 of the SF3 dataset specifies the education status for men and
women age 25 and older for each census block group, based on the last
grade completed.  To obtain the total number of adults without a high
school degree, we added together the numbers who had completed grades
below a high school senior.  Thus, the number of people without a high
school degree equals the sum of the number of males with no schooling,
the number of females with no schooling, the numbers of males and
females who have completed nursery school through 4th grade, up to the
numbers of males and females who have completed some high school but not
received a high school degree.

The number of adults without a high school degree as a fraction of the
total population was assumed to be the same for each block in the block
group.  Thus, the number of adults without a high school degree in each
block was computed as follows:

) ∕ N(t,bg)	(3)

where:

	N(nhs,b/bg) =	number of adults without a high school diploma, in block
“b” of block group “bg”

	N(t,b/bg) =	total number of people in block “b” of block group
“bg”

	N(nhs,bg) =	number of adults without a high school diploma in block
group “bg”

	N(t,bg) =	total number of people in block group “bg”

Equation (1) was then used to generate risk distributions based on the
block-level results, and Equation (2) was used to compute the average
risk for adults without a high school diploma. 

Household Income

	Table P52 of the SF3 dataset estimates the numbers of households in
each block group with income for the year 1999 in various ranges,
generally divided into $5,000 increments (e.g. $10,000 to $14,999,
$15,000 to $19,999, etc.).  The median national income for 1999 was
about $42,000 per year.  Therefore, in order to determine the number of
households with incomes under the median income, we added the estimates
for the ranges below that level.  We assumed that the household incomes
in the $40,000 to $44,999 increment were evenly distributed over this
range.  Therefore, 40% of the households in the $40,000 to $44,999
income range were assumed to be below the national median income of
about $42,000.  The following equation was used to estimate the fraction
of households below the national median income within each census block
group:

		F(nm,bg) = [C<10 + C10-15 + …. + C35-40 + (0.4×C40-45)] ∕ CT	(4)

where:

	F(nm,bg) =	fraction of households in block group “bg” with incomes
below the median national income

	C<10 =	number of households with incomes under $10,000

	C10-15 =	number of households with incomes from $10,000 to $14,999

	C35-40 =	number of households with incomes from $35,000 to $39,999

	C40-45 =	number of households with incomes from $40,000 to $44,999

	CT =	total number of households in block group “bg”

	The fraction of people living in households below the median income for
each block within the block group was assumed to be the same as the
fraction of households below the median income for the block group.

		N(nm,b/bg) =  F(sm,bg) × N(t,b/bg)	(5)

where:

	N(nm,b/bg) =	number of people in block “b” of block group “bg”
living in households below the national median income

	F(nm,bg) =	fraction of households in block group “bg” below the
national median income

	N(t,b/bg) =	total number of people in block “b” of block group
“bg”

Equation (1) was then used to generate risk distributions based on the
block-level results, and Equation (2) was used to compute the average
risk for people living in households below the national median income. 
It must be noted that this approach neglects any potential relationship
between household size and income level within a particular block group.
 However, it is expected to provide a reasonable indication of the risk
level of people living below the national median income, relative to the
population as a whole. 

Poverty Level

	Table P87 of the SF3 dataset estimates the total number people in each
block group living below the poverty level, as well as the numbers of
people below the poverty level in different age groups.  The current
study did not include an analysis of poverty status by age group, only
of the total population below the poverty line.  The fraction of people
below the poverty line was assumed to be the same for each block in the
block group.  Thus, the population below the poverty line in each block
was computed as follows:

) ∕ N(T,bg)	(6)

where:

	N(p,b/bg) =	number of people below the poverty line in block “b” of
block group “bg”

	N(T,b/bg) =	total number of people in block “b” of block group
“bg”

	N(p,bg) =	number of people below the poverty line in block group
“bg”

	N(T,bg) =	total number of people in block group “bg”

Equation (1) was then used to generate risk distributions based on the
block-level results, and Equation (2) was used to compute the average
risk for people living below the poverty level. 

3.5	Disability

Table P42 of the SF3 dataset estimates the total number of people in
each block group with disabilities, as well as the numbers of people
with disabilities in different age groups.  This analysis includes the
total number of working aged people (16 through 64 years of age) with a
disability.  The fraction of working aged people with a disability was
assumed to be the same for each block in the block group.  Thus, the
population of working aged people with a disability in each block was
computed as follows:

) ∕ N(T,bg)	(7)

where:

	N(d,b/bg) =	number of working aged people with a disability in block
“b” of block group “bg”

	N(T,b/bg) =	total number of people in block “b” of block group
“bg”

	N(d,bg) =	number of working aged people with a disability in block
group “bg”

	N(T,bg) =	total number of people in block group “bg”

Equation (1) was then used to generate risk distributions based on the
block-level results, and Equation (2) was used to compute the average
risk for working aged people with a disability.

3.6	Linguistic Isolation

Table P20 of the SF3 dataset estimates the fraction of households in
linguistic isolation in each block group.  For this analysis, the
fraction of people living in linguistic isolation for each block within
the block group was assumed to be the same as the fraction of households
in linguistic isolation for the block group.  Thus, the population of
linguistically isolated people in each block was computed as follows:

		N(li,b/bg) =  F(li,bg) × N(t,b/bg)	(8)

where:

	N(li,b/bg) =	number of people in block “b” of block group “bg”
living in linguistically isolated households

	F(li,bg) =	fraction of households in block group “bg” in linguistic
isolation

	N(t,b/bg) =	total number of people in block “b” of block group
“bg”

Equation (1) was then used to generate risk distributions based on the
block-level results, and Equation (2) was used to compute the average
risk for people living in linguistic isolation.

Results

	The distribution of estimated lifetime inhalation cancer risks greater
than or equal to1 in a million for different racial and ethnic groups
among the population living near wool fiberglass facilities is shown in
Table 2.  For comparison purposes, Table 2 provides the nationwide
percentages of the various demographic groups.  Detailed demographics
data and analyses used to create Table 2 can be found in Appendix A of
this document. 

	The results of the demographic analysis presented in Table 2 indicate
that there are approximately 850,000 people exposed to a cancer risk
greater than or equal to 1-in-1 million.  The demographic results for
the population potentially impacted by wool fiberglass emissions
indicate that the Minority, African American, Other and Multiracial, and
Hispanic or Latino percentages are higher than their respective national
percentages (by 11, 7, 5, and 2 percentage points, respectively).  With
respect to age categories, there is a modest percentage increase (of 4
percentage points) for the “ages 19 to 64” demographic compared to
the national demographic percentage for that age group.  With respect to
the population below the poverty level potentially impacted by wool
fiberglass emissions, there is a modest percentage increase (of 2
percentage points) compared to the national percentage.  With respect to
the population over 25 without a high school diploma, there is a small
percentage increase (of 1 percentage point) compared to the national
percentage.  With respect to the working age population with a
disability, there is a modest percentage increase (of 2 percentage
points) compared to the national percentage.  The percentages for the
other demographic categories are similar to or below the respective
national percentages.

Table 2.  Summary of Demographic Assessment of Risk Results for the
Wool Fiberglass Source Category





Emissions Basis

Demographic Group



Total	Minority	African 

American	Other and

Multiracial	Hispanic

or Latino	Native 

American	Ages 0 

to 18	Ages 19 

to 64	Ages 65 

and up	Below the Poverty Level	Over 25 Without a HS Diploma	

Working Age with

Disability	

Linguistic Isolation

Nationwide Demographic Breakdown	n/a	285,339,128	25%	12%	12%	14%	0.9%
27%	60%	12%	13%	13%	12%	5.4%

	Maximum 

Risk 

(in 1 million)	Population With Cancer Risk Greater Than or Equal to 1 in
1 million

Source Category	40	848,782	36%	19%	17%	16%	0.7%	25%	64%	12%	15%	14%	14%
5.7%

Notes:



Source Category emissions were estimated based on Information Collection
Request (ICR) data from March of 2010.



Minority population is the total population minus the white population.



Population figures are for the population residing within 50 km from the
center of these facilities whose cancer risks are estimated to be
greater than or equal to 1 in a million.





Uncertainty Discussion

	

	Our analysis of the distribution of risks across various demographic
groups is subject to the typical uncertainties associated with census
data (e.g., errors in filling out and transcribing census forms), which
are generally thought to be small, as well as the additional
uncertainties associated with the extrapolation of census-block group
data (e.g., income level and education level) down to the census block
level.  

	The uncertainties in these risk estimates include the same
uncertainties in emissions data sets, in air dispersion modeling, in
inhalation exposure and in dose response relationships that are
associated with our source category risk estimates.

g

Æ

g

p

q

r

s

Þ

Ā 

ᘍ詨ꄧ㔀脈䩃$ᬀý

þ

ÿ

	

b

¹

㄀$摧➊¡ċ̀$☊଀ņ愀$摧⠥3

摧睋tᐇഀࣆȀͰ⒆ȀĀ ᐀A

B

\

]

^

_

`

a

b

c

d

€

‚

ƒ

†

‡

—

˜

™

³

´

µ

¶

·

¸

¹

º

»

×

Ø

ᄀ脈栖睋tࡕ洁H渄H甄ĈḀØ

Ù

Ú

Ý

Þ

ë

ì

í

kd¸

摧⠥3欀ᡤ

欀쵤

@

h

h

h

ꀂ吅Ȇ㄀$摧㾗Z

Æ

hži

D

„Tᄆ겄ㇹ$葞ٔ葠怜摧穵

huz

huz

hži

huz

huz

$༂咄ᄆ겄ㇹ$葞ٔ葠怜摧䥧A
ഀெ̀֠ٔ܈ༀ咄ᄆ겄ㇹ$葞ٔ葠怜摧穵

摧穵

huz

huz

huz

hži

摧穵

hži

hži

huz

huz

hži

hži

摧榞

Ѐ摧榞

hži

  hži

hži

 hži

hži

hži

@

摧榞

摧榞

hži

tainties, to EPA decision makers and the public as additional analyses
to inform RTR decisions.Appendix A



.	EC/R.  2006.  Modeling for the Residual Risk and Technology Review
Using the Human Exposure Model 3 – AERMOD Version.  Prepared by EC/R
Incorporated for the U.S. Environmental Protection Agency, Research
Triangle Park, NC.  

. 	EC/R.  2008.  HEM-3 User’s Guide.  Prepared by EC/R Incorporated
for the U.S. Environmental 	Protection Agency, Research Triangle Park,
NC.    HYPERLINK "http://www.epa.gov/ttn/fera/human_hem.html" \l "guide"
 http://www.epa.gov/ttn/fera/human_hem.html#guide 

.	Census.  2002.  LandView 5 on DVD [electronic resource] : a viewer for
EPA, Census and USGS data and maps. U.S. Census Bureau, Washington, D.C.

.	Census.  2008.  Virgin Islands Summary File. U.S. Census Bureau,
Washington, D.C.    HYPERLINK "http://www.factfinder.census.gov" 
www.factfinder.census.gov 

.	Census.  2004.  Census DVD 2000 Long Form SF3, Release 2.2. 
Geolytics, Inc., East Brunswick, NJ.    HYPERLINK
"http://www.geolytics.com"  www.geolytics.com 

.	Census.  2008.  SF3 Data for Puerto Rico.  U.S. Census Bureau,
Washington, D.C.    HYPERLINK "http://www.factfinder.census.gov" 
www.factfinder.census.gov 

7.	HEM-3 generally uses the coordinates given by the census for the
internal point, or “centroid” of each 	block.  However, when the
footprint of an industrial facility includes the block centroid, the
model is 	designed to identify the highest-risk point outside of the
facility’s footprint.

  PAGE   \* MERGEFORMAT  iv 

  PAGE   \* MERGEFORMAT  ii 

A-  PAGE   \* MERGEFORMAT  1 

