Responses to Significant Comments on the
2008 Proposed Rule on the
National Ambient Air Quality Standards
for Lead
(May 20, 2008; 73 FR 29184)
Docket Number OAR-2006-0735
U.S. Environmental Protection Agency
October 2008
-------
Table of Contents
Frequently Cited Documents
I. INTRODUCTION 1
II. RESPONSES TO SIGNIFICANT COMMENTS ON PROPOSED Pb STANDARDS 2
A. Primary Pb Standard 2
1. General Comments on Need for Revision 2
a. Support for Revising the Current Standard 2
b. Support for Retaining the Current Standard 3
2. Comments on Elements of Proposed Primary Pb Standard 3
a. Indicator 3
b. Averaging Time and Form 4
c. Level for Pb-TSP-based Standard 6
i. General Comments on Level 6
ii. Comments on Public Health Policy Goal 11
iii. Comments on Air-to-blood Ratio 13
iv. Comments on Concentration-Response Function for IQ Loss 17
v. Comments on Exposure and Risk Considerations 26
d. Comments on Adequate Margin of Safety 28
3. Additional Comments on the Interpretation of Scientific Evidence 29
4. Additional Comments on the Exposure and Health Risk Assessment 32
B. Secondary Pb Standard 40
C. Comments Related to Data Handling (Appendix R) 40
1. Use of "standard conditions" Pb-TSP data collected prior to January 1, 2009 without
adjustment to represent "local conditions" 40
2. Data Completeness Tests 41
3. Criteria and Formula for site-specific scaling factors 42
4. Criteria for exclusion of data from comparison to the NAAQS based on the influence
of an exceptional event 43
D. Comments Related to Monitoring 44
1. Existing Sampling and Analysis Methods 44
2. Proposed Pb-PMi0 Federal Reference Method 45
3. FEM criteria 46
4. Quality Assurance 46
5. Adequacy of Existing Network Design Requirements 47
6. Source-oriented Monitoring Requirement 47
7. Nonsource-oriented Monitoring Requirement 51
8. Monitoring Near Roadways 52
9. Use of Pb-PMio Monitors in lieu of Pb-TSP Monitors 52
10. Required Timeline for Monitor Installation and Operation 53
11. Sampling Frequency 53
12. Monitoring for the Secondary Standard 54
13. Cost of Monitoring Network and Funding Issues 54
III. RESPONSES TO COMMENTS RELATED TO IMPLEMENTATION 55
-------
Many comments on implementation issues are addressed in section VI of the preamble.
Significant comments on specific issues not addressed in the preamble are addressed in this
section 55
A. Nonattainment Area Boundaries 55
B. Nonattainment Area SIP Submittals 60
C. Emissions Inventory Requirements 60
D. RACM and RACT for Lead Nonattainment Areas 64
E. Attainment Demonstration and Modeling Requirements 68
F. Transportation Conformity 68
G. Transition from the currentNAAQS to aRevised LeadNAAQS 69
IV. RESPONSES TO SIGNIFICANT COMMENTS RELATED TO EXCEPTIONAL
EVENTS INFORMATION SUBMISSION SCHEDULE 70
V. RESPONSES TO LEGAL, ADMINISTRATIVE, AND PROCEDURAL ISSUES AND
NONGERMANE COMMENTS 70
References 76
Appendix A 1
in
-------
Frequently Cited Documents
The following documents are frequently cited throughout EPA's response to comments, often by
means of the short names listed below:
Criteria Document (CD):
Environmental Protection Agency (2006) Air Quality Criteria for Lead.
Volumes I and II. Washington, DC, EPA/600/R-5/144aF and
EPA/600/R-5/144bF. Available online at: http://www.epa.gov/ncea/
Preamble to the final rule:
Preamble to the Final Rule on the Review of the National Ambient Air
Quality Standards for Lead; to be published in the Federal Register in
September or October 2008.
Proposal notice: National Ambient Air Quality Standards for Lead: Proposed Rule. 73 FR
29184, May 20, 2008.
Advance Notice of Proposed Rulemaking (ANPR):
National Ambient Air Quality Standards for Lead: Advance Notice of
Proposed Rulemaking. 72 FR 71488, December 17, 2007.
Staff Paper: Environmental Protection Agency (2007a) Review of the national ambient
air quality standards for lead: assessment of scientific and technical
information. OAQPS staff paper. (Final) November 2007. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; EPA
report no. EPA- 452/R-07-013. Available online at:
http ://www. epa. gov/ttn/naaqs/standards/pb/s_pbcrsp. html
Final Risk Assessment Report:
U.S. Environmental Protection Agency. (2007b) Lead: Human Exposure
and Health Risk Assessments for Selected Case Studies, Volume I. Human
Exposure and Health Risk Assessments—Full-Scale and Volume II.
Appendices. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. EPA-452/R-07-014a and EPA-452/R- 07-014b.
Available online at:
http://www.epa.gov/ttn/naaqs/standards/pb/s_pb_cr_td.html
Pilot-phase Risk Assessment Report:
ICF International (2006) Lead Human Exposure and Health Risk
Assessments and Ecological Risk Assessment for Selected Areas. Pilot
Phase. Draft Technical Report (with appendices). Prepared for the U.S.
EPA's Office of Air Quality Planning and Standards, Research Triangle
Park, NC. December. Available online at:
http://www.epa.gOv/ttn/naaqs/standards/pb/s pb cr td.html
IV
-------
Responses to Significant Comments on the 2008 Proposed Rule on
the National Ambient Air Quality Standards for Lead
I. INTRODUCTION
This document, together with the preamble to the final rule on the review of the national
ambient air quality standards (NAAQS) for lead (Pb), presents the responses of the
Environmental Protection Agency (EPA) to the thousands of public comments received on the
2008 Pb NAAQS proposal notice (72 FR 37818). All significant issues raised in timely public
comments have been addressed. Where comments were submitted after the close of the public
comment period, EPA responded to the extent practicable.
Comments were received from EPA's Children's Health Protection Advisory Committee,
the American Academy of Pediatrics, the American Medical Association, the American Thoracic
Society, two organizations of state and local air agencies (National Association of Clean Air
Agencies [NACAA]and Northeast States for Coordinated Air Use Management [NESCAUM]),
approximately 40 State, Tribal and local government agencies, approximately 20 environmental
or public health organizations or coalitions, approximately 20 industry organizations or
companies, and approximately 6200 private citizens (roughly 150 of whom were not part of one
of several mass comment campaigns). Due to the large number of comments that addressed
similar issues, this response-to-comments document does not generally cross-reference each
response to the commenter(s) who raised the particular issue involved, although commenters are
identified in some cases where they provided particularly detailed comments that were used to
frame the overall response on an issue.
The responses presented in this document are intended to augment the responses to
comments that appear in the preamble to the final rule or to address comments not discussed in
the preamble to the final rule. Although portions of the preamble to the final rule are
paraphrased in this document where useful to add clarity to responses, to the extent any
ambiguity is introduced by this paraphrasing, the preamble itself remains the definitive statement
of the rationale for the revisions to the standards adopted in the final rule.
In many instances, particular responses presented in this document include cross
references to responses on related issues that are located either in the preamble to the Pb NAAQS
final rule, or in this Response to Comments document. In other instances the comment is
appropriately addressed by the Agency's discussion in other parts of the record. All issues on
which the Administrator is taking final action in the Pb NAAQS final rule are addressed in the
Pb NAAQS rulemaking record.
Accordingly, this Response to Comments document, together with the preamble to the Pb
NAAQS final rule and the information contained in the Criteria Document (EPA, 2006a), the
Staff Paper (EPA, 2007a), the Advance Notice of Proposed Rulemaking, and the Notice of
Proposed Rulemaking should be considered collectively as EPA's response to all of the
significant comments submitted on EPA's 2008 Pb NAAQS proposed rule. This document
-------
incorporates directly or by reference the significant public comments addressed in the preamble
to the PbNAAQS final rule as well as other significant public comments that were submitted on
the proposed rule.
Consistent with the final decisions presented in the notice of final rulemaking, comments
on the following topics are addressed in section II: the primary Pb standard (section II. A), the
secondary Pb standard (section II.B), data handling procedures (section II.C), and monitoring-
related issues (section II.D). Comments on implementation-related issues are addressed in
section III. Comments on issues related to the exceptional events rule are addressed in section
IV. Section V includes responses to legal, administrative, procedural, or misplaced comments.
II. RESPONSES TO SIGNIFICANT COMMENTS ON PROPOSED Pb STANDARDS
A. Primary Pb Standard
1. General Comments on Need for Revision
General comments based on relevant factors that either support or oppose any change to
the current Pb primary standard are addressed in section II.B of the preamble to the final rule
and/or in section II. A. 1 below. Specific comments on the proposed primary standard, including
comments on the indicator, averaging time and form, and level are addressed in sections II.C.I,
II.C.2, or II.C.3, respectively, in the preamble to the final rule and/or in section II.A.2 below.
Additional comments about the health effects evidence and the results of the human exposure
and health risk assessments are addressed in sections II.A.3 or II.A.4 below.
a. Support for Revising the Current Standard
The vast majority of public comments received on the proposal asserted that, based on
the available scientific information, the current Pb standard is insufficient to protect public health
with an adequate margin of safety and revisions to the standard are appropriate. Among those
calling for revisions to the current standards are medical groups, including the American
Academy of Pediatrics (AAP), the American Thoracic Society, and the American Medical
Association (AMA), as well as medical doctors and academic researchers. Similar conclusions
were also submitted in comments from local public health organizations, including Coalition to
End Childhood Lead Poisoning, St. Louis Lead Prevention Coalition, and Physicians for Social
Responsibility, as well as in letters to the Administrator from EPA's Children's Health
Protection Advisory Committee (CHPAC) (Marty, 2008a, 2008b). Environmental groups also
commented in support of revising the standard, including the Sierra Club, and the Natural
Resources Defense Council (NRDC). All of these medical, public health and environmental
commenters stated that the current Pb standard needs to be revised and that an even more
protective standard than proposed by EPA is needed to protect the health of sensitive population
groups. Some 6000 individual commenters also expressed such views.
The vast majority of State and local air pollution control authorities, as well as tribal
governments and tribal air agencies who commented on the Pb standard supported revision of the
current Pb standard. State organizations, including the National Association of Clean Air
-------
Agencies (NACAA), and the Northeast States for Coordinated Air Use Management
(NESCAUM) urged that EPA revise the Pb standard.
In supporting adoption of a more stringent NAAQS for Pb, these commenters variously stated
that:
• the current evidence indicates Pb effects at much lower exposure levels than when the
current standard was set;
• the current evidence indicates Pb effects in multiple systems (e.g., neurological effects in
children, cardiovascular and renal effects in adults);
• there is now evidence of a greater air-to-blood ratio than that understood when the standard
was set; and
• EPA's risk assessment indicates risks under current NAAQS of a magnitude clearly
harmful to public health.
Comments received in support of revising the current standard are addressed in section II.B.2 of
the preamble to the final rule.
b. Support for Retaining the Current Standard
Three industry commenters (National Association of Manufacturers, Non-Ferrous
Founders' Society, and the Wisconsin Manufacturers and Commerce) indicated support for
retaining the current standard. In supporting this view, these commenters variously stated that:
• a revised NAAQS would require action by Pb air sources while other nonair or historic
Pb sources of exposure contribute much greater risks to children's health;
• a revised NAAQS will "obstruct innovation" and have an "adverse impact on achieving
environmental goals";
• reduction of the Pb standard will not provide meaningful benefits to public health; and
• risks associated with current NAAQS are due to non-complying sources.
These comments received in support of retaining the current standard are generally addressed in
Section II.B.2 of the preamble to the final rule. We additionally note here that EPA disagrees
with commenters regarding the significance of health risk associated with air-related Pb
exposures allowed by the current standard, and that, under the Clean Air Act, EPA may not
consider the costs of compliance in determining what standard is requisite to protect public
health with an adequate margin of safety.
2. Comments on Elements of Proposed Primary Pb Standard
a. Indicator
The majority of public comments recommended retaining the indicator as Pb-TSP, with
some commenters qualifying their recommendation that Pb-TSP should be retained only if the
level for the standard is set at or above 0.10 or 0.20 |ig/m3. A few commenters recommended
revision of the indicator to Pb-PMio (or Pb-PM2.s) regardless of the level of the standard. These
comments on indicator are described and addressed in section Il.C.l.b of the preamble to the
-------
final rule. Specific indicator-related comments not addressed in the preamble are described
below.
(1) Comment: One commenter, in indicating support for retaining the current indicator of Pb-
TSP, referenced two papers (Donguk and Namwon, 2004; Park and Paik, 2002) which
they described as reporting that 35 to 70% and 11 to 49%, respectively, of airborne Pb
associated with particular industries was in particle sizes greater than 10 micrometers in
diameter.
Response: As discussed in section II.C.I.2 of the preamble to the final rule, EPA agrees
with this commenter that Pb occurs in some areas in the U.S. in particle sizes that would
not be captured by PMi0 samplers, and accordingly EPA has decided to retain Pb-TSP as
the indicator. EPA notes, however, that the studies cited by this commenter provide
occupational exposure measurements for industries in Korea, rather than ambient air
measurements for locations in the U.S.
b. Averaging Time and Form
Comments received from public health and environmental organizations on the issue of
averaging time for the Pb primary standard recommended revision of the averaging time to
monthly, as did most comments from state and local air pollution control authorities and from
private citizens. Among these commenters who also commented on form, most recommended a
form of maximum monthly average, with a few recommending the form of 2nd maximum
monthly average. Some of these commenters additionally recommended that the averaging time
be derived as a rolling 30-day average. Another group of commenters comprised of industry
associations and businesses and some state and local air pollution control authorities opposed a
revision from quarterly to monthly averaging time. Some of these commenters additionally
recommended that the averaging time be derived as a rolling 3-month average. These comments
are summarized and addressed in section II.C.2.b of the preamble for the proposed rule. Specific
aspects to some of these comments with regard to interpretation of the evidence are discussed
below.
(1) Comment: A few industry commenters, most particularly, the International Lead and Zinc
Research Organization (ILZRO), state that the evidence with regard to toxicodynamics
and toxicokinetics of Pb in the body indicates that "longer averaging times are
appropriate". The rationale provided included statements (without scientific citation)
regarding the time period pertinent to neurological effects such as neurocognitive and
neurobehavioral effects. For example, one industry group (ILZRO) generally states that
effects associated with a time frame of weeks are limited to frank effects such as
encephalopathy, and implies that neurological effects of interest in this review occur
"over extended periods of cognitive development." Additionally, this commenter asserts
that from a toxicokinetic standpoint, longer averaging times are appropriate (citing
Leggett, 1993). For example, this commenter stated that the uptake rate of lead by the
brain is "extremely low." They further stated that "at least 6 months is required for lead
in the central nervous system to equilibrate with environmental exposures" and that from
-------
a "dose to target tissue" perspective a longer averaging time is appropriate as they state
the "removal half-time" of Pb is "approximately two years."
Response: EPA disagrees with several aspects of these comments and the commenters'
interpretation of the evidence. As discussed in the preamble section II.C.2, the response
of blood Pb to environmental exposures to Pb is quite rapid and the evidence concerning
neurocognitive effects indicates the importance of relatively short exposures. With
regard the comments, particularly those from ILZRO, regarding short as compared to
longer exposures, EPA first notes that in considering the appropriate averaging time, we
are not, as implied by the framing of the comments, identifying the maximum allowable
duration for a single exposure at the level of the standard. Rather, we are considering the
shortest time period during which the level of the standard may not be exceeded, in the
context of an ongoing, yet varying, ambient exposure to a persistent pollutant.
Neurological effects in children, such as neurocognitive and neurobehavioral effects, and
not encephalopathy have been identified as key effects in this review. As discussed in the
preamble section II.C.2 (and in the proposal, ANPR, and Staff Paper and Criteria
Document), the evidence for these effects indicates the importance of exposures on the
order of 1 to 3 months. In several epidemiologic studies examining the relationship
between neurocognitive effects and blood Pb levels, the strongest associations were
observed with concurrent blood Pb levels (e.g., Chen et al., 2005; Lanphear et al., 2005).
With regard to toxicokinetics of Pb in the body, EPA disagrees with the commenter's
characterization of brain uptake of Pb as "extremely low." EPA notes that while the cited
publication by Leggett (1993) reports that the author's model predicts 6 months to reach a
peak level in the brain compartment after bolus administration, the author reports that
within two months approximately 80% of steady state brain levels of lead are reached
(Leggett, 1993). The uptake half-time varies by age in the Leggett (1993) paper from
only 0.9 to 3.7 days. Further, contrary to the commenter's statement, the cited study
(Leggett, 1993) does not indicate a compartment for the full CNS, only the brain. Thus,
EPA concludes that given the rapid accumulation of lead from the plasma into the brain
and subsequent slow removal from the brain, from a toxicokinetic standpoint, and the
considerations described in the preamble section II.C.2.b, an averaging time longer than a
few months (e.g., a 3-month period) is not appropriate.
(2) Comment: One industry commenter (American Smelting and Refining Company
[ASARCO]) recommended that unless a provision was included to prevent counting a
single month more than once, EPA should not use a rolling average form in order to
avoid "double counting of exceedances".
Response: EPA disagrees with the commenter that for an averaging time longer than a
month, air Pb concentrations during any one month should only contribute to a single
statistic that is compared to the level of the standard. The use of a rolling average has
neither the purpose nor effect of "double counting ... exceedances." The purpose of the
rolling 3-month average, as discussed in the preamble, is to provide an average that is
more representative of air quality over the 3-month period. An exceedance of the
standard is only determined by reference to the 3-month mean, and each month, whether
a "high" or "low" month, contributes to three 3-month means. Moreover, EPA notes that
-------
the form of the standard is a not-be-exceeded (maximum) form measured over a 3-year
span, so only a single exceedance of the standard is sufficient to violate the standard,
regardless of whether there are additional contemporaneous violations.
c. Level for Pb-TSP-based Standard
A large number of comments were received on the proposed range of levels for the
primary Pb standard. The vast majority expressed support for a revision to a more health
protective standard at or below the range of levels proposed by EPA; a small number expressed
support for more modest adjustment or revision to a level ranging above the proposed range up
to 0.50 |ig/m3; and a few commenters expressed opposition to any appreciable modification of
the current Pb standard. These comments are summarized and addressed in section Il.C.S.b of
the preamble to the proposed rule. In addition to the discussion contained in that section, EPA
provides the following responses to specific issues related to the level for the primary Pb
standard in sections II.A.2.c.i through II.A.2.c.iv below. The subsequent sections (sections
II.A.2.C.V and II.A.2.c.vi) respectively address additional specific comments on health evidence
considerations and exposure and risk considerations related to consideration of the level for the
primary standard. Additional specific comments on interpretation of the scientific evidence and
the health risk assessment are included in sections II. A.3 and II.A.4 below.
/'. General Comments on Level
Specific aspects of general comments on level for the primary Pb standard not addressed
in the preamble are addressed here.
(1) Comment: Several commenters (e.g., NRDC, Missouri Coalition for the Environment [MO
Coalition]) stated that EPA's proposed range did not include a level protective of
vulnerable and susceptible subgroups. In particular, these commenters identify minority
and low-income children as a sensitive subpopulation for Pb exposures. In discussing
these subgroups, some commenters note factors such as nutritional deficiencies as a
factor contributing to susceptibility and cite the evidence of higher blood Pb levels in
these populations as evidence that minority and low-income children are more likely to
be exposed to high levels of lead. One commenter stated that, "Setting the NAAQS
without reference to the nutritional deficiencies ... of minority and low-income ... fails to
fulfill EPA EJ CAA obligations" (MO Coalition, p. 22). Another commenter stated that
EPA must set the NAAQS in the lower part or below the proposed range "to protect the
many, poor, minority, and urban individuals whose blood Pb already contains
dangerously high levels of lead" (NRDC, p.22), referring to various EPA documents
from this review as providing information pertinent to this point.
Response: EPA notes, as discussed in the proposal and this preamble, that due to
additional scientific information about Pb since the 1978 standard was established
(particularly the lack of an accepted safe level for Pb exposure), EPA is not using the
1978 approach of setting the NAAQS based on total Pb exposure, but rather is basing the
NAAQS upon air-related Pb exposure. EPA considers the sensitive subpopulation for
air-related Pb to be those children more highly exposed to air-related Pb (e.g., due to
-------
proximity to sources) as compared to children on average. As discussed in the preamble,
the Administrator selected the level, form, indicator, and averaging time with full
consideration of the need to provide sufficient protection for sensitive groups with an
adequate margin of safety.
EPA agrees with commenters that on average in the U.S., minority and low-income
children have higher blood Pb levels than the general population, as shown by the
NHANES information referenced in section II.A.2.b of the preamble and described in
section II.B.l.a of the proposal (e.g.,
http://www.epa.gov/envirohealth/children/body burdens/bl-table.htm). EPA also agrees
that these higher average blood Pb levels are indicative of these subpopulations being
sensitive with regard to total Pb exposures. However, EPA notes that this evidence does
not provide information regarding a potential for increased sensitivity to air-related Pb
exposure. Due to the nonlinear dose-response relationship between IQ loss and blood Pb,
EPA concludes that the population with lower blood Pb levels from both air and nonair
sources would be expected to have a greater incremental sensitivity to Pb impacts on IQ
(i.e., greater risk of IQ loss per incremental unit of blood Pb). Thus, a standard that is
selected to provide protection from the incremental amount of blood Pb associated with
air-related Pb to populations with lower total blood Pb levels will also necessarily
provide protection from the air-related Pb associated with exposures at the level of the
standard to populations with higher total blood Pb levels. As explained in section
Il.C.S.b of the preamble, the air-related IQ loss framework focuses on children exposed at
the level of the standard, children living near air sources of Pb who are likely to be most
highly exposed. In concentrating on this highly exposed group of children, EPA has
focused on providing sufficient protection for the appropriate subpopulation for this
review.
(2) Comment: Some industry commenters (American Petroleum Institute [API], Association of
Battery Recyclers [ABR], Doe Run Resources Corp.1) stated that any level for the
NAAQS derived using the air-related IQ loss evidence-based framework should be
viewed as pertaining to an averaging time, for the standard, of one year. In support of
this premise, commenters generally reference the risk assessment, with one commenter
stating that the levels in the table presenting air-related IQ loss estimates in the proposal
table "can only be interpreted as annual average air lead levels in order to be consistent
with the Risk Assessment" (API). Another commenter additionally states that the risk
assessment "was used to support the proposed range of Lead NAAQS standards" (ABR).
Based on the commenters' premise of levels for an annual average-based standard, these
commenters state that the levels should be converted to ones more appropriate to the
selected averaging time/form (e.g., maximum quarterly average or second maximum
monthly), and they recommend the use of air quality data statistics for that purpose.
Response: EPA disagrees with the commenters' premise and finds no basis to interpret
the levels for the evidence-based framework as pertaining to an annual averaging time for
the standard. Rather, EPA concludes they are appropriately interpreted with an averaging
time no longer than quarterly. To interpret them as the commenters have as longer than
Variously referred to throughout this document as "Doe Run" or "Doe Run Resources Corp.".
-------
quarterly averages, ignores the body of evidence on the response of blood Pb and
associated effects to Pb exposure described in detail in the Criteria Document, and
summarized in the Staff Paper, ANPR, proposal and preamble to the final rulemaking.
As discussed in section II. C.2 of the preamble, while the evidence is not precise as to the
duration of exposures that trigger neurocognitive effects, for which IQ loss is the
indicator used in the framework, these studies do not indicate that a duration as prolonged
as a year is required; rather, they indicate the importance of durations of approximately
one to three months. For example, as described in the Criteria Document and
summarized in the Staff Paper, ANPR, proposal and preamble to the final rule, among the
blood Pb metrics analyzed in these studies, concurrent blood Pb (i.e., blood Pb measured
at the time of IQ test) has the strongest association with effects (Chen et al., 2005 and
Lanphear et al., 2005)2 and, the concurrent blood Pb metric is most strongly related to a
child's exposure episodes within the past few (e.g., one to three) months, rather than
exposures as long as a year in the past (e.g., CD p. 4-25).
Further, the evidence regarding transfer of air Pb along the multiple exposure pathways to
the blood, while also imprecise, indicates the significance of durations appreciably
shorter than a year's length. The studies for air-to-blood ratios, described in sections
II.A.2.a.iii and II.C.3, generally do not specify air concentration durations, and
consequently could not support a presumption that levels derived from the framework
reflect a full year exposure at a particular level (with air-to-blood ratios as one of the
inputs). While we are uncertain as to the precise duration for any one air-to-blood ratio
and the ways in which an air-to-blood ratio may vary with the duration of the air Pb
concentration, we cannot find a basis in the evidence regarding transfers of air Pb to
blood Pb for the commenter's statement that the Table 7 levels pertain to an annual
average form for the standard. Rather, EPA finds clear evidence to support the
importance of time periods on the order of 1-3 months affecting air-related blood Pb
levels and exposure pathways.
• For example, evidence from the time of leaded gasoline shows children's blood
Pb levels responded to leaded gasoline sales with a time lag of one month.
(Schwartz and Pitcher, 1989)
• Additionally, various studies have observed blood Pb levels to exhibit seasonal
patterns, perhaps related to seasonality in exposure variables, (e.g., Rabinowitz et
al 1985)
• One of the studies cited in the proposal, and published since 1986 CD, reported on
air blood Pb levels near a smelter and reported a reduction in blood Pb levels in
response to significantly reduced air Pb levels of 3-month duration. (Hilts, 2003)
• A dustfall study described in CD reported a relatively rapid response of indoor
dust Pb loading to ambient airborne Pb, on the order of weeks.3 (Caravanos et al.,
2006)
The young ages at which effects have been observed further indicate the significance of exposures shorter than a
year (e.g., Tellez-Rojo et al 2006; Canfield et al 2003). Additional health evidence demonstrates the sensitivity of
the early years of life and increased vulnerability to specific types of effect during some developmental periods (e.g.,
prenatal), the length of which indicates an importance of exposures much shorter than annual.
3 Similarly, one of the commenters (Doe Run Resources Corp.), in comments submitted on the ANPR, stated that
"... due to the short residence time of dust in a house (on the order of months rather than years), it is unlikely that
-------
• Dust Pb modeling analyses performed as part of the quantitative risk assessment
provide an estimate of approximately four months as the time over which an
increase in air Pb will reach 90% of the final steady state change in indoor dust
Pb. (USEPA, 2007b)
• A study of changes in air, dust and blood Pb levels associated with smelter Pb
emissions reports dust Pb levels, in terms of quarterly averages which show
temporal variability by as much as a factor of 3 from quarter to quarter. (Hilts,
2003)
EPA further notes that the commenters' conclusion that the IQ loss estimates from the
risk assessment pertain to annual average standard levels is not accurate. The risk
assessment provided IQ loss estimates associated with just meeting the current and
alternative standards with maximum quarterly or maximum monthly averaging
times/forms. It did not include any assessment of just meeting alternative standards with
an annual averaging time.
EPA did use relationships from air quality monitoring data between annual average Pb
concentrations and maximum quarterly or maximum monthly average Pb concentrations
in conducting these risk assessments because the blood Pb model used in the assessment
does not accept air quality inputs of a temporal scale shorter than a year. Such air quality
factors were used as a method for obtaining an annual average air Pb concentration (for
input to the blood Pb model) that might be expected to be associated with just meeting
the quarterly or monthly standards being assessed, to reflect the variability in air Pb
concentrations that occur over time scales less than a year as a result of temporal changes
in meteorology and source and emission characteristics. Commenters misinterpreted
EPA's use of such factors as implying that the risk assessment results pertained to annual
average standards at the alternate levels assessed. Based on that misinterpretation, the
commenters stated that EPA needed to use such air quality factors to "adjust" the level of
the standard to reflect a quarterly or monthly averaging time. However, for the reasons
stated above, no such adjustment would be appropriate. EPA further notes that the use of
such factors in the exposure simulation in the risk assessment is completely distinct from
and unrelated to the evidence with regard to the duration over which air Pb
concentrations contribute to health effects, which is the consideration pertinent to
selecting the averaging time for the standard.
(3) Comment: Several commenters stated that the record does not support setting a NAAQS for
Pb of zero in this review. Some of these commenters further stated that they agree with
EPA's interpretation that the CAA does not require EPA to establish a "risk-free"
standard, and stated that this interpretation is consistent with Congress' intent in enacting
these provisions of the CAA and with case law interpreting the CAA.
Response: EPA agrees that the current record does not support setting a NAAQS for Pb
of zero in this review, and for the reasons discussed in the preamble the final standard is
yearly average air lead levels are the best predictor of observed dust lead levels." In fact, the Doe Run Resources
Corp. comment went on to recommend that a monthly average air Pb is a more appropriate predictor of dust Pb than
an annual average air Pb.
-------
not set at a level of zero. EPA appreciates the views of commenters on whether a zero, or
risk-free, standard would ever be justified.
(4) Comment: Two commenters (Teck Cominco Alaska and Alaska Department of
Environmental Conservation) indicated we should consider issues related to
bioavailability in setting the NAAQS for Pb. More specifically, comments from Teck
Cominco state that the chemical form of lead in the ore and concentrate from their Red
Dog Mine, lead sulfide, has low bioavailability, that was considered in a risk assessment
conducted by Teck Cominco through the use of a lower value for the absolute
bioavailability input to the IEUBK model than the EPA default. Teck Cominco stated
that EPA should develop a NAAQS for lead sulfide compounds that reflects a lower
bioavailability of these compounds, however, they did not provide specific suggestions
for such standards. The Alaska Department of Environmental Conservation (ADEC)
suggested EPA study toxicity and bioavailability of Pb to see if some forms of Pb should
be addressed more urgently than others, noting concern with ore dust that "might be
considered of low bioavailability but over time ... becomes more bioavailable" and
noting that there are many old mine sites in Alaska around which Pb may have relatively
high bioavailability.
Response: EPA recognizes the points made by both commenters. As discussed in the
Criteria Document (e.g., CD, section 4.2.1), the form of lead in soil can affect Pb
solubility in the gastrointestinal (GI) tract and potentially absorption from the GI tract.
The potential for absorption (or bioavailability) of ambient Pb can vary among Pb from
different sources and the bioavailability of ambient Pb (e.g., from the same source), once
released to the environment, can vary over time. For example, the bioavailability of
galena, the term for commonly mined Pb ore which contains Pb sulfide, may be
somewhat low upon initial release into the environment (CD, section 4.2.1), but there is
information indicating a much higher bioavailability after some time in soil (USEPA,
2006b ; Casteel 2005 ). EPA notes that, particularly in light of evidence indicating
changes in Pb sulfide bioavailability over time, the current air quality criteria do not
provide a basis for setting a separate NAAQS for Pb sulfides based on bioavailability.
EPA also notes that the NAAQS are set with applicability to all ambient air in the U.S.,
such that the primary Pb standard provides protection in areas across the U.S., regardless
of site-specific Pb aspects. In considering the evidence on air-related Pb and associated
health effects, EPA recognizes variability in the oral absorption of air-related ambient Pb
that stems from multiple factors, including but not limited to bioavailability of different
forms of Pb in different matrices. Other factors include those related to behavioral,
physiological and dietary characteristics of the exposed public. EPA considered the
variability contributed by all of these factors in setting the primary Pb standard that is
requisite to protect public health with an adequate margin of safety.
(5) Comment: One industry commenter (Teck Cominco Alaska) stated that "the proposed
NAAQS for lead assumes that all ambient lead is the result of anthropogenic activity"
and that EPA should consider levels of naturally occurring Pb when setting the NAAQS.
In support of their statement, the commenter states that naturally occurring lead in soil,
10
-------
resulting from "surface exposure of mineralized rock" "can be released into the
atmosphere ... during higher wind conditions".
Response: Consistent with EPA's past practice, EPA considered as part of the risk
assessment those risks associated with naturally occurring levels of Pb in ambient air.
Accordingly, information regarding airborne Pb resulting from natural sources, including
windborne soil particles from areas free of anthropogenic activity, is discussed in the
Staff Paper (section 2.5) and the proposal (section II.A.3), and is briefly summarized in
section II. A. 1 of the preamble for the final rule. This information has lead EPA to
conclude that such nonanthropogenic Pb in air (i.e., on the order of 0.00005 |ig/m3) is
insignificant, particularly in comparison to the contributions from exposures to nonair
sources. In support of their view on this issue, the commenter did not provide any
estimates (e.g., of air Pb concentrations in areas where naturally exposed, naturally
occurring Pb in the earth may contribute to atmospheric air Pb) in contradiction to EPA's
characterization. EPA notes that issues of nonanthropogenic contributions to ambient
concentrations may be relevant to implementation (e.g., exceptional events), but are not
part of the evidence-based approach used as the primary focus in setting the standard.
/'/'. Comments on Public Health Policy Goal
Comments were received on various aspects of the public health policy goal proposed for
use with the air-related IQ loss evidence-based framework. These comments are described and
addressed in section Il.C.S.b of the preamble to the final rule.
(1) Comment: One industry commenter (ILZRO) disagrees with EPA's focus on IQ loss. They
state that the argument that blood Pb in a population "will decrease the number of gifted
children and increase the number of children with low intelligence", is misleading and
incorrect, stating that "IQ scores merely assign ordinal rankings to individuals within
society. There is no 'unit of cognitive function' attached to an IQ point - it is merely an
indication of an individual's performance on the IQ test relative to expected societal
norms...". The commenter goes on to note that "This is not to say that there may be no
adverse impact of lead exposure from a societal perspective, but it is difficult to evaluate
or define in a quantitative fashion." They also state that they offer no alternatives to the
use of IQ.
Response: While EPA recognizes that IQ represents a relative ranking of individuals
compared to a mean performance of the study population on one or more of several
standardized tests of intelligence, EPA disagrees with the commenter that shifts in the
intelligence of a population do not have significant implications on the number of gifted
and handicapped children. In this case EPA recognizes IQ response to blood Pb as a
useful, well studied, quantitative metric reflecting impact on neurocognitive function and
neurodevelopment. Further, EPA considers it an appropriate metric for consideration of a
public health policy goal for the NAAQS. For example, in discussing the public health
significance of Pb-related IQ impacts, the Criteria Document stated the following:
In regard to neurodevelopment, although a two- or three-point decline in
IQ might not be consequential for an individual, it is important to
11
-------
recognize that this figure represents the central tendency of the
distribution of declines among individuals. Thus, some individuals might
manifest declines that are much greater in magnitude, while others
manifest no decline at all, reflecting interindividual differences in
vulnerability. Moreover, the import of a decline for an individual's well-
being is likely to vary depending on the portion of the IQ distribution. For
an individual functioning in the low range due to the influence of
developmental risk factors other thanPb, a Pb-associated decline of
several points might be sufficient to drop that individual into the range
associated with increase risk of educational, vocational, and social
failure.
(2) Comment: In objecting to EPA's focus for the air-related IQ loss framework on population
mean, one commenter stated that EPA did have the means to focus on a higher percentile
of the population. In support of this view, the commenter stated that the findings of the
performance evaluation performed by EPA on the blood Pb estimates indicated that upper
percentile blood Pb estimates in the risk assessment compared well with NHANES
estimates, thus indicating that EPA has methods (e.g., use of GSD as was done in the risk
assessment) that might be used to derive upper percentile estimates of air-related IQ loss
using the evidence-based framework. Additionally, this commenter suggested that EPA
consider using the upper bounds of the confidence intervals for the concentration-
response slopes.
Response: EPA agrees with the commenter that the total blood Pb estimates for the 95th
percentile in the general urban case study were found to compare well to the NHANES
IV data, providing some confidence in our ability to estimate total blood Pb (as described
in section 3.5.2.3 of the Risk Assessment Report). EPA notes, however, that it is
estimates of air-related blood Pb (and air-related IQ loss) with which we are concerned in
the context of the air-related IQ loss evidence-based framework. As noted in the
proposal (sections II.C.2.h, II.C.3.a and II.C.3.b), EPA recognizes limitations in the data
and methods on which to base estimates for the air-related component of total blood Pb
and associated IQ loss at the upper percentiles, and, accordingly, greater uncertainty in
such estimates. EPA disagrees with the commenter's suggestion to use confidence
intervals (CIs) associated with the concentration-response functions or slope for the
models reported in the epidemiological studies to provide some sense of the magnitude of
air lead-related IQ loss for upper percentiles of the subpopulation exposed at a level of
the standard. EPA notes that the 95% confidence interval of the slope parameter
specifies the range of estimates for the slope within which the true value for the study
population is projected to lie with 95% confidence, based on use of a two-sided t-test of
no effect at the 5% significance level given the model assumptions are true. We note that
the 95% confidence interval represents a characterization of uncertainty associated with
the best estimate of the slope for the study population, not the uncertainty for the 95th
percentile of the population. The upper range of the 95% confidence interval for the
concentration-response coefficients does not reflect the effect estimate for air-related
blood Pb levels on IQ loss for the subpopulation that is more greatly affected (i.e., upper
percentile estimates), as the commenter has implied. Rather, the size of the confidence
12
-------
intervals in the epidemiological studies represent factors specific to those studies,
including study population size, blood lead measurement error, collinearity of covariates,
as well as the variability of the association between total blood Pb levels and IQ loss.
Thus, the 95% confidence intervals reflect several other variables specific to the study
design which are unrelated to the upper percentile estimates of air lead-related IQ loss.
(3) Comment: Some commenters stated that EPA should take into account the effects of total Pb
exposure on children when setting the public health goal for a revised NAAQS (MO
Coalition, pi5). They note that, "Because a child may already experience significant IQ
loss and other neurocognitive effects due to non-air lead exposures, the effects due to air-
related lead exposures are of even greater significance."
Response: EPA acknowledges that a segment of the exposed population may experience
blood Pb levels that are significantly higher than typical levels in the U.S. due to
exposure to non-air related sources of Pb and that these children may experience degrees
of IQ loss greater than the average U.S. child due to this exposure. However, EPA
disagrees with the commenter that the impacts of air-related exposure will be even
greater for children with significant non-air related Pb exposure. In the case of the child
who experiences significant exposure to Pb due to non-air sources in addition to ambient
air-related Pb exposure (producing a relatively larger total Pb exposure), the IQ loss for
each increment of air-related Pb exposure will likely be less than it is for a child with
lower total Pb exposure. This phenomenon reflects non-linearity in the relationship
between IQ loss and blood Pb (with greater incremental IQ loss occurring at lower blood
Pb levels). For this reason, given a focus on ambient air Pb exposure and risk, it is
appropriate for the evidence-based framework to focus on children with total Pb exposure
closer to the current U.S. average, since these children will have greater response to each
increment of air-related Pb exposure than would children with higher overall blood Pb
levels.
/'/'/'. Comments on Air-to-blood Ratio
Several commenters (e.g., NACAA, NRDC, CHPAC) recommend greater weight be
given to CASAC's advice supporting use of a ratio above the proposal range (e.g., closer to 1:9
to 1:10), some commenters (e.g., NESCAUM, New York Department of Health [NY DOH])
recommend use of a ratio at the upper end of the proposal range (i.e., 1:7), and three industry
commenters indicate support for the proposal range (1:3 to 1:7), with one commenter identifying
1:5 as supported by the "weight of the scientific evidence" (Doe Run Resources Corp.). These
comments are described and addressed in section Il.C.S.b of the preamble to the final rule.
(1) Comment: A number of commenters (e.g., CHPAC, NACAA, NY DOH, Physicians for
Social Responsibility, NRDC) recommended that EPA consider estimates of air-to-blood
ratios higher than the range in the proposal (which was 1:3 to 1:7), citing evidence of
estimates ranging up to 1:16 (NRDC). In support of higher ratios, one commenter (NY
DOH) cites Schwartz and Pitcher (1989) as suggesting a ratio of approximately 1:9
reflecting all air Pb exposure pathways. This commenter also cites ratios up to 1:8.5 from
the Brunekreef et al., (1984) meta-analysis of studies that included adequate quality
13
-------
control measures and adjustment for co-variates. Several commenters also cite the
exposure and risk assessment as providing support for air-to-blood Pb ratios up to 1:9 and
higher (NY DOH, NRDC). Several commenters (e.g., NACAA, NRDC, CHPAC) also
recommend that greater weight be given to CAS AC's interpretation of evidence which
supports use of ratios above the proposal range (e.g., closer tol :9 to 1:10), particularly
noting today's lower air and blood Pb levels. Some (e.g., NESCAUM, NY DOH)
recommend the upper end of the proposal range (1:7) as being well supported by the
evidence or as representing the middle of a supportable range.
Response: EPA generally agrees with commenters regarding the evidence of air-to-blood
ratios ranging up to approximately 1:10. Specifically, we recognize that a number of
studies identified by the commenters including Schwartz and Pitcher (1989) and
Brunekreef et al., (1984) support higher air-to-blood Pb ratios in the range of 1:7 to 1:10,
while also reflecting key study design elements (i.e., quality control measures and
adjustment for co-variates) that increase our confidence in the air-to-blood Pb ratios
supported by these studies. Regarding the Hayes et al. (1994) study in particular, while
EPA agrees that the study, also cited by CASAC, presents an air-to-blood ratio greater
1:10, we are not relying on this study in our decision as it has not been reviewed as part
of the air quality criteria (as described in Section 1C of the preamble). We also agree that
the risk assessment provides support for higher air-to-blood Pb ratios (in the range of 1:7
to 1:10). We also note, that these ratios from the risk assessment reflect blood Pb
response to changes in a subset of air-related pathways (with all nonair and some air
pathways held constant) and accordingly would not be affected by nonair blood Pb
confounding and might have been higher if all air-related pathways were assessed.
(2) Comment: Three industry commenters indicate support for proposal range (1:3 to 1:7), with
one identifying 1:5 as supported by "weight of the scientific evidence" (Doe Run
Resources Corp.). One of the commenters also asserts that the upper end range for the
air-to-blood ratios should be 1:6 to 1:8 and not 1:10 or higher.
• One commenter (ABR) asserts that the approach used by CASAC in deriving an air-to-
blood ratio of 1:9 to 1:10 based on Schwartz & Pitcher (1989) utilizes a flawed method to
relate gasoline usage to ambient air Pb levels. The commenter offers an alternate
approach for deriving a ratio based on Schwartz and Pitcher (1989). The commenter's
approach utilizes a regression relating gasoline usage to ambient air Pb levels that is
derived from data presented in the 1986 AQCD for Pb and results in an air-to-blood ratio
of 1:7.8.
• One of the commenters (ABR) asserts that Pb emissions from gasoline may behave
differently in terms of the relationships between air Pb, Pb deposition and blood Pb,
serving to overestimate air-to-blood Pb ratios, relative to what is found with other sources
of Pb in air. In supporting this assertion, the commenter cites Hayes et al. (1994) and
states that that study reports no significant relationship between air Pb and blood Pb when
time is included as a variable. They also point to Brunekreef et al. (1983) and state that
that study reports that deposition resulting from auto emissions was high when ambient
air Pb levels remained low. The commenter concludes from this that there is a weak
association between blood Pb and ambient air Pb resulting from leaded gasoline which
14
-------
makes it inappropriate to use Schwartz and Pitcher (1989) to estimate an air-to-blood Pb
ratio.
Some industry commenters (ABR, ILZRO) disagreed with EPA's interpretation of the
evidence of higher air-to-blood ratios. Specifically, they assert that the higher air-to-
blood ratios reported in the literature (including those summarized in Brunekreef 1984)
were derived without appropriate adjustment for nonair contributions to blood Pb and
therefore may be biased high. One commenter (ILZRO) further stated that the higher
ratios (e.g., at or near 1:10) are based on exposures scenarios "prone to exaggeration" due
to confounding by changes in nonair exposures (e.g., Pb solder, Pb paint, and dietary Pb).
Furthermore, one commenter (ABR) provided an analysis based on IEUBK blood Pb
simulation intended to illustrate how a lack of adjustment for nonair contributions can
produce higher air-to-blood ratios at lower blood Pb levels.
Two commenters (Association of Battery Recyclers and Doe Run Resources Corp.) also
suggested that the air-to-blood ratios presented in the risk assessment are over-stated due
to underlying errors in the indoor dust models used to model both the urban case studies
and the primary Pb smelter (subarea) case study. These comments are addressed in
Section II. A.4 below.
Response: The alternate approach advanced by one commenter related to air-to-blood Pb
ratio calculations combines the same factor relating blood Pb and gasoline usage (from
Schwartz and Pitcher, 1989) used by CAS AC (Henderson, 2007a) with a regression
model relating gasoline usage and ambient air Pb levels (by quarter) to derive the air-to-
blood Pb ratio. EPA considers both the CASAC approach and the alternate approach
presented by the commenter to generally represent conceptually sound strategies for
translating the relationship between gasoline usage and blood Pb (provided in the
Schwartz and Pitcher, 1989 study) to air-to-blood Pb ratios. In addition, EPA notes that
both approaches support air-to-blood ratios in the range of 1:7 to 1:10.
Regarding the assertion by the commenter that there is a weak association between blood
Pb and ambient air Pb derived from gasoline Pb, which makes it inappropriate to rely on
a ratio derived from Schwartz and Pitcher (1989), EPA disagrees with the commenter's
view, noting that the body of evidence regarding this relationship is robust (e.g., USEPA,
1986, sections 11.3.6 and 11.6). As stated in the 1986 Criteria Document, "there is
strong evidence that changes in gasoline lead produce large changes in blood lead"
(USEPA, 1986, p. 11-187). We additionally recognize CASAC's acceptance of this
study and note commenters' use of this study to independently derive a ratio of
approximately 1:8.
Specifically regarding the commenter's assertion that the Hayes et al., (1994) study
shows that a significant relationship does not exist between air Pb and blood Pb when
time is included as a variable in regression modeling, EPA notes that Hayes et al. (1994)
clearly states: "the overall decline in blood lead levels is associated with a corresponding
decline in air lead levels, and average air lead levels by quarter are highly correlated with
median blood lead levels by quarter...". While the addition of a sequentially ordered
variable for each quarter did remove the statistically significant association between air
Pb and blood Pb, the authors clearly note that they believed this provided additional
15
-------
support for the role of Pb in dust, soil and food (contributed by gasoline Pb emissions) in
determining overall blood Pb, as compared with direct inhalation alone (Hayes et al
1994). EPA notes, however, that it is not relying on this study for an estimate of air-to-
blood ratio (as discussed in section II.C.3.b(iii) of the Preamble to the final rule). In
questioning the linkage between gasoline-related ambient air Pb and blood Pb, the
commenter also points to Brunekreef et al. (1983) as showing that, while there are high
levels of deposition associated with gasoline Pb emissions, ambient air Pb levels are low,
pointing again (in their interpretation) to the unusual dynamics of gasoline Pb emissions.
However, Brunekreef et al. (1983) describe a number of factors, which could account for
the lower ambient air Pb levels seen in the suburban and urban study areas, including the
ambient air Pb measurement methods, location of the study areas, wind conditions, and
topography. These factors may make the air-to-blood ratios reported in this study more
relevant to areas similar to those in the study. However, EPA does not believe that these
factors call into question the well-established linkage between gasoline Pb and ambient
air Pb, or argue for considering gasoline Pb to represent a fundamentally different type of
ambient air Pb source, compared with other potential air Pb sources.
EPA disagrees with the comment that studies providing higher air-to-blood ratios (in the
range of 1:7 to 1:10) do not include control for potential confounders to blood Pb. As
noted in the proposal (section II.B.l.c) and in the 1986 Criteria Document, Brunekreef et
al. (1983) provides an air-to-blood Pb ratio of 8.5 that has been adjusted for a variety of
factors, including nationality, age, drinking water, milk consumption-calcium intake,
house age, number of rooms (crowding), paint status in home, parental education,
parental job attainment/income, mouthing behavior, pets, parent occupation/hobbies
related to Pb exposure, and geographic area. Additionally, Schwartz and Pitcher (1989)
addresses the issue of potential confounding, by providing evidence that reductions in
children's blood Pb were directly associated with decreases in the use of leaded gasoline
and not by simultaneous decreases in exposure to Pb paint, Pb in drinking water or
dietary Pb. A more recent study described in the proposal (Hilts, 2003) includes an
analysis that provides control for potential confounders, including alternate sources of Pb
exposure, through study design (i.e., by following a similar group of children located
within the same study area over a period of time). In addition, this study was conducted at
a time (late 1990's through 2001) after the period when significant reductions in non-air
Pb (e.g., solder used in cans) had taken place, thereby reducing the potential for temporal
confounding by these non-air sources of Pb exposure. The study authors report a ratio of
1:6 from this study (Hilts, 2003) and additional analysis of the data by EPA for the initial
time period of the study resulted in a ratio of 1:7.
Regarding the lEUBK-based analysis submitted by the commenter that they assert
illustrates the potential for confounding by non-air related Pb sources in deriving air-to-
blood Pb ratios. EPA agrees that this analysis illustrates the potential for confounding
when air-to-blood ratios are derived based on total blood Pb without attempting to
determine the fraction of total blood Pb that results from air-related Pb. It is specifically
for this reason that, in selecting studies to inform the identification of an air-to-blood Pb
ratio range, EPA focused on studies that address this issue of potential confounding by
nonair Pb sources (see previous paragraph). In addition, in obtaining air-to-blood Pb
16
-------
ratios from the risk assessment, EPA has also focused on identifying the fraction of total
blood Pb that is specifically related to air Pb and excludes non-air related Pb to the extent
feasible given the design of the analysis.
Lastly, EPA disagrees with the comments that air-to-blood ratios derived from the risk
assessment may be overstated due to errors in the dust Pb models employed. As
discussed more fully in section II.A.4 below, the dust Pb models used in the risk
assessment for the different case studies were derived from available data for Pb in house
dust for the different scenarios. While (as discussed in section II.A.4 below) EPA
recognizes limitations in the models, EPA disagrees with the commenters conclusions
that these limitations result in consistent bias in resultant dust Pb estimates.
In summary, EPA believes that the evidence supports consideration of ratios ranging up
to approximately 1:10 and that the results of the risk assessment add further support to this
air-to-blood ratio range.
iv. Comments on Concentration-Response Function for IQ Loss
The majority of commenters, which include public health organizations, national
organizations of air pollution control agencies, environmental organizations, and state, tribal and
local health and environmental agencies recommend the use of C-R function slopes from
analyses of children with lower blood Pb levels, with some of these additionally suggesting
alternate approaches to identifying the most relevant slopes. Industry commenters recommend
use of the median value from the second set described in the proposal. These comments are
described and addressed in section Il.C.S.b of the preamble to the final rule. Specific aspects of
these comments not discussed in the preamble are described and addressed below.
(1) Comment: A few industry commenters (ILZRO and ABR) state that EPA's conclusion of a
nonlinear dose-response relationship between blood Pb concentrations and
neurobehavioral or cognitive effects (e.g., IQ) is inappropriate because the supralinear
behavior of the relationship between the lognormally distributed variable (blood Pb
concentration) and the normally distributed variable (IQ) may simply reflect the expected
statistical outcome of regression analyses between such distributions. One industry
commenter (ILZRO) states that smaller number of observations at lower blood Pb levels
exerts extra influence on shape of function. In support of their views, commenters
variously cite Bowers and Beck (2006, 2007a, 2007b).
Response: This comment reflects the conclusions stated by Bowers and Beck (2006) in a
theoretical analysis examining the relationship between blood Pb and IQ. This
theoretical analysis inversely matched complementary percentiles of a lognormal
distribution of blood Pb concentrations with a normal distribution of IQ (e.g., matched
the 90th percentile blood Pb concentration with the 10th percentile IQ), which resulted in
a perfectly supra-linear relationship between blood Pb levels and IQ as would be
expected from their methodology. The authors concluded that "we expect to see the
supra-linear IQ-blood lead slope in all such studies because it is a requirement of the
shape of the blood lead concentration and IQ distributions" (Bowers and Beck, 2006).
17
-------
Multiple comments on this analysis have been published. Critiques of this analysis
addressed the appropriateness of matching complementary percentiles; the assumption of
normality for the IQ data, IQ covariates, and the IQ effects of Pb; and Bower's and
Beck's interpretation of their analysis (Hornung et al., 2006; Jusko et al., 2006; Bergdahl,
2006, 2007; Svendsgaard et al, 2007).
A discussion of the original Bowers and Beck paper and some of the published comments
on that analysis was presented in the Criteria Document (section 8.5.1). This discussion
notes that the analyses of Bowers and Beck (2006) are based on several assumptions,
including that the blood Pb concentrations are log-normally distributed, the IQ values are
normally distributed, and that the two distributions have an inverse relationship. The CD
states that, while the conclusions drawn by Bowers and Beck may be true under these
specified conditions, the authors' assumptions are not generally met in epidemiologic
analyses, as is described using the example of the Lanphear et al. (2005) pooled analysis.
In their response to the comments noted above that were published after the Criteria
Document was completed, Bowers and Beck (2007b) clarified that "our primary point
remains that modeled (not true) supralinear dose-response relationships are an expected
outcome of the type of analysis done in epidemiological studies, and are not, in
themselves, evidence of non-linear dose-response biological mechanisms." Further, they
stated that the main point of their original publication "was that there are instances where
the statistical constraints imposed by the distributional properties of blood lead
concentration data and IQ data do form the basis for the shape of the dose-response
relationship and one should not automatically eliminate this possibility in the
interpretation of non-linear dose-response relationships that are found" (Bowers and
Beck, 2007a).
In conclusion, EPA finds that the analysis conducted by Bowers and Beck is too
theoretical and narrow to contribute significantly to the interpretation of the
epidemiological literature on the relationship between blood Pb concentrations and
neurocognitive effects. EPA agrees with Bowers and Beck regarding the need to
carefully interpret epidemiological findings, and has done so in the formulation of
conclusions in the Criteria Document by considering all the available relevant literature
on this matter.
(2) Comment: One commenter from an industry group implies that EPA's conclusion of
supralinearity in the Pb-IQ loss C-R function is based on a linear slope reported by
Lanphear et al. (2005) for the analysis of children with blood Pb levels that never
exceeded 7.5 ug/dL (ILZRO). The commenter further states that EPA has not considered
conclusions of others such as CDC with regard to the Lanphear et al. (2005) study
findings.
Response: In considering these comments, EPA notes that the conclusion in this review
with regard to the nonlinear relationship between IQ loss and blood Pb is not based on a
single analysis. The Criteria Document contains an extensive discussion on the body of
evidence indicating the occurrence of steeper C-R slopes for IQ loss with lower blood Pb
18
-------
levels (CD, sections 6.2, 6.10 and 8.5.1). In several studies with cohorts of children with
varying population mean blood Pb levels, a trend of steeper C-R slopes for IQ loss with
lower blood Pb levels has been observed. In addition, in other studies which compared
C-R slopes below and above a blood Pb level cutoff point (generally 10, 7.5, or 5 |ig/dL )
within the study cohort, steeper C-R slopes were observed below the cutoff point. EPA
also notes the strong evidence provided by the pooled international analysis of 1333
children by Lanphear and others (2005), supported by further analyses by Rothenberg
and Rothenberg (2006), which is discussed in the Criteria Document (CD, section
6.2.13). In this analysis, various models, including the linear model, cubic spline
function, the log-linear model, and the piece-wise linear model, were investigated. The
shape of the C-R relationship was determined to be nonlinear, with the log-linear model
providing a better fit for the data compared to the linear model. The additional analysis
of that data set which segregated data based on peak blood Pb levels above and below 7.5
|ig/dL, also indicated significantly greater incremental Pb-associated intellectual
decrement for children with maximal blood Pb levels below 7.5 |ig/dL compared to those
with maximal blood Pb levels at or above 7.5 |ig/dL, providing further support for the
conclusion of nonlinearity. In considering the evidence with regard to nonlinearity, we
have explicitly considered the CDC review mentioned by the commenter (CD, pp. 6-64
to 6-65), noting the conclusion of the CDC review regarding the weight of evidence
support for "an inverse association between blood Pb levels <10 ug/dL and the cognitive
function of children" and the limitations recognized by the CDC review with regard to
the available studies that specifically examined the effect of blood Pb levels <10 ug/dL.
Further, as noted in the Criteria Document, there have been multiple studies published
since the CDC review that evaluate the effect of blood Pb levels <10 ug/dL, reducing
concern for the potential limitations cited by CDC.
(3) Comment: A few industry commenters (e.g., ILZRO) state that estimates of effects at blood
Pb levels less than 10 |ig/dL are inherently unstable due to relatively limited size of study
populations. In stating this view, the commenter focuses on the pooled international
analysis by Lanphear et al (2005), noting the relatively smaller number of observations in
that dataset with concurrent blood Pb levels below 10 ug/dL as compared to above 10
ug/dL . They additionally note that most of the children in this less than 10 ug/dL group
are from the Rochester cohort and none are from the Port Pirie, Australia cohort, stating
that the differential contributions from the cohorts to the lower blood Pb component of
the full dataset weakens conclusions made with regard to lower blood Pb levels. Further,
they state that as most of the population in the full pooled dataset has much higher blood
Pb levels, that portion of the population somehow serves to inflate the correlation
coefficient (r2) and reduce p-values and also reduce confidence intervals for the model
coefficients (beta estimates).
Response: As indicated in the previous response, EPA notes that the conclusion in this
review with regard to the nonlinear relationship between IQ loss and blood Pb is not
based on a single analysis, but rather based on the full body of evidence. The steeper C-
R slopes for IQ loss with lower blood Pb levels are observed both across different studies
with cohorts of children with varying population mean blood Pb levels and within studies
which compared slopes below and above a blood Pb level cutoff point. In response to the
19
-------
specific comment that the C-R slope estimates from the low blood Pb level subset of the
Lanphear et al. (2005) study is due to the varying contributions from the seven different
cohorts, EPA notes that the Lanphear et al. (2005) study had conducted sensitivity
analyses to examine the stability of their models. First, they evaluated the results of
applying a random-effects model (with sites random) rather than a fixed-effects model.
Lanphear et al. (2005) observed that results from the random-effects model compared to
the fixed-effects model (only a 3.7% difference) suggested that there was little
heterogeneity in effects across the different sites. Second, they examined the effect of
any one site on the overall model by calculating the blood Pb coefficient after omitting
each one of the seven cohorts at a time. Once again, there was little change in the blood
Pb coefficient (ranging from a 2.6% decrease to an 8.9% increase), providing further
evidence of the stability of the model and indicating that the results of the pooled analysis
did not depend on the data from any single study. Therefore, EPA does not agree with
the commenter that the data in the Lanphear et al. (2005) study were used
inappropriately, resulting in inflated r2 values and significance levels.
(4) Comment: A few commenters (e.g., ILZRO, ABR, BCI) cautioned against relying on
Lanphear et al. (2005), particularly the slope estimate from the subgroup analysis which
included children whose maximal blood Pb levels were below 7.5 |ig/dL. They state that
the slope from this analysis, which is the high end of first set of "steeper" slopes
identified in the proposal, is an "outlier". That is, the commenter stated that the linear
slope for the relationship of IQ with blood Pb levels from the Lanphear et al. (2005)
analysis of children whose maximal blood Pb levels did not exceed 7.5 ug/dL should not
be considered in identifying an estimate for the concentration-response relationship
between blood Pb and IQ. The commenter notes that the sample size for this analysis
included only 103 children, with majority representation (-67%) from the Rochester
cohort and minority representation from the Boston and Yugoslavia studies.
Response: While EPA recognizes that the slope from the analysis of children with peak
blood Pb levels below 7.5 ug/dL is notably higher than slopes from other analyses
involving children with somewhat similar blood Pb levels, EPA disagrees with the
commenters' view regarding the analysis supporting this slope (see response to previous
comment). Further, in identifying a C-R slope for use with the air-related IQ loss
evidence-based framework, EPA has relied on consideration of the four slopes from four
different analyses from four different studies of children with blood Pb levels closest to
those in the U.S. today, as described in section II.C.3.b of the preamble. The analysis
from Lanphear et al. (2005) of children with peak blood Pb levels below 7.5 ug/dL is
included among these four analyses. Given the general similarity of the blood Pb levels
in these four analyses, EPA concludes that it is not appropriate to single one slope out,
but rather has given equal weight to the full group. As described in section II.C.3.b of the
preamble, the median from these slopes (1.75 IQ points loss per 1 |ig/dL blood Pb) is
used to avoid undue influence from any one study.
(5) Comment: A few industry commenters state that confounders can contribute to "the
exaggeration of an observed non-linear slope to the concentration-response curve" (ABR,
p. 13). They cite the CDC (2005) analysis in support of this point. One commenter
20
-------
(ILZRO) additionally states that the handling of confounders in Lanphear et al. (2005)
contributes to exaggeration of the slope in the nonlinear model reported by those authors.
Response: Some commenters have indicated that residual confounding may have
contributed to the non-linear C-R relationship between blood Pb levels and IQ. The CDC
work group provided a hypothetical example in which this may be the case. In their
hypothetical example, children in a high exposure setting and children in a low exposure
setting experienced the same absolute change in IQ associated with an unidentified
factor. In the high exposure group, this change in IQ was observed across a change in
blood Pb level from 10 to 20 ug/dL while in the low exposure group, this change in IQ
was observed across a change in blood Pb level from 1 to 2 ug/dL. As the work group
explains, the IQ loss-blood Pb slope derived in this hypothetical scenario is steeper in the
lower blood Pb level setting than in the higher blood Pb level setting. One interpretation
of how such an example might occur is that there is a change in IQ attributable to an
unidentified factor that is the same regardless of the distribution of the blood Pb levels.
This implies that the unidentified factor is not related to the exposure of interest (i.e.,
blood Pb levels) and therefore cannot be a confounder of the relationship between IQ and
blood Pb levels.4 Another interpretation of this example is that the impact of confounding
is uneven along the distribution of the blood Pb levels. The only situation in which an
unidentified factor would contribute to the log-linear C-R relationship observed between
IQ and blood Pb levels is if this factor has a linear relationship with IQ but a non-linear
relationship with blood Pb levels. Examples of such factors are unknown. Given that the
non-linear relationship was observed in several different studies that adjusted for various
different potentially confounding factors (e.g., home environment, parental IQ, parental
education, indicators of SES, birth weight), it is unlikely that the non-linear relationship
is solely driven by residual confounding. A related comment states that the handling of
confounders in Lanphear et al. (2005) contributed to the exaggeration of the non-linear C-
R function. The Lanphear et al.(2005) analysis initially examined ten factors individually
and in combination with the other covariates to assess potential confounding of the
relationship between IQ loss and blood Pb levels. Consideration was given to the
stability of the parameter estimates as each additional term was added. For the final
multiregression model, five covariates were chosen: site, HOME score, birth weight,
maternal IQ, and maternal education. The addition of child's sex, tobacco exposure
during pregnancy, alcohol using during pregnancy, maternal age at delivery, marital
status, and birth order were not found to alter the effect estimate. Therefore, EPA does
not agree that the non-linear C-R function in the Lanphear et al. (2005) study was due to
the methodology used to select and adjust for potential confounders.
(6) Comment: Some industry commenters (e.g., ABR, ILZRO) disagree with EPA's
interpretation of the evidence for nonlinearity, particularly for blood Pb levels below 10
|ig/dL, stating that some studies (e.g., studies by Tellez-Rojo et al. [2006], Jusko et al.
[2007] and Surkan et al. [2007]) do not provide statistical support for the existence of a
4 For something to be a confounder, it must be related to both the outcome of interest (IQ in this case) and the
exposure of interest (blood Pb level), regardless of whether the exposure of interest is related to the outcome of
interest. A confounder can either make it seem like there is a relationship between the outcome and exposure when
there isn't at all, or make the relationship seem stronger or weaker than it actually is.
21
-------
nonlinear C-R function. In making this comment, commenters additionally state that,
contrary to the statement by CASAC in their July 2008 letter to EPA, Surkan et al.
(2007), Jusko et al. (2007), and Solon et al. (2008) do not provide support for steep C-R
function at low blood Pb levels or evidence of statistically significant nonlinearity at low
blood Pb levels. They further state that the study by Solon et al. (2008) focused on
children which are chronically malnourished making them more susceptible to
neurodevelopmental effects and making this study inappropriate to this U.S. context.
Further, with regard to Solon et al (2008), one commenter (ABR) states that the reported
risk estimate must be the result of error.
Response: First, EPA notes that, as with all NAAQS reviews, we have considered the
evidence in this review as a whole, and in assessing findings of individual studies, we
consider them within the context of the larger body of evidence. Further, we note that
three of the four studies cited by these commenters were published subsequent to the
completion of the Criteria Document. We have provisionally reviewed these studies in
conjunction with other relevant recent studies, published since 2006 and identified in
routine review of journals, (e.g., Miranda et al., 2007 and Jusko et al., 2008)5 in the
context of the findings of the 2006 Criteria Document.
EPA disagrees with the comment that analyses by Tellez-Rojo et al. (2006), which are
reviewed in the 2006 Criteria Document, do not support the general conclusion of
nonlinearity in the relationship between IQ and blood Pb at low levels. This study
observes a statistically significant loglinear relationship with blood Pb at 24 months in a
group of children for whom the mean blood Pb is 4.28 |ig/dL. Further, this study
separately performed linear regression analyses on two subsets of the full dataset. From
these analyses, the authors report a steeper slope for their study subgroup of young
children with individual blood Pb levels below 5 |ig/dL (n=193, for which the slope of-
1.71 was statistically significant, p=0.01) than those with blood Pb levels between 5 and
10 |ig/dL (n=101, for which the slope of-0.94 was not statistically significant, p=0.12).
While these slopes were not found to be significantly different, EPA notes the
consistency of the difference in slopes with the nonlinear relationship observed across the
full data set. EPA also notes the similarity of the slope for children with blood Pb levels
below 5 |ig/dL, for which the mean blood Pb level is 2.9 |ig/dL, and the slopes for other
subgroups analyses for similarly low blood Pb levels. For example, the slope reported by
Canfield et al. (2003) for children with peak blood Pb levels below 10 |ig/dL (mean of
3.32 |ig/dL) is -1.79 and that reported by Bellinger and Needleman (2003) for a similarly
selected group of children (mean blood Pb level of 3.8 |ig/dL) is -1.56. This is contrasted
with the slope reported by Tellez-Rojo et al. (2006) for all children with blood Pb levels
below 10 |ig/dL (mean 4.28 |ig/dL ) of-1.04, and other analyses of subgroups with mean
blood Pb levels between 7 and 10 |ig/dL for which still lower slopes are reported (see
CD, Table 8-7 and 6-1). Thus, EPA concludes that the findings of this study are
consistent with and supportive of the conclusion of nonlinearity in the IQ loss-blood Pb
relationship described in the Criteria Document, with this study providing particular
5 The complete list of studies on neurotoxic effects of Pb published subsequent to the Criteria Document that have
been considered in light of these comments is provided in Appendix A.
22
-------
support for this conclusion at low blood Pb levels (e.g., with analyses for mean and
individual blood Pb levels below 5 jig/dL).
With regard to the study by Jusko et al. (2008; published after completion of the CD),
EPA also disagrees with commenters. This study examined the relationship between
blood lead levels and IQ loss in children followed from 6 months to 6 years of age. Their
mean peak and concurrent blood Pb levels were 11.4 |ig/dL and 5.0 |ig/dL, respectively.
They conducted categorical analyses that compared the difference in IQ scores for three
categories of individual blood Pb levels, <5 |ig/dL, 5.0-9.9 |ig/dL, and >10 |ig/dL. In the
analyses of IQ loss with both peak and concurrent blood Pb, the authors observed a
statistically significant decrease in full-scale IQ across the increasing blood Pb categories.
Although mean blood Pb levels for the three categories are not reported, the mean
concurrent blood Pb level across the complete data set was 5 |ig/dL. We note that such
categorical analyses, while supportive of the general conclusion regarding an inverse
relationship of IQ with concurrent blood Pb, do not provide an assessment of the shape of
the concentration-response relationship. The authors did examine the shape of the C-R
function between IQ loss and peak blood Pb levels using individual-level data and
observed a non-linear relationship. That is, they observed that the slope of the IQ loss-
blood Pb relationship was steeper at lower than at higher peak blood Pb levels. Thus,
while the shape of the C-R function for the relationship with concurrent blood Pb levels
was not examined in this study, the nonlinear finding with regard to peak blood Pb levels
is consistent with our conclusions of nonlinearity in the IQ loss blood Pb relationship.
With regard to the study by Surkan et al. (2007; published after completion of the CD),
EPA notes that this paper does not include the type of analysis appropriate to draw
conclusions regarding the slope or shape of the C-R function. Rather than an analysis of
association across the full range of blood Pb levels studied, this study includes only a
much less powerful categorical analysis in which the IQ scores are compared among
three groups: children with blood Pb levels 1-2 |ig/dL (n=286), 3-4 |ig/dL (n=71) and 5-
10 |ig/dL (n=32). The paper reports a significant difference in full-scale IQ between
children with blood Pb levels of 1-2 |ig/dL and those with blood Pb levels of 5-10 |ig/dL.
Additionally, although a statistically significant difference in performance was reported
between children with blood Pb levels 1-2 |ig/dL and those with blood Pb levels 3-4
|ig/dL on one of the IQ subtests (digit span) which has been shown previously to be
affected by Pb, the authors did not find a significant difference in full-scale IQ between
these two groups. This type of categorical analysis is highly sensitive to how subjects are
grouped into categories. The small difference in the blood Pb levels in the two categories
(only 1-2 |ig/dL difference) likely contributed to this finding. In summary with regard to
this study, however, EPA disagrees with commenters that it is contradictory to EPA
conclusions of nonlinearity in the concentration-response relationship as it does not
include analyses suitable to examine the issues regarding the shape of the concentration-
response function.
The study by Solon et al. (2008; published after completion of the CD) focuses on
children with somewhat higher blood Pb levels than those of the other two studies (mean
of ~7) and examines the potential role of folate concentration in modifying the effect of
23
-------
Pb on intelligence. The authors report a significant association of MDI scores with blood
Pb levels, with a coefficient of-3.32 for this relationship for the theoretical case of a
folate level of zero. Application of the mean folate level for the study (225 ug/ml) to this
coefficient yields a coefficient for MDI of-1.07 IQ point loss per 1 |ig/dL, a value
comparable to coefficients reported for studies of similar blood Pb level children. Given
this comparability, EPA disagrees with the commenter's view that the findings of this
study are inconsistent with other studies within the body of evidence.
EPA also notes findings of other more recent studies not considered in the CD, which
examined neurocognitive and neurobehavioral effects at low blood Pb levels. These
include the study by Miranda et al. (2007), which observed a decline in reading and math
scores at blood Pb levels down to 2 |ig/dL. This study further found a non-linear
relationship between blood Pb levels and these neurocognitive outcomes, i.e., steeper
slope with lower blood Pb levels. Another recent study is that by Braun et al. (2006)
which observed associations between individual blood Pb levels below 5 |ig/dL and
increased risk of attention deficit hyperactivity disorder. These two studies provide
further evidence - among studies published subsequent to the Criteria Document - that
other neurotoxic effects of lead are also observed at lower blood Pb levels, and also, in
the case of Miranda et al. (2007) of a supralinear relationship between blood Pb and
neurocognitive outcomes.
In summary, after consideration of the studies cited by commenters, EPA concludes that
the more recent studies they cited provide only limited information with regard to the
shape of the C-R curve and, in light of other recent studies and those reviewed in the
Criteria Document, do not warrant reopening the air quality criteria for Pb to reconsider
the evidence relied upon in this review of a nonlinear concentration-response relationship
between blood Pb and IQ at blood Pb levels below 10 |ig/dL. Having concluded that
these more recent studies do not warrant reopening the air quality criteria review for Pb,
EPA is not relying on them in this review for the reasons stated in the preamble (section
1C). Furthermore, EPA believes that the conclusion of nonlinearity is well founded in
the evidence described in the Criteria Document.
(7) Comment: One industry commenter stated that the Lanphear et al. (2005) paper contains
errors and cannot be relied upon (ABR pp. 10-11). For example, comments from ABR
note an error regarding which EPA published a technical memo (Jan 26, 2007), describe a
potential 2nd error in Table 4 of the paper that they note they cited in March 2008
comments to EPA, which they state has not been resolved, and also describe what they
claim is a potential 3rd error regarding Figure 3 of the paper.
Response: EPA agrees with the commenter that two errors have been identified with
regard to Table 4 in the Lanphear et al. (2005) publication. However, EPA notes the
February 21, 2007 email from the R. Hornung, one of study authors, that is in the docket
for this rulemaking, provides a corrected version of this table that addresses both of these
issues. In addition, EPA identified typographical errors in two numbers associated with
confidence intervals reported at the top of the 1st column on p. 897 of the publication, In
reporting this information in the CD, EPA corrected these errors (CD, p. 6-70). Further,
24
-------
none of the errors identified by EPA affect aspects of the study on which EPA has relied
in this review.
EPA considers the commenters statements about Figure 3 to be without sound basis.
Figure 3 illustrates the log-linear model for concurrent blood Pb level and IQ, along with
the mean IQ and 95% confidence intervals on the mean for 5 subgroups of the full
dataset. The commenters state that the confidence intervals shown in this figure are a
function of sample size and they suggest that based on their interpretation of information
in the paper for the sample size of the group with peak blood Pb levels below 10 |ig/dL,
the <5 |ig/dL concurrent blood Pb level group is much smaller than the 5-10 |ig/dL
concurrent blood Pb, and accordingly the confidence interval bars should be much
different in size. As the commenters state, the confidence intervals are a function of
sample size, which is likely to differ notably among the study groups (although the
authors do not report these sample sizes), however they are not simply a direct function
of sample size. The standard deviation also affects the width of the confidence interval,
and may vary in the different blood Pb categories. Thus, there is no basis to conclude
that the confidence intervals displayed in Figure 3 are incorrect, although EPA's
conclusions drawn regarding this study did not depend upon this figure. Additionally, in
considering the Lanphear et al. (2005) publication, EPA notes the analysis by Rothenberg
and Rothenberg (2005) of the same dataset. In the latter publication, the authors state
that the data set "was analyzed with the original model specifications, including log-
transformed BPb, using multiple regression analyses" and they obtained the same effect
estimates for the loglinear model as those reported by Lanphear et al. (2005). These
findings add to EPA's confidence in our consideration of the Lanphear et al. (2005)
publication. Procedural concerns on this point raised by commenters and EPA's response
are described in section V. A below.
(8) Comment: One commenter recommended use of a non-linear function to estimate IQ loss at
different standard levels, stating that use of average linear slopes underestimate impact
(NYDOH- pp. A-3 to A-4).
Response: When considering multiple changes in blood Pb level over a broad range of
blood Pb levels, such as is done in the risk assessment, EPA notes that a non-linear
function such as the log-linear model from Lanphear et al. (2005) may provide better
estimates - across a broad range of blood Pb levels - than use of a single average linear
slope.
However, for purposes of estimating IQ loss at the lower mean blood Pb levels that are of
interest to EPA in considering estimated air-related IQ loss associated with the revised
level for the standard, EPA has concluded that it is appropriate to utilize linear slopes
derived directly from analyses with mean blood Pb levels near these levels. The median
blood Pb level for the full Lanphear et al. (2005) pooled analysis is 9.7 |ig/dL, which is
substantially higher than the mean levels of interest. As the nonlinear model based on
this data set is influenced by the full range of data points including those at the higher
blood Pb levels, a linear slope derived from the low end of the log-linear model will
necessarily not be as robust as a linear slope derived directly from data points in that
25
-------
range. Accordingly, as described in section II.C.3.C of the preamble, in applying the
evidence-based framework in consideration of a revised level for the standard, rather than
relying on the Lanphear et al. (2005) log-linear C-R function, EPA has relied on the
median of four analyses from four different studies for which the study group mean blood
Pb levels were closest to the lower blood Pb levels of today's U.S. children.
v. Comments on Exposure and Risk Considerations
With regard to considering how the quantitative exposure and health risk assessments
should factor into a decision on the standard level, EPA notes that the comments generally fell
into two groups, with one arguing that the risk assessment provided support for lower levels for
the standard and the other arguing that the risk assessment indicated that substantial revision
provided negligible public health benefit. These comments are described and addressed in
section Il.C.S.b of the preamble to the final rule and below.
(1) Comment: Some commenters (e.g., New York Department of Health, NRDC, MO Coalition
for the Environment) state that more weight should be placed on the risk assessment in
selecting a level for the standard and that the air-related risk estimates for both the
median and 95th percentile of simulated populations support selection of a standard level
at or below 0.2 |ig/m3. In support of this view, some commenters note CASAC support
for the risk assessment. Additionally, one commenter asserted that EPA should not
dismiss the results of the risk assessment for reasons of uncertainty, without stating why
greater uncertainty is ascribed to the risk assessment than to the evidence-based
framework (e.g., without a "transparent evaluation" of the relative uncertainties of the
risk assessment vs. the evidence-based frameworks"). Several commenters (e.g.,
Physicians for Social Responsibility) also point to the IQ loss incidence results of the risk
assessment as supporting a standard level at or below 0.02 |ig/m3. In support of this
view these commenters note results of the risk assessment with regard to the estimates for
a standard of 0.02 |ig/m3, of reduction in number of children with IQ losses of 7 points or
more and cite studies discussing research on the impact of Pb exposure-related cognitive
impairment on earnings potential.
Response: EPA did consider the degree of uncertainty associated with both the risk
assessment and the evidence-based considerations. Uncertainties associated with both
approaches are thoroughly discussed in both the proposed rule and the preamble to the
final rule. In the case of the risk assessment, specific sources of uncertainty are described,
while in the case of the evidence based analysis, the basis for selection of individual
analysis elements such as the air-to-blood Pb ratio is discussed in detail, including
sources of uncertainty. The uncertainties associated with the evidence-based framework
derive from uncertainties in the evidence itself, and what can be projected directly from
the evidence. The uncertainties in the risk assessment include many of the same
evidence-based uncertainties, as the risk assessment is largely premised on the same body
of evidence (e.g., the development of a concentration-response function), as well as
additional uncertainties such as those related to modeling of air dispersion, lead
deposition, lead intake through ingestion and inhalation, and assignment of lead exposure
to air-related and non-air related pathways, (see sections 2.4.6, 3.5, 4.3 and 5.3.3 of
26
-------
Volume I of the Risk Assessment Report for details on characterization and discussion of
uncertainty related to the risk assessment). Specifically with regard to the IQ loss
incidence results referenced by the commenter and the potential public health benefit
associated with reducing the number of children with total Pb-related IQ loss above seven
points, EPA notes that we considered all of the risk metrics generated as part of the risk
assessment in considering the alternate standard levels. We further note, however, that
this particular category of IQ loss incidence results is subject to considerable uncertainty
due to its focus on risk for upper percentiles of the population, as discussed in section
4.2.7 of the Staff Paper. Thus, as discussed further in sections II.B.3 and II.C.S.c of the
preamble, the Administrator, in his decision-making for this review, gave primary
consideration to the available evidence and the evidence-based framework, but also took
the exposure and risk assessments into consideration in his decisions, noting that they
provide some further perspective on the potential magnitude of air-related IQ loss.
2) Comment: Several industry commenters (ABR and the Doe Run Resources Corp.)
recommended that more weight be placed on the risk assessment, which they state
indicates little benefit to public health associated with levels in the lower part of the
proposed range (e.g., "current conditions" in the risk assessment's urban case studies).
One commenter (the Doe Run Resources Corp.) further noted that the reduction in
incidence of children with greater than 1 IQ point lost due to Pb exposure under
alternative standard levels (compared against current conditions) is relatively small and
does not support setting the NAAQS at these alternative lower standard levels.
Response: Because the majority of the residential populations modeled in the location-
specific urban case studies have current conditions that are well below a 0.2 |ig/m3
standard level, risk estimates for lower standard levels will show only modest risk
reductions. However, this does not mean that a subset of an urban population would not
experience more significant risk reductions at standard levels below current conditions.
EPA has noted that the general urban case study can reasonably represent smaller
subpopulations within urban areas that experience ambient air Pb concentrations at the
standard level being evaluated. Therefore, risk results for this case study (for a particular
standard) provide perspective on the degree of risk reduction that could be realized by
urban populations or subpopulations experiencing ambient air Pb concentrations at the
standard level of interest. As can be seen by reviewing risk results generated for the
general urban case study (and comparing them to those generated for the location-specific
case studies - proposal notice, Table 4), the degree of risk reduction seen for the general
urban case study is larger across alternate standard levels. In addition, the more
appropriate risk metric to consider in terms of the risk assessment is not the degree of risk
reduction in moving between standard levels, but rather, the degree of air-related risk
remaining at any particular standard level. That metric describes the risk faced by a
population or subpopulation when a given standard is just met, which is the metric most
relevant to selecting a standard that protects public health with an adequate margin of
safety.
Regarding the comment concerning the population incidence results and particularly the
small shift in the numbers of children with greater than 1 point total Pb-related IQ loss
27
-------
under alternate standard levels (proposal notice, Table 5), it is important to note that the
mean Pb-related IQ loss for the three location-specific urban case studies under all
standard levels evaluated is above one IQ point. This means that we would expect to see
little shift in the numbers of children with greater than one IQ point loss (since most
children are above that level of IQ loss under all standard levels). Essentially, a focus on
the shift in number of children with at least one IQ point loss emphasizes the extreme
low-end of the risk distribution for all urban case studies. EPA notes that another risk
metric related to evaluating population incidence in terms of magnitude of IQ loss is the
number of children with greater than 7 points Pb-related IQ loss (proposal notice, Table
6). In contrast to the number of children with more than one IQ point loss, as shown in
Table 5 of the proposal notice, Table 6 focuses on the upper-end of the risk distribution
for urban children. By comparing the estimates in Tables 5 and 6 in the proposal notice, it
can be seen that the various alternate standard levels result in a larger shift in the number
of children with greater than 7 points Pb-related IQ loss than in the number of children
with greater than 1 point Pb-related IQ loss.
d. Comments on Adequate Margin of Safety
Some commenters indicated concern that EPA has not described how the proposed range
of standards provides an adequate margin of safety, as required by the CAA. This comment is
addressed in section II.C.S.c. An additional comment is responded to in this section.
(1) Comment: One commenter (Bayview Hunters Point Community Advocates) stated that in
considering the level for the primary Pb standard, EPA failed to consider that EPA's
socio-demographic analysis (Pekar et al., 2008) indicated that people living closer to
monitors registering higher Pb levels tend to be poorer and more predominately African
American than the county or national averages. The commenter stated that EPA could
have estimated IQ loss in the risk assessment stratified by race and income. In the view
of the commenter, such information about environmental inequities should be used in
determining the adequate margin of safety for the standard.
Response: While as the commenter notes, EPA's socio-demographic analysis (Pekar et
al., 2008) indicated that people living closer to monitors registering higher Pb levels tend
to be poorer and more predominately African American than the county or national
averages, EPA does not agree with the commenter that there were sufficient data
available to stratify remaining air-related risk for each alternative standard level by race
or income in the quantitative risk assessment, particularly in light of the use of a
nonlinear C-R function. The data that would have been necessary to support more
refined population-level modeling of ambient air-related Pb risk^br specific sensitive
subpopulations—such as demographic information about children (e.g., particular
income or racial groups) at the appropriate geographic scale, as well as behavioral
exposure factors specific to those groups —is not presently available. Thus, within the
quantitative risk assessment, EPA did not characterize risk by these subgroups and for
purposes of the socio-demographic analysis, EPA simply identified characteristics of
populations living in proximity to Pb sources or monitors.
28
-------
As explained in section Il.C.S.b of the preamble, EPA's reliance on the air-related IQ loss
framework in effect focuses on identifying a level for the Pb NAAQS that is requisite to
protect the group of children who are living near air-related sources of lead who are
likely to be most highly exposed. Thus, to the extent that minority or low-income
populations are in fact concentrated near stationary sources of air-related Pb emissions,
they would be included in the subpopulation of children on which the Administrator has
focused his attention through use of the air-related IQ loss framework. The revised Pb
NAAQS are set at a level which, in the judgment of the Administrator, protects the health
of the sensitive subgroup of more highly exposed children (along with other groups) with
an adequate margin of safety. To the extent the commenter is suggesting the standard
should be set more stringent than necessary to protect the health of sensitive groups with
an adequate margin of safety, EPA disagrees.
3. Additional Comments on the Interpretation of Scientific Evidence
Specific comments on the EPA's interpretation of the scientific evidence not discussed in
the preamble to the final rule are described and addressed in this section.
(1) Comment: Several commenters stated that EPA's consideration of the health effects
associated with low blood Pb levels should include more explicit recognition of
neurological effects beyond IQ loss and effects beyond childhood. Such comments
regarding health effects that were raised by commenters included reference to studies
reviewed as part of the CD as well as a number of studies published subsequent to the
2006 Criteria Document. The effects noted from these more recent studies include the
following:
• evidence of association of blood Pb levels with attention deficit hyperactivity disorder
in children (with reference to Braun et al., 2006 and Nigg et al., 2008);
• evidence of association of childhood blood Pb levels with aggressive and criminal
behavior in early adulthood (with reference to Wright et al., 2008);
• evidence of changes in brain volume at 19-24 years of age (with most affected areas
including those responsible for executive functions, mood regulation and decision-
making), associated with childhood Pb exposure (Cecil et al., 2008);
• evidence of association of blood Pb levels well below 10 |ig/dL_with cardiovascular
mortality of adults (with reference to Menke et al., 2006); and
• irreversibility of Pb-related neurological effects (with reference to Miranda et al. 2007
and Jusko et al., 2008).
Response: As an initial matter, EPA has considered a range of health effects associated
with low blood Pb levels beyond IQ loss and effects beyond childhood. These are
described in detail in the CD and summarized in the Staff Paper, ANPR, proposal and
final notice. While EPA has focused on IQ loss in selecting a level for the standard, the
revised standard is judged to protect, with an adequate margin of safety, the health of
children and other at-risk populations against an array of adverse health effects, with
neurological effects in children among the most notable.
29
-------
Discussion of the more recent studies cited by commenters was not included in the CD or
in the Staff Paper, as they were published after the CD was finalized. EPA's provisional
consideration of these studies in conjunction with consideration of other relevant recent
studies6, published since 2006 and identified in routine review of journals, in the context
of the findings of the Criteria Document, concludes that the new information and findings
do not materially change any of the broad scientific conclusions regarding the neurotoxic
effects and cardiovascular effects of Pb exposure made in the 2006 Lead Criteria
Document. Consistent with the conclusions reported in the CD, these studies, along with
additional recent studies, have observed adverse effects of lead exposure on
neurocognitive and neurobehavioral outcomes other than IQ, including academic
achievement, behavioral problems, attention-deficit / hyperactivity disorder, conduct
disorder, and criminal arrests. The study by Menke et al. (2006) observed an association
between blood Pb levels and increased all-cause and cardiovascular mortality in adults,
providing further supportive evidence for an association of Pb with cardiovascular
morbidity and mortality, as reported in the CD. These studies add to the overall weight of
evidence and focus on findings of neurocognitive and neurobehavioral effects beyond IQ,
as well as findings of cardiovascular effects, in most cases involving study groups with
lower blood Pb levels than were available for review in the CD. Having concluded that
these studies do not warrant reopening the air quality criteria review for Pb, EPA is not
relying on them in this review for the reasons stated in the preamble (section 1C). These
studies will be fully considered in the next review of the Pb NAAQS.
(2) Comment: Several commenters stated that EPA was proposing to use 9 |ig/dL as a baseline
blood Pb level and provided recommendations for alternate views.
Response: EPA is not clear what the commenters are referring to. In the proposal, EPA
has described the evidence of effects that extend to some of the lowest blood Pb levels
assessed, including mean blood Pb levels below 2 ug/dL. EPA has not proposed to use 9
|ig/dL as a baseline or acceptable level.
(3) Comment: Some industry commenters (ABR, p. 12, ASARCO) state that EPA, in
considering the evidence for the impact of Pb (and particularly air-related Pb) on IQ has
not adequately considered whether the past reduction in air Pb has resulted in a benefit to
children's IQ. One commenter (ABR, p!2) further cites an editorial by CDC scientists
(Brown and Rhoads, 2008) as concluding that, while blood Pb levels have dropped since
the late 1970s, there has been no observed increase in IQ by the amount that might be
predicted by some concentration-response (C-R) curves under consideration in EPA
proposed rule.
Response: EPA notes that there is evidence of increased IQ scores over the past several
decades, this evidence is generally termed the "Flynn effect" (Flynn, 1994). This
evidence indicates that IQ scores have increased by an average of 3 points per decade in
the United States. Due to these observed increases in IQ scores, IQ tests are routinely re-
standardized to ensure that subjects are not scored against inaccurate norms. For
6 The complete list of studies on neurotoxic effects of Pb published subsequent to the Criteria Document that have
been considered in light of these comments is provided in appendix A.
30
-------
example, the Wechsler Intelligence Scale for Children (WISC) has been renormalized
three times, in 1974, 1991, and 2003, since it was originally developed in 1949.
Hypothesized explanations for the observed Flynn effect include environmental changes,
societal changes, and improvements in nutrition.
Additionally, as described in section II.A.2.C of the preamble to the final rule, the
evidence indicates that the C-R function is nonlinear, such that to estimate IQ loss
prevented by the drop in blood Pb levels from the late 1970s to today, a C-R function
should be chosen from studies with blood Pb levels comparable to those of the late 1970s
(e.g., mean of 15 jig/dL). Among those considered in the proposal, the Lanphear et al.
(2005) study has the highest blood Pb levels, with a median of 9.7 |ig/dL. Application of
the average linear slope of-0.20 IQ points per |ig/dL derived from this study (estimated
for the range of data from from the 5th to the 95th percentile values of 2.5 to 33.2 |ig/dL)
would yield IQ gains on the order of 2-3 points for a decrease in the population mean
blood Pb level from 15 to ~2 |ig/dL over a 30 year period. This estimate is well within
the increase in IQ scores observed by the Flynn effect (increase of 9 points over 30
years). EPA concludes, as have others (e.g., Nevin, 2000) that the decline in blood Pb
levels likely contributed to the observed Flynn Effect. EPA additionally notes that other
variables, including other environmental changes, societal changes, and improvements in
nutrition also likely contribute to this effect.
(4) Comment: One commenter (American Farm Bureau) recommends that EPA clarify that
agricultural related activities are not significant sources of Pb exposure.
Response: Based on the information currently available to EPA, we agree with the
commenter that we do not have information indicating that agricultural sector activities
are significant sources of Pb in human exposure pathways. While diet can be a
significant source of Pb, particularly for adults, and Pb emitted into the air can make its
way into dietary pathways, as described in the Criteria Document, Staff paper and
Proposal (section II. A), agricultural activities have not been identified as a significant
source of Pb to dietary pathways.
(5) Comment: One commenter (ILZRO) states that EPA incorrectly describes the IARC
classification of Pb as a probable human carcinogen, noting that it is "inorganic lead
compounds" which are classified by IARC as "probably carcinogenic to humans", while
organo-lead compounds are assigned by IARC to Category 3 ("inadequate data to
classify"), with metallic lead categorized as "possibly carcinogenic to humans"
(Category 2B).
Response: EPA agrees with the commenter that different forms of lead have been given
three different classifications by IARC, with the most prevalent form of ambient Pb,
inorganic Pb compounds, being classified as "probably carcinogenic to humans" (class
2A). As the commenter notes, organic Pb compounds are classified by IARC as "not
classifiable as to carcinogenicity to humans" (class 3) and metallic Pb is classified as
"possibly carcinogenic to humans" (class 2B).
-------
(6) Comment: One commenter (ILZRO) cited an assessment recently submitted to the European
Union (EU) for their review in support of the commenter's statement that the effects of
Pb on blood pressure and cardiovascular disease are overstated by EPA and should not be
utilized as endpoints of concern for low-level human Pb exposure.
Response: EPA notes that the assessment cited by the commenter is not available to the
public. It was submitted to the EU and is currently under review by that body. The
Criteria Document identifies cardiovascular effects in adults as among the effects best
substantiated as occurring at low blood Pb concentrations (e.g., as low as 5 to 10 |ig/dL
or possibly lower) and a category of effect that is "clearly of greatest public health
concern" (CD, p. 8-60). EPA notes, however, that we have identified neurological and
neurobehavioral effects in children as those of greatest concern in revising the primary Pb
standard.
4. Additional Comments on the Exposure and Health Risk Assessment
Comments related to the exposure and risk assessment conducted for Pb that are not
discussed in the preamble to the final rule (e.g., in sections II.B and II.C.3) are described and
addressed in this section. EPA notes that the Admininstrator placed greater weight on the
evidence-based framework in selecting the standard than on the risk assessment.
(1) Comment: Two commenters from industry (ABR and the ILZRO) asserted that the use of
proportional roll-up and roll-down procedures for the location-specific urban case studies
(to predict air quality conditions approaching the current standard and conditions under
alternate lower standard levels, respectively) were not appropriate. Specifically, they
argued that ambient monitoring trends in urban areas demonstrated that ambient air Pb
levels tended to decrease disproportionately, with areas near existing industrial Pb
sources (i.e., areas near source-oriented monitors) experiencing greater percentage
reductions than areas more distant from sources. Furthermore, they noted that past
experience with urban areas developing State Implementation Plans under the current Pb
NAAQS suggested that areas in the vicinity of industrial Pb sources were the focus of
emissions reduction actions, resulting in larger reductions in ambient air Pb near those
locations compared with areas further from these sources. For these reasons, the
commenters asserted that the use of proportional roll-up procedures to simulate
conditions under the current NAAQS overstate likely associated Pb-related risk.
Similarly, the use of a proportional roll-down approach to evaluate alternate standard
levels likely over-states potential risk reduction.
Response: EPA acknowledges that there is significant uncertainty associated with the use
of proportional roll-down and roll-up procedures for simulating ambient air Pb levels
under alternate standard levels and the current NAAQS. Furthermore, EPA does not
dispute the observation made by the commenters regarding the evidence from available
monitoring data in some urban locations as well as experience with earlier efforts to
attain the Pb NAAQS. However, given the limited monitoring data currently available in
urban areas, EPA believes that application of a proportional roll-down or roll-up
approach is reasonable and that efforts to develop regional-specific non-proportional
32
-------
strategies could be subject to greater uncertainty than the approach used in the risk
assessment.
Given the current uncertainty in simulating either roll-down or roll-up procedures to
model alternate standard levels for larger urban areas, EPA believes that this supports
placing greater emphasis on risk results from the general urban case study. This case
study does not rely on a roll-down or roll up approach. This case study represents a
smaller urban area expected to have ambient air Pb levels in-line with the standard level
being evaluated, thereby avoiding the need to apply roll-down or roll-up procedures.
There is uncertainty in this case study as well, but it derives from the modeling of
representative gradients of air quality in a hypothetical urban area that meets a certain
level of the standard, as compared to rolling ambient levels up or down for any one
specific city.
2) Comment: One commenter (ABR) completed a sensitivity analysis looking at the potential
impact on overall blood Pb levels from reductions in outdoor soil Pb levels associated
with reductions in ambient air Pb. The sensitivity analysis utilized regression-derived
relationships relating ambient air Pb to indoor dust Pb and outdoor soil Pb, with both
relationships obtained from USEPA, 1989. The results of their analysis suggested that the
change in outdoor soil Pb levels would result in a relatively small impact on total blood
Pb levels for children. Therefore, they asserted that the uncertainty resulting from EPA
not having considered the potential impact of air Pb deposition on outdoor soil, and any
resulting bias toward underestimating risk estimates, is small and unlikely to be policy
relevant.
Response: The sensitivity analysis completed by the commenter utilizes a regression-
based relationship between ambient air and outdoor soil Pb that reflects an underlying
dataset that may be significantly different from conditions encountered in present-day
urban residential settings. The dataset underlying the regression model (see Tables A-3
and B-3 in USEPA, 1989) reflects primarily conditions near larger industrial point
sources of Pb in the 1970's and 1980's and consequently (a) reflects ambient air Pb levels
that are an order of magnitude or more higher than alternate standard levels evaluated in
the commenter's sensitivity analysis and (b) includes residential areas with significant
historical Pb impacts to soil, primarily from smelter facilities, which may in many
instances not be representative of conditions in present day urban residential locations.
Therefore, EPA believes that use of the regression model in the sensitivity analysis
reflects an extrapolation of that model to conditions beyond the dataset used in its
derivation and is therefore subject to significant uncertainty. There is the potential that
urban areas under present or future conditions (i.e., lower ambient air Pb levels with a
different particle size mix) could display different soil deposition and buildup dynamics
from those reflected in the USEPA (1989) regression model.
3) Comment: Two commenters (Kentucky Division of Air Quality and ABR) asserted that the
risk assessment overstates risk for urban case studies. The Kentucky DAQ noted that
"the air quality scenario for the urban case studies assumes ambient air Pb concentrations
higher than those currently occurring in nearly all urban areas nationally" (partial quote
33
-------
from 73 FR 29210 cited in their comments), leading them to conclude that risk estimates
generated for the urban case studies are overestimated. ABR asserted that risks for the
urban case studies were overstated due to errors in the indoor dust model used by EPA
for the urban case studies. Specifically, the commenter asserted that the indoor dust Pb
model underestimates the contribution from nonair Pb sources due to a mathematical
quirk associated with the nonlinearity of the model. They assert that the nonair
contributions to Pb in housedust predicted by the model are significantly lower than
levels seen in the literature and they argued that this underestimation results in residual
indoor dust Pb being incorrectly assigned to air-related indoor dust Pb by the model, such
that air-related indoor dust Pb estimates are biased high. Furthermore, the commenter
suggested that a sensitivity analysis completed by EPA looking at two versions of the
model, one as used in the risk assessment and an alternate model with fixed background
dust Pb levels, shows that the model as used in the risk assessment overestimates risk
reduction benefits associated with alternate standard levels.
Response: EPA disagrees with the commenter's assertion that air Pb concentrations in the
urban case studies are higher than conditions in nearly all urban areas nationally. The
statement from the proposal cited by the commenter pertains to the current NAAQS
scenario in which ambient air Pb levels in the urban case studies are increased to just
meet the current NAAQS. As stated in the NPR, this scenario is hypothetical and
represents ambient air Pb levels which are significantly higher than current conditions in
most urban areas, based on available monitoring data. In contrast, the "current
conditions" air quality scenarios for the urban case studies are based on monitoring data
specific to the urban case study areas. Additionally, alternative NAAQS scenarios
simulated for the location-specific urban areas involved air Pb concentrations in those
locations lower than those described as "current conditions" which represented recent air
quality data.
Regarding the comment that risk for the urban case studies are overestimated due to
errors in indoor dust Pb modeling, EPA acknowledges that the hybrid model has non-
linearity in the underlying conversion of indoor dust Pb loading to concentration and that
this results in an estimate of background non-air indoor dust Pb that varies and is
dependent on the ambient air Pb level being modeled. However, the non-linear
conversion of indoor dust Pb loading to concentration is based on an underlying HUD
dataset that reflects a non-linear relationship between these two variables. Therefore, this
aspect of non-linearity in the indoor dust Pb model is supported by the literature.
Regarding the assertion by the commenter that nonair contributions to Pb in housedust
predicted by the indoor dust model are significantly lower than levels seen in the
literature, EPA notes that the literature does not currently support the identification of a
nonair indoor dust Pb level. Available datasets characterizing indoor dust Pb levels in the
residential setting reflect housing subject to some degree of ambient air Pb impact and
consequently indoor dust Pb concentrations presented in these studies reflect a
combination of background and air-sourced Pb impacts. Furthermore, EPA notes that
observations made by the commenter of houses in areas not adjacent to point sources
having indoor dust Pb concentration values in the hundreds (ppm), likely reflect older
34
-------
housing with increased potential for indoor dust Pb impacts from Pb paint. It was not the
intent in developing the indoor dust Pb model to focus on older housing vintage, but
rather to model a background (non-air) indoor dust Pb prediction that would represent
broadly the U.S. housing stock, including both newer and older houses. It is likely that
this non-air indoor dust Pb value would be lower than the value presented by the
commenter (although as noted here, it is not possible to identify that value from the
literature).
As the commenter noted, EPA conducted a sensitivity analysis to assess the influence of
non-linearity in the indoor dust model on estimates of total and air-related exposure and
risk (USEPA, 2007b). This sensitivity analysis included an alternate approach for
estimating a fixed background (non-air) indoor dust Pb level. While this alternate
approach did result in noticeably lower recent air-related exposure and risk estimates for
each simulated standard level, total risk (for each standard level) was the same as that
estimated using the original formulation of the indoor dust Pb model (i.e., the total
change in exposure and risk between standard levels was the same for both model
formulations). In addition, it is important to point out a key limitation in this alternate
formulation of the indoor dust Pb model, which undermines the validity of this approach.
By setting ambient air Pb levels to zero and then solving for indoor dust Pb, the alternate
formulation used the steepest part of the loading-to-concentration curve to conduct the
key step of translating indoor dust Pb loading to equivalent concentration. In reality, we
would expect to always have a mixture of indoor dust Pb loadings from non-air and air
sources, thereby resulting in a larger total loading value, which would in turn be
translated into an indoor dust Pb concentration at a flatter portion of the curve (see
Section 5.3.3.4 of Volume I of the Risk Assessment Report).
EPA does acknowledge the uncertainty in its modeling of indoor dust concentrations, and
recognizes that this is one of the several sources of uncertainty in the risk assessment.
4) Comment: A few commenters representing industry (ABR and the Doe Run Resources Corp.)
stated that EPA did not distinguish air-related Pb from nonair Pb in indoor dust for the
primary Pb smelter subarea case study. One commenter states that this lack of separation
of air from nonair sources to indoor dust Pb overstates the risk reduction associated with
reductions in air Pb concentrations (as might be associated with alternative standard
levels).
Response: EPA agrees that we did not develop separate estimates of air-related and
nonair indoor dust Pb concentrations for this case study for reasons described in II.C.2.h
of the proposal. However, this has no effect on the magnitude of differences in total Pb-
related risk between alternative standards. The risk assessment first estimated total Pb-
related risk and then partitioned those estimates of Pb-related risk to the pathways of
interest in this review, with various limitations affecting how completely that partitioning
could be done (as described in the proposal section II.C.2.e). Thus, not separating
exposure pathways for air-related dust Pb from nonair-related dust Pb did not affect the
derivation of total Pb-related risk estimates or the differences in magnitude of these
estimates associated with different alternative standards.
35
-------
5) Comment: One commenter (City of El Paso, Texas) stated that the risk assessment
underestimated risk associated with the current NAAQS by underestimating the soil Pb
concentrations that can result over time. Specifically, they assert that even the soil Pb
levels reflected in modeling risk for the primary Pb smelter, for the remediation zone
closest to the facility, do not fully reflect the contamination in residential areas that would
be expected from this type of source. The commenter points out that the soil Pb levels
used in the risk assessment for the subarea near the smelter, which reflect remediation
activities, are significantly lower than levels in similar areas that have not undergone
remediation. Furthermore, they argue that maternal blood Pb levels in areas with
historically elevated ambient air Pb levels may be significantly higher than the central-
tendency levels used in the risk assessment for that contribution to childhood blood Pb.
Response: EPA agrees there is uncertainty in the soil Pb concentration estimates,
including those for the primary Pb smelter case study. As the commenter notes, the soil
Pb estimates for areas near the smelter were based on post-remediation measurements of
soil Pb contamination and EPA recognizes that this approach may produce an
underestimate of Pb exposure (for the soil ingestion pathway) for the current NAAQS
scenario. In response to the comment regarding higher maternal blood Pb levels, the
application of a GSD reflecting national-scale variation in child blood Pb levels is likely
to capture the relative contribution of these elevated maternal blood Pb levels to the
broader distribution of blood Pb levels. However, EPA acknowledges that it has not
specifically modeled the subpopulation of children whose mothers have blood Pb levels
above the average national maternal level.
6) Comment: A number of commenters (Bayview Hunters Point Community Advocates, MO
Coalition for the Environment, CARS, Sierra Club, NRDC) stated that EPA's risk
assessment did not consider the fact that minority and low income groups are
disproportionately exposed to Pb. In making this statement, one commenter (MO
Coalition), cited the NHANES findings for these subpopulations. Further, one
commenter (CARB) recommends using different maternal blood Pb levels specific to
African American and Hispanic women as inputs necessary for adequately estimating
childhood blood Pb levels for those subpopulations.
Response: EPA agrees with commenters that low-income subpopulations and some
minority subpopulations have higher blood Pb levels, as evidenced by the NHANES data
cited in the proposal. As noted in the proposal and in the risk assessment report, blood
Pb levels are influenced by a range of exposure pathways (nonair-, as well as, air-related)
and other factors related to human behavior, nutrition, and genetic factors. In planning
the risk assessment and in considering at-risk populations in the Risk Assessment, Staff
Paper, ANPR and proposal notice, EPA has recognized the disproportionately higher
blood Pb levels of low income and African-American children. These portions of the
U.S. population are considered to be represented in the risk assessment by the upper
percentile estimates of total blood Pb and related IQ loss. Rather than developing
separate blood Pb and IQ loss estimates for individual subpopulations of children such as
those identified by the commenters, EPA characterized this variability in the risk
36
-------
assessment by estimating blood Pb distributions using a geometric standard deviation
based on the U.S. population, based on advice from CASAC on the analysis plan and
draft assessments. While this precludes EPA from estimating blood Pb and IQ loss
specific to those subpopulations, this method does provide estimates of blood Pb and IQ
loss for the upper ends of the distribution for populations exposed to air concentrations
meeting the level of the current standard and alternative standards, which may, as the
commenters indicate, include notable representation by these subpopulations.
EPA additionally notes that the GSD approach employed in the risk assessment,
consistent with CASAC advice, was adopted in recognition of the limited data and tools
for developing subpopulation-specific estimates of blood Pb and IQ loss. While data are
available for some key inputs, such as maternal blood Pb levels for the subpopulations
noted by one of the commenters, data are quite limited to specify all needed exposure
related factors specifically for special subpopulations, including those identified by the
commenters.
Finally, EPA recognizes that the higher blood Pb levels identified in these subpopulations
are indicative of these subpopulations being sensitive subpopulations with regard to Pb
exposures overall. However, EPA notes this evidence does not indicate increased
sensitivity to air-related Pb alone. For example, due to the nonlinear dose-response
relationship of IQ loss with blood Pb, EPA concludes that it is that portion of this
population with lower blood Pb and lower exposure to nonair sources that would be
expected to be more sensitive to air-related Pb impacts on IQ per incremental unit
increase in blood Pb.
7) Comment: One industry commenter (ABR) asserted that air-to-blood ratios derived in the risk
assessment for the urban case study are over-stated due to underlying errors in the indoor
dust model used. Specifically, they contend that the higher air-to-blood ratios generated
at lower ambient air Pb levels for this case study result from non-linearities in the hybrid
indoor dust model (i.e., higher indoor dust Pb concentrations being estimated per unit
ambient air Pb at lower ambient air Pb levels) which they asserted are not supported by
science.
Response: The non-linearities reflected in the hybrid indoor dust Pb model, which
contribute to the higher air-to-blood Pb ratios at lower ambient air Pb levels, are based on
empirical data (as detailed below) and therefore are scientifically supported. Non-
linearity in the hybrid indoor dust model generates higher indoor dust concentrations per
unit ambient air Pb at lower air Pb levels, primarily because of non-linearity in the
conversion of indoor dust Pb loading to concentrations. EPA notes that the hybrid model
generates initial dust predictions in terms of loading and these need to be converted to
concentrations for use in blood Pb modeling. The procedure for converting estimated
indoor dust Pb loading to indoor dust Pb concentrations within the hybrid model involves
the use of regression equations derived from empirical data characterizing dust Pb levels
in residential housing in the U.S. (see section G.3.4.1 and Attachment G-l in Appendix G
of Volume II of the Risk Assessment Report). The regression models used in completing
the loading-to-concentration conversion are non-lineaer reflecting the underlying data
37
-------
used in their derivation. Therefore, the non-linearity in the hybrid indoor dust Pb model is
scientifically supportable since it reflects empirical data representative of residential dust
Pb characteristics in the U.S., including the relationship between indoor dust Pb loading
and concentration for different housing vintages.
8) Comment: Two commenters from industry (ABR and Doe Run Resources Corp.) asserted
that air-to-blood ratios derived in the risk assessment for the primary Pb smelter case
study subarea are overstated due to errors in characterizing indoor dust Pb exposure for
this case study. Specifically, both commenters argued that EPA did not separate out
nonair dust Pb from air-related indoor dust Pb. The commenters note that this precluded
EPA from deriving air-to-blood ratios based on recent air. Furthermore, one of the
commenters (the Doe Run Resources Corp.) suggested that a temporal mismatch in
relating indoor dust Pb measurements and ambient air Pb levels for the regression-based
model used in the primary Pb smelter case study also results in EPA overstating the air-
to-blood Pb ratios. Furthermore, they suggested that a more appropriate approach would
have been to match indoor dust Pb sampling data to ambient air Pb measurements from
the nearest ambient air Pb monitor (with the indoor dust sampling data also being
matched to the monitoring data in terms of the time period, or month). The commenter
contends that this regression model would have predicted lower indoor dust Pb levels for
a given ambient air Pb level (for the primary Pb smelter case study) than those estimated
using EPA's approach.
Response: The comment asserts that failure to differentiate air-related and background
sources of indoor dust Pb results in inflated air-to-blood ratios for the primary Pb smelter
(subarea) case study. EPA acknowledges that due to modeling and data limitations, we
were not able to parse out background sources of indoor dust Pb (e.g., indoor Pb paint)
from ambient air-related Pb impacts. EPA was clear in stating the uncertainty and
potential high-bias that this limitation meant for the policy-relevant risk results. However
EPA does not agree that this limitation in the indoor dust model resulted in biased air-to-
blood ratios. The category of air-to-blood Pb ratios given emphasis in the Staff Paper was
generated by comparing total blood Pb levels estimated for adjacent standard levels, i.e.,
comparing the absolute change in blood Pb level to the associated difference in ambient
air Pb levels between adjacent standard levels - see Section 5.2.5.2 of Volume I of the
Risk Assessment Report. Because these air-to-blood ratios do not depend on the fraction
of blood Pb that is considered air-related, but only on the difference in total predicted
blood Pb between standard levels, these ratios are not affected by limitations in
differentiating air-related indoor dust and associated exposure from background dust. The
only factor that would be important in impacting these ratios would be the performance
of the model in predicting levels of total indoor dust Pb under alternate standard levels.
Regarding the comment that there is a potential temporal mismatch in the ambient air Pb
and indoor dust Pb datasets used in deriving the regression-based indoor dust Pb model
for the primary Pb smelter case study, EPA acknowledges that some uncertainty is
introduced into the model due to difference in the timing of the two datasets. However,
the alternate strategy advanced by the commenter - regressing indoor dust Pb
measurements directly on ambient monitor data for the same month - is likely to severely
38
-------
under-represent any potential spatial relationship between indoor dust and ambient air Pb.
The use of monitoring data without any form of spatial interpolation as the basis for
characterizing the spatial gradient in ambient air Pb levels will result in a relatively
undifferentiated spatial surface, compared to the use of an air concentration surface based
on air dispersion modeling, as was done with the EPA's model. While EPA's regression
model may have uncertainty related to a temporal mismatch in the indoor dust and
ambient air Pb datasets, which EPA recognizes could introduce bias into the model, the
alternate approach of using monitoring data is subject to considerably greater uncertainty
and downward bias in the relationship between ambient air Pb and indoor dust Pb and
therefore, can not be considered as a preferred approach for developing a site-specific
indoor dust model for the primary Pb smelter case study.
9) Comment: One commenter (Kentucky Division of Air Quality), asserted that children's Pb
exposures from toys and jewelry containing Pb is significant and was not considered in
the risk assessment. Further, they state that EPA's limitations in separating blood and
risk estimates between air-related and non-air sources have contributed to EPA deriving
"risk attributed to ambient air Pb levels [that] is overly conservative and is not able to be
used in determine [sic] the appropriate NAAQS for Pb".
Response: EPA recognizes that in certain instances Pb associated with toys and jewelry
may contribute to significant Pb exposure. While the risk assessment did not explicitly
model risks associated with ambient air Pb for the subset of children experiencing
coincident exposure to high levels of Pb contained in toys and related items, the risk
assessment did provide general representation for these children to the extent that they
are represented in the set of children surveyed by NHANES IV. Specifically, the GSD
used in modeling the distribution of blood Pb levels, and hence IQ loss, within the study
populations included in the risk assessment was developed from the NHANES IV
dataset. Children with elevated blood Pb due to a variety of sources, including
contaminated toys, would be reflected in the GSD, therefore these children would also be
reflected broadly in the population exposure and risk distributions developed in the risk
assessment. However, EPA did not explicitly model the distribution of risk separately for
these subpopulations, due primarily to limitations in the exposure input data required to
support this level of refined modeling. EPA does acknowledge that in those instances
where a child has been exposed to elevated Pb levels through contaminated toys or
jewelry, such exposure may dominate overall Pb exposure and associated risk of IQ loss.
However, the frequency of this type of higher Pb exposure related to toys and jewelry
within the general child population is not currently known.
10) Comment: One commenter (California Air Resources Board) stated that the risk assessment
should have considered prenatal exposures and associated risk in modeling IQ loss for the
case studies.
Response: In modeling risk for child populations in the risk assessment, EPA included
consideration for prenatal exposure (i.e., maternal Pb contribution) as a component of
overall Pb uptake in determining lifetime average and concurrent blood Pb levels for the
7 year old child. Specifically, prenatal exposure is simulated within IEUBK as an initial
39
-------
contribution to child blood Pb. Additionally, while EPA recognizes the potential for
adverse health effects resulting from prenatal exposure to Pb, the health endpoint on
which we focused in the quantitative health risk assessment for this review is
developmental neurotoxicity in children, with IQ decrement as the risk metric. This
reflects EPA's consideration of CAS AC advice on the priority endpoint for the risk
assessment (as stated in Staff Paper section 4.1.2).
B. Secondary Pb Standard
The public comments received on the proposed secondary standard for Pb were very
general in nature basically expressing support for revising the current secondary to be identical to
the revised primary standard. Public comments on issues related to the secondary Pb standard
are described and addressed in section III.B.2 of the preamble to the final rule.
C. Comments Related to Data Handling (Appendix R)
1. Use of "standard conditions" Pb-TSP data collected prior to January 1, 2009 without
adjustment to represent "local conditions"
(1) Comment: Several commenter addressed EPA's proposal that in the future lead
concentration data be reported in terms of local temperature and pressure conditions, and
that data collected before January 1, 2009 and already reported on the basis of standard
temperature and pressure be comparable to the NAAQS without adjustment to local
conditions, with the state having the option to withdraw the data and make those
adjustments if it wishes. One commenter argued that Pb concentrations should continue,
as in the past, to be reported in terms of standard temperature and pressure conditions and
that only those values should be compared to the level of the NAAQS. In support of this
view, this commenter claimed generally that ambient air Pb concentrations used in
deriving relationships between air Pb concentrations and blood Pb levels were in terms of
standard temperature and pressure. Another commenter expressed a similar but less
specific concern about consistency between the conditions for reporting concentrations
and the logic used by the Administrator to set the level of the NAAQS. A commenter
gave the example of monitoring conducted at 0 degrees C (32 degrees F), for which the
concentration reported on the basis of local conditions would be 9 percent higher than if
reported on the basis of standard conditions.
Response: As stated in the proposal, EPA believes that a concentration expressed in
terms of local conditions of temperature and pressure is more closely related to the rate of
and total deposition of lead to the ground and other surfaces, because that concentration
is directly proportional to the amount of lead in the air within a fixed distance of the
surface onto which deposition occurs. Thus, concentrations expressed in terms of local
conditions are the better indicator of the risks which the NAAQS seeks to limit, and thus
comparing these concentrations to the NAAQS allows the NAAQS to protect public
health more effectively and more consistently across areas in the U.S., given differences
in local temperature and pressure. EPA believes that the disparity, if any, attributable to
the use of historical data that may have been based on standard conditions in air-to-blood
40
-------
relationships when selecting the level of the NAAQS is less than the 9 percent anecdote
cited by the commenter, because most data used to relate air concentrations and blood
lead levels were collected at temperature and pressures much closer to standard
conditions, so the adjustment factor would be less. EPA notes that the disparity, if any, is
small, particularly compared to the magnitude of the reduction in the level of the standard
and to other uncertainties which the Administrator has considered in selecting the
elements of the revised NAAQS. Although it is not possible for all the historical data to
be re-adjusted due to the lack of information on temperature and pressure conditions for
each measurement day in each study, EPA considered the potential for a disparity
between historical monitoring data reported at STP and the requirement to report data for
the revised NAAQS at local conditions in setting the standard.
2. Data Completeness Tests
(1) Comment: One commenter stated that the imputation procedures in the proposed
Appendix R are similar to those that states have been required to use for PM2 5. This
commenter believes that there is little to be concerned about, except in the highly unusual
cases where virtually every scheduled sample in a single calendar quarter is missing. This
commenter was concerned that the proposed imputation procedures may not satisfactorily
resolve a situation in which a large portion of the scheduled sampling fails to yield valid
data, because the proposed diagnostic tests for use in the case of missing data are quite
conservative. One test substitutes high estimates for the missing data to see if
nevertheless the mean concentration is below the NAAQS, and the other test substitutes
low estimates for the missing data to see if nevertheless the mean concentration is above
the NAAQS. The commenter is concerned that these tests may leave some situations
unresolved, and recommends that that there be room for an alternative treatment in such
cases.
Response'.The final rule addresses this concern by including the proposed provision
which allows data from periods that do not meet the completeness criteria to be
considered valid (and complete) with the approval of, or at the initiative of, the
Administrator, who may consider factors such as monitoring site closures/moves,
monitoring diligence, the consistency and levels of the valid concentration measurements
that are available, and nearby concentrations in determining whether to use such data.
(2) Comment: In the proposed rule, EPA invited comment on also incorporating into the
final rule two other possible tests that could allow a NAAQS exceedance determination
to be made on the basis of monthly data that is not at least 75% complete. The first test
would compare the monthly (or quarterly) mean to a fraction of the NAAQS level, for
example 50 percent, and consider the month to be exceedance-free if the (incomplete)
monthly mean were less than that fraction of the NAAQS. The other test would use a
statistically based confidence interval for the monthly (or quarterly) mean to test for a
high probability of either an exceedance or non-exceedance with the NAAQS for that
month. The commenter noted that the preamble to the proposed rule stated that EPA was
planning to analyze data sets using hypothetical data to explore these two possible
41
-------
approaches, and reserved comment on this issue until this analysis became available for
review.
Response: This issue is addressed in section IV.C of the preamble to the final rule.
(3) Comment A state environmental agency proposed, in the context of a second maximum
monthly mean concentration form for the NAAQS, that when in spite of everyone's best
efforts completeness requirements are not met for two months in a 3-year period, the next
highest monthly average of the remaining 33 months in the 3-year period should be used
as a valid design value.
Response: While the comment was made in the context of a different averaging
time/form than was selected in the final rule, the general sense of the recommendation
can also be taken to apply to the selected averaging time/form (a rolling 3-month mean
concentration) as a recommendation that the completeness criteria overlook a small
number of incomplete 3-month means and allow a finding of compliance with the
NAAQS if the highest valid 3-month mean is below the NAAQS. EPA recognizes that a
failure to meet completeness requirements can occur despite everyone's best efforts to
collect samples on the scheduled days or on allowed make up days. As a practical matter,
the inability to make a finding of attainment due to data incompleteness in the most
recent 3-year period could affect states and sources only where a site has recorded a
NAAQS violation in a previous 3-year period. If there has been no previous violation,
there is no practical effect from the lack of a compliance finding for the 3-year period that
does not meet completeness requirements. In a situation in which there has been a prior
violation, EPA believes it would be inappropriate to overlook incompleteness in order to
make a finding of attainment. Instead, no new finding should be made until complete
data for a 3-year period does show attainment.
3. Criteria and Formula for site-specific scaling factors
(1) Comment: Comments on the subject of scaling factors to relate Pb-PMio measurements
to Pb-TSP concentrations were generally negative towards EPA's proposed requirement
for the development of site-specific scaling factors. Also relevant to this issue are the
comments regarding whether Pb-TSP or Pb-PMio should be the indicator for the
NAAQS. Many commenters argued that Pb-PMio should be selected as the indicator for
the NAAQS only if the level of the NAAQS were set at or below the low end of the range
proposed by EPA, i.e., at or below 0.10 |ig/m3.
Response: These comments are summarized in more detail and responded to in section
IV.D of the preamble to the final rule.
(2) Comment: A monitoring agency submitted an analysis of data from a monitoring site
near Detroit at which two Pb-TSP monitors and two high-volume Pb-PMio monitors
operate simultaneously. The data analysis concluded that the precision levels of these
monitoring methods were not greatly different, and that the ratio of concentrations of Pb-
TSP and Pb-PMio was highly variable.
42
-------
Response: These data were part of the data examined by EPA in preparing the proposed
and final rules. The findings stated by the commenter are consistent with EPA's findings
based on data from this and other collocation sites.
(3) Comment: A monitoring agency submitted a brief analysis of collocated Pb-TSP and Pb-
PMio data from a monitoring site near a secondary Pb smelter in Tampa, FL, for the
period of January through June 2008. This data set included two Pb-TSP samplers and
one Pb-PMio sampler. The analysis indicated a mean Pb-TSP concentration during this
period of 0.4 micrograms per cubic meter and a mean Pb-PMi0 of 0.2 micrograms per
cubic meter. This factor of two difference was consistent with the slope of a regression
between daily measurements of the two parameters. The analysis also showed close
agreement between the duplicate measurements of Pb-TSP.
Response: These data appear not to have been submitted to AQS at the time these
comments were reviewed for this document. The close agreement between duplicate
measurements of Pb-TSP is consistent with EPA's stated opinion at the time of the
proposal that the Pb-TSP sampler is reasonably precise and is suitable, but not ideal, for
purposes of implementing the NAAQS. The factor of two difference between the two
sizes of Pb is about the same as the highest factor in the data examined by EPA when
developing the proposal, and confirms EPA's conclusion that monitoring for Pb-PMio
near industrial sources of Pb may result in measurements that do not indicate the actual
concentration of Pb-TSP, the indicator selected in the final rule. EPA notes that the
graph submitted by the commenter appears to be mislabeled as a comparison of Pb-PMi0
and Pb-TSP concentrations, when it actually is a comparison between the two Pb-TSP
measurements.
4. Criteria for exclusion of data from comparison to the NAAQS based on the influence
of an exceptional event
(1) Comment: One commenter addressed the topic of excluding Pb concentration data that
has been affected by an exceptional event. In the commenter's view, the proposal did not
address what an exceptional event would be for lead, in particular for resuspension of
lead from lead-contaminated soils. The commenter argued that local construction
activities or agitation of barren lead-contaminated soils by vigorous winds could lead to
significant re-suspension of lead particles that in turn could adversely impact local
monitors. A state air quality agency would have little control over these activities, and for
local construction, no notice, yet these conditions could significantly contribute to a
violation. The commenters said that the final rule should set forth criteria for exceptional
events unique to lead.
Response: 40 CFR 50.14, Treatment of air quality monitoring data influenced by
exceptional events, contains general criteria for determining whether an event is
exceptional and whether data may be excluded when making comparisons to the
NAAQS. EPA believes that the provisions of 50 CFR 50.14, and the illustrative
examples in the preamble to the final rule explaining that section, are suitable for
43
-------
application to lead and that no special critiera or further explication are needed or
appropriate. EPA notes that local construction activities or agitation of barren soils by
vigorous winds can also lead to re-suspension of PM2.5 and PM10 and cause
exceedances of those NAAQS, so these possibilities do not make the lead NAAQS
uniquely different.
D. Comments Related to Monitoring
1. Existing Sampling and Analysis Methods
(1) Comment: We received a number of comments on our proposal to continue the use of high-
volume TSP samplers as the sampling method for Pb. In their comments on the
proposed rule, CAS AC reiterated their concerns over the measurement uncertainty due to
effects of wind speed and wind direction on sampling efficiency. These concerns were
discussed in detail in our proposed rule, and as such are not reiterated here. However,
CAS AC stated that if the final level of the NAAQS were to be set at 0.10 |ig/m3 or above,
then the high-volume Pb-TSP sampler should be used. Some public commenters also
stated similar concerns with the performance of the Pb-TSP sampler. A large number of
other commenters stated that the high-volume TSP sampler should continue to be the
sampler for determining compliance with the Pb NAAQS. They expressed concerns that
PMio samplers would not capture ultra-coarse particles (i.e., particulate matter with an
aerodynamic diameter greater than 10 jam), and could greatly underestimate Pb
concentrations in the ambient air, especially near Pb sources.
Response: This issue is addressed in V.A.I the preamble to the final rule.
(2) Comment: We received several comments supporting the need for the development of a low-
volume Pb-TSP sampler. However, in our consultation with CAS AC's AAMM
Subcommittee, we were cautioned against finalizing a new low-volume Pb-TSP FRM
without an adequate characterization of the sampler's performance over a wide range of
particle sizes.
Response: We agree with the interest for a low-volume Pb-TSP sampler and the desire for
such a sampler to be adequately characterized prior to being promulgated as a new FRM.
Accordingly, we plan to further investigate the possibility of developing a low-volume
FRM in the future.
(3) Comment: One commenter suggested allowing the use of low-volume Pb-TSP samplers at
special purpose monitoring sites (SPM). The data could be used to determine the need
for an official SLAMS site and would provide data for the possible approval of new
FRM/FEM methods.
Response: We agree with the commenter, and point out that monitoring agencies may use
non-FRM/FEM methods at non-required sites. However, we note that these data from
these non-FRM/FEM monitors can not be compared to the NAAQS.
44
-------
2. Proposed Pb-PMi 0 Federal Reference Method
(1) Comment: The CAS AC AAMM Subcommittee agreed with our proposed use of the PMi0c
sampler. Other comments on our proposed use of the low-volume PMioC sampler for the
Pb-PMio FRM were in support of the PMi0c as an appropriate sampler for the FRM.
Response: As discussed in the preamble in section V.A.2, Pb-PMlO monitoring is
permitted in some limited situations that meet certain specified conditions. For those
situations, we are promulgating the Pb-PMi0 FRM based on the use of the low-volume
oc sampler.
(2) Comment: We received comments on our proposed use of XRF as the analysis method for the
Pb-PMio FRM, including comments from CASAC's AAMM Subcommittee during the
peer review of the proposed FRM. Several commenters agreed with our proposed use of
XRF as the analysis method, citing several of the advantages we identified in the
preamble to the proposed rule. However, several other commenters suggested that
Inductively-Coupled Plasma-Mass Spectrometry (ICP-MS) would be a more appropriate
analysis method for the FRM.
The AAMM Subcommittee and other commenters raised concerns with the potential for
measurement bias due to non-uniform filter loadings. They noted that the analysis beam
of the XRF analyzer does not cover the entire filter collection area; therefore, it is
possible for the measurement to be biased if the Pb particles deposit more (or less) on the
edge of the filter as compared to the center of the filter.
Response: This issue is addressed in the preamble to the final rule.
(3) Comment: Several commenters suggested that Inductively-Coupled Plasma-Mass
Spectrometry (ICP-MS) would be a more appropriate analysis method for the FRM.
Advantages identified with ICP-MS included the analysis of the entire filter deposit and a
higher sensitivity (i.e., lower MDL.) In addition, a number monitoring agencies noted
that their laboratories were already equipped for ICP-MS making ICP-MS less costly
than XRF for them.
Response: This issue is addressed in section V.A.2 of the preamble to the final rule. We
expect that other analysis methods, including ICP-MS are likely to be approved as FEM
to compliment the PMio-FRM. Following approval, monitoring agencies will have the
option of choosing between XRF, or other approved FEMs.
(4) Comment: One commenter suggested that high-volume PMio samplers should be considered
for the FRM or FEM because the method collects a larger sample which should result in a
higher precision, allow for split analyses, and allows for options for repeat analysis.
Response: We do not plan to accept high-volume Pb-PMio samplers as either an FRM or
FEM. While we agree that high-volume Pb-PMio samplers do collect a larger sample
that would lead to lower detection limits, they have a less precise cut point which may be
45
-------
affected by wind-speed. High-volume samplers also have less precise flow-control
capabilities than low-volume samplers and most current high-volume samplers do not
possess the ability to actively control flow to actual temperature and pressure conditions.
(5) Comment: One commenter suggested that XRF may not be an adequate analysis method due
to poor detection capabilities. The commenter indicates that XRF is used as a "screening
method" for analysis of metals in solids.
Response: We are aware of XRF instruments that are used for screening level analyses,
however, these instruments would not meet the FRM analysis method description. The
XRF analysis method defined in the FRM is for a research grade laboratory analysis
method that is capable of precisely measuring Pb at levels well below the level of the
final Pb NAAQS.
(6) Comment: One commenter suggested that detailed guidance was needed on how to perform
XRF analysis.
Response: We believe the final FRM provides adequate details concerning the details of
how the analysis is to be completed. Therefore, we believe that further detailed guidance
is not required at this point.
3. FEM criteria
(1) Comment: One commenter suggested that the proposed MDL requirement, 1 percent of
the NAAQS, was overly stringent, and that an MDL of 5 percent would be sufficient.
Another commenter suggested that an MDL at 10 percent would be more achievable.
Response: This issue is addressed in the preamble to the final rule. See section V.A.3.
(2) Comment: We received two comments supporting the development and consideration of
the use of continuous Pb monitors.
Response: We agree that the FEM criteria should allow for the development of
continuous Pb monitors as FEM. As such, we have revised the FEM criteria to
accommodate the potential of continuous Pb FEM.
4. Quality Assurance
(1) Comment: We received one comment on the proposed QA requirements specifically
addressing the overall sampling and analysis bias. The commenter was concerned that
the proposal to implement one independent performance evaluation audit (similar to the
PM2.5 Performance Evaluation Program (PEP)) and then augment that sample with four
samples from collocated precision site would be inadequate. The commenter suggested
that in order for the audit program to be successful it would require the same independent
laboratory be used by all monitoring agencies across the country.
46
-------
Response: This issue is addressed in the preamble to the final rule. See section V.A.4.
5. Adequacy of Existing Network Design Requirements
(1) Comment: We received numerous comments that the existing monitoring network and
network design requirements were not adequate for a lowered NAAQS. We did not
receive any comments stating that the existing network was adequate.
Response: We agree with the comments that the existing network design requirements
need to be updated. We are finalizing new network design requirements as described in
the preamble to the final rule, see section V.B.
6. Source-oriented Monitoring Requirement
(1) Comment: We received several comments supporting the need for monitoring near Pb
sources. Alternatively, one commenter suggested that near source monitoring is not
necessary because "the EPA and the State already know where and what the problems
are" and "EPA should ... develop control standards to deal with the problem ..."
Response: We note that under CAA sections 107 and 110 states must submit State
Implementation Plans (SIPs) that include measures to attain the NAAQS as expeditiously
as possible. These measures can require control of emissions at sources contributing to
an exceedence of the NAAQS. Thus, monitoring near Pb sources is needed to ensure
controls of Pb sources contributing to violations of the NAAQS.
(2) Comment: We received a comment that the methods used in developing the emission
thresholds estimated ambient impacts over different averaging periods, and that the
emission thresholds should be recalculated for all methods using the final averaging
period. We recognized this issue in our memorandum documenting the analysis, and we
have recalculated the estimate of the lowest Pb emission rate that under reasonable worst-
case conditions could lead to Pb concentrations exceeding the NAAQS, based on the final
level and form of the standard.
We also received comments on the approach used in developing the proposed emission
thresholds that would trigger consideration of the placement of a monitoring site near a
Pb source. Commenters expressed concerns that the approach overestimated the potential
impact of Pb sources, and would result in either unnecessary burden on monitoring
agencies or worse yet, monitoring agencies would install and operate monitors at sources
that had little to no potential to exceed the NAAQS. Several commenters suggested
various alternative levels, including a threshold of 1 ton or higher, basing their
recommendations on concerns such as the reliability of data in the NEI. Other
47
-------
commenters suggested that EPA was in the best position to determine which sources had
the potential to exceed the NAAQS.
Response: This issue is addressed in the preamble to the final rule, see section V.B.3.
(3) Comment: We received several comments supporting the need for monitoring waivers,
and one comment that did not support waivers. Those in favor of the waivers pointed out
that, as discussed above, many Pb sources will result in much lower Pb impacts than the
"worst case" Pb source. They argued that the states need flexibility in meeting the
source-oriented monitoring requirements, and agreed that it is appropriate to focus on
sites near those Pb sources with the greater potential to result in Pb concentrations that
exceed the Pb NAAQS. The commenter who cautioned against the allowance of
monitoring waivers expressed concerns that modeling techniques have uncertainty that
could result in waivers being granted when actual Pb concentrations could exceed the
NAAQS.
Response: We agree that it is appropriate to allow for monitoring waivers as proposed,
see section V.B.3. We took the uncertainty of modeled data into account when proposing
to limit waivers to situations where the modeled data indicated maximum impacts would
be 50 percent of the NAAQS, rather than at 100 percent of the NAAQS, and we believe
this provides an appropriate margin of safety.
(4) Comment: We received comments questioning the need to restrict the provision of
waivers to sites near sources emitting less than 1000 kg/yr when it is possible for sources
impact to be well below the level of the Pb NAAQS.
Response: We agree it is possible for sources greater than 1000 kg/yr to have an impact
less than 50 percent of the NAAQS under certain conditions. We also acknowledge the
need for flexibility in implementing the Pb NAAQS monitoring network. As such, we
have reconsidered our proposed restriction limiting waivers to those for sources emitting
less than 1000 kg/yr, and we are not finalizing a restriction on the size of sources near
sites eligible for a waiver from the source-oriented monitoring requirement.
(5) Comment: We received comments on relying on the NEI to identify Pb sources with
emissions greater than the emission threshold. In general, several commenters said better
data should be used to identify Pb sources emitting above the emission threshold.
Several commenters expressed concerns with the accuracy of the National Emission
Inventory (NEI), and recommended allowing states to use "the best available
information" on emissions from Pb sources. Some commenters pointed to differences in
Pb emissions data reported in the Toxics Release Inventory and the NEI as evidence that
the NEI was inaccurate. One commenter said current practices to reduce toxic emissions
are not reflected in the NEI and wanted the opportunity to update the information.
Commenters said EPA should correct the errors in the NEI or allow states to submit
revised local data that more accurately reflect Pb emissions before emissions inventory
data are used to determine which sources exceed the threshold.
48
-------
Response: This issue is addressed in the preamble to the final rule, see section V.B.3.
(6) Comment: We received comments that the proposed source-oriented monitoring
requirements did not address situations where multiple sources contribute to Pb
concentrations at one location.
Response: This issue is addressed in the preamble to the final rule, see section V.B.3.
(7) Comment: We received two comments that source-oriented monitors should be located at
the maximum estimated Pb concentration without consideration for the potential for
population exposure, and six comments that source-oriented monitors should be located
in an area where population exposure occurs. In their comments on the proposed rule,
one commenter argued that monitors "should be located in or around only those Pb point
sources with a nearby population base" because "air Pb concentrations have regulatory
importance largely in those areas where significant groups of children are exposed for
considerable time periods." They argue that as an example "a rural road going by a lead
mining facility is an unlikely place that children will spend considerable amounts of
time" and as such "placing a monitoring site on such a road would have de minimis, if
any, value." Another commented that "monitors should be located near playgrounds,
sports fields, long-established highways, and the like."
Response: This issue is addressed in the preamble to the final rule, see section V.B.3.
(8) Comment: We received a comment that in our calculation of the emission threshold we
should have used the maximum of the four methods rather than the average of the four
methods in determining the emission threshold.
Response: The threshold used in the final regulation is not derived directly from the
emissions calculations, for further discussion see the preamble.
(9) Comment: We received a number of comments that the emission threshold was not
realistic and was based on worst case assumptions. One commenter suggested the
threshold should be reconsidered based on modeling performed for the risk assessment.
Response: The emission threshold is intended to represent a reasonable worst case
impact from an emission source. States are allowed, and encouraged, to request a waiver
of the monitoring requirement based on more refined modeling which includes site
specific characteristics such as stack heights, temperatures, etc.
(10) Comment: We received a comment that we should not require monitoring of Pb impacts
due to combustion sources for several reasons including inadequate inventories, an
inability of States to control such sources through the SIP process, and because this issue
would be similar across many areas of the country. The commenter suggested that EPA
conduct Pb monitoring at selected sites to determine if it was likely that these sources
would result in violations of the NAAQS.
49
-------
Response: While combustion of traditional fuels was identified as a potentially
significant source of Pb emissions nationwide, we expect that states generally would be
able to meet the requirements to receive a waiver for most of the combustion sources
because these types of facilities typically have tall stacks, high exhaust velocities, and hot
exhausts, which would result in significant dispersion and reduced ambient impacts.
(11) Comment: A number of commenters suggested that the emphasis on source-oriented
monitoring was a departure from network designs for other NAAQS pollutants such as
ozone and PM2 5.
Response: One of the primary objectives for all NAAQS monitoring networks is to
identify locations of maximum concentrations as well other objectives as described in the
proposal. As such the EPA is finalizing both source and nonsource-oriented monitoring
requirements as preamble.
(12) Comment: One commenter suggested that EPA develop different emission thresholds for
a number of different source categories (e.g., airports and combustion sources).
Response: We do not have sufficient data to develop different emission thresholds for
different source categories at this time. However, we plan to work with monitoring
agencies to develop information on a number of different source categories such as
airports that will be useful in estimating the impact on Pb concentrations from these
sources.
(13) Comment: One commenter suggested that the EPA needs to develop a "regulatory off
ramp" for when source-oriented monitors can be shut down.
Response: Once a monitor has collected 3-years worth of data, the data can be used to
request a waiver of the source-oriented monitoring requirement if the data shows that
concentrations were less than 50% of the NAAQS.
(14) Comment: One commenter suggested that if a facility increased its operations, that only
the increased emissions should be modeled, and the highest concentration from the
existing monitoring data could be used to predict the impact on Pb concentrations.
Response: We believe that the most recent emissions data available for a facility should
be used in a scientifically justifiable means (e.g., dispersion modeling) to determine if Pb
monitoring is required. If a facility increased its operations resulting in increased
emissions above emissions data contained in the NEI, for example, the highest level of
emissions should be evaluated as part of the monitoring demonstration.
(15) Comment: One commenter suggested that certain types of Pb sources (including
smelters, battery manufacturers, and mines with tailings piles) should not be allowed to
receive a monitoring waiver.
50
-------
Response: We agree that these source types should be given additional scrutiny by
monitoring agencies, but believe that under certain circumstances it is possible that these
source types may afee have minimal impact on Pb concentrations. As such, we do not
believe it is appropriate to categorically disqualify these source types from being capable
of receiving a monitoring waiver.
(16) Comment: One commenter suggested that the EPA should consider presumptively not
requiring monitoring for facilities in source categories that already have extensive lead
controls in place.
Response: The presence or lack of controls does not necessarily affect the impact of Pb
emissions on ambient Pb concentrations. That is to say, emissions of 1.0 tons per year
from a well-controlled facility would likely have a similar impact on Pb concentrations as
1.0 tons per year of emissions from an uncontrolled facility.
(17) Comment: One commenter suggested it can be very difficult, and in some places
impossible, to find suitable locations for monitoring stations and to gain access to the
areas.
Response: Our final monitoring requirements allow for the consideration of logistics
when siting a required Pb monitor.
(18) Comment: One commenter suggested that EPA needs to provide guidance on probe
heights.
Response: Probe height requirements are clearly defined in the existing 40 CFR part 58,
Appendix E requirements.
7. Nonsource-oriented Monitoring Requirement
(1) Comment: One state and several tribes commented that the proposed population limit
would result in no required non-source oriented monitors in low population states and
tribal lands. One commenter expressed concerns that the population limit was too high,
and would result in environmental justice concerns since many poor communities would
not be monitored.
Response: This issue is addressed in the preamble to the final rule, see section V.B.3.
(2) Comment: One commenter suggested that NCore multi-pollutant sites would be an
appropriate location for nonsource-oriented sites.
Response: We agree that a few of the planned NCore sites that are located in CBS As of
500,000 people or more could be appropriate locations for nonsource-oriented sites
especially when such sites represent areas near schools, parks, or other areas where
children may live and play. However, the final Pb network design states that non-source
51
-------
Pb monitors should be located to estimate neighborhood scale Pb concentrations in urban
areas impacted by re-entrained dust from roadways, closed industrial sources which
previously were significant sources of Pb, hazardous waste sites, construction and
demolition projects, and other fugitive dust sources of Pb. The NCore network design
discourages placement of stations near sources that are not representative of relative wide
areas, so relatively few of the NCore stations are likely to be ideal non-source Pb
monitoring locations.
(3) Comment: One commenter suggested that the nonsource-oriented threshold of 1,000,000
was too high and would result in areas where 60% of children live in poverty with no
monitors.
Response: As discussed in the staff paper, based on the analysis of the existing
monitoring network, we do not expect areas away from Pb sources to show Pb
concentrations in excess of the NAAQS. However, we have decreased the population
threshold for requiring non-source Pb monitors from CBS As of 1,000,000 people to
CBSAs of 500,000 people, bringing in approximately 50 more CBSAs into the
requirement. Furthermore, we are requiring Pb monitors near Pb sources regardless of the
size of the community, or its economic status.
8. Monitoring Near Roadways
(1) Comment: The majority of commenters agreed with our finding that the available data on
Pb concentrations near roadways do not indicate the potential for exceedances of the
proposed range of Pb NAAQS levels and requirements for monitors near roadways were
not needed to ensure compliance with the NAAQS. However, one commenter argued
that our finding that activity on roadways would not likely contribute to air Pb
concentrations in exceedence of the proposed levels for the standard was based on data
from monitors that did not represent the maximum impact from roadways.
Response: This issue is addressed in the preamble to the final rule, see section V.B.3.
9. Use ofPb-PMio Monitors in lieu ofPb-TSPMonitors
(1) Comment: Several commenters suggested an approach for the use of Pb-PMio monitors
as an alternative to the proposed use of scaling factors. As suggested by the commenters,
Pb-PMio monitoring would be allowed in certain instances. Specifically, Pb-PMio
monitoring would be allowed where estimated Pb concentrations were predicted to be
less than 50 percent of the NAAQS and where Pb in ultra-coarse particulate was expected
to be low. Again as suggested by the commenter, if at some point in the future the
monitor were to show that Pb-PMio concentrations exceeded 50 percent of the NAAQS,
the monitoring agency would be required to replace the Pb-PMio monitor with a Pb-TSP
monitor.
Response: This issue is addressed in the preamble to the final rule, see section V.B.2.d.
52
-------
(2) Comment: One commenter raised concerns that a PMio sampler may collect PM greater
than 10 microns in diameter and may result in an over estimate of PMi0.
Response: We have chosen Pb-TSP as the indicator for the Pb NAAQS. As such, this
concern does not apply to this rule making. While Pb-PMio is being allowed as a
surrogate for Pb-TSP in certain circumstances, any potential over sampling of ultra-
coarse particulate would just move the estimated concentration closer to the actual Pb-
TSP level that would have been measured by a Pb-TSP sampler had one already in
operation.
10. Required Timeline for Monitor Installation and Operation
(1) Comment: We received several comments from monitoring agencies regarding the
proposed timeline for monitor installation. Commenters supported the need for a
staggered network deployment, especially if a large number of new monitors would be
required. Two commenters argued that even the proposed two-year deployment would
not provide enough time for monitoring agencies to site and install the number of
monitors needed.
Response: In response to these comments, EPA is permitting a tiered network
deployment process over two years. This issue is addressed in the preamble to the final
rule, see section V.B.3. We believe that two years is sufficient to site and deploy the
approximately 235 new monitors that will be required by this rule.
(2) Comment: We received comments from a number of states expressing concerns that the
number of required monitors would fluctuate year to year due to changes in the actual Pb
emissions inventory.
Response: We have tied the monitoring requirement to the Pb emissions estimates in the
"most recent" NEI. Since the NEI is updated once every 3 years, this would mean that
States would need to reassess the Pb monitoring requirement once every 3 years. We
also expect States to perform network assessments every 5-years as required by 58.10(d).
Such assessments should include an evaluation of Pb network changes that may be
required due to changes in the NEI or the availability of other information that supports
re-evaluation of potential source impacts on ambient Pb concentrations.
11. Sampling Frequency
(1) Comment: We did not receive any comments on our proposed sampling frequency for a
NAAQS based on a quarterly average. We did receive 4 comments supporting a move to
a 1 in 3 day sampling frequency if the final Pb NAAQS was based on a monthly
averaging time, 3 that supported maintaining the 1 in 6 day sampling frequency despite a
move to a monthly averaging time, and 2 comments that supported daily sampling if the
final Pb NAAQS was based on a monthly averaging time.
53
-------
Response: The final NAAQS is based on a rolling 3-month average. The statistical and
practical monitoring considerations are the same for a 3-month average as with a calendar
quarterly average. As such, we are maintaining the current l-in-6 day minimum sampling
frequency as proposed.
(2) Comment: Two commenters supported allowing for less frequent sampling if the
measured Pb concentrations were less than 50% of the standard.
Response: We are maintaining the current l-in-6 day sampling frequency. Since we are
not moving to a 1 in 3 day sampling frequency we do not believe it is necessary, or
appropriate to allow for a reduction from the 1 in 6 day sampling schedule.
12. Monitoring for the Secondary Standard
(1) Comment: We received one comment supporting our proposed reliance on the
IMPROVE network Pb-PM2 5 data for tracking trends in Pb concentrations in rural areas.
We did not receive any other comments on additional monitoring needs to support the
secondary Pb NAAQS.
Response: We are not finalizing any additional requirements for Pb monitoring
specifically for the secondary Pb NAAQS.
13. Cost of Monitoring Network and Funding Issues
(1) Comment: We received numerous comments that raised concerns with the cost of the
proposed Pb monitoring network, and need for flexibility in implementing the monitoring
network. Several monitoring agencies questioned where the money for the new Pb
requirements would come from, and suggested that EPA fund the monitoring network
either partially or fully through 103 monitoring grants.
Response: While the CAA prohibits us from considering costs when setting the Pb
NAAQS, we may consider costs when establishing implementing requirements such as
monitoring requirements. As such, we have attempted to provide substantial flexibility to
monitoring agencies responsible for implementing the monitoring requirements, while
still ensuring that the monitoring network will be sufficient to determine compliance with
the final Pb NAAQS. Nonetheless, the final monitoring requirements will result in an
increase in monitoring expenses to some monitoring agencies. We will work with the
state, local, and tribal monitoring agencies to appropriately address these issues.
54
-------
III. RESPONSES TO COMMENTS RELATED TO IMPLEMENTATION
Many comments on implementation issues are addressed in section VI of the preamble.
Significant comments on specific issues not addressed in the preamble are addressed in this
section.
A. Nonattainment Area Boundaries
(1) Comment: Several commenters state that they support the use of the county, or smaller,
boundaries for nonattainment areas dominated by single large emitting sources. These
commenters state that they support this position on the assumption that EPA will have
sufficient monitoring and modeling data to determine with confidence that areas outside
of the nonattainment boundary do not, in fact, have ambient lead concentrations
exceeding the revised lead NAAQS. The commenter recommends that EPA adopts the
alternative boundary presumption, the Metropolitan Statistical Area (MSA), for urban
areas. Such areas likely have a combination of factors contributing to their nonattainment
status, and the larger area should allow for a more holistic SIP.
Response: The EPA agrees with the commenter that lead emissions do not generally
transport over long distances (as compared, e.g., to fine paniculate matter). In the
proposed rule, EPA proposed to presumptively define the boundary for designating a
nonattainment area as the perimeter of the county associated with the air quality
monitor(s) which records a violation of the standard. In the proposed rule, EPA also
stated that, at the revised level of the standard, EPA expects stationary sources to be the
primary contributor to violations of the NAAQS, although we also believe that nearby
area sources may also contribute to concentrations of lead emissions that may affect a
violating monitor. In light of the possibility that a number of smaller sources may
collectively contribute to concentrations in excess of the NAAQS, EPA believes that
adopting the county boundary as the presumptive boundaries for lead nonattainment areas
is appropriate. However, as stated in the proposed rule, a state, Tribe, or EPA may
conduct additional area-specific analyses that could lead to the boundary for an area
either being increased or decreased from the presumptive county boundary. In situations
where a single source, rather than multiple sources, is causing a NAAQS violation, the
EPA believes that a state may well be able to use area-specific analyses to identify
whether a nonattainment area that is smaller than the county boundary is appropriate.
(2) Comment: The commenter states that they have serious concerns over EPA's use of
"presumptive boundaries" for nonattainment purposes. The preamble to the lead
NAAQS proposal notes that the EPA is proposing county boundaries as the presumptive
boundaries for lead nonattainment areas and is taking comment on the use of the MSA
boundaries. The commenter further states that they have asserted in the past and
continues to insist that the OMB has defined the metropolitan areas for statistical
purposes to include the collection, tabulation, and publication of data by federal agencies
for geographic areas to facilitate the uniform use of comparability of data on a national
scale. The commenter further states that for EPA to default to a presumptive boundary for
"consistency" purposes stifles the creatively to improve air quality as expeditiously as
possible to bring clean air to the public and rewards those who chose to wait. EPA's
55
-------
broad brush approach discourages initiatives by local areas, counties, and states to be
proactive.
Response: As stated previously, the EPA believes that lead emissions do not generally
transport over long distances (as compared, e.g., to fine paniculate matter). In the
proposed rule, EPA proposed to presumptively define the boundary for designating a
nonattainment area as the perimeter of the county associated with the air quality
monitor(s) which records a violation of the standard. EPA solicited comment on the use
of MSAs as presumptive boundaries, but is not adopting that approach in the final rule.
In the proposed rule, EPA also stated that, at the revised level of the standard, EPA
expects stationary sources to be the primary contributor to violations of the NAAQS,
although we also believe that nearby area sources may also contribute to concentrations
of lead emissions that may affect a violating monitor. In light of the possibility that a
number of smaller sources may collectively contribute to concentrations in excess of the
NAAQS, EPA believes that adopting the county boundary as the presumptive boundaries,
and the starting point for setting the boundaries for an area, for lead nonattainment areas
is appropriate. However, as stated in the proposed rule, a state, tribe, or EPA may
conduct additional area-specific analyses that could lead to the boundary for an area
either being increased or decreased from the presumptive county boundary. In situations
where a single source, rather than multiple sources, is causing a NAAQS violation, the
EPA believes that a state may well be able to use area-specific analyses to identify
whether a nonattainment area that is smaller than the county boundary is appropriate. The
commenter further states that for EPA to use a presumptive boundary for "consistency"
purposes stifles the creatively to improve air quality as expeditiously as possible to bring
clean air to the public and rewards those who chose to wait. To the contrary, EPA
believes that by adopting a presumptive boundary, and by providing states with the
opportunity to submit recommendations based upon qualitative information, which
substantiates the deviation from the presumptive boundary provides flexibility in terms of
setting the correct boundaries for the areas and helps to target the appropriate sources to
control.
(3) Comment: The commenter states that they oppose EPA's suggestion that it can use the MSA
as the presumptive boundary for the designation of lead nonattainment areas. The
commenter states that the use of the MSA would mean that a single violating monitor
results in the entire MSA being designated as being nonattainment unless a regulated
entity or other party establishes that an exception is warranted. Because lead
concentrations, unlike ozone or PM-2.5, are effectively the result of direct emissions, the
long range transport concerns of ozone and PM-2.5 are not present. Thus, rather than
"presuming" that contribution to the lead problem is regional in nature, the EPA should
restrict the area to no larger than the county boundary or defer to the recommendations
made by the state or Tribal government.
Response: The EPA agrees with the commenter that using the MSA as the presumptive
boundary for designating areas for the Pb NAAQS would be inappropriate due to the fact
that lead emissions are usually deposited within short ranges of the initiating sources.
Therefore, EPA is finalizing the proposed option of using the county as the presumptive
56
-------
boundary for designating areas, and provides for the use of several factors to make
determinations where deviating from the presumptive boundary is necessary.
(4) Comment: Several commenters stated that they believe that nonattainment boundaries will
need to be developed using specific data. As the preamble discusses, lead will be
associated with stationary emission sources, and thus where nonattainment area
boundaries need to be established, they will be inherently associated with these sources,
not political boundaries or consensus based boundaries. The commenters agree with this
premise, stating that in their experience with lead nonattainment, the issue was source
oriented in nature. The commenters narrowed down the nonattainment boundary to the
city blocks that were affected through use of modeling and monitoring. The commenter
states that based upon their experience with the two options provided in the proposed
rule, they support using the county boundary as the presumptive boundary for designating
areas, with the use of the 8 factors to support the changing of the boundaries. The
commenter further states that they do not support the use of MSA as the presumptive
boundaries for designation.
Response: As stated previously, at the revised level of the standard, EPA expects
stationary sources to be the primary contributor to violations of the Pb NAAQS, although
we also expect that in some areas a number of smaller sources may collectively
contribute to concentrations in excess of the NAAQS. MSAs are frequently composed of
a number of counties. Recognizing that lead emissions, particularly ultracoarse particles,
deposit relatively short distances from the proximity of their initial source, EPA believes
that adopting the county boundary as the presumptive boundary for lead nonattainment
areas is more appropriate than using the much larger MSA boundary. Furthermore, as
stated in the proposed rule (and the previous response), a state, Tribe, or EPA may
conduct additional area-specific analyses that could lead to the boundary for an area
either being increased or decreased from the presumptive boundary.
(5) Comment: The commenter states that in regard to designations, EPA believes that a key
factor in establishing lead boundaries is "to include both the area judged to be violating
the standard as well as the source areas that are determined to be contributing to these
violations." The commenter states that EPA must, however, define "how" it will
establish which sources are contributing to monitored lead nonattainment. While
modeling is often used for this assessment, EPA has not yet established a modeling
significance level for lead. Absent lead significance levels, use of modeling to determine
lead nonattainment areas is not a viable option.
Response: As stated in the preamble to the final rule, EPA believes that the county
boundary should be used as the presumptive boundary for designating areas for the Pb
NAAQS. In cases where there is a determination that areas should be either increased or
decreased from the presumptive boundary, the EPA is finalizing factors that should be
used to substantiate any deviation from the presumptive boundary. EPA will, as
appropriate, review and revise guidance, regulation, or policy related to modeling and
designations following the promulgation of the NAAQS.
57
-------
(6) Comment: The commenter states that EPA should refrain from employing the same
methodologies used to establish ozone and PM-2.5 nonattainment area boundaries when
designating lead nonattainment areas. This methodology is inappropriate in determining
the health threat from lead to the local population. Unlike ozone and PM-2.5, lead
concentrations are not driven by complexed chemistry or long range transport issues.
The proposed methodologies will encompass large nonattainment areas. In contrast,
existing monitoring data shows sharp lead gradients surrounding known lead emissions
sources. Bounding a nonattainment area using MSA should only be considered if there is
irrefutable evidence that area source emissions within the entire MSA are contributing to
a violation of the lead NAAQS.
Response: As stated previously, the EPA believes that lead emissions do not generally
transport over long distances (as compared, e.g., to fine particulate matter or 8-hour
ozone). In the proposed rule, EPA proposed to presumptively define the boundary for
designating a nonattainment area as the perimeter of the county associated with the air
quality monitor(s) which records a violation of the standard. In the proposed rule, EPA
also stated that, at the revised level of the standard, EPA expects stationary sources to be
the primary contributor to violations of the NAAQS, although we also believe that nearby
area sources may also contribute to concentrations of lead emissions that may affect a
violating monitor. In light of the possibility that a number of smaller sources may
collectively contribute to concentrations in excess of the NAAQS, EPA believes that
adopting the county boundary as the presumptive boundaries, and the starting point for
setting the boundaries for an area, for lead nonattainment areas is appropriate. However,
as stated in the proposed rule, a state, tribe, or EPA may conduct additional area-specific
analyses that could lead to the boundary for an area either being increased or decreased
from the presumptive county boundary. The EPA is finalizing the factors relevant to
such an analysis as described in the proposed rule because we believe that they will allow
for both the State as well as EPA in some cases to more accurately define the appropriate
boundaries for an area. Also as stated previously, the state may in addition to the factor
analysis also choose to submit information to recommend lead nonattainment boundaries
using any one, or a combination of the following techniques, the results of which EPA
would consider when making a decision as to whether and how to modify the Governors'
recommendations: (1) qualitative analysis, (2) spatial interpolation of air quality
monitoring data, or (3) air quality simulation by dispersion modeling. In situations where
a single source, rather than multiple sources, is causing a NAAQS violation, the EPA
believes that a state may well be able to use area-specific analyses to identify whether a
nonattainment area that is smaller than the county boundary is appropriate. On the other
hand, where it appears that emissions from one or more sources are contributing to
nonattainment throughout an MSA, the site-specific analysis may result in the boundaries
of the nonattainment area overlapping with those of the MSA.
(7) Comment: The commenter states that in the proposed rule EPA states that states may conduct
additional area specific analysis using factors that "closely resemble the factors identified
in recent EPA guidance for the 1997 8-hour ozone and PM-2.5 NAAQS, and the 2006
PM-2.5 NAAQS nonattainment area boundaries that could lead EPA to depart from the
presumptive boundary. These factors, as listed in the proposal are: (1) emissions, (2) air
58
-------
quality, (3) population density and the degree of urbanization, (4) expected growth, (5)
meteorology, (6) geography/topography, (7) jurisdictional boundaries, and (8) level of
control of emission sources. The commenter states that 4 of these factors may be
inappropriate for nonattainment areas that are based on source oriented monitors. (1)
Population density and degree of urbanization, (2) expected growth, and (3) jurisdictional
boundaries have absolutely no relationship to current or expected lead emissions from the
specific lead sources that may be causing high ambient lead levels. Also, if lead
emissions from specific source or sources that were the basis for locating the source
oriented monitor are found to be causing or significantly contributing to NAAQS
violations, then the level of control of these emissions should not be a factor in
nonattainment boundaries. The commenter therefore states that the only appropriate
factors that should be considered in determining the boundaries for lead nonattainment
areas are: (1) emissions, (2) air quality, (3) meteorology, and geography/topography
Response: As stated in the preamble to the final rule, EPA is finalizing the use of the
factors as proposed for making determinations related to the deviation from the
presumptive boundary for designating areas for the Pb NAAQS. The EPA believes that
all of the factors as proposed provide useful information in making a determination
concerning whether the boundaries for a nonattainment area should be either increased or
decreased. Depending on the circumstances in each case where a violation is observed,
certain factors may be more or less important in determining an appropriate boundary.
EPA believes that the commenter is incorrect in stating that the factors or "population
density and degree of urbanization", "expected growth", and "jurisdictional boundaries"
are not important in terms of making decisions related to the appropriate boundaries for
designating lead areas. As stated in the preamble to the final rule, EPA generally expects
contributions to violations of the lead standard to be the result of emissions from larger
stationary point sources, however, smaller area sources may also contribute to these
violations. In these cases, "population density and degree of urbanization" and "expected
growth" may bear a direct relationship to actual and possible future lead source
emissions, and emissions caused by re-entrainment of lead embedded in soil. Also, while
not necessarily bearing a direct relationship to emissions, "jurisdictional boundaries" may
be an important consideration for practically managing air quality, and therefore may be
relevant to establishing a boundary where nonattainment area requirements can be
managed. Therefore, EPA is finalizing the guidance identifying factors for deviating
from the presumptive boundary as provided in the proposed rule. We note also that EPA
does not view this list of factors as necessarily inclusive. Boundary determinations may
also be influenced by additional relevant factors identified by states when making the
boundary recommendations, or by the public in commenting on the state- or EPA-
developed boundary recommendations.
(8) Comment: The commenter states that the maximum impact from the secondary lead smelter
and the battery manufacturing plant in Tampa, Florida are close in proximity to their
plants. Emissions from fugitive sources, short stacks and unenclosed activities seem to be
the biggest contributors to off-site problems. Given their low release height, the affected
air generally should only be in the immediate vicinity of the source. As such, it does not
appear to be a good public policy to make nonattainment designations in all cases using
59
-------
county boundaries or MSA. The key to making any designation would be the scale of the
offending monitor.
Response: As discussed in the preamble and in response to other comments, EPA
recognizes that nonattainment may be a result of a number of sources, or of a single large
source. Therefore, EPA believes it is appropriate to begin with the county boundary, and
then consider adjustments to that boundary as appropriate.
B. Nonattainment Area SIP Submittals
(1) Comment: The commenter states that the proposed rule would allow States [and Tribes] the
statutory maximum 18 months after designation, per the CAA section 191(a), to submit
nonattainment SIPs. For reasons set forth elsewhere in the comment letter, existing
nonattainment and maintenance areas likely have the greatest need for immediate action
to bring ambient lead concentrations down to the revised NAAQS levels. Because of
history of excessive exposure of young children in these areas, and the likely familiarity
of affected States with at least some of the relevant concerns, the nonattainment SIP
submittal deadline for current nonattainment and maintenance areas (assuming that they
are also designated nonattainment under the revised lead NAAQS) should be nine months
after their nonattainment designation.
Response: Section 191(a) states "Any State containing an area designated or
redesignated under section 107(d) as nonattainment with respect to the national primary
ambient air quality standards for ... lead ... shall submit to the Administrator, within 18
months of the designation, an applicable implementation plan meeting the requirements
of this part." EPA does not interpret this language as granting authority to require states
to submit SIPs sooner than 18 months following designation, although states are certainly
allowed to do so. Furthermore, given the time needed for the state to adopt the
appropriate control measures for the sources within the nonattainment area in order for
the area to demonstrate attainment, EPA believes that the time period provided under
section 191(a) is appropriate. We also believe that requiring the submittal of the SIP
within the 18 months provided under section 191(a) allows sufficient time for the
adoption of control measures and provides sufficient time for emissions reductions to be
obtained in order for the area to demonstrate attainment within the 5 year period allowed
under subpart 5 of the CAA for the areas to attain the standard.
C. Emissions Inventory Requirements
(1) Comment: One commenter stated that the existing reporting requirements contained in the
CERR are sufficient to develop lead SIPs for most areas. Problematic areas that are
substantially influenced by local sources may require additional inventory information,
but should be evaluated on a case-by-case basis. Any additional inventory reporting
requirements should be identified during the SIP development process, in cooperation
with the EPA regional office, and should be addressed through the CERR/AERR.
60
-------
Response: The EPA acknowledges that problematic areas may require information in
addition to that required by the CERR. EPA agrees that the need for additional data may
vary by nonattainment area, and thus it may be important for states to determine what
additional data needs may exist for their lead nonattainment areas as part of the SIP
development process in consultation with the EPA regional office. EPA notes we are
updating the SIP emission inventory requirements set forth in 40 CFR §51.117(e) in this
rule. In addition, we note that the Air Emissions Reporting Requirements (AERR) rule,
which will replace the CERR, has been proposed but not yet promulgated. The AERR is
expected to be a means by which the Agency will implement additional data reporting
requirements for the Pb NAAQS SIP emission inventories..
(2) Comment: One commenter stated that EPA should develop additional guidance on emission
inventories related to the SIP submittal because the requirements under the CERR and the
AERR may not be enough to adequately address the emissions inventory requirements
related to the attainment demonstration for the SIP. Commenters also stated that EPA
should provide guidance on what should be the base year for the emissions inventory for
the nonattainment SIP submittal.
Response: As discussed earlier in this section, EPA acknowledges that requirements
under the CERR may not be adequate to address the emission inventory requirements for
lead SIPs. We are evaluating the need for additional guidance to states on lead SIP
emission inventory development. Existing guidance is presented in a document titled
"Emission Inventory Guidance for Implementation of Ozone and Particulate Matter
National Ambient Air Quality Standards and Regional Haze Regulations" (EPA-454/R-
05-001, updated November 2005). EPA anticipates that if we determine additional
guidance on developing lead SIP emission inventories is warranted, it will be presented
as an update to this document. The EPA will review, and as appropriate, revise or update
regulations, policies, and guidance related to the implementation of the revised Pb
NAAQS following the promulgation of the NAAQS.
(3) Comment: One commenter stated that states currently work with regional offices in
developing nonattainment area inventories and that this approach should be encouraged.
The commenter further indicated that states should be allowed to start with the National
Emissions Inventory (NEI) and customize their nonattainment area inventories to analyze
nonattainment problems.
Response: The EPA encourages the states to continue to work closely with the EPA
Regional Offices in developing their nonattainment area emissions inventories as well as
any enhancements that need to be made to the NEI. The EPA condones and encourages
the use of the NEI as a tool to assist states in developing their nonattainment area SIP
emissions inventory. States, however, should be reminded that the nonattainment area
SIP emissions inventory is required pursuant to 40 CFR 51.117(e) and must be approved
by EPA pursuant to the CAA and is subject to the public hearing requirements pursuant
to section 110(a)(2).
61
-------
(4) Comment: Several commenters stated that EPA should revise 40 CFR 51.117(e)(l), relating
to the emissions reporting threshold level for lead nonattainment area SIPs. The current
threshold level as stated in 51.117(e)(l), requires that the point source inventory on
which the summary of the baseline lead emissions inventory is based must contain all
sources that emit 5 or more tpy of lead.
Response: The EPA agrees with the commenter that the requirement for the emissions
inventory reporting threshold for lead nonattainment SIPs, as stated in 40 CFR
51.117(e)(l), should be revised to reflect the stringency of the revised Pb NAAQS.
Accordingly, the EPA is setting the threshold level of the emissions inventory reporting
requirement at 0.5 tpy consistent with the threshold for analysis of RACT/RACM control
measures.
(5) Comment: In general, several commenters said better data should be used to quantify Pb
emissions from sources for implementation purposes. Several commenters (Alaska
Department of Environmental Conservation; Georgia Environmental Protection Division;
New York State Department of Environmental Conservation; Pennsylvania Department
of Environmental Protection; Tennessee Department of Environment and Conservation,
Division of Air Pollution Control; and American Foundry Society) expressed concerns
with the accuracy of the National Emission Inventory (NEI), and recommended allowing
states to use "the best available information" on emissions from Pb sources. Some
commenters pointed to differences in Pb emissions data reported in the Toxics Release
Inventory and the NEI as evidence that the NEI was inaccurate. One industry commenter
(American Foundry Society) said current practices to reduce toxic emissions are not
reflected in the NEI and wanted the opportunity to update the information. Commenters
said EPA should correct the errors in the NEI or allow states to submit revised local data
that more accurately reflect Pb emissions before emissions inventory data are used to
determine which sources exceed the threshold.
Response: The EPA agrees that the most accurate Pb emissions information, based on
scientifically justifiable methods and data, should be used when making decisions about
implementing the Pb NAAQS. This may include supplemental datasets that could include
sources not contained in the NEI. We acknowledge that many of the NEI emission
estimates likely would be improved with more site specific data (e.g., emissions test
data). For this reason we specified that one option available to monitoring agencies
seeking a monitoring waiver is to demonstrate that actual emissions are less than the
emission threshold.
(6) Comment: One commenter said the lead and lead compound reporting requirements under the
Consolidated Emissions Reporting Rule (CERR) were insufficient concerning both the
threshold and timeline. The commenter said the reporting threshold for lead under the
CERR is much higher than the proposed monitoring thresholds so the EPA should use
some other inventory instead of the NEI to determine which sources exceed the threshold.
The commenter also said sources above the RACT threshold should be required to report
Pb emissions annually instead of every 3 years.
62
-------
Response: The EPA recognizes the commenter's concerns that the emissions inventory
reporting threshold is too high; however, the emission inventory reporting threshold for
Pb and the timeline for reporting emission inventories to EPA for the National Emission
Inventory are set under the CERR and are apart from the Pb NAAQS. . EPA notes we
are updating the SIP emission inventory requirements set forth in 40 CFR §51.117(e) in
this rule. In addition, we note that the Air Emissions Reporting Requirements (AERR)
rule, which will replace the CERR, has been proposed but not yet promulgated. The
AERR is expected to be a means by which the Agency will implement additional data
reporting requirements for the Pb NAAQS SIP emission inventories..
(7) Comment: One commenter (Georgia Environmental Protection Division) said EPA emission
factors should be improved for significant source categories of Pb emissions. The
commenter said the absence of or below-C ratings on AP-42 emission factors for
facilities such as Pb smelters and cement kilns showed that emissions estimates to be
used for monitoring determinations may be insufficient.
Response: Response: The EPA agrees that the available lead emissions factor in AP-42
may be less than optimum for use in ambient air monitoring determinations and that
improved emissions factors may help siting of monitoring stations. Nevertheless States
should obtain the most current lead emissions information for those facilities that are
expected to adversely impact the ambient air quality so that the lead monitoring sites are
placed appropriately. At this time, the EPA has no specific plans to update the factors
mentioned by the commenter. EPA is establishing a process to allow States and Industry
to supply quality assured source test data for future improvements in the emissions
factors. It is suggested that future emissions test information be fully documented
electronically to facilitate improved emissions factors development.
(8) Comment: One commenter (Georgia Environmental Protection Division) said EPA should
share the methodologies its technical experts used to adjust 2002 NEI data that EPA
believed significantly overestimated Pb emissions from sources.
Response: EPA made adjustments to the industrial process emission estimates for those
sources that appeared to be emitting very high levels of Pb after consultation with the
EPA Regional Offices and State and local agencies. Boiler emissions were adjusted
using the procedure described in the May 1, 2008 technical memo to the docket
(Document ID EPA-HQ-OAR-2006-0735-5160). In addition, EPA refined its estimates
for airport-specific lead emissions, as explained in another memo to the docket
(Document ID EPA-HQ-OAR-2006-0735-5483).
(10) Comment: Two commenters said EPA needs to provide tribes with updated guidance on
developing emission inventories for Pb.
Response: EPA is committed to working with tribes that want to develop emission
inventories for tribal lands. EPA has training opportunities and programs such as the
63
-------
Tribal Emissions Inventory Software Solution (TEISS) available for tribes that choose to
develop an emission inventory.
D. RACM and RACT for Lead Nonattainment Areas
(1) Comment: One commenter states that certain facility types, such as secondary lead smelters
and lead-acid battery manufacturing plants should be required to implement RACT
regardless of their claimed emissions. The commenter states that most emission
estimates use once a year stack test, and assumptions about fugitive emissions and upsets,
that are not necessarily accurate. The commenter states that if there is a monitored
violation of the standard, and these types of facilities contribute, then at a minimum they
should be required to implement RACT. This would close the door on operators who
attempt to seek RACT exemptions using best case stack tests and not correctly
quantifying fugitive emissions.
Response: EPA believes that the appropriate requirement is for States to do a
RACM/RACT analysis for point sources within the nonattainment area that meet the
threshold level as finalized in the Pb NAAQS rulemaking. As explained in the preamble,
EPA believes that it is appropriate to set the recommended threshold for the RACT
analysis at 0.5 tpy. As discussed in the preamble, while EPA is today setting the
threshold level for sources that the state should include in its RACM/RACT analysis at
0.5 tpy, the state's control technology analysis should also include, as appropriate,
sources which actually emit less than the threshold level of 0.5 tpy of lead, or lead
compounds, in the area or other sources in the area that are reasonable to control, in light
of the attainment needs and feasibility of controls for the affected area. In some cases this
may mean that controls must be placed on sources such as those suggested the
commenter (secondary lead smelters, lead-acid battery manufacturers, and those sources
identified in 40 CFR 51.117(a)(l)). The state must provide appropriate analysis which
demonstrates timely attainment in the area in light of the attainment needs for the affected
area. As stated in the preamble, this may also mean doing an analysis on those sources
that are at or below the threshold level being identified for RACM/RACT.
(2) Comment: Several commenters stated that the implementation proposals for the new Pb
NAAQS level fail to consider the relative bioavailability (RBA) of different lead forms,
even though the much lower RBA of lead sulfide (PbS or galena), which is a predominant
form at lead mining sites, means that it has a much lower potential for adverse health
effects. The commenter further states that despite the widespread recognition that lead
sulfide (galena) has low bioavailability, the implementation plans do not differentiate
between it and other lead forms with higher RBA, or make any other adjustment for
varying degrees of RBA for different Pb forms, in determining whether a particular
emissions source is in compliance with the NAAQS. The comments further state that
rather, and apparently based on the erroneous premise that all Pb forms present
equivalent health risks, any and all Pb emissions are counted for compliance purposes.
The commenter further states that, this overestimates the amount of Pb that could
possibly present a health risk by failing to recognize that galena in particular is not
bioavailable through either ingestion or inhalation, and thus does not contribute to
64
-------
possible health effects. The commenters further state that because the potential health
effects related to Pb exposure vary with KB A of the Pb source, consideration of KB A in
determining compliance would be appropriate.
Response: For the reasons discussed above (in section II.A.2.c.i), the current air quality
criteria do not provide a basis for setting a separate NAAQS for Pb sulfides based on
bioavailability, and therefore EPA did not set such a standard. In light of that conclusion,
there is no basis on which to consider KBA in determining compliance with the NAAQS.
EPA must determine compliance with the standard in accordance with the indicator,
form, averaging time, and indicator selected by the Administration to protect public
health with an adequate margin of safety, pursuant to CAA section 109(b).
(3) Comment: Those seeking to implement Reasonably Available Control Measures ("RACM")
should be directed by EPA to start their consideration with far more than the 1993
Addendum to the General Preamble referenced at 73 Fed. Reg. 29271. That document
(published at 58 Fed. Reg. 67748, 67750-52 (Dec. 22, 1993)) focused principally on
stationary point sources emitting more than 5 tons per year. While it properly addressed
concerns regarding fugitive dust from such sources and possible releases into the air from
historic emissions, much more information has since been developed on control of the
"area sources" that now are the focus of concern.
Furthermore, it is incorrect for the Agency to imply, as it does at 73 Fed. Reg. 29271, that
consideration of "the impact and reasonableness of the measures" mandated as RACM is
a concern only where municipal or other governmental entity resources are involved, or
that the economic feasibility of RACT for area sources should be governed by what
stationary sources have been able to achieve. Nor is it appropriate for EPA to judge the
cost-effectiveness of all elements of revised SIPs submitted in response to a revised lead
NAAQS on the basis of cost per ton of reduction. This measure is sensible for industrial
point sources, but where major contributions to children's exposure are coming from
previously deposited lead in soil or paint-related house dust, cost effectiveness needs to
be measured by the reductions achieved in the immediate areas of those releases.
The Agency appears to properly recognize (at 73 Fed. Reg. 29272) that it would be
incongruous to require RACT analyses only where sources release more than five tons of
lead per year when the threshold of concern for monitoring is expected to be set between
200 kg/year and 600 kg/yr. In response to the request for comment as to what release
level should trigger a RACT analysis, BCI provides this two-fold response: for stationary
sources and fugitive sources associated with them, RACT analysis is only appropriate as
to those facilities at which an exceedance of the NAAQS is measured or reasonably
anticipated. As to "area sources," however - i.e., roadsides, public parks, etc. - the
threshold for RACT analysis should be the soil contamination level EPA has projected to
be likely to create an exceedance in that particular immediate area. Based on analyses
discussed by CASAC in its submissions to EPA, these exceedances appear likely to occur
even in the immediate areas of roadsides and parks.
65
-------
Response: The EPA's historic definition of RACT is the lowest emissions limitation that
a particular source is capable of meeting by the application of control technology that is
reasonably available considering technological and economic feasibility. RACT applies
to the "existing sources" of lead in an area including stack emissions, industrial process
fugitive emissions, and industrial fugitive dust emissions (e.g., on-site haul roads,
unpaved staging areas at the facility, etc) (see section 172(c)(l)). The EPA's previous
guidance for implementing the pre-existing Pb NAAQS recommends that stationary
sources which emit a total of 5 tpy of lead or lead compounds, measured as elemental
lead, be the minimum starting point for RACT analysis (see 58 FR 67750, December 22,
1993). As explained in the preamble, the EPA believes that it is appropriate to set the
recommended threshold for the RACT analysis at 0.5 tpy. While EPA is today
recommending a threshold of 0.5 tpy for sources that the state should include in its
RACM/RACT analysis, EPA also agrees with commenters that the state's control
technology analysis should also include sources which actually emit less than the
threshold level of 0.5 tpy of lead or lead compounds in the area, or other sources in the
area that are reasonable to control, in light of the attainment needs and feasibility of
controls for the affected area. In addition, and as stated in the preamble to the final rule as
it relates to RACM/RACT control analysis, EPA still believes that states should start
should their consideration controls for the affected nonattainment area with the guidance
provided in the 1993 Addendum to the General Preamble referenced at (published at 58
Fed. Reg. 67748, 67750-52 (Dec. 22, 1993)) taking into consideration EPA's guidance
concerning the threshold 0.5 tpy for a RACM/RACT analysis as stated in the preamble to
the final rule.
In addition, as stated in the proposed rule, EPA believes that the regulations, policies, and
guidance currently in place for the implementation of the pre-existing Pb NAAQS are
still appropriate to address the issues required to implement the revised Pb NAAQS. The
EPA believes that these guidance, policies, and regulations should be used by states,
local, and Tribal governments to implement the revised Pb NAAQS at this time. The
EPA will, as appropriate, review, and revise or update policies, guidance, and regulations
to provide for effective implementation of the Pb NAAQS.
The EPA also believes that in identifying the range of costs per ton that are reasonable,
information on benefits per ton of emission reduction can be useful as one factor to
consider. It should be noted that such benefits estimates are subject to significant
uncertainty and that benefits per ton vary in different areas. Nonetheless this information
could be used in a way that recognizes these uncertainties. If a per ton cost of
implementing a measure is significantly less than the anticipated benefits per ton, this
would be an indicator that the cost per ton is reasonable. If a source contends that a
source-specific RACT level should be established because it cannot afford the technology
that appears to be RACT for other sources in its source category, then the source should
support its claim by providing detailed and verified information regarding the impact of
imposing RACT on:
fixed and variable production costs ($/unit),
product supply and demand elasticity,
• product prices (cost absorption vs. cost pass-through),
66
-------
expected costs incurred by competitors,
• company profits, and
• employment costs.
(4) Comment: Given the current capabilities of dispersion models and whole atmosphere models
(such as CMAQ and CAMx), the affect of various lead emission reduction strategies can
be analyzed to determine their effectiveness at reducing ambient lead levels. The
annualized cost of the control measure can be combined with the results of this modeling
analysis to yield a cost effectiveness metric in dollars per microgram per cubic meter
($/|ig/m3). This $/|ig/m3 metric is a better indicator of the cost effectiveness of a control
measure toward attaining the lead NAAQS than a simple $/ton metric which ignores the
effectiveness of a measure in reducing ambient lead levels. Therefore, dollars per
microgram per cubic meter ($/|ig/m3) should be used in lieu of, or at least in addition to
dollars per ton ($/ton) when evaluating the economic feasibility of a lead emission
reduction technology.
Response: As stated in the preamble to the final rule, the EPA still believes that in
determining appropriate emission control levels, the state should consider the collective
public health benefits that can be realized in the area due to projected improvements in
air quality. Because EPA believes that RACT requirements will be met where the state
demonstrates timely attainment, and areas with more severe air quality problems typically
will need to adopt more stringent controls, RACT level controls in such areas will require
controls at higher cost effectiveness levels ($/ton) than areas with less severe air quality
problems. In identifying the range of costs per ton that are reasonable, information on
benefits per ton of emission reduction can be useful as one factor to consider. It should be
noted that such benefits estimates are subject to significant uncertainty and that benefits
per ton vary in different areas. Nonetheless this information could be used in a way that
recognizes these uncertainties. If a per ton cost of implementing a measure is
significantly less than the anticipated benefits per ton, this would be an indicator that the
cost per ton is reasonable. The EPA, however, agrees with the commenter that the dollars
per microgram per cubic meter ($/|ig/m3) may be a reasonable approach to consider in
lieu of, or at least in addition to dollars per ton ($/ton) when evaluating the economic
feasibility of a lead emission reduction technology. The EPA believes that the decision to
use of this metric should be left to the state, or tribal government to decide, taking into
consideration the persistent nature of Pb. However, the appropriateness of the conclusion
related to the results of dollar per ton or $/ug/m3 determination will be assessed by the
EPA during the review of the nonattainment area SIP for the area during the EPA review
of the SIP for approval.
(5) Comment: As with the monitoring issues discussed above, the Proposed Rule's discussion of
RACM, RACT, RFP, and other attainment planning requirements (73 Fed. Reg. at
29270-73) gives inadequate attention to non-point sources of air lead emissions. Where
the principal driver for reducing the NAAQS is the potential effect of lead exposure on
children's intellectual development, and where it is undisputed that reentrainment of
historic emissions is a significant contributor to current emissions, efforts to reach
67
-------
attainment must give considerable attention to such "area sources" as playgrounds, parks,
and similar areas.
Response: EPA expects that exceedances of the revised Pb NAAQS will be primarily
attributable to industrial sources of Pb, and not to reentrainment of historic emissions in
areas such as playgrounds and parks. The EPA further believes that it is appropriate to
set the recommended threshold for the RACT analysis at 0.5 tpy. While EPA is today
recommending a threshold of 0.5 tpy for sources that the state should include in its
RACM/RACT analysis, EPA also agrees with commenters that the state's control
technology analysis should also include sources which actually emit less than the
threshold level of 0.5 tpy of lead or lead compounds in the area, or other sources in the
area that are reasonable to control, in light of the attainment needs and feasibility of
controls for the affected area. The EPA also believes that the regulations, policies, and
guidance currently in place for the implementation of the pre-existing Pb NAAQS are
still appropriate to address the issues required to implement the revised Pb NAAQS. The
EPA believes that these guidance, policies, and regulations should be used by states,
local, and Tribal governments as a starting point to begin implementation of the revised
Pb NAAQS. The EPA will, as appropriate, review, and revise or update policies,
guidance, and regulations to provide for effective implementation of the Pb NAAQS.
E. Attainment Demonstration and Modeling Requirements
(1) Comment: One commenter stated that the proposal cites 40 CFR 51.117 as requiring the use
of dispersion modeling for the demonstration of attainment in vicinity to point sources.
This is appropriate for source oriented lead nonattainment areas. However, the use of
whole atmosphere models (such as CMAQ or CAMx), which can model emissions from
a large number of point, area, and mobile sources should be acceptable for use in urban-
oriented nonattainment areas.
Response: Since Pb is an inert pollutant, Gaussian models such AERMOD (or CALPUFF
when long range transport is of concern) are used due to factors such as ease of use and
costs. However, selection of the best air quality model for the individual application is
encouraged in consultation with the EPA Regional Office.
F. Transportation Conformity
(1) Comment: One commenter stated that they support EPA's finding that transportation
conformity does not apply to the revised Pb NAAQS, since lead additives in gasoline
have been eliminated.
Response: The EPA agrees with the commenter that transportation conformity does not
apply to the revised Pb NAAQS. (See also the preamble for a discussion of this issue).
68
-------
G. Transition from the current NAAQS to a Revised Lead NAAQS
(1) Comment: The commenters states that EPA proposes that, for the current nonattainment
areas, that current NAAQS would remain in effect until the approval of new SIPs. We
suggest, instead that the revised NAAQS take effect upon designation under the revised
NAAQS. We understand that the existing SIP will remain in effect, and enforcement,
until a new SIP under the revised NAAQS is approved. Having the revised NAAQS in
effect will focus the area as promptly and realistically as possible on the extent to which
ambient air exceeds the revised lead NAAQS and spur the development of strong,
effective SIP measures as promptly as possible.
Response: EPA notes that the revised Pb NAAQS will take effect 60 days after
publication of the final rule in the Federal Register. Thus, during the transition both the
old and the new NAAQS will be in effect, as reflected by the revisions to 40 CFR § 50.12
in addition to the promulgation of 40 CFR 50.16. EPA expects this approach will enable
a smooth transition to the revised standard while also addressing the concerns identified
by the commenters.
(2) Comment: One commenter states that in the proposed rule, EPA describes the rationale for
transitioning from the current Pb NAAQS to a new or revised Pb NAAQS. EPA
proposes to revoke the existing standard one year following the promulgation of
nonattainment designations for the new standard. With final designations for a newly
revised standard scheduled for September 2011, the existing standard would be revoked
in September 2012. The commenter stated that they support this approach.
The commenter further states that, for areas in current non-attainment with the pre-
existing Pb NAAQS, EPA proposes to apply the new Pb NAAQS and revoke the pre-
existing Pb NAAQS only after "the affected area submits, and EPA approves, an
attainment demonstration which addresses the attainment of the revised Pb NAAQS".
Under this approach, areas in nonattainment with the pre-existing standard could be
subject to two NAAQS-the old standard as well as the new standard- at the same time
until EPA approves the SIP submittal demonstrating attainment of the revise Pb
NAAAQS. The commenter states that the period of potential overlap could range from at
least 3 years from the September 2011 designation to a much longer and possibly
indeterminate period of time if a state is unable to submit a SIP that demonstrates
attainment with the revised standard or EPA issues a FIP for the area. The commenter
states that this raises a question that EPA should address, when a state is unable to submit
a SIP to demonstrate attainment with the revised Pb NAAQS, what is the transition
process.
Response: As stated in the preamble to the final rule, the EPA believes that Congress
generally did not intend to permit states to relax levels of pollution control when EPA
revises a standard until the new or revised standard is implemented. Therefore, we
believe that controls that are required under the current Pb NAAQS, or that are currently
in place under the current Pb NAAQS, should generally remain in place until new
designations are established and, for current nonattainment areas, new attainment SIPs
69
-------
are approved for any new or revised standard. As a result, EPA proposed that the current
Pb NAAQS should stay in place for one year following the effective date of designations
for any new or revised NAAQS before being revoked, except in current nonattainment
areas, where the existing NAAQS will not be revoked until the affected area submits, and
EPA approves, an attainment demonstration for the revised Pb NAAQS. Accordingly,
the CAA mechanisms, including sanctions, that help ensure continued progress toward
timely attainment would remain in effect for the existing Pb NAAQS, and would apply to
existing Pb nonattainment areas. In cases where the state does not submit a SIP which
demonstrates attainment of the revised Pb NAAQS, the pre-existing (1978) Pb NAAQS
as well as the revised Pb NAAQS would remain in effect for the affected area until the
state submits and EPA approves an attainment demonstration for the revised Pb NAAQS.
IV. RESPONSES TO SIGNIFICANT COMMENTS RELATED TO EXCEPTIONAL
EVENTS INFORMATION SUBMISSION SCHEDULE
One comment was received regarding the exceptional events information submission
schedule. This comment is described and addressed in section VII of the preamble of the final
rule.
V. RESPONSES TO LEGAL, ADMINISTRATIVE, AND PROCEDURAL ISSUES AND
NONGERMANE COMMENTS
A number of comments were received that addressed a wide range of issues including
legal, administrative, and procedural issues, as well as issues that are not germane to the setting
of the NAAQS. Many legal issues are addressed generally throughout the preamble to the final
rule. Specific responses to other comments are presented below.
In addition, EPA also notes that in CAS AC's July 2008 advice to the Agency on the
proposal (Henderson, 2008b), CASAC expressed concern, with respect to the air-related IQ loss
evidence-based framework, that "[a]ll other previous analyses, risk/exposure assessments, staff,
CASAC and public recommendations appear to have been set aside, with this single new meta-
analysis used as the exclusive basis for the proposed NAAQS level."
EPA fully agrees with CASAC that it would be inappropriate to set aside all of the other
information, evidence, and input before the Agency and instead rely exclusively on this
evidence-based framework. However, EPA did not do so in the proposal, and is not doing so in
this final rule. Instead EPA's proposal carefully considered the entire body of evidence and
information, in an integrated fashion, giving appropriate weight to each part of that body of
evidence and information. (See proposal section H.E.3., discussed in preamble section III.C.3.a.)
In the proposal EPA discussed at length its analysis of the evidence, including two different
evidence-based frameworks. EPA also discussed at length its evaluation of the risk assessment,
as well as the advice and recommendations received from CASAC and the public. While EPA
placed primary weight in the proposal on the guidance derived from the air-related IQ loss
evidence-based framework, EPA did not rely on it solely or exclude other information or
70
-------
evidence. For example, while EPA did not give primary weight to the information from the risk
assessment, EPA did determine that it was generally supportive of the estimates obtained from
the evidence based framework. (73 FR 29184, 29243).
(1) Comment: A number of commenters, including states and tribes, expressed concerns about
EPA's new NAAQS review process, variously citing CASAC's concern that the new
process weakened the scientific foundations for the NAAQS review and the need for
EPA to return to the practice of issuing Staff Papers. One commenter (API) supported
the issuance of an ANPR, finding it helpful in reviewing the proposed rule.
Response: EPA appreciates receiving the views of commenters on the revised NAAQS
process. The current NAAQS review process is consistent with the requirements of the
CAA and was developed after consultation with CASAC and public comment to
improve the efficiency of the process while ensuring that the Agency's decisions are
informed by the best available science and timely advice from CASAC and the public. In
addition, EPA notes that while this review included certain aspects of the new process,
such as the issuance of an ANPR, a Staff Paper was also issued in this review.
(2) Comment: One commenter (ABR) stated, and another (BCI) agreed, that EPA cannot rely
upon the Lanphear et al. (2005) study in promulgating the final NAAQS standard because
the underlying data have not been reviewed by EPA or made publicly available. In
support of this argument, ABR cites a number of cases which stand for the proposition
that under the APA an agency must disclose the technical studies and data on which the
proposed rule relies.
Response: EPA notes that revisions to the NAAQS are promulgated under section 307(d)
of the Act, and the APA rulemaking provisions generally do not apply to such
rulemakings. See CAA section 307(d)(l). When this precise question was raised in a
challenge to the 1997 PM NAAQS, the U.S. Court of Appeals for the D.C. Circuit looked
to the specific language of CAA 307(d) and concluded that the "Clean Air Act imposes
no ... obligation [to obtain and publicize data underlying published studies on which the
Agency relies]; it merely directs EPA to include in any notice of proposed rulemaking
'data, information, and documents ... on which the proposed rule relies.'" American
Trucking Associations, Inc. v. EPA, 283 F.3d 355, 372 (D.C. Cir. 2002). The court
found that since EPA was relying on the published studies, not the underlying data, it was
unnecessary to docket the underlying data. The court explicitly endorsed EPA's view that
imposing a requirement on EPA to obtain data for published studies would be
impractical, unnecessary, and would make plainly relevant scientific information
unavailable to EPA for use in standard-setting.
EPA continues to believe, for the reasons stated in the notice of the final rulemaking for
the PM NAAQS in 1997 (62 FR 38652, 38689), that it would unnecessarily and
improperly limit EPA's scientific review to interpret the CAA as requiring that data
underlying studies be included in the docket, even where (as here) EPA has never been in
possession of, or reviewed, the raw data underlying a study. As was the case for the PM
NAAQS reviewed in American Trucking, EPA has placed in the docket all data,
71
-------
information, and documents on which it relied in promulgating this rule. EPA placed in
the docket the Lanphear et al (2005) publication, as well as documentation of correction
of two errors with regard to Table 4 of that publication. These errors were identified as a
result of EPA's examination of the published study in the course of the risk assessment,
and were corrected by the study authors. EPA recognized at that time that the errors did
not affect the results of the risk assessment (Pekar, 2007). In addition, EPA identified
typographical errors in two numbers associated with confidence intervals reported at the
top of the 1st column on p. 897 of the publication. In reporting this information in the
CD, EPA corrected these errors (CD, p. 6-70). Further, as EPA notes in response to
comment (7) in section II.A.2.c.iv above, none of the errors identified by EPA affect
aspects of this study on which EPA has relied in this review. ABR has alleged in its
comments that there are additional, uncorrected mathematical errors in Figure 3 of the
Lanphear et al. (2005) study. As discussed elsewhere in this Response to Comments
(section II.A.2.c.iv), EPA has no reason to believe that Figure 3 of this published, peer-
reviewed study contains the errors suggested by the commenter, and further notes that
conclusions drawn regarding this study did not depend on this figure. Furthermore, even
assuming the items identified by ABR are errors, EPA does not believe they would rise to
the level of fraud, abuse or scientific misconduct warranting review of the raw data. EPA
notes that this study was generally consistent with a large body of other evidence
demonstrating associations between exposure to Pb and neurocognitive decrement in
children. EPA does not consider ABR's comments to provide a basis for doubting the
overall, fundamental validity of the study's conclusions. The public had adequate
opportunity to comment on the strengths and weaknesses of each study, including
Lanphear et al. (2005). EPA does not consider its reliance on this study, its lack of
review of the underlying data, and the lack of docketing of the underlying data, to be an
error, either procedural or substantive.
(3) Comment: Some commenters stated that EPA has failed to meet the requirements of E.O.
12898 on Environmental Justice. One commenter reached this conclusion because the
Agency did not "legitimately determine, acknowledge, assess, evaluate, analyze, or
address the disproportionate adverse impacts of lead exposure on poor and minority
populations" (Sierra Club, p. 2). In the commenter's view, the Agency disregarded
available literature, failed to make a proper assessment of disproportionate exposure and
sensitivity to lead among minority and low-income populations, and proposed no
remedies for these disparate impacts. Another commenter agreed, stating that EPA had
failed to meet its legal obligation to carry out an environmental justice assessment of
proposed NAAQS alternatives (NRDC, p. 22). A third commenter stated that "the
proposed rule ... essentially ignores environmental justice" (Bayview Hunters Point
Community Advocates p. 4) and suggests that, given that the standards under
consideration by the Agency in the proposal were not risk free, EPA should have
considered the distribution of the risks that would remain under alternative standard
levels, and whether those risks would be disproportionately borne by communities that
have traditionally been subject to discrimination.
Response: The NAAQS must afford requisite protection with an adequate margin of
safety, including for sensitive subpopulations as well as to the general populace. See,
72
-------
Section IB. of the preamble. Minority and low-income populations are often such
sensitive subpopulations. The Pb NAAQS established in today's final rule are nationally
uniform standards which in the Administrator's judgment are requisite to protect public
health with an adequate margin of safety. As discussed in section II of the preamble to
the final rule and in other comment responses, the Administrator expressly considered the
available information regarding air-related Pb exposure and health effects among
sensitive populations, including low income and minority populations, in making these
determinations.
In accordance with E.O. 12898, EPA has considered whether the decisions promulgated
in the final rule may have disproportionate negative impacts on minority or low-income
populations. This rule establishes national ambient air quality standards for lead that are
significantly more protective than the current standard and is not expected to have
disproportionate negative impacts on minority or low-income populations. EPA did
conduct a quantitative analysis of the socio-demographics of populations living near
ambient air Pb monitors and Pb sources emitting more than one ton of Pb per year. This
analysis was necessarily limited by the lack of data on certain key parameters which
prevented the Agency from assessing actual exposures and risks to these adjacent
populations. However, EPA believes that the revised Pb standards will reduce health
risks precisely in the areas subject to the highest ambient air concentrations of Pb,
including areas immediately adjacent to Pb-emitting sources.
EPA notes that, as discussed elsewhere in the preamble and this response to comments,
we concluded that the approach used in 1978 was no longer appropriate in light of
scientific developments and the Administrator set the standard in this review based on
protecting sensitive groups from air-related Pb risk. EPA further notes that populations
with the greatest total exposure to Pb may not be the populations most sensitive to air-
related Pb, but a standard set to protect sensitive populations (including populations with
greater exposure and increased susceptibility) from air-related Pb risks with an adequate
margin of safety would also provide sufficient protection for other populations as well. In
the Administrator's judgment these national standards will protect public health,
including the health of sensitive groups, with an adequate margin of safety. To the extent
any of the commenters is suggesting E.O. 12898 requires additional quantitative analysis
or assessment of environmental justice issues related to revising the Pb NAAQS, or that
the standard should be set more stringent than necessary to protect the health of sensitive
and other groups with an adequate margin of safety, EPA disagrees.
(4) Comment: Some commenters provided comments on the cost or economic impact of
monitoring, implementation, or compliance associated with a revised Pb NAAQS. For
example, one commenter, on behalf of Doe Run Resources Corp., submitted a lengthy
comment on the domestic and international lead markets, and the possible effect of the
revised Pb NAAQS on those markets (based on assumed shutdowns of Pb primary and
secondary lead smelting facilities), and other commenters indicated that monitoring costs
would be burdensome unless borne by EPA.
73
-------
Response: As noted in section IB of the preamble, the Clean Air Act bars consideration
of costs in setting the NAAQS, and accordingly EPA has not considered costs, including
the costs or economic impact of monitoring, implementation or compliance, in revising
the Pb NAAQS.
(5) Comment: Some commenters stated that EPA is obligated to consider studies that EPA
indicated it would not rely upon because the studies in question were not included in the
Criteria Document and Staff Paper, which have undergone CAS AC and public review.
One commenter (MO Coalition) stated that a study by Hayes et al. (1994) is highly
relevant, had been timely submitted to EPA as part of the review and cited by CAS AC,
and must be considered by EPA under CAA section 307(d)(6)(C). The commenter also
questioned EPA's statement in the proposal (made in reference to Hayes et al. (1994) and
a second study) that "EPA is not basing its proposed decisions on these two studies, but
notes that these estimates are consistent with other studies that were included in the 1986
and 2006 Criteria Documents and accordingly considered by CASAC and the public"
(MO Coalition, pp. 27-28).
Response: For the reasons stated in section 1C of the preamble, EPA is not relying upon
studies that were not included in the Criteria Document and Staff Paper, which have
undergone extensive critical review by EPA, CASAC and the public. EPA believes that
even where a study has been identified in public comment on the drafts of the CD or Staff
Paper, if the study is not cited in the CD or Staff Paper, EPA does not consider it to have
undergone the expected intensive review by CASAC, EPA and the public, just because it
was included in a public comment on a draft of the CD, and therefore should not be relied
upon for this review. This approach is consistent with EPA's practice in prior NAAQS
reviews and its interpretation of the requirements of section 109 of the CAA. EPA does
not consider the language of section 307(d)(6)(C), which specifies the record for judicial
review, to be inconsistent with its interpretation of section 109. Likewise, a citation by
CASAC in a letter to the Administrator on the proposal is not an adequate substitute for
the review associated with the development of the Criteria Document and Staff Paper.
Accordingly, EPA considered the study by Hayes et al. (1994) and other "new" studies
only provisionally, in conjunction with other "new" studies, and in the context of the air
quality criteria to determine whether the "new" studies, considered in context, materially
change any of the broad scientific conclusions regarding the health effects and exposure
pathways of ambient air Pb made in the air quality criteria. The results of EPA's
provisional consideration of these studies (including the study by Hayes et al.[1994]) was
that they do not materially change those broad scientific conclusions and thus do not
warrant reopening the air quality criteria review. As EPA notes in the preamble, there are
strengths and limitations for the Hayes et al. (1994) study which may affect the specific
magnitudes of the ratios reported in that study, but the study's findings and trends are
generally consistent with the air quality criteria and analyses considered in this review.
(6) Comment: One commenter states that "there are several errors in the Agency's data base that
relate to lead acid battery manufacturing facilities and the nation's sole remaining
primary lead smelter" and that "[Corrections should be made before EPA bases any
conclusions on the erroneous data" (BCI p. 3).
74
-------
Response: EPA appreciates receiving the information on Pb emissions inventories
submitted by the commenter, and acknowledges that we have continued to refine our
information about Pb emissions inventories during the course of this review based on
additional information. EPA notes, however, that neither the revision of the NAAQS nor
other decisions made in the final rule depend on the data identified for correction by the
commenter.
75
-------
References
Bellinger, D.C. and Needleman, H.L. (2003) Intellectual impairment and blood lead levels
[letter]. N. Engl. J. Med. 349: 500.
Bergdahl IA. (2006) What is the meaning of non-linear dose-response relationships between
blood lead concentrations and IQ? Neurotoxicology. 2006 Dec;27(6):1125.
Bergdahl IA. (2007) Another fundamental error in "What is the meaning of non-linear dose-
response relationships between blood lead concentrations and IQ?" became obvious in
the authors' response to comments. Neurotoxicology. 2007 May;28(3):705-6.
Bowers, TS; Beck, BD. (2006) "What is the meaning of non-linear dose-response relationships
between blood lead concentrations and IQ?" Neurotoxicology 27:520-524.
Bowers, TS; Beck, BD. (2007a) Response to comments by Bergdahl, Hornung etal., Jusko etal,
and Svendsgaard etal on "What is the meaning of non-linear dose-response relationships
between blood lead concentration and IQ?" Neurotoxicology 28: 197-201.
Bowers, TS; Beck, BD. (2007b) "Response to second comment by Bergdahl on "What is the
meaning of non-linear dose-response relationships between blood lead concentration and
IQ?" and our earlier response.'" Neurotoxicology 28 :706.
Braun JM, Froehlich TE, Daniels JL, Dietrich KN, Hornung R, Auinger P, Lanphear BP. (2008)
Association of environmental toxicants and conduct disorder in U.S. children: NHANES
2001-2004. Environ Health Perspect. 2008 Jul;116(7):956-62.
Braun JM, Kahn RS, Froehlich T, Auinger P, Lanphear BP. (2006) Exposures to environmental
toxicants and attention deficit hyperactivity disorder in U.S. children. Environ Health
Perspect. 2006 Dec; 114(12): 1904-9.
Brunekreef, B.; Noy, D.; Biersteker, K.; Boleij, J. (1983) Blood lead levels of Dutch city
children and their relationship to lead in the environment. J. Air Pollut. Control Assoc.
33:872-876.
Brunekreef, B. (1984) The relationship between air lead and blood lead in children: a critical
review. Science of the total environment, 38: 79-123.
Canfield, R.L.; Henderson, C.R., Jr.; Cory-Slechta, D.A.; Cox, C.; Jusko, T.A.; Lanphear, B.P.
(2003a) Intellectual impairment in children with blood lead concentrations below 10 ug
per deciliter. N. Engl. J. Med. 348: 1517-1526.
Canfield, R.L., Kreher, D.A., Cornwell, C., and Henderson, C.R., Jr. (2003b) Low-level lead
exposure, executive functioning, and learning in early childhood. Child Neuropsychol. 9:
35-53.
76
-------
Caravanos, J.; Weiss, A.L; Blaise, M.J.; Jaeger RJ.(2006b) A survey of spatially distributed
exterior dust lead loadings in New York City. Environ Res 100: 165-172.
Casteel, S.W.; Tessman, R.; Brattin W.J.; Wahlquist A.M. (2005) Relative Bioavailability of
Lead in House Dust and Soil from the Herculaneum Lead Smelter Site in Herculaneum,
Missouri. Prepared for Black & Veatch Special Projects Corporation; May.}
Centers for Disease Control and Prevention, (2005). Preventing lead poisoning in young
children: A statement by the Centeres for Disease Control and Prevention. Atlanta GA:
U.S. Department of Health and Human Services, Public Health Service. August.
Chen A, Cai B, Dietrich KN, Radcliffe J, Rogan WJ. (2007) Lead exposure, IQ, and behavior in
urban 5- to 7-year-olds: does lead affect behavior only by lowering IQ? Pediatrics. 2007
Mar;119(3):e650-8.
Chen, A.; Dietrich, K.N.; Ware, J.H.; Radcliffe, J.; Rogan, WJ. (2005) IQ and blood lead from 2
to 7 years of age: Are the effects in older children the residual of high blood lead
concentrations in 2-year-olds? Environ. Health Perspect. 113: 597-601.
Chiodo LM, Covington C, Sokol RJ, Hannigan JH, Jannise J, Ager J, Greenwald M, Delaney-
Black V. (2007) Blood lead levels and specific attention effects in young children.
Neurotoxicol Teratol. 2007 Sep-Oct;29(5):538-46.
Donguk P., Namwon P. (2004) Exposure to Lead and its Particle Size Distribution, Journal of
Occupational Health, 46: 225-229, 2004.
Flynn, J. R. (1994). IQ gains over time. In R. J. Sternberg (Ed.), Encyclopedia of human
intelligence (pp. 617-623). New York: Macmillan.
Hayes, E.B.,; McElvaine, M.D.; Orbach, H.G.; Fernandez, A.M.; Lyne, S.; Matte, T.D. (1994)
Long-term trends in blood lead levels among children in Chicago: Relationship to air lead
levels. Pediatrics 93:195-200.
Henderson, R. (2007a) Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory
Committee, to Administrator Stephen L. Johnson. Re: Clean Air Scientific Advisory
Committee's (CAS AC) Review of the 1st Draft Lead Staff Paper and Draft Lead
Exposure and Risk Assessments. March 27, 2007.
Hilts, S.R. (2003) Effect of smelter emission reductions on children's blood lead levels. Sci.
Total Environ. 303: 51-58.
Hornung R, Lanphear B, Dietrich K. (2006) Response to: "What is the meaning of non-linear
dose-response relationships between blood lead concentration and IQ?".
Neurotoxicology. 2006 Jul;27(4):635.
ICF International (2006) Lead Human Exposure and Health Risk Assessments and Ecological
Risk Assessment for Selected Areas. Pilot Phase. Draft Technical Report (with
appendices). Prepared for the U.S. EPA's Office of Air Quality Planning and Standards,
77
-------
Research Triangle Park, NC. December. Available online at:
http://www.epa.gov/ttn/naaqs/standards/pb/s_pb_cr_td.html
Jusko TA, Lockhart DW, Sampson PD, Henderson CR Jr, Canfield RL. (2006) Response to:
"What is the meaning of non-linear dose-response relationships between blood lead
concentrations and IQ?". Neurotoxicology. 2006 Dec;27(6): 1123-5.
Jusko TA, Henderson CR, Lanphear BP, Cory-Slechta DA, Parsons PJ, Canfield RL. (2008)
Blood lead concentrations < 10 microg/dL and child intelligence at 6 years of age.
Environ Health Perspect. 2008 Feb;l 16(2):243-8.
Lanphear, B.P.; Hornung, R.; Khoury, J.; Yolton, K.; Baghurst, P.; Bellinger, D.C.; Canfield,
R.L.; Dietrich, K.N.; Bornschein, R.; Greene, T.; Rothenberg, S. J.; Needleman, H.L.;
Schnaas, L.; Wasserman, G.; Graziano, J.; Roberts, R. (2005) Low-level environmental
lead exposure and children's intellectual function: an international pooled analysis.
Environ. Health Perspect. 113: 894-899.
Leggett, R.W. (1993) An age-specific kinetic model of lead metabolism in humans. Environ
Health Persp 101(7):598-616.
Miranda ML, Kim D, Galeano MA, Paul CJ, Hull AP, Morgan SP. (2007) The relationship
between early childhood blood lead levels and performance on end-of-grade tests.
Environ Health Perspect. 2007 Aug; 115(8): 1242-7.
Nevin, R. (2000) How lead exposure relates to temporal changes in IQ, violent crime, and unwed
pregnancy. Environ Res. 2000 May;83(l):l-22.
Nigg JT, Knottnerus GM, Martel MM, Nikolas M, Cavanagh K, Karmaus W, Rappley MD.
(2008) Low blood lead levels associated with clinically diagnosed attention-
deficit/hyperactivity disorder and mediated by weak cognitive control. Biol Psychiatry.
2008 Feb 1;63(3):325-31.
Park D., Paik N. (2002) Effect on Blood Lead of Airborne Lead Particles Characterized by Size,
Annals of Occupational Hygiene, 46(2): 237-243, 2002.
Pekar, Z. (2007) Correction to Errors Identified in Lanphear et al. 2005 Pooled Analysis Study
and Implications for Pilot Risk Assessment. From: Memoradum to the Lead NAAQS
Review Docket. EPA-HQ-OAR-2006-0735. January 26, 2007.
Pekar, Z., Sasser, E., and Ashley, J. (2008) Analysis of Socio-Demographic Factors for
Populations Living Near Pb TSP Monitors and Larger Pb Point Sources. From:
Memoradum to the Lead NAAQS Review Docket. EPA-HQ-OAR-2006-0735. May 20,
2008.
Rothenberg, S.J.; Rothenberg, J.C. (2005) Testing the dose-response specification in
epidemiology: public health and policy consequences for lead. Environ. Health Perspect.
113:1190-1195.
78
-------
Schwartz, J., and Pitcher, H. (1989) The relationship between gasoline lead and blood lead in the
United States. J Official Statistics 5(4):421-431.
Solon O, Riddell TJ, Quimbo SA, Butrick E, Aylward GP, Lou Bacate M, Peabody JW. (2008)
Associations between cognitive function, blood lead concentration, and nutrition among
children in the central Philippines. J Pediatr. 2008 Feb;152(2):237-43.
Surkan PJ, Schnaas L, Wright RJ, Tellez-Rojo MM, Lamadrid-Figueroa H, Hu H, Hernandez-
Avila M, Bellinger DC, Schwartz J, Perroni E, Wright RO. (2008) Maternal self-esteem,
exposure to lead, and child neurodevelopment. Neurotoxicology. 2008 Mar;29(2):278-85.
Surkan PJ, Zhang A, Trachtenberg F, Daniel DB, McKinlay S, Bellinger DC. (2007)
Neuropsychological function in children with blood lead levels <10 microg/dL.
Neurotoxicology. 2007 Nov;28(6): 1170-7.
Svendsgaard D, Kim JY, Kotchmar D, Rothenberg SJ. (2007) A conclusion regarding: "what is
the meaning of non-linear dose-response relationships between blood lead and IQ?".
Neurotoxicology. 2007 Jan;28(l): 196-7.
Te'llez-Rojo, M.M.; Bellinger, D.C.; Arroyo-Quiroz, C.; Lamadrid-Figueroa, H.; Mercado-
Garci'a, A.; Schnaas-Arrieta, L.; Wright, R.O.;Herna'ndez-Avila, M.; Hu, H. (2006)
Longitudinal associations between blood lead concentrations < 10 ug/dL and
neurobehavioral development in environmentally-exposed children in Mexico City.
Pediatrics 118: e323-e330.
U.S. Environmental Protection Agency. (1986) Air quality criteria for lead. Research Triangle
Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; EPA report no. EPA-600/8-83/028aF-dF. 4v. Available from: NTIS,
Springfield, VA; PB87-142378.
U.S. Environmental Protection Agency. (1989) Review of National Ambient Air Quality
Standard for Pb: Exposure Analysis Methodology and Validation. Research Triangle
Park, NC: Office of Air Quality Planning and Standards. EPA-450/2-89-011. June.
U.S. Environmental Protection Agency. (2006a) Air Quality Criteria for Lead. Volumes I and II.
Washington, DC, EPA/600/R-5/144aF and EPA/600/R-5/144bF. Available online at:
http://www.epa.gov/ncea/
U.S. Environmental Protection Agency. (2006b) Final Report. Regional Applied Research
(RARE) Project: Final Report for investigation of changes in lead relative bioavailability
following weathering of ore concentrate-soil mixtures from the herculaneum, Missouri,
Smelter Site. Prepared by MSE Technology Applications, Inc. For USEPA, National
Risk Management Research Laboratory, Cincinnati, OH and USDOE. November.
U.S. Environmental Protection Agency. (2007a) Review of the national ambient air quality
standards for lead: assessment of scientific and technical information. OAQPS staff
paper. (Final) November 2007. Research Triangle Park, NC: Office of Air Quality
79
-------
Planning and Standards; EPA report no. EPA- 452/R-07-013. Available online at:
http://www.epa.gov/ttn/naaqs/standards/pb/s_pb_cr_sp.html
U.S. Environmental Protection Agency. (2007b) Lead: Human Exposure and Health Risk
Assessments for Selected Case Studies, Volume I. Human Exposure and Health Risk
Assessments—Full-Scale and Volume II. Appendices. Office of Air Quality Planning and
Standards, Research Triangle Park, NC. EPA-452/R-07-014a and EPA-452/R- 07-
014b. Available online at: http://www.epa.gov/ttn/naaqs/standards/pb/s_pb_cr_td.html
U.S. Environmental Protection Agency. (2007). Correction to Errors Identified in Lanphear et al.
2005 Pooled Analysis Study and Implications for Pilot Risk Assessment. Office of Air
Quality Planning and Standards (OAQPS), USEPA, (Jan 26, 2007)
Wang HL, Chen XT, Yang B, Ma FL, Wang S, Tang ML, Hao MG, Ruan DY. (2008) Case-
Control Study of Blood Lead Levels and Attention-Deficit Hyperactivity Disorder in
Chinese Children. Environ Health Perspect. doi: 10.1289/ehp. 11400 available via
http://dx.doi.org/ [Online 5 June 2008].
Wright JP, Dietrich KN, Ris MD, Hornung RW, Wessel SD, Lanphear BP, HO M, and Rae MN.
(2008) Association of Prenatal and Childhood Blood Lead Concentrations with Criminal
Arrests in Early Adulthood. PLoS Medicine, doi.: 10.1371/journal.pmed.0050101
available via http://dx.doi.org/ [Online 27 May 2008].
80
-------
Appendix A. Studies on neurotoxic effects of Pb cited by public commenters that were
published after closure of 2006 Air Quality Criteria Document and other relevant recent studies
published since that time that were identified in routine review of journals. These studies were
provisionally considered by EPA, as discussed in section 1C of the preamble and in this
document, in responses to comment number 6 in section I. A.2.c.iv and comment number 1 in
section II.A.3, respectively.
Braun JM, Kahn RS, Froehlich T, Auinger P, Lanphear BP. Exposures to environmental
toxicants and attention deficit hyperactivity disorder in U.S. children. Environ Health Perspect.
2006 Dec; 114(12): 1904-9.
Braun JM, Froehlich TE, Daniels JL, Dietrich KN, Hornung R, Auinger P, Lanphear BP.
Association of environmental toxicants and conduct disorder in U.S. children: NHANES 2001-
2004. Environ Health Perspect. 2008 Jul;l 16(7):956-62.
Brown MJ, Rhoads GG. Guest Editorial: Responding to Blood Lead Levels < 10 ug/dL.
Environmental Health Perspectives, February 2008, Vol 116 (2), p. A60-61.
Cecil KM, Brubaker CJ, Adler CM, Dietrich KN, Altaye M, Egelhoff JC, Wessel S, Elangovan I,
Hornung R, Jarvis K, Lanphear BP. Decreased brain volume in adults with childhood lead
exposure. pLoS Med. 2008 May 27;5(5)
Chakraborty BM, Lee HS, Wolujewicz M, Mallik J, Sun G, Dietrich KN, Bhattacharya A, Deka
R, Chakraborty R. (2008) Low Dose Effect of Chronic Lead Exposure on Neuromotor Response
Impairment in Children is Moderated by Genetic Polymorphisms. J. Hum. Ecol., 23(3): 183-194
Chen A, Cai B, Dietrich KN, Radcliffe J, Rogan WJ. Lead exposure, IQ, and behavior in urban
5- to 7-year-olds: does lead affect behavior only by lowering IQ? Pediatrics. 2007
Mar;119(3):e650-8.
Chiodo LM, Covington C, Sokol RJ, Hannigan JH, Jannise J, Ager J, Greenwald M, Delaney-
Black V. Blood lead levels and specific attention effects in young children. Neurotoxicol Teratol.
2007 Sep-Oct;29(5):538-46
Froehlich et al 2007. Prevalence, recognition and treatment of attention deficit/hyperactivity
disorder in a national sample of U.S. children. Arch Pediatr Adolesc Med 161 (9): 857-64
Jedrychowski et al 2008. Prenatal low-level lead exposure and developmental delay of infants at
age 6 months (Krakow inner city study). Int J Hyg Environ Health 211: 345-351.
Jusko TA, Henderson CR, Lanphear BP, Cory-Slechta DA, Parsons PJ, Canfield RL. Blood lead
concentrations < 10 microg/dL and child intelligence at 6 years of age. Environ Health Perspect.
2008Feb;116(2):243-8.
A-l
-------
Menke A, Muntner P, Batuman V, Silbergeld EK, Guallar E. Blood Lead Below 0.48 umol/L
(10 ug/dL) and Mortality Among US Adults. Circulation, Journal of the American Heart
Association, published online Sep 18, 2006. p. 1385-1394.
Miranda ML, Kim D, Galeano MA, Paul CJ, Hull AP, Morgan SP. The relationship between
early childhood blood lead levels and performance on end-of-grade tests. Environ Health
Perspect. 2007 Aug; 115(8): 1242-7.
Nevin. 2007. Understanding international crime trends: the legacy of preschool lead exposure.
Environ Res 104(3): 315-336.
Nigg JT, Knottnerus GM, Martel MM, Nikolas M, Cavanagh K, Karmaus W, Rappley MD. Low
blood lead levels associated with clinically diagnosed attention-deficit/hyperactivity disorder and
mediated by weak cognitive control. Biol Psychiatry. 2008 Feb 1;63(3):325-31.
Solon O, Riddell TJ, Quimbo SA, Butrick E, Aylward GP, Lou Bacate M, Peabody JW.
Associations between cognitive function, blood lead concentration, and nutrition among children
in the central Philippines. J Pediatr. 2008 Feb;152(2):237-43.
Surkan PJ, Zhang A, Trachtenberg F, Daniel DB, McKinlay S, Bellinger DC.
Neuropsychological function in children with blood lead levels <10 microg/dL.
Neurotoxicology. 2007 Nov;28(6): 1170-7.
Surkan PJ, Schnaas L, Wright RJ, Tellez-Rojo MM, Lamadrid-Figueroa H, Hu H, Hernandez-
Avila M, Bellinger DC, Schwartz J, Perroni E, Wright RO. Maternal self-esteem, exposure to
lead, and child neurodevelopment. Neurotoxicology. 2008 Mar;29(2):278-85.
Wang HL, Chen XT, Yang B, Ma FL, Wang S, Tang ML, Hao MG, Ruan DY. Case-Control
Study of Blood Lead Levels and Attention-Deficit Hyperactivity Disorder in Chinese Children.
Environ Health Perspect. 116(10). p. 1401.
Wright JP, Dietrich KN, Ris MD, Hornung RW, Wessel SD, Lanphear BP, HO M, and Rae MN.
Association of Prenatal and Childhood Blood Lead Concentrations with Criminal Arrests in
Early Adulthood. PLoS Medicine. doi.:10.1371/journal.pmed.0050101 available via
http://dx.doi.org/ [Online 27 May 2008].
A-2
------- |