oEPA
United States
Environmental Protection
Agency
Office of Water
4304T
EPA-822-R-20-004
January 2020
EPA RESPONSE
TO PUBLIC COMMENTS ON
2017 DRAFT ALUMINUM
AMBIENT WATER QUALITY
CRITERIA (2020)
-------
EPA-822-R-20-004
EPA RESPONSE TO PUBLIC COMMENTS ON 2017 DRAFT ALUMINUM AMBIENT WATER
QUALITY CRITERIA (2020)
January 2020
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF WATER
OFFICE OF SCIENCE AND TECHNOLOGY
HEALTH AND ECOLOGICAL CRITERIA DIVISION
WASHINGTON, D C.
li
-------
Table of Contents
INTRODUCTION 1
TOPIC 1: Comments regarding acute toxicity data 2
TOPIC 2: Comments regarding alum (aluminum sulfate) used for pollution control 8
TOPIC 3: Comments regarding the Aluminum Criteria Calculator 13
TOPIC 4: Comments regarding aluminum not being a priority pollutant 19
TOPIC 5: Comments regarding BLM Approach 20
TOPIC 6: Comments regarding chronic toxicity data 21
TOPIC 7: Comments regarding compliments to Aluminum AWQC development 32
TOPIC 8: Comments regarding the document in general 50
TOPIC 9: Comments regarding the Endangered Species Act 75
TOPIC 10: Comments regarding exposure routes 81
TOPIC 11: Comments regarding other general issues 82
TOPIC 12: Comments providing information on external research 95
TOPIC 13: Comments regarding the lack of marine criteria 98
TOPIC 14: Comments regarding the MLR (multiple linear regression) models 99
TOPIC 15: Comments regarding mussel toxicity data 142
TOPIC 16: Comments regarding plant toxicity data 147
TOPIC 17: Comment regarding Multi-Sector General Permit (MSGP) 149
TOPIC 18: Comments regarding implementation issues 150
TOPIC 19: Comments regarding implementation issues with measuring aluminum 176
TOPIC 20: Comments regarding EPA policies 223
TOPIC 21: Comments regarding the regulatory burden of aluminum criteria 234
TOPIC 22: Comments regarding a request for an extension on the comment period 236
in
-------
INTRODUCTION
Section 304(a) (1) of the Clean Water Act, 33 U.S.C. § 1314(a)(1), directs the Administrator of
the EPA to publish water quality criteria that accurately reflecting the latest scientific knowledge
on the kind and extent of all identifiable effects on health and welfare that might be expected
from the presence of pollutants in any body of water. In support of this mission, the EPA is
updating water quality criteria to protect aquatic life from the potential effects of aluminum in
freshwater environments. The 2018 aluminum criteria document provides a scientific evaluation
of ecological effects and is not a regulation. The recommended limit on the level of aluminum in
freshwater that will still be protective of aquatic life depends on a site's water quality parameters.
Studies have shown that three water chemistry parameters, pH, dissolved organic carbon (DOC),
and total hardness, influence the toxicity of aluminum by affecting the bioavailability of
aluminum in the water to aquatic species. Unlike the fixed criteria values in the EPA's 1988
criteria document, the 2018 updated criteria use a Multiple Linear Regression (MLR) model to
normalize the toxicity data. The criteria are then generated though a criteria calculator following
the 1985 Guidelines calculation procedures based on site pH, DOC, and total hardness levels.
This allows users to develop aluminum criteria for fresh waters that appropriately reflect local
water chemistry parameters.
The EPA submitted its Draft Aquatic Life Ambient Water Quality Criteria for Aluminum - 2017
for public comment on July 28, 2017. The request for scientific views on the draft was open for
90 days (60 days plus a 30-day extension). As of October 26, 2017, four hundred and twenty
comments from sixty-nine commenters were received (note: one entry was repeated). The EPA
considered scientific views from the public on this draft document as well as any new data or
information received. This report documents the EPA's response to public comments on the
2017 draft aluminum criteria document.
The following tables divide the comments into common topics for ease of the reader (e.g.,
chronic toxicity data, Aluminum Criteria Calculator, mussel toxicity data, etc.). Comments are
summarized and the EPA's responses to the public comments are provided. The EPA completed
the 2018 Final Aquatic Life Ambient Water Quality Criteria for Aluminum considering these
comments and noted in the table where the document was edited, when applicable.
1
-------
TOPIC 1: Comments regarding acute toxicity data
(o 1111111-111
Nil in her
(Or^iini/iilioii)
Public ( cim moil I on lopic 1: Rciiiinlinii iiculc (o\ici(\ d;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rilcrhi Document
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of C. dubia and Daphnia magna data from European
Aluminium Association 2009
The 2009 data from the European Aluminium Association include
a series of acute C. dubia tests under varying pH, hardness, and
DOC conditions, as well several tests testing the effects of test
solution aging on aluminum toxicity.
As part of this series ofpH, hardness, and DOC
manipulation tests, the results for the pH 6, 120 mg/L
hardness, 0 mg/L DOC test conditions were not included
in Appendix A. The LC50 for this test was 2007.7 jug/L
and should be included in the acute dataset unless
sufficient reasons are provided.
The LC50 of 2007.7 |ig/L was not included because of a poor
concentration-response relationship displayed in the raw data
so the data was deemed unacceptable for use.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of Ceriodaphnia dubia data from ENSR 1992d
ENSR 1992d tested the toxicity of aluminum to C. dubia under four
different hardness conditions, 26, 46, 96, and 194 mg/L; all four
results were deemed acceptable for criteria derivation.
The results from the highest hardness tests were not
included in the SMA V derivation because "a more
definitive value is available, or value is considered an
outlier" (footnote c of Appendix A). This test resulted in
an LC50 of >99,600 jug/L. And while this value is higher
than any other test, it empirically demonstrates the
relative insensitivity of C. dubia under high pH and
hardness conditions and high aluminum loadings,
conditions which are not well represented by the other
acceptable data.
Other acute studies, such as from the European
Aluminium Association, under comparable high pH and
hardness conditions similarly did not demonstrate
toxicity. However, the highest aluminum concentrations in
those tests were far lower compared to the ENSR study
and so the resulting LC50s (e.g., LC50 of >5,000 jug/L at
pH 7.88, 120 mg/L hardness, and 0.5 mg/L DOC) carried
over into the SMA V calculation may overestimate the
toxicity under these conditions.
It would be helpful for EPA to provide further discussion
on how unbounded tests were deemed acceptable for
inclusion in the criteria calculation, per the decision rule
Thank you for your comment. The results from the highest
hardness tests from ENSR (1992d) were included in the final
SMAV calculation for Ceriodaphnia dubia, as described in
the final 2018 criteria document on page 44.
Regarding the use of unbounded toxicity values, use of
"greater than" values follows the "decision rule" as described
in the final aluminum criteria document (Section 3.1), as
follows: "greater than" (>) low chronic values and "less than"
(<) high chronic values were not used in the calculation of the
SMCV; but "less than" (<) low chronic values and a "greater
than" (>) high chronic values were included in the SMCV.
This approach was also followed for acute SMAV
calculations. The methodology is based on the finding that
"greater than" values for concentrations of low magnitude,
and "less than" values for concentrations of high magnitude
do not generally add significant information to the toxicity
analysis. In the 2018 Final Aluminum Criteria document in
Section 3.1, All Species Mean Acute Value (SMAV)
calculations were re-evaluated to verify that they adhere to
the decision rule. This approach to the use of "greater than"
values was initially described in the 2013 Aquatic Life
Ambient Water Quality Criteria for Ammonia in Freshwater
and has continued to be applied in subsequent criteria.
Appendix A
2
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 1: Rciiiinlinii iiculc l«i\icil> (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
in Section 3.1.1, and how studies were determined to be
outliers. For the ENSR study, it appears the unbounded
result decision rule would not apply because this is a
' 'greater than' high acute value. " Inclusion of these pH
8.1 test data would "add significant information " because
no toxicity was observed with very high loadings. EPA
should provide clarification on whether this decision rule
does or does not apply to the lower unbounded LC50s
under similar test conditions (e.g., >5000 fig/L).
This represents a larger issue with the C. dubia acute
dataset, where in 23 of the 52 acceptable test results
(44%) for the SMAV calculation, an LC50 could not be
calculated. This may be problematic because using the
Aluminum Criteria Calculator V. 1.0 spreadsheet, under
most water quality conditions, Ceriodaphnia are one of
the four most acutely sensitive genera. However, the test
concentrations used to test the sensitivity of C. dubia were
insufficient to elicit toxicity in nearly half of the tests,
likely overestimating the sensitivity of this species.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Review of the Acute Studies Incorporated into the Draft Criteria
GE1 reviewed the acute toxicity studies that were deemed
acceptable by EPA for the purpose of deriving freshwater aquatic
life criteria, as presented in Sections 3.1 and Appendix A of the
draft criteria document. Our review included comparison of the
EC/LC50s endpoints reported in draft criteria document with those
found in the original studies, evaluation of whether the
inclusion/exclusion of studies were consistent with the 1985
Guidelines, and review of whether the test conditions for each
study were accurately reported.
Number of studies used
Section 3.1.1 states that the dataset of acceptable acute data
includes 118 toxicity tests encompassing 20 freshwater species,
representing 18 genera. Within the "Acute Dataset" tab of the
spreadsheet included with the criteria document (Aluminum
Criteria Calculator V.1.0), only the results of 94 toxicity tests,
encompassing 19 freshwater species, representing 18 genera are
presented. It is understood that some data that were deemed
acceptable were not ultimately included in the Species Mean Acute
Value (SMAV) calculation for a number of reasons provided in
Sections 3.1 and 5.1 (e.g., results were considered outliers).
Thank you for your suggestions. Additional rows were added
to the Aluminum Criteria Calculator so that the "Acute
Dataset" tab will match Appendix A and "Chronic Dataset"
tab will match Appendix C. These additional rows were in
fact not used in the SMAV/SMCV calculations so they were
originally omitted for ease of development of the calculator.
Aluminum Criteria
Calculator "Acute
Dataset" and "Chronic
Dataset" tabs
3
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 1: Rciiiinlinii iiculc (o\ici(\ d;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
Furthermore, ire noted some differences in toxicity data between
Appendix A in the criteria document, and the "Acute Dataset" tab
(also labeled Appendix A) in the Aluminum Criteria Calculator
V.1.0 spreadsheet. Some of these differences are discussed further
below; we suggest EPA provide additional justification where
needed to ensure the acceptable toxicity datasets are consistent
and defensible.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of data from Call et aL 1984
Not all available data from the Call et al. study were utilized by
EPA for the acute database, including:
The acute database included toxicity results for snail
(Physa sp.), stonefly (Acroneuria sp.), fathead minnow
(Pimephales promelas), and green sunfish (Lepomis
cyanellus). Rainbow trout (Oncorhynchus mykiss) data
were included, but were not used for the SMA V derivation
because these were static tests, and flow-through data
were available from Gundersen et al. 1994.
However, channel catfish (Ictalurus punctatus) and
yellow perch (Perca flavescens) data were not included
with no explanation provided by EPA for their rejection.
The criteria document does mention channel catfish as an
example of a recreationally important species and data
from this species should be considered for its
acceptability.
The channel catfish and yellow perch test results were not
used because each test employed only two exposure
concentrations (plus a control), and only six fish per
treatment and the data are not acceptable for criteria
derivation. These studies and their deficiencies are identified
in Appendix J.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
The 2009 study also included a series of seven acute D. magna
tests with variable pH and DOC.
Six of these studies were deemed acceptable, though only
two were included in the SMAV calculation. It is
presumed the others were not included because no
toxicity was observed and the highest concentrations
tested were also relatively low (e.g., 500 jug/L), and thus
would qualify as a " 'greater than' low acute value "
exclusion per Section 3.1.1. If so, further clarification
would be helpful.
The one LC50 that could be calculated that was retained
for SMA V calculation, 795 jug/L, does not correspond to
any of the D. magna results we can observe from the
original report. In the original study, two tests, both
tested at pH 8, 165 mg/L hardness, 0 mg/L DOC, had
measurable toxicity, resulting in LC50s of 787.8 jug/L and
720.8 jug/L, respectively. It is unclear whether the LC50
presented in Appendix A refers to either of these studies,
Use of "greater than" values follows the "decision rule" as
described in the final aluminum criteria document (Section
3.1), as follows: "greater than" (>) low chronic values and
"less than" (<) high chronic values were not used in the
calculation of the SMCV; but "less than" (<) low chronic
values and a "greater than" (>) high chronic values were
included in the SMCV. This approach was also followed for
acute SMAV calculations. The methodology is based on the
finding that "greater than" values for concentrations of low
magnitude, and "less than" values for concentrations of high
magnitude do not generally add significant information to the
toxicity analysis. This approach to the use of unbounded
values was initially described in the 2013 Aquatic Life
Ambient Water Quality Criteria for Ammonia in Freshwater
and has continued to be applied in subsequent criteria.
All seven studies reported were evaluated. The missing value
(720.8 |ig/L) displayed a poor concentration-response
relationship, so it was deemed unacceptable for use.
Appendix A
Appendix K
4
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 1: Rciiiinlinii iiculc l«i\icil> (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
and it is unclear why all seven D. magna lesls, rather
than just six, were not evaluated for acceptability.
The D. magna LC50 of500 jug/L that was included for
SMAV derivation should be marked as unbounded (i.e.,
"> ") in the "Acute Dataset" tab of the criteria document
spreadsheet.
The 795 |ig/L value is a recalculated value of the author-
reported value of 787.8 |ig/L. This was recalculated because
the raw data reported also indicated a less than optimal
concentration-response. The recalculated value (using TRAP)
was used instead of the author reported value because it is
more appropriate and better fit the empirical data.
Thank you for this correction. This was an error and the LC50
of 500 |ig/L that was included for SMAV in the Aluminum
Criteria Calculator was corrected as listed as >500 ng/L.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use ofC. dubia data from European Aluminium Association 2010
The 2010 data collected by the European Aluminium Association
were for the purpose of evaluating the effects of buffers and the
presence of phosphate on aluminum toxicity, to aid in method
development for aluminum toxicity exposures. Application of these
results for criteria derivation is not recommended in some cases,
and no acute LC50s were presented in the original studies.
Phosphate was a confounding factor in two of these
exposures. Phosphate may competitively bind with
aluminum, providing a protective effect against aluminum
toxicity to invertebrates. These data should not be
included for criteria derivation.
The use of synthetic buffers (e.g., HEPES, MES) is
important for stabilizing pH over the exposure duration.
For example, this study compared to toxicity of twoAl test
solutions pre-adjusted to pH 6, one using the synthetic
buffer of MES (data were not included in Appendix A of
the criteria document) and the other using HCl (data
were included). In the MES-buffered solution, pH
changed at most by 0.02 SU over the duration, while in
the HCl adjusted solution, pH changed by over 1.0 SU in
each of the test treatments. Indeed, the results of the HCl-
adjusted solution in Appendix A, the test condition is
listed as 7.08, though the starting pHwas <6 in each of
the treatments.
Given this large potential for pH drift in the unbuffered
tests, and its potential effect on aluminum speciation,
these results should not be considered for criteria
derivation. However, it may be helpful to include a
discussion of these patterns elsewhere in the text (e.g.,
5.1.1) to further emphasize the importance of pH control
LC50s were calculated for many of these studies, and where
appropriate, included in Appendix A.
Tests conducted with a phosphate buffer were removed as
you suggested. Thank you for your comment. The EPA
agrees that phosphate may competitively bind with aluminum
and these data should not be included for criteria derivation.
Tests conducted where the exposure solution was not
buffered are retained because the pH drift was not well
explained for many of studies. In addition, if only pH
buffered tests are retained, the database for aluminum criteria
development would be very limited. Additional text has been
added to the document regarding pH drift during the test
exposure.
Section 2.3
Appendix A
5
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I oil lopic 1: Rciiiinlinii iiculc 1 o \ i c i I > d;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
in the selection of acceptable toxicity tests with aluminum.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of data from Lamb and Bailey 1981,1983
Acute toxicity tests using the midge, Paratanytarsus dissimilis
(Lamb and Bailey 1981,1983) were included, but we recommend
that EPA reconsider their inclusion for the reasons indicated
below.
The original report states that after 96 hours, "no
apparent effects" were observed, and that "the larvae,
including controls, were generally active and they
exhibited typical movements and food searching. "
The endpoint measured after 96 hours is not clearly
defined in the original study and no statistics are
provided to determine whether any "apparent effects "
were statistically significant. We ask EPA to review
whether these data are acceptable for inclusion for
criteria calculation. The results of study suggest that this
species is relatively insensitive to aluminum, and removal
of this species would not affect the acute criterion
outcome.
The LC50 is a greater than value due to the reasons stated.
These are not reasons for exclusion as recommended by the
1985 Guidelines. Use of "greater than" values will follow the
approach described in the "decision rule" as described in the
2018 Final Aluminum Criteria document in Section 3.1 as
follows: "greater than" (>) low chronic values and "less than"
(<) high chronic values were not used in the calculation of the
SMCV; but "less than" (<) low chronic values and a "greater
than" (>) high chronic values were included in the SMCV
(U.S. EPA 2013). This approach was also followed for acute
SMAV calculations. This approach to the use of unbounded
values was initially described in the 2013 Aquatic Life
Ambient Water Quality Criteria for Ammonia in Freshwater
and has continued to be applied in subsequent criteria.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of smallmouth bass (Micropterus dolomieui) data from Kane
and Rabeni 1987
The results from three exposures at pH 5.05, 6.75, and 7.45,
respectively, were deemed acceptable to be included in Appendix
A. Toxicity was only observed in the pH 5.05 test, and only this
result was used as the basis of the SMAV/GMAV calculation.
It is presumed that the results from the pH 6.75 and pH
7.45 exposures were not included because they were
unbounded and would qualify as a " 'greater than' low
acute value " exclusion per Section 3.1.1. Confirmation of
this exclusion would be helpful.
As discussed earlier, we believe it is questionable to use
the MLR to normalize data outside the range pH 6 to 8.
This is because different forms of aluminum dominate
outside this range and the mechanisms of toxicity are
likely to differ as well.
Under a number of water quality scenarios, smallmouth
bass is one the four most acutely sensitive species, and
thus the inclusion of the results from this one pH 5.05
study has large effects on the ultimate acute criterion.
Given that the acute effects of aluminum for this species
have not been well characterized at circumneutral pH, it
may be questionable to use this one study to predict
As the commenter presumed, the Micropterus dolomieui test
result at pH 7.45 was not used to calculate the SMAV for the
species as specified by the "greater than" decision rule
(Section 3.1.1). The test result at pH 6.25 is used in the
SMAV calculation in the final version of the AWQC.
The pH of toxicity test waters for the MLR in the 2018 final
document for Pimephales promelas toxicity test data ranged
6.0-8.12 for pH. The EPA included some tests beyond these
pH values for criteria derivation. The criteria calculator can
be also used to address all waters within a pH range of 5.0 to
10.5. This approach was taken so that the recommended
criteria can be provided for, and will be protective of, a
broader range of U.S. natural waters. Extrapolated criteria
values outside of the empirical pH data tend to be lower
values and will be more protective of the aquatic environment
in situations where pH plays a critical role in aluminum
toxicity.
No edits.
6
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 1: Rciiiinlinii iiculc 1 o \ i c i I > d;il;i
I'.I'A Response
Revision l.ociilion in
20IX Aliiniiiiiiin
Crilcriii Document
toxicity over the full range ofpll conditions for which the
criteria would apply.
7
-------
TOPIC 2: Comments regarding alum (aluminum sulfate) used for pollution control
Comment
Nil in Ikt
(Oi'Uiini/iilioii)
Public C oiiimoil 1 on l opic 2: Rciiiudinii ilium ciliiiiiiiiiini
sull'iilc) used lor pollution control
I'.I'A Response
Ke\ision Location in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
4. Exclusion for aluminum additions to address high priority
Waste Load Allocations
Several NPDES permittees are involved in programs that add alum
(aluminum sulfate) to receiving waters as part of an "offset"
program designed to achieve compliance with TMDL waste load
allocations for phosphorus. The alum is used to counteract
increased levels of phosphorus, which is often the controlling
nutrient. Excess phosphorus can increase algae, impair aesthetics
and recreation, create odor problems, and promote the formation
of unwanted byproducts during drinking water treatment.
Cyanotoxins also may present a significant risk to aquatic
organisms. The alum effectively sequesters the reactive mobile
phosphorus in the waterways. The following nutrient TMDLs may
potentially consider or are using aluminum compounds to control
phosphorus.
[TABLE]
In the absence of identifiable adverse effects from the addition of
alum, we request that the standards provide an explicit exception
to the criteria when a significant beneficial use (e.g., nutrient
control, protection of drinking water) is achieved by the aluminum
addition.
The EPA's 2018 aluminum criteria provide recommendations
for states and authorized tribes to protect aquatic life from
potential effects of aluminum. The implementation
documents that the EPA is developing are intended to provide
assistance to states and authorized tribes that adopt into the
water quality standards criteria based on or similar to the
EPA's recommended criterion. The implementation guidance
will describe state flexibilities in implementing the aluminum
criteria. The implementation documents are also intended to
provide assistance to other stakeholders and the public. The
EPA recognizes that there are several aspects of the
recommended criteria that will benefit from technical support
documents to enhance implementation of state and tribal
criteria and is planning to develop such documents and make
them available.
No edits.
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
On March 6, 2017, EPA Region 1 issued a Potable Water
Treatment Facility General Permit (PWTF GP) for Massachusetts
& New Hampshire. This permit stated that discharge limits for
Aluminum would be included in this permit. Many Public Water
Systems use alum (aluminum sulfate) as a coagulant in their
drinking water treatment process and we feel it will be difficult for
them to achieve the current numeric limit while maintaining their
current treatment processes. Many of the receiving waters in New
England, including many high quality, pristine waterways, already
have natural background levels of Aluminum that exceed the
current national water quality standard that is used as the basis
for numeric permit limits. The high levels of background Aluminum
in waters generally considered to be very clean suggest that the
current standard is grossly inaccurate and unnecessarily
overprotective.
For Public Water Systems, coagulant changes (such as to iron-
based coagulants) could be both a costly and lengthy process
8
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 2: Rciiiirdinii iiliini cilu milium
siill'iilc) used lor pollution control
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
(rilcrhi Document
which may have significant operational impacts. Public H ater
Systems that change coagulants to meet Aluminum limits may have
problems with other Safe Drinking Water Act requirements; they
may have to reassess their lead and copper corrosion control
program as one example. In some instances, change in coagulants
have resulted in sudden highly elevated lead levels. It simply does
not make sense to have Public Water Systems potentially
compromising public health, or having to make costly investments
to change coagulants or treatment processes, so that they do not
exceed an arbitrary water quality standard, which even if
exceeded, does not appear to be causing environmental harm in
our area. Nor does it make sense for wastewater treatment plants
serving communities across the state to spend their limited funds
trying to reduce Aluminum in treated discharges with no
environmental benefit to be gained.
We have reviewed EPA's proposal and believe that the changes
proposed are beneficial and should move forward, however, we do
offer the following comments for EPA's consideration before the
new criteria is finalized:
EPA-HQ-OW-
2017-0260-0053
(Abdul Alkhatib,
Director,
Massachusetts Water
Works Association
(MWWA))
On March 6, 2017, EPA Region 1 issued a Potable Water
Treatment Facility General Permit (PWTF GP) for Massachusetts
& New Hampshire. This permit stated that discharge limits for
Aluminum would be included in this permit. Many Public Water
Systems use alum (aluminum sulfate) as a coagulant in their
drinking water treatment process and we feel it will be difficult for
them to achieve the current numeric limit while maintaining their
current treatment processes. For Public Water Systems, coagulant
changes (such as to iron-based coagulants) could be both a costly
and lengthy process which may have significant operational
impacts. Public Water Systems that change coagulants to meet
Aluminum limits may have problems with other Safe Drinking
Water Act requirements; they may have to reassess their lead and
copper corrosion control program as one example. In some
instances, change in coagulants have resulted in sudden highly
elevated lead levels. It simply does not make sense to have Public
Water Systems potentially compromising public health, or having
to make costly investments to change coagulants or treatment
processes, so that they do not exceed an arbitrary water quality
standard, which even if exceeded, does not appear to be causing
environmental harm in our area. Nor does it make sense for
wastewater treatment plants serving communities across the state
9
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 2: Rciiiirdinii iiliini cilu milium
sull'iilc) used lor pollution control
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
(rilcrhi Document
to spend their limited funds trying to reduce Aluminum in treated
discharges with no environmental benefit to be gained.
EPA-HQ-OW-
2017-0260-0074
(Timothy F. Moore,
Risk Sciences, on
behalf ofLake Elsinore
and Canyon Lake
Nutrient TMDL Task
Force administered by
the Lake Elsinore San
Jacinto Watershed
Authority (LESJWA))
Thank you for the opportunity to review EPA's Draft Updated
Aquatic Life Ambient Water Quality Criteria for Aluminum in
Freshwater (EPA-822-P-17-001) published in July of 2017. The
following comments are submitted on behalf of the Lake Elsinore
and Canyon Lake Nutrient TMDL Task Force ("Task Force ")
administered by the Lake Elsinore San Jacinto Watershed
Authority (LESJWA).
Lake Elsinore and Canyon Lake are located in western Riverside
County, California. Both lakes are on the state's 303(d) list of
impaired waters due to excess algae caused by elevated nutrient
concentrations. State and federal authorities established a TMDL
for these lakes in 2005. The task Force was formed as a
collaborative partnership among local stakeholders (principally
MS4 permittees, agricultural operators and POTWs) to comply
with the TMDL by implementing large-scale water quality
improvement projects in the watershed and in the lakes.
Two of the most effective water quality improvement projects rely
on judicious use of aluminum sulfate (aka "Alum") to reduce
phosphorus loading in both lakes. The Task Force is deeply
concerned that, for reasons described below, the proposed water
quality criteria may severely restrict future applications of
aluminum-based compounds such as Alum to waters of the U.S.
The unintended consequence would be to reduce rather than
enhance protection of designated beneficial uses. For this reason
the Task Force recommends that the draft criteria be revised to
distinguish between beneficial and detrimental forms of aluminum.
EPA-HQ-OW-
2017-0260-0074
(Timothy F. Moore,
Risk Sciences, on
behalf ofLake Elsinore
and Canyon Lake
Nutrient TMDL Task
Force administered by
the Lake Elsinore San
Jacinto Watershed
Authority (LESJWA))
The Task Force supports the Multiple Linear Regression (MLR)
approach that EPA used to develop the draft aluminum criteria
and believes it represents a significant improvement over the
304(a) criteria that was promulgated 30 years ago. The MLR
adjusts for several water chemistry factors, such as pH, hardness
and dissolved organic carbon (DOC), that have been shown to
mitigate the potential for aluminum toxicity. However, the model
does not yet include a similar adjustment for phosphorus - an
equally important mitigating factor that governs the potential
toxicity of aluminum.
Aluminum readily binds with phosphorus to form aluminum
10
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 2: Rciiiirdinii iiliini cilu milium
siill'iilc) used lor pollution control
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Criloriii Document
phosphate a fad already acknowledged in the draft criteria
document. [Draft Criteria @ pg. 13] This chemical bond occurs
quite rapidly and is virtually insoluble (Ksp = 6.3 x 10'19) under
natural stream conditions. [Ksp = Solubility Product Coefficient
(https://en.wikipedia.org/wiki/Aluminum_phosphate)] When alum
is properly applied, the resulting aluminum phosphate molecule is
inert and no longer bioavailable. For this reason, alum is
increasingly used to minimize phosphorus concentrations in
wastewater discharges and, more recently, to mitigate some of the
excess phosphorus contributed by the natural sediments in lakes
and reservoirs. [Draft Criteria @ pg. 3 and pg. 7] The Task Force
uses alum in both ways.
In Canyon Lake, alum is applied to remove and sequester
phosphorus from the water column. This program, which has been
underway for nearly five years, is expected to help assure
compliance with the TMDL targets for phosphorus and
chlorophyll-a by the 2020 deadline. Without this program, it is
unlikely that Canyon Lake would ever achieve the TMDL targets
because lake bottom sediments are, by far, the dominant source of
phosphorus to the water column and this phosphorus has a very
long half-life (10-15years). [Anderson, MA. Technical
Memorandum: Estimate Rate at Which Phosphorus is Rendered
No Longer Bioavailable in Sediments of Canyon Lake and Lake
Elsinore. Dec. 31, 2011] Alum is the only cost effective method for
addressing these significant non-point source loads.
Lake Elsinore is the largest freshwater lake in southern California.
Until recently, limited rainfall and natural evaporation caused the
lake to dry-up every 25-30years. Today, approximately 6 mgd of
recycled water is added to Lake Elsinore to offset evaporation.
Various aluminum-based compounds are used to reduce
phosphorus concentrations during the wastewater treatment
process. Without these compounds, the recycled water would be
unable to comply with the TMDL's waste load allocation for
phosphorus and could no longer be legally discharged to Lake
Elsinore. Without recycled water, there is nothing to prevent Lake
Elsinore from disappearing completely during the recurring
droughts that commonly afflict this area. At such times, all of the
designated recreational and aquatic habitat uses will be lost.
Alum can only be applied in accordance with a NPDESpermit.
11
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 2: Rciiiudinii iiliini cilu milium
sull'iilc) used lor pollution control
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Criloriii Document
[Federal Water Polliilion Control Ad, §311(et. seq.j and 40 CTR
116.4] The Task Force is concerned that some of the statements
made in the draft criteria document may make may make it
virtually impossible for state authorities to permit such
applications in the future. EPA referenced only one field study on
the use of alum to control phosphorus and emphasized that this
particular study reported significant adverse effects on
invertebrate populations following the alum application. [Draft
Criteria @ pg. 63 (referring to Barbiero et al 1988)] This is
somewhat misleading because the authors of the study warned that
these adverse effects may have been due to the unusual way alum
was continuously applied for 35 consecutive days which, in turn,
resulted in over-saturation and incomplete complexation. The
authors also concluded that: "since continuous application of
aluminum sulfate exposes downstream communities to continuous,
fresh solutions of aluminum in which polymerization of the
hydroxide and complexation with organics are incomplete, the
response of affected communities would be expected to differ from
those exposed to a single alum application treatment such as a
lake treatment. " [Barbiero, R., R.E. Carlson, G.D. Cooke &A. W.
Beals. The Effects of Continuous Application of Aluminum Sulfate
on Lotic Benthic Invertebrates. Lake and Reservoir Management.
4:2pgs. 63-72 (1988)]
EPA-HQ-OW-
2017-0260-0074
(Timothy F. Moore,
Risk Sciences, on
behalf ofLake Elsinore
and Canyon Lake
Nutrient TMDL Task
Force administered by
the Lake Elsinore San
Jacinto Watershed
Authority (LESJWA))
In short, the aluminum in alum is a special case and should be
treated as such. Perhaps it would be best to regulate alum
applications under F1FRA using the registration and labeling tools
that EPA purposely designed to balance the risks and benefits of
using potentially toxic substances in the environment. Alum is
already on the 4B list of "other inert ingredients for which EPA
has sufficient information to reasonably conclude that the current
pattern of use in pesticide products will not adversely affect public
health or the environment. " [https://www.epa.gov/pesticide-
registration/categorized-lists-inert-ingredients-old-lists]
Alum has been used to purify drinking water for more than 2,000
years. Today, it is used by thousands of permitted dischargers to
enhance wastewater treatment and protect the environment. It is
essential that EPA distinguish between the beneficial and
detrimental forms of aluminum in order to avoid unintended
consequences when the proposed 304(a) criteria is later used to
establish state water quality standards and related waste discharge
requirements.
12
-------
TOPIC 3: Comments regarding the Aluminum Criteria Calculator
( omilKMII
Number
(()ri*;ini/;ilion)
Public C onimoil 1 on l opic 3: Rciiiudinii (lie Aluminum
( riloriii ( ;ilcul;Kor
I-'.I'A Response
Kc\ision Location in
20IX Aluminum
( riloriii Document
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed-
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQfWQD))
WDEQ/WQD also appreciates the look-up tables provided in
Appendix K and the accompanying Aluminum Criteria Calculator
V.1.0 spreadsheet used to determine criteria values and taxa
sensitivity based on various pH, hardness and DOC values.
Though these tools are also helpful, WDEQ/WQD questions
whether they can be easily adjusted to delete species/genera in
order to facilitate site-specific criteria development. WDEQ/WQD
requests that EPA describe how to approach site-specific
aluminum criteria using these tools and provide a user manual
detailing the various functions and capabilities of the spreadsheet
calculator.
Thank you for your suggestion. The Aluminum Criteria
Calculator will still be locked to ensure version control.
Please work with your local EPA Region and the EPA
Headquarters' staff to develop site-specific criteria values
(i.e., add/delete species/genera) on a case-by-case basis, when
appropriate.
Additionally, another tab will be added to the Aluminum
Criteria Calculator that provides instructions.
Aluminum Criteria
Calculator new tab
entitled "Read Me"
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC))
3. EPA has provided states with an interactive aluminum criteria
calculator to assist when generating revised aluminum criteria per
revised acceptable acute and chronic studies. The calculator's
upper limits for pH and hardness result in parameters which
extend outside of the model's input capacity to provide the user
with modelled output parameters that are certain to be protective.
ADEC questions the use of the aluminum criteria calculator when
the model will allow the user to enter parameters that extend
beyond the range of empirical data used for model development. In
addition, ADEC would like EPA to provide further clarification on
the certainty of the values found in Appendix K, which provide
criteria for various water chemistry conditions that are outside of
the model input parameters and how states should justify their use
of these parameters for site-specific criteria development.
Since the draft document was released, additional toxicity test
were conducted with Ceriodaphnia dubia and Pimephales
promelas thereby expanding the water chemistry empirical
data used for model development.
As a result, the water chemistry bounds for the 2018 criteria
were thus expanded, with details and rationale provided in the
criteria document and summarized below. The criteria
calculator can be used to address waters within a pH range of
5.0 to 10.5. For hardness values, the criteria calculator allows
entry of values between 0.01 and 430 mg/L total hardness;
criteria magnitudes will not increase or decrease by
increasing the hardness above 430 mg/L total hardness (as
CaC03). For DOC, the criteria calculator will not extrapolate
below the lowest empirical DOC of 0.08 mg/L and upper
limit of the empirical MLR models will be bounded at a
maximum 12.0 mg/L DOC in the criteria calculator; criteria
magnitudes will not increase or decrease by increasing the
DOC above 12.0 mg/L.
The pH of toxicity test waters ranged from 6.0-8.7. The EPA
has determined that for pH users may extrapolate beyond
these values for criteria derivations. The criteria calculator
can be used to address all waters within a pH range of 5.0 to
10.5. Thus, criteria values for pH input values beyond the
range of the underlying empirical pH data used for model
development (pH 6.0 to 8.7) can be generated using the
criteria calculator. (This is also reflected in the criteria lookup
tables in Appendix K of the 2018 Final Aluminum AWQC
document.) The EPA took this approach for pH so that the
Text, tables and MLR
equations edited to
incorporate new
toxicity data
throughout the
document.
13
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 3: Rciiiudinii (lie Alii iiiiini in
(rilorhi (;ilcul;Kor
I.PA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
rccoimiicndcd criteria arc a\ a liable lor prolccliv c ol a broader
range of U.S. natural waters. Extrapolated criteria values
outside of the empirical pH data tend to be more protective of
the aquatic environment (i.e., lower criteria values) in
situations where pH plays a critical role in aluminum toxicity.
However, criteria values generated outside of the range of the
pH conditions of the toxicity tests underlying the MLR
models are more uncertain than values within the pH
conditions of the MLR toxicity tests, and thus should be
considered carefully and used with caution.
The total hardness of toxicity test waters underlying the MLR
models ranged from 9.8 to 428 mg/L. Since a decrease in
total hardness tends to increase aluminum toxicity, the EPA
has determined it is reasonable to extrapolate on the lower
bound of the hardness data to enable generation of lower
criteria at low hardnesses beyond the limit of the empirical
data. Thus, hardness input values in the criteria calculator can
be entered that are less than 9.8 mg/L down to a limit of 0.01
mg/L. This is consistent with existing EPA approaches to low
end hardness (U.S. EPA 2002). However, criteria values are
bounded at the approximate upper limit of the empirical MLR
models' underlying hardness data, at a maximum of 430
mg/L total hardness (as CaC03). The user can input hardness
values into the criteria calculator that are greater than 430
mg/L for total hardness, but the criteria magnitude will reach
its maximum value at 430 mg/L total hardness (as CaC03),
and criteria magnitudes will not increase or decrease by
increasing the hardness above 430 mg/L total hardness (as
CaC03). This is also consistent with existing EPA guidance
on high end hardness "caps" (U.S. EPA 2002). (These total
hardness bound approaches are also reflected in the criteria
lookup tables in Appendix K of the 2018 Final Aluminum
AWQC document.) The EPA took this approach so that the
recommended criteria can be provided for, and will be
protective of, a broader range of U.S. natural waters. Criteria
values generated beyond the lower bound of the hardness
conditions of the toxicity tests underlying the MLR models
are more uncertain than values within the hardness bounds of
the MLR toxicity test data.
The DOC of toxicity test waters ranged from 0.08 to 12.3
14
-------
( omilKMII
Number
(Oi^;ini/;ilion)
Public C oiii moil I oil lopic 3: Rciiiirdinii (lie Aluminum
Crilcriii (;ilcul;ilor
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
nig,L. Since mobl iialurul w aler^ contain mjiiic DOC, llie
lower bound of the empirical toxicity test data (0.08 mg/L) is
the lowest value that can be entered into the criteria
calculator; thus no extrapolation below the lowest empirical
DOC of 0.08 mg/L is provided. The criteria values generated
with the criteria calculator are bounded at the upper limit of
the empirical MLR models' underlying DOC data: at 12.0
mg/L DOC. The user can input DOC values greater than 12.0
mg/L into the calculator, but criteria magnitudes will not
increase or decrease by increasing the DOC above 12.0 mg/L.
This is also reflected in the criteria lookup tables in Appendix
K of the 2018 Final Aluminum AWQC document. This is
consistent with the existing approach for hardness (U.S. EPA
2002) to provide for protection of aquatic organisms through
the use of protective, conservative values when water
chemistry conditions are beyond the upper limits of the
empirical toxicity test data.
Please work with your local EPA Region and Headquarters'
staff to regarding any refinements sought for situations where
water chemistry for a particular water falls outside the bounds
of the model.
EPA-HQ-OW-
2017-0260-0062
(John St. Clair,
Rosebud Mining
Company)
Another concern with the proposed Criteria for Aluminum is the
limitations of the Aluminum Criteria Calculator V.l.O.xlsx. The
calculator does not allow for hardness, DOC or pH values outside
a certain range. It is unclear how the limits for aluminum will be
established for water chemistries outside the calculator range.
This limitation will directly impact discharges that contain pH
variance above 9.0 in impaired streams. Typically pH variance up
to 10.0 are given to discharges for the treatment of manganese and
in receiving streams with suppressed pH levels due to legacy AMD
discharges. While the goal with pH variances is to improve water
quality, this benefit may be impacted by restrictions placed on
aluminum levels.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development.
The bounds for pH of the models ranged from 6.0-8.7. The
EPA is allowing the user to extrapolate beyond the pH values
used to generate the MLR models. The criteria calculator can
be used to address all waters within a pH range of 5.0 to 10.5.
Text, tables and MLR
equations edited to
incorporate new
toxicity data
throughout the
document.
EPA-HQ-OW-
2017-0260-0046
(Jennifer Wigal,
Program Manager,
Water Quality
Standards &
Assessments, Oregon
Department of
Environmental
Quality)
Implementation
Since the criteria values cannot be determined without use of the
spreadsheet calculator provided by EPA, it is critical that EPA
provide a calculator capable of receiving input for more than 20
sets of the input parameters at a time. States have a need to
calculate site-specific criteria values for hundreds of samples when
assessing aluminum for Integrated Reporting purposes. A
calculator that has room to input at least 500 sets of input samples
Another tab was added to the Aluminum Criteria Calculator
(Over 20 Scenarios). This tab will allow the user to enter
input data for 500 samples.
Aluminum Criteria
Calculator new tab
"Over 20 Scenarios"
15
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 3: Rciiiirdinii (lie Aluminum
Crilcriii (;ilcul;ilor
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
is needed. DEQ hopes that EPA will make a new calculator
incorporating these improvements available as soon as possible.
In conclusion, DEQ agrees that the new 304(a) criteria guidelines
and calculator for aluminum is an improvement over the 1988
guidelines. We recognize there is a lack of available data to
develop the criteria to more fully reflect diverse environmental
conditions and species responses. EPA should seek to expand the
boundaries of the model for all parameters with toxicity data that
accounts for additional species across a more representative range
of the natural water conditions that are likely to be encountered in
the states.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Appendix A spreadsheet
Comment: suggest adding footnote identifier for "a" associated
with Method.
Footnote added to the "Acute Dataset" tab (Appendix A) of
the Aluminum Criteria Calculator.
Aluminum Criteria
Calculator "Acute
Dataset" tab
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Appendix A spreadsheet
Comment: Add dilution water description to each test, as this data
is available from the references.
Dilution water information was added to the Aluminum
Criteria Calculator.
Aluminum Criteria
Calculator "Acute
Dataset" tab
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Appendix C spreadsheet
Comment: Add dilution water description to each test, as this data
is available from the references.
Dilution water information was added to the Aluminum
Criteria Calculator.
Aluminum Criteria
Calculator "Chronic
Dataset" tab
16
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic 3: Rciiiirdinii (lie Aluminum
Crilcriii (;ilcul;ilor
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Appendix C spreadsheet
Comment: Please provide references column for each study.
References were added to the Aluminum Criteria Calculator
"Chronic Dataset" tab.
Aluminum Criteria
Calculator "Chronic
Dataset" tab.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
(I) The EPA's spreadsheet tool should be revised so that it provides
a warning when an input value is outside the MLR range but it
should not censor or change such values on its own without
additional user-authorization.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development, as noted in the
response above.
The EPA chose to apply the criteria value bounding
approaches selected so that recommended criteria can be
provided and that they will be protective of, a broader range
of U.S. natural waters.
Text, tables and MLR
equations edited to
incorporate new
toxicity data
throughout the
document.
17
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic 3: Rciiiirdinii (lie Alii iiiiini in
Crilcriii C;ilcul;ilor
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Application of the Criteria from pH 5.0 to 9.0
A key revision included in the updated criteria is the expansion of
the pH range over which the criteria apply. The current Nationally
Recommended criteria for aluminum (EPA 1988) apply from 6.5 to
9.0, while the updated criteria extend the range down to pH 5.0.
This is significant for several reasons, The Nationally
Recommended Water Quality Criteria for pH (EPA 1986) for
freshwater is 6.5 to 9. Does it make practical sense to apply any
metals criteria to pH values outside the range used in the pH
criteria? It may help for EPA to provide additional explanation
regarding the regulatory significance of any aluminum criteria
outside this pH range to help states determine how to implement
these criteria.
The speciation of aluminum changes considerably from pH 5 to 6,
which also affects the mode by which aluminum elicits toxicity to
aquatic organisms. At pH 6, insoluble aluminum hydroxides are
expected to dominate which may smother gill surfaces thereby
limiting respiratory exchange. At pH < 6, dissolved ionic and
monomeric species of Al are more abundant, and will affect
organisms by a different mechanism by binding to gill tissues and
disrupting ionoregulatory function. Furthermore, aluminum
solubility increases at pH >8 compared to circumneutral
conditions, and the speciation of dissolved aluminum is dominated
by the aluminate anion, rather than either cationic forms or
neutral hydroxides which dominate a lower pH. The mechanisms
of toxicity at these elevated pH levels are less well understood.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. As a result, the
water chemistry bounds for the 2018 criteria were thus
expanded, with details and rationale provided in the criteria
document and summarized below. The criteria calculator can
be used to address waters within a pH range of 5.0 to 10.5.
No edits.
18
-------
TOPIC 4: Comments regarding aluminum not being a priority pollutant
( omilKMII
Number
(()ri*;ini/;ilion)
Public C onimoil 1 on l opic 4: Rciiiudinii ;ilniiiiiniin mil hciuii ;i
priority pcillu 1 ;in 1
I'.I'A Response
Kc\ision Location in
20IX Aluminum
( riloriii Document
EPA-HQ-OW-
2017-0260-0036
(Barry N. Burnell,
Water Quality Division
Administrator, State of
Idaho Department of
Environmental Quality
(DEQ))
Thank you for the opportunity to provide comments to the EPA on
the Draft Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater. DEQ understands that while EPA has
provided this update, aluminum is not considered a priority
pollutant, and that states are not currently required to consider
these recommended criteria.
Thank you for your comments. You are correct that EPA has
not identified aluminum as a priority pollutant. However, 40
CFR 131.20 states ".. .if a State does not adopt new or revised
criteria for parameters for which EPA has published new or
updated CWA section 304(a) criteria recommendations, then
the State shall provide an explanation when it submits the
results of its triennial review to the Regional Administrator
consistent with CWA section 303(c)(1)..."
No edits.
EPA-HQ-OW-
2017-0260-0069
(Julia Young, Water
Quality Standards
Coordinator, Kansas
Department of Health
and Environment
(KDHE))
The Kansas Department of Health and Environment, Bureau of
Water (KDHE) appreciates the opportunity to comment on the
Draft Aquatic Life Ambient Water Quality Criteria for Aluminum
2017 (draft aluminum criteria guidance).
Comments on Proposed Standards:
1) KDHE supports the development of criteria using site-specific
water chemistry (aluminum, pH, hardness and DOC), because it
will allow more realistic criteria limits to be established, than the
one-size-fits-all approach of the 1988 aluminum freshwater
aquatic life criterion. The fact that adoption is optional and not
mandatory because it is not a priority pollutant is also
appreciated.
Thank you for your comments. EPA has not identified
aluminum as a priority pollutant, and therefore states are not
required to develop state water quality standards for that
pollutant. 40 CFR 131 20 states ".. .if a State does not adopt
new or revised criteria for parameters for which EPA has
published new or updated CWA section 304(a) criteria
recommendations, then the State shall provide an explanation
when it submits the results of its triennial review to the
Regional Administrator consistent with CWA section
303(c)(1)..."
No edits.
EPA-HQ-OW-
2017-0260-0074
(Timothy F. Moore,
Risk Sciences, on
behalf ofLake Elsinore
and Canyon Lake
Nutrient TMDL Task
Force administered by
the Lake Elsinore San
Jacinto Watershed
Authority (LESJWA))
Finally, the Task Force recommends that EPA add an
"Implementation" section to the draft document. This section
should note that Aluminum is not a Priority Pollutant metal like
those covered by the National Toxic Rule or the California Toxics
Rule. It should also explain the range of alternatives available for
integrating the proposed 304(a) criteria into state water quality
standards, including the option to implement it through existing
narrative standards.
A sub-section of the Implementation chapter should be devoted to a
discussion of how to permit the use of alum in the context of TMDL
compliance programs. Of particular concern is whether the
chronic criteria (CCC) should even be applied to individual alum
applications in lakes and reservoirs. Similarly, EPA should
carefully consider whether the 1 hour exposure assumption (CMC)
or once-in-three-years exceedance interval are appropriate where
alum is being used to bind and sequester phosphorus.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water quality
standards under CWA section 303(c). The EPA does not
include implementation sections in criteria documents,
because the criteria recommendations are based strictly on
scientific determinations regarding toxicity.
The separate implementation documents that the EPA is
developing are intended to provide assistance to states and
authorized tribes that adopt into the water quality standards a
criterion based on or similar to the EPA's recommended
criterion.
No edits.
19
-------
TOPIC 5: Comments regarding BLM Approach
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Com nun 1 on Topic 5: Rciiiirdinii lil.M iippronch
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
( rilerhi Document
EPA-HQ-OW-
2017-0260-0020
(Jon Tack, Chief,
Water Quality Bureau,
Iowa Department of
Natural Resources
(DNR))
3. BLM Approach
The draft criteria use regression models to characterize the impact
of water chemistry (pH, dissolved organic carbon and hardness)
on aluminum toxicity. The Biotic LigandModel has shown to be
the more accurate approach to predict metal toxicity. Iowa
suggests that the EPA also provide the BLM model option for
states to use.
The Aluminum AWQC are recommendations. States may
choose other scientifically defensible methods to develop
aluminum criteria.
We do not agree that the BLM is a more accurate approach
than a MLR model to predict metal toxicity. Current research
indicates that the MLR and Biotic Ligand models have
comparable performance in predicting aquatic toxicity for
several chemicals, as long as both models are well-
constructed and are supported with sufficient data. For
example, Brix et al (2017) concluded that the MLR and BLM
models' performance for copper were comparable across a
wide range of water chemistries and species (Environ. Sci.
Technol., 2017, 51(9): 5182-5192). Furthermore, the
aluminum BLM we are familiar with does not include all the
new available data we have included and has not been
finalized at this time.
No edits.
20
-------
TOPIC 6: Comments regarding chronic toxicity data
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Comment on Topic (>: Rciiiirdinii chronic 1 o \ i c i I > ri;il;i
T.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rilcriii Document
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
Overall Use of the Chronic Database
The chronic database is limited and has serious deficiencies. US
EPA pretends that the limited chronic database is meaningless, but
it directly affects the calculation of the FCV. Moreover, the studies
for the three most sensitive species have fundamental flaws and
inconsistencies that must be resolved. The normalization of the
data based on the MLR is also questionable. To complete the
normalization process, US EPA compiled the pH, hardness, and
DOC concentrations for the studies in the chronic database. DOC
was measured for only thirteen of the twenty-nine chronic values.
Of these thirteen, more than half were for C. dubia in the
Gensemer study utilized in the development of the MLR, where
DOC was held constant at 0.5 ug/l. The overall range of DOC in
the chronic database was <0.5 mg/1 to 1.9 mg/1, which is very
limited compared to the range of DOC concentrations in the MLR.
Even though hardness is known to have a mitigating effect, the
highest hardness represented in the chronic database is 220 jug/l
for the fathead minnow. Only nine of the twenty-nine chronic
values were based on water with hardness >100 mg/1, even though
many streams have much higher hardness concentrations. US EPA
should reconsider the use of the chronic database and determine
whether the acute to chronic ratio offers a more reliable chronic
criterion.
The EPA disagrees with the commenter. To clarify how the
criteria were developed, EPA notes that there are 2 different
aspects of toxicity data supporting the criteria: 1) the MLR
data normalization data set used to describe how the
bioavailability of aluminum varies across water chemistries;
this MLR data set was expanded as noted since the 2017 draft
criteria to encompass a wider range of ambient water
chemistry conditions, and 2) the sensitivity distribution data,
which is the ecotoxicity dataset normalized with the MLR
model. These normalized data are then applied in the criteria
calculator, following the 1985 Guidelines methods, to
determine the criteria for a given set of water chemistry
conditions.
Normalization of the database with the MLR equations
(relative to pH, total hardness and DOC concentrations)
utilizes the most current scientific information available for
aluminum. Since publication of the draft, additional toxicity
tests were conducted with C. dubia and P. promelas, thereby
expanding the water chemistry empirical data used for model
development. The MLR models applied in criteria document
were developed by an independent expert in the field of
modeling metal toxicity and the model was published in a
peer-reviewed journal. The most important information for
understanding the effects of water chemistry on toxicity is
captured via the MLR model's underlying toxicity dataset,
not the range of conditions in the toxicity tests used to
develop the sensitivity distribution for the criteria calculator.
The range of conditions captured through the MLR's
underlying toxicity dataset for the 2018 final criteria are: total
hardness ranged from 9.8 to 428 mg/L; DOC ranged from
0.08 to 12.3 mg/L; pH of ranged from 6.0-8.7.
The current chronic sensitivity distribution database has a
sufficient number of diverse studies to support criteria
derivation, as recommended by the 1985 Guidelines (8-
family MDR satisfied). The 1988 aluminum freshwater
chronic dataset included 2 species of invertebrates and one
fish species grouped into 3 genera. The 2018 criteria update
includes new chronic data for an additional 9 species and
consists of 8 invertebrate and 4 fish species grouped into 12
No edits.
21
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic (>: Reiiiirdinii chronic Io\ici 1 > ri;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
genera. Willi llic addition of one sludv from Appendix H, die
Minimum Data Requirements (MDRs) for direct calculation
(using a sensitivity distribution, as described in the 1985
Guidelines) of the Final Chronic Value (FCV) were fulfilled.
Use of the ACR method over the 8-family MDR approach
would introduce more uncertainty to the derived chronic
criterion, not less.
All toxicity studies used to derive the criteria are
scientifically sound. The studies were subjected to a two-
level quality review within the EPA, as all studies in criteria
documents always are: first, through the ECOTOX database
scientific quality screen, and second, through the EPA Office
of Water rigorous quality control review as described in the
1985 Guidelines and supporting materials.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WYCA))
Atlantic Salmon and Brook Trout
The two most sensitive species in the US EPA chronic database are
both salmonids, both of which have very limited ranges within the
United States. If these two studies are excluded from the chronic
database, the FCV increases from 394 jug/l to 816 jug/l even when
N (the number of GMCVs in the dataset) is reduced from 12 to 10.
Clearly, these two studies strongly affect the calculated criteria.
The Cleveland brook trout study indicated greater aluminum
toxicity at pH 6.55 (Exposure B) than a nearly identical study
conducted at pH 5.65 (Exposure A). This directly contradicts the
expected results and yielded normalized chronic values that
deviated by 1,000 jug/l. Instead of questioning the disparate
results, US EPA averaged the two chronic values. The control in
Exposure B had higher mortality (10.8%) than most of the test
exposures, indicating an independent factor could have affected
the results of the study. If the Exposure B results are excluded, then
brook trout would no longer be among the four most sensitive
species in the chronic database.
The issue with the Atlantic salmon study is more fundamental. US
EPA selected the normalized chronic value for biomass because it
is the "most sensitive endpoint. " (Draft Aluminum Criteria,
Appendix C, footnote d). However, the biomass endpoint was
calculated on a wet weight basis. The dry sample weight should
have been selected for the biomass endpoint. The toxicity
relationship does not exist on a dry weight basis. In fact, the dry
We disagree that it is scientifically defensible to remove the
two most sensitive species from the dataset used to derive the
national ambient water quality criteria for aluminum. EPA
considers all available reliable data in development of
national ambient water quality criteria. The peer-reviewed
methodology used to derive the criteria considers data for all
aquatic species found across the U.S., not just the two most
sensitive species. Further, species included in the sensitivity
distribution serve as surrogates for other species in their
genera for which chemical-specific toxicity data are not
available, due to genetic conservation of important toxicity
response traits in species.
We disagree in the characterization of the range of the two
species (atlantic salmon and brook trout) most sensitive to
aluminum as limited in the US. The atlantic salmon is
ecologically and commercially important. The brook trout is
the state fish of nine US states, including West Virginia. If a
state or authorized tribe chooses to modify the criterion to
reflect absence of one or more species and all surrogates, then
a new criteria value can be derived for relevant waters. Please
work with your local EPA Region and Headquarters' staff to
develop site-specific criteria that consider any modification
of the criteria's toxicity database.
The two brook trout studies conducted at pH 5.65 (Exposure
A) and pH 6.55 (Exposure B) did yield different normalized
No edits.
22
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ComiihmiI on Topic (>: Reiiiirdinii chronic loxicih (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
weight for the highest exposure concentration was greater than the
dry weight of the control sample. If the EC20 survival endpoint is
utilized, then Atlantic salmon would no longer be among the four
most sensitive species in the chronic database.
EC2o effect levels, but some variability is expected in aquatic
toxicological studies. We disagree with the assertion that
Exposure B results should be excluded from the criteria
derivation. In the Cleveland et al. (1989) paper, the maximum
control mortality reported for Exposure B was 7.5 percent,
which is well below the 20 percent maximum allowed for
chronic tests in the 1985 Guidelines; therefore, these results
should not be excluded. Because the values are less than 10-
fold different, the values were averaged following the 1985
Guidelines methods.
We disagree that the dry weight should have been selected for
the biomass endpoint. The dry weight data reported by
McKee et al. (1989) did not exhibit a dose-response
relationship, whereas the wet weight did. The wet weight,
therefore, was used to calculate the EC20 for the test. And as
noted in Appendix C of the document, Buckler et al. (1995)
appears to be a republication of McKee et al. (1989),but does
not report the most sensitive endpoint and therefore only the
most sensitive endpoint (biomass) was used for calculation of
the SMCV.
EPA-HQ-OW-
2017-0260-0047
(Kathleen M. Roberts,
Executive Director,
North American
Metals Council
(NAMC))
Determination of Acceptable Data for Use in Model Development
NAMC requests that the results of the Gensemer et al. (2017)
seven-day P. promelas tests be included in the chronic toxicity
database as these short-term chronic data have been shown to
predict reliably early life stage (ELS) chronic toxicity.[Gensemer,
R, GondekJ, Rodriguez P, Arbildua JJ, Stubblefield W, Cardwell
A, Santore R, Ryan A, Adams W, Nordheim E. (2017). Evaluating
the effects of pH, hardness, and dissolved organic carbon on the
toxicity of aluminum to freshwater aquatic organisms under
circumneutral conditions. Environ Toxicol Chem. Accepted Author
Manuscript, doi: 10.1002/etc.3920.] Specific studies were
performed with aluminum to insure the accuracy of the seven-day
studies. This will improve the robustness of the database.
The 7-day P. promelas values will not be included as core
chronic data in the sensitivity distribution used to derive the
criterion for aluminum because the exposure duration is too
short compared to the other tests used in the sensitivity
distribution, thus making relative sensitivity difficult to
determine. Seven-day chronic tests were used in the MLR
normalization studies because they are used solely to
characterize the effects of water chemistry on toxicity for the
same species, not to evaluate relative taxa sensitivity.
No edits.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Inclusion/Exclusion of data
Use of 7-day fathead minnow toxicity tests in chronic database
The draft criteria document does not reflect the availability and
use of the short-term chronic Pimephales promelas tests conducted
under varying pH, hardness, and DOC conditions (Gensemer et al.
2017). Although it is known that longer-term ELS tests are
preferred for criteria derivation, short-term chronic data have
been shown to reliably predict early-life stage chronic toxicity test
23
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic (>: Reiiiirdinii chronic loxicih (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
results for metals (Xorberg and Mount 1985, Xaddy et al. 2007).
Although EPA notes this (page 28), citing DeForest et al. 2017,
this was also noted in Gensemer et al. (2017) who presented data
from a short-term chronic test in test conditions identical to those
used in the Cardwell et al. (2017) P. promelas ELS test. The
results are very comparable, with a 7-day biomass EC2o of624.1
(409.8-950.5) fig total Al/L and a biomass EC2o of500.8 (237.2 -
1057.2) jug total Al/L in the ELS test. Based upon both the
expansive dataset (pH, hardness, DOC) of short-term chronic tests
(Gensemer et al. 2017) which were used in the development of the
MLR, we recommend that EPA re-evaluates these studies for
possible inclusion to expand the chronic toxicity test database.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of data from Gensemer et aL 2017
Chronic data presented in this study summarizes a series of pH,
hardness, and DOC manipulation tests for fathead minnow (P.
promelas), cladocerans (C. dubia and D. magna) and the green
alga (P. subcapitata). These data were used to support the
development of the aluminum BLM and the vertebrate and
invertebrate MLRs. Some of these studies were already included in
the draft criteria document under the name of the testing
laboratory (e.g., The Center for the Ecotoxicology and Chemistry
of Metals) or sponsor of the research (e.g., European Aluminium
Association). However, it does not appear that all available and
acceptable data from this study were included in the chronic
database, particularly those ultimately published in Gensemer et
al. 2017 and DeForest et al. 2017.
The additional chronic cladoceran studies were added to the
final aluminum criteria document. These studies were not
included as not all publications were available when the 2017
draft criteria were being developed.
However, the 7-day P. promelas values will not be included
as core chronic data in the sensitivity distribution used to
derive the criterion for aluminum because the exposure
duration is too short compared to the other tests used in the
sensitivity distribution, thus making relative sensitivity
difficult to determine.
Appendix C
Aluminum Criteria
Calculator "Chronic
Dataset" tab
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
C. dubia tests, run under varying conditions of pH, hardness, and
DOC (also referenced as European Aluminum Association 2010)
were found to be unacceptable for inclusion in the chronic
database. These data are described in Appendix H as an
"Unmeasured chronic exposure" and only NOEC-LOECs are
presented. However, total aluminum was indeed measured in these
tests. And while EC20s were not calculated in the original
laboratory reports, Gensemer et al. 2017 presents the full
EC 10/20/5Os for these chronic studies. The species mean chronic
value (SMCV) for C. dubia should, therefore, include these
additional data.
The additional chronic cladoceran studies were added to the
final aluminum criteria document. These studies were not
included since the publications and associated data were not
all available when the 2017 draft criteria were being
developed.
Appendix C
Aluminum Criteria
Calculator "Chronic
Dataset" tab
24
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ComiihmiI on Topic (>: Reiiiirdinii chronic loxicih (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
Criteria Development
The draft 2017 aluminum criteria updates the 1988 aluminum
criteria by incorporating new toxicity data for existing and
additional aquatic taxa. To derive the acute and chronic criteria,
EPA followed the 1985 Guidelines for Deriving Numerical
National Water Quality Criteria for the Protection of Aquatic
Organisms and Their Uses (hereafter referred to as "1985
Guidelines"). According to the 1985 Guidelines, acute and chronic
criteria are to be derived from toxicological studies that have been
screened for acceptable assessment endpoints, measures of effect,
study duration and organismal life stage. Though exceptions are
presented for specific taxa (e.g., daphnids, cladocerans,
salmonids), the 1985 Guidelines reiterate that "the agreement of
the data within and between species should be considered. "
WDEQ/WQD has noted several inconsistencies within and
between species toxicological data presented in the draft criteria
document. For instance, when calculating chronic criteria, EPA
selected studies that identified aluminum concentrations at which
certain assessment endpoints were observed in 20 percent of test
organisms (i.e., EC20). Acceptable assessment endpoints were
defined as declines in either biomass, egg numbers, population
size, emergence rates and/or survival. WDEQ/WQD is concerned
with the use of differing assessment endpoints since each endpoint
represents a different aspect of organismal fitness and therefore a
different level of aluminum susceptibility. As a result, a
considerable amount of uncertainty and/or variability may have
been introduced into species and genera mean values and
ultimately criteria values. WDEQ/WQD noted similar
inconsistencies for other aspects of selected studies, including the
chemical salts used, test duration and organismal life stage.
WDEQ/WQD recommends that EPA standardize data
requirements when possible and elaborate on how data
inconsistencies may influence the final recommended criteria.
The most sensitive endpoint available for each chronic test
was used for criteria derivation (although when available,
biomass is preferred over growth). Each endpoint selected
relates to the organism/species long-term survival, growth, or
reproduction. Adverse impacts (reduced fitness) on any of the
endpoints used could potentially result in long-term impacts
on the species. The 1985 Guidelines does utilize a diversity
of chronic test endpoints.
The chemical salts used, test duration and organismal life
stage all follow EPA Guidelines recommendations.
No edits.
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
In addition to meeting data quality standards, the 1985 Guidelines
also require that the toxicity data represent eight diverse
taxonomic groups. These minimum data requirements (MDRs)
ensure that final criteria incorporate varying levels of taxonomic
sensitivity within the targeted aquatic community. When
developing the draft aluminum criteria, EPA was able to meet the
eight MDRs for acute criterion derivation but only seven of the
eight MDRs were met for the chronic criterion. EPA decided to use
The study was not included in Appendix C (acceptable
chronic data) because the test pH was only marginally lower
than 5 (4.68-4.70). All other test conditions, procedures and
results were acceptable for criterion derivation. Satisfying the
eight-family MDR to develop the chronic criterion is superior
to using the alternate acute to chronic ratio procedure.
No edits.
25
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic (>: Reiiiirdinii chronic loxicih diilii
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
qualitative information found in a tree frog study to fulfill the
remaining chronicMDR (i.e., an additional chordate) despite the
study's inability to meet data quality standards. EPA considered
the tree frog's inclusion as justified since its toxicity value did not
affect the final chronic value.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Chronic Toxicity Data
There are some differences in the chronic Al toxicity data compiled
in the draft criteria document and the recent publications by
Cardwell et al. (2017) and DeForest et al. (2017). These
differences are matters of interpretation and professional
judgment, so we are not necessarily recommending that the
USEPA adjust any of the toxicity values in the draft criteria
document. However, we thought it would be useful to document the
basis for these differences.
Salvelinus fontinalis (brook trout): The USEPA used two
EC20s to define the sensitivity of S. fontinalis toAl: one
from a test at pH 5.65 and one from a test at pH 6.55. In
Cardwell et al. (2017) and DeForest et al. (2017), only
the test at pH 6.55 was used, as the pH 5.65 test was
considered to be too low to be appropriate for criteria
development. This results in different species mean
chronic values (SMCVs) for S. fontinalis.
Hyalella azteca (amphipod): The USEPA used two 28-d
biomass EC20s to define the sensitivity of H. azteca: one
from Cardwell et al. (2017) and one from Wang et al.
(2017). In Cardwell et al. (2017), however, the most
sensitive endpoint reported was reproduction (based on a
42-d exposure), and in Wang et al. (2017) the most
sensitive endpoint was dry weight (based on a 28-d
exposure). Cardwell et al. (2017) and DeForest et al.
(2017) used the 42-d EC20 based on reproduction to
define the sensitivity of H. azteca.
Lampsilis siliquoidea (mussel): The USEPA used the
biomass EC20 based on Wang et al. (2017), while
Cardwell et al. (2017) and DeForest et al. (2017) used the
slightly more sensitive mean dry weight endpoint from
that study.
Lymnaea stagnalis (snail): The USEPA used a 30-d
biomass EC20 to define the sensitivity ofL. stagnalis,
which was independently derived based on data reported
in OSU (2012b) and Cardwell et al. (2017). In contrast,
Cardwell et al. (2017) and DeForest et al. (2017) used the
Thank you for noting the differences in the studies that you
have highlighted. If aluminum reduced survival and growth,
the product of these variables (biomass) was analyzed (when
possible), rather than analyzing them separately as in USEPA
2013 recommendations (United States Environmental
Protection Agency. 2013. Aquatic life ambient water quality
criteria for ammonia - freshwater. EPA-822-R-13-001.
Office of Water, Washington, DC). Biomass addresses both
survival and growth impacts simultaneously. The rationale
for each endpoint selected is detailed in the final aluminum
criteria document.
No edits.
26
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic (>: Reiiiirdinii chronic loxicih (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
30-d EC20 based on wet weight, the most sensitive
reported endpoint in that study, to define the sensitivity
of L. stagnalis. We recommend that the basis for
USEPA's biomass endpoint calculation be provided since
that is not available from the original study report (OSU
2012b) and paper (Cardwell et al. 2017).
As a general comment, which relates to several of the
species-specific decisions above, the USEPA states that
the biomass endpoint was used to define the sensitivities
of species where tests included both the survival and
growth endpoints, rather than using the most sensitive
endpoint. Their rationale was for consistency with the
criteria for ammonia (USEPA 2013). However, the basis
for this decision is not apparent in the 2013 ammonia
document. We suggest USEPA provide clarification or a
basis for using an endpoint other than the most sensitive.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Gensemer et al. 2017 - 21-day Daphnia magna chronic test
Appendix H of the document states that the 21-day D. magna
chronic test reported in Gensemer et al. (2017) was excluded
because of unmeasured chronic exposures. This is incorrect, both
total measured Al and EC2oS based on the measured values are
reported in the publication. This toxicity test should be included in
the chronic database.
The additional chronic cladoceran study was added to the
document.
Appendix C
Aluminum Criteria
Calculator "Chronic
Dataset" tab
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Determination of acceptable data for use in model development
The Association notes that there are several discrepancies in the
selection of study data that was used by EPA to develop both the
acute and chronic MLR models. These are more specifically
detailed in the appended GE1 letter report, and need to be
addressed by the EPA prior to finalizing updated aluminum
criteria. In particular, the Association requests that the results of
the Gensemer et al. (2017) 7-day P. promelas tests be included in
the chronic toxicity database as this short term chronic data has
been shown to reliably predict early life stage (ELS) chronic
toxicity. This will improve the scientific accuracy and reliability of
the database.
The 7-day P. promelas values will not be included as core
chronic data in the sensitivity distribution used to derive the
criterion for aluminum because the exposure duration is too
short, thus making relative sensitivity difficult to determine.
No edits.
27
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ComiihmiI on Topic (>: Reiiiirdinii chronic loxicih (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Review of the Chronic Studies Incorporated into the Draft
Criteria
GEL reviewed the chronic toxicity studies that were deemed
acceptable by EPA for the purpose of deriving freshwater aquatic
life criteria, as presented in Sections 3.2 and Appendix C of the
draft criteria document.
General
The table presented in the "Chronic Dataset" tab of the Aluminum
Criteria Calculator spreadsheet of the criteria document, titled
"Appendix C. Acceptable Chronic Toxicity Data of Aluminum to
Freshwater Aquatic Animals", does not include a column for
references. To aid the reader in understanding the source of the
data retained for the chronic criteria derivation, EPA should
review the database included in this spreadsheet to ensure all the
studies are properly referenced, and then provide the references in
the final version of the Aluminum Criteria Calculator spreadsheet.
The database was reviewed and references were added to the
"Chronic Dataset" tab (Appendix C) of the Aluminum
Criteria Calculator.
Aluminum Criteria
Calculator "Chronic
Dataset" tab
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of data from OSU 2012
A number of vertebrate and invertebrate toxicity studies were
conducted by Oregon State University (andpublished in Cardwell
et al. 2017) to help address data gaps for ecotoxicity for
freshwater species under circumneutral conditions.
The EC20 result for great pond snail (Lymnaea stagnalis)
are presented in the draft criteria document as 745.7 jug/L
for biomass. These data, as published by Cardwell et al.,
only assessed snail survival and wet weight, with a
reported EC20 of 1148.5 jug/L for wet weight. EPA should
clarify why an endpoint different than what was reported
in the original studies was used and how the EC20 for
this endpoint was derived.
The endpoint reported in Cardwell et al. (2017) was wet
weight of the great pond snail, Lymnaea stagnalis. However,
the EPA used biomass as the endpoint. If aluminum reduced
survival and growth, the product of these variables (biomass)
was analyzed (when possible), rather than analyzing them
separately. The biomass endpoint was used when available if
growth effects were the most sensitive. This approach is as
per USEPA 2013 (U.S. EPA (United States Environmental
Protection Agency). 2013. Aquatic life ambient water quality
criteria for ammonia - freshwater. EPA-822-R-13-001.
Office of Water, Washington, DC). For purposes of
consistency in calculating the biomass endpoint, the Lymnaea
stagnalis data from Table 3-8 of OSU 2012b were used to
calculate the biomass using EPA's TRAP program.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of data from Wang et ah 2017
Wang et al. presented the chronic toxicity results for mussel
(Lampsilis siliquoidea) and amphipod (Hyalella azteca).
The EC20 presented in the draft criteria document for
mussel, 169 jug/L, differs from the EC20 of 163 jug/L as
published in the original study. The chronic database
should be updated to reflect this.
The EC20 of 169 |ig/L is the biomass reported in the study.
The EC20 of 163 |ig/L is for dry weight. Biomass was chosen
over growth endpoints for chronic values.
If aluminum reduced survival and growth, the product of
these variables (biomass) was analyzed (when possible),
rather than analyzing them separately. The biomass endpoint
was used when available if growth effects were the most
sensitive.
No edits.
28
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ComiihmiI on Topic (>: Reiiiirdinii chronic loxicih (l;il;i
I.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Short-term chronic (i.e., 7-day) P. promelas tests
conducted under varying pH, hardness, and DOC
conditions were also not included. While early-life stage
tests for P. promelas are preferred for criteria derivation,
short-term chronic data have been shown to reliably
predict early-life stage chronic toxicity (Norberg and
Mount 1985, Naddy et al. 2007). EPA noted this similarity
on page 28, citing DeForest et al. 2017, but we also noted
this similarity in Gensemer et al. 2017. One of the test
conditions in a short-term chronic P. promelas test
presented by Gensemer et al. (2017), pH 6, 120 mg/L
hardness, and 0 mg/L DOC, were identical to the test
conditions used in the 33-d early-life stage test conducted
by Cardwell et al. 2017 (included in Appendix C of the
criteria document). The results between these two studies
were comparable, with EC20s (and 95% CI) for the
biomass endpoint of624.1 (409.8-950.5) jug/L in the 7-
day test and 500.8 (237.2 - 1057.2) jig/L in the early-life
stage test.
To help illustrate the similarity of the 7-day and ELS P.
promelas tests, we used the MLR to normalize all of the 7-
day test results, and recalculate the GMCV under for
different water quality conditions with the data included
(Table 1).
[Table 1]
Not surprisingly, the recalculated GMCVs were extremely similar
(only ca. 2% different) to GMCVs calculated in the draft EPA
criteria (Table 1). Therefore, we recommend that all of short-term
chronic results from Gensemer et al. (2017) should be evaluated
and considered for the possible inclusion to improve the scientific
reliability of the chronic toxicity database.
The 7-day P. promelas values will not be included as core
chronic data in the sensitivity distribution used to derive the
criterion for aluminum because the exposure duration is too
short, thus making relative sensitivity difficult to determine.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
A 21-day D. magna chronic test was also reported in Gensemer et
al. (2017) that should be considered in the chronic database. While
Appendix H described this test as an "Unmeasured chronic
exposure, " total aluminum was measured and EC20s were
reported in Gensemer et al. 2017.
The additional chronic cladoceran study was added to the
document and the study was cited as Gensemer et al. 2018 (in
addition to the European Aluminum Association 2010
citation).
Appendix C
Aluminum Criteria
Calculator "Chronic
Dataset" tab
29
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ComiihmiI on Topic (>: Roiiiirdinii chronic loxicih (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Use of data from McCauley et al. 1986
Of the two chronic C. dubia results presented in Appendix
C, the EC20s could only be calculated for one of these.
The other is reported as a maximum allowable toxicant
concentration (MATC) of <1,100 fig/L. In the original
report, the MATC for this test is estimated as 1600 jug/L.
EPA should provide explanation for this difference,
though this will not affect the resulting criteria as this
result was not retained for SMCV calculation.
An EC2o could only be calculated for the Lake Superior water
test. The UW lab-water test missed the endpoint (no
treatment with insignificant effects). Thus, an EC20 is not
available for this test (neither TRAP model EC20 is
recommended for this test).
No edits.
EPA-HQ-OW-
2017-0260-0056
(Chris Burbage, Ph.D.,
Environmental
Scientist, Hampton
Roads Sanitation
District (HRSD),
Virginia Beach, VA)
Comment 5: Use of non-native invasive vertebrate in calculation
of chronic criteria.
HRSD requests that EPA omit toxicity data related to the
vertebrate Zebrafish (Danio rerio) in the calculation of the
aluminum chronic freshwater criteria.
HRSD supports the intent of the 1985 AWQC Guidelines to the use
of data representing the diversity of species found in the United
States, because a diverse group of test subjects is more
representative of ecosystems as a whole. The addition of a non-
native vertebrate species in the calculation of the aluminum
chronic criteria, though helpful in meeting MDRs, is unacceptable.
HRSD requests documentation confirming the naturally occurring
geographic distribution of D. rerio in the continental United
States, and hence, justification of the use of data for this species in
the calculation of the freshwater aluminum chronic criteria. IfD.
rerio is in fact a non-native species HRSD requests that its use in
the calculation of the above stated chronic criteria be justified or
removed from the Final Chronic Value database.
[Cited References]
The USEPA determined it was appropriate to include the
zebrafish (Danio rerio) in the acceptable chronic toxicity
database. While the zebrafish was originally non-native,
zebrafish populations are now established and reproducing in
the United States. See USGS fact sheet:
https://nas.er.usgs.gov/queries/FactSheet. aspx?speciesID=504
In addition, zebrafish is a commonly used test species that
provides information for other non-tested organisms.
Zebrafish are used to fulfill the "second fish family"
requirement for aluminum per the 1985 guidelines minimum
data requirements. It serves as a representative for other,
numerous, untested fish species in the U.S. Further, zebrafish
was ranked 6th in sensitivity in the 2018 aluminum chronic
data set, thus its chronic value is not included in the numeric
criteria calculations, but is included only in the "N," count of
the number of genera in the data set. Inclusion of zebrafish
for surrogacy increases criteria values by increasing the "N"
in the criteria calculation denominator. Finally, zebrafish are
included in analyses in other EPA programs, e.g., Office of
Pesticide Programs, for the purposes of including all
available quality data to serve as surrogates, given the
sparseness of data relative to the number of untested species
in U.S. waters. Inclusion in the aluminum criteria is
consistent with this practice.
No edits.
30
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ComiihmiI on Topic (>: Reiiiirdinii chronic 1 o\ici 1 > d;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
One of the most sensitive species tested, zebrafish (Danio rerio)
are not native to North America, yet were included in the
acceptable chronic toxicity database. While zebrafish are an
invaluable model organism for ecotoxicological studies, they may
not represent the sensitivity of native fishes to aluminum as
recommended in the 1985 Guidelines. EPA should either
reconsider including this species, or provide justification for their
inclusion as the results from this species may have a large impact
on the resulting chronic criterion.
The USEPA determined it was appropriate to include the
zebrafish {Danio rerio) in the acceptable chronic toxicity
database. While the zebrafish was originally non-native,
zebrafish populations are now established and reproducing in
the United States. See USGS fact sheet:
https://nas.er.usgs.gov/queries/FactSheet. aspx?speciesID=504
In addition, zebrafish is a commonly used test species that
provides information for other non-tested organisms.
Zebrafish are used to fulfill the "second fish family"
requirement for aluminum per the 1985 guidelines minimum
data requirements. It serves as a representative for other,
numerous, untested fish species in the U.S. Further, zebrafish
was ranked 6th in sensitivity in the 2018 aluminum chronic
data set, thus its chronic value is not included in the numeric
criteria calculations, but is included only in the "N," count of
the number of genera in the data set. Inclusion of zebrafish
for surrogacy increases criteria values by increasing the "N"
in the criteria calculation denominator. Finally, zebrafish are
included in analyses in other EPA programs, e.g., Office of
Pesticide Programs, for the purposes of including all
available quality data to serve as surrogates, given the
sparseness of data relative to the number of untested species
in U.S. waters. Inclusion in the aluminum criteria is
consistent with this practice.
No edits.
31
-------
TOPIC 7: Comments regarding compliments to Aluminum AWQC development
( omilKMII
Number
(()ri*;ini/;ilion)
Public C onimoil 1 oil l opic n: Rciiiudinii compliments In
Aluminum AWQC (lc\clopnicn(
IIP A Response
Kc\ision l.ociilion in
20IX Aluminum
( rilcriii Document
EPA-HQ-OW-
2017-0260-0008
(F. Paul Calamita,
Chairman, AquaLaw
PLC on behalf of North
Carolina Water
Quality Association et
al.)
JOINT COMMENTS OF THE NORTH CAROLINA WA TER
QUALITY ASSOCIATION SOUTH CAROLINA WATER
QUALITY ASSOCIATION WEST VIRGINIA MUNICIPAL
WATER QUALITY ASSOCIATION ASSOCIATION OF
MISSOURI CLEANWATER AGENCIES REGARDING THE
DRAFT UPDATED AQUATIC LIFE AMBIENT WATER QUALITY
CRITERIA FOR ALUMINUM IN FRESHWATER
Please accept the following comments on the Environmental
Protection Agency's (EPA 's) draft updated aquatic life ambient
water quality criteria for aluminum in freshwater ("Aluminum
Criteria"), Docket No. EPA-HQ-OW-2017-0260, on behalf of the
North Carolina Water Quality Association, South Carolina Water
Quality Association, West Virginia Municipal Water Quality
Association, and Association of Missouri Clean Water Agencies.
The North Carolina Water Quality Association, South Carolina
Water Quality Association, West Virginia Municipal Water Quality
Association, and Association of Missouri Clean Water Agencies
are incorporated associations of owners and operators of
Publically Owned Treatment Works throughout their respective
states.
All of the members of these four associations appreciate EPA's
efforts to develop updated aluminum criteria. It is critically
important to POTWs that applicable water quality criteria
accurately reflect the water quality goals for which they are
designed without being unnecessarily stringent.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0008
(F. Paul Calamita,
Chairman, AquaLaw
PLC on behalf of North
Carolina Water
Quality Association et
al.)
Our joint comments on EPA's draft Aluminum Criteria are as
follows:
EPA's original ambient water quality criteria for aluminum in
freshwater provided a single concentration to all water bodies
using a pH range of 6.5 and 9.0 while ignoring hardness and
dissolved organic content (DOC). This approach overlooked the
fact that aluminum toxicity can be significantly ameliorated by
conditions within individual water bodies. EPA promulgated the
new aluminum criteria in response to comments from different
industries that the previous, one-size-fits-all aluminum criteria for
acute and chronic aluminum concentrations were both difficult to
achieve and unnecessarily low.
Thank you for your comment.
No edits.
32
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\olopnicn(
I-IPA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
The proposed Aluminum Criteria represent dynamic criteria which
more accurately reflect aluminum's "real world" toxicity. Recent
research indicates that aluminum's observed toxic effects were
caused by freely-dissolved aluminum ions, concentrations of which
depend upon the chemical characteristics of a given water body.
Although the availability of aluminum ions in fresh water vary due
to many other factors, aluminum toxicity is generally proportional
to a site's DOC, hardness, and pH. EPA's proposed aluminum
better criteria reflect this scientific reality and, accordingly, are
more scientifically robust.
In summary, we thank and support EPA for promulgating these
criteria so that aluminum limits can be appropriately tailored to
waters nationwide, rather than imposing an overly conservative,
one-size-fits-all criterion. This is a much smarter and appropriate
way to provide full environmental protection with substantially
reduced regulatory burdens.
EPA-HQ-OW-
2017-0260-0010
(Adam D. Link,
Director of
Government Affairs,
California Association
of Sanitation Agencies
(CASA))
CASA and its members have long advocated for responsible
rulemaking for protection ofpublic health and the environment
alike. The proposed aluminum criteria update represents
significant progress since the original 1988 document. Basing
evaluations upon an expanded data set and the incorporation of
key water quality characteristics enables appropriate, site-
substantive assessments of potential aluminum toxicity. In
addition, use of the multiple linear regression model (MLR)
approach provides a balance between model accessibility and
robustness. CASA commends USEPA for its commitment to the
protection of aquatic life through use of the best available science.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0021
&
EPA-HQ-OW-
2017-0260-0022
(Daryll Joyner,
Administrator, Water
Quality Standards
Program, Florida
Department of
Environmental
Protection (DEP))
The Florida Department of Environmental Protection (DEP) has
reviewed the Environmental Protection Agency's (EPA) Draft
Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater. The general methodology for deriving
the criteria for aluminum (Al) in freshwater appears to be
reasonable and consistent with standard methods used for other
toxic pollutants. However, we have a few concerns regarding the
development and implementation of the proposed Al criteria. We
respectfully submit the following comments and suggestions.
Thank you for your comment; substantive comments are
addressed in detailed responses.
No edits.
33
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\elopnicn(
I'.PA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
CASQA strongly supports updating the recommended aluminum
criteria to reflect the latest science concerning the aquatic effects
of aluminum. In the sections below we discuss the problems caused
by the current (1988) criteria and also our suggestions and
comments regarding the proposed new criteria. We are very
concerned that many waterways, both natural and impacted by
human activity, will be classified as impaired by aluminum when in
fact no impairment exists.
Thank you for your comment; substantive comments are
addressed in detailed responses.
No edits.
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
Comments and suggestions on the Aluminum Notice
CASQA strongly supports the use of site-specific water chemistry
to develop criteria appropriate for the waterbody being evaluated.
The use of the site-specific parameters for pH, dissolved organic
carbon (DOC), and hardness will result in water quality criteria
that more accurately reflect the risk of aluminum toxicity to
aquatic organisms. We have the following comments and
suggestions:
Thank you for your comments; additional substantive
comments are addressed in detailed responses. The EPA
agrees that the use of site-specific water chemistry data for
developing aluminum criteria is desirable, and the 2018 final
aluminum criteria was developed on this basis. The
Aluminum Criteria Calculator V.2.0 enables site-specific
criteria derivation that addresses local water chemistry.
No edits.
EPA-HQ-OW-
2017-0260-0028
(Joshua D. Schimmel,
Executive Director,
Springfield Water and
Sewer Commission
(SWSC))
As the second largest public water system in Massachusetts, our
mission is to provide a continuous supply of potable water to out
250,000 customers. The SWSC currently holds an individual
NPDESpermit that expires on November 30, 2017. The SWSC
supports EPA's efforts to update its recommended aluminum
criteria to reflect the latest science, as the current criteria have not
been revised since 1988. The new criteria focus on aluminum
toxicity and bioavailability of aluminum to invertebrates and
vertebrates rather than a set value, which would allow for system-
specific responses without compromising protection of our
environment.
The 1988 criteria were based on the impacts of dissolved
aluminum concentrations on eight species of invertebrates and
seven species of fish, for a total of 15 species. The new criteria are
improved as it is based on the effects of Total Recoverable
Aluminum in studies of eleven species of invertebrates, eight
species of fish, and one frog species. The new criteria also take
site-specific ambient water quality into account, including the
presence of dissolved organic carbon, hardness, and pH, which
influences the bioavailability of aluminum to aquatic species. In
our own receiving water, Cooks Brook, background levels of Total
Recoverable Aluminum often exceed the current 87 jug/L national
water quality standard, but as the brook's pH is near neutral,
aluminum is less available to the fish and aquatic life. These new
parameters in the criteria would allow higher levels of aluminum
Thank you for your comment. To clarify, the 1988 criteria
were applicable for total recoverable aluminum, not dissolved
aluminum. It is accurate that the 2018 final aluminum criteria
recommendations, also applicable for total recoverable
aluminum, involve the use of site-specific ambient water
chemistry data, specifically pH, hardness and DOC, for
criteria calculations, to provide appropriately protective
aluminum criteria.
No edits.
34
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\olopnicn(
I-IPA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
in our dischargers without compromising toxicity limits.
Without the site-specific parameters of the new criteria, the SWSC
could be forced into the position of being unable to comply with
the Clean Water Act while also complying with the Safe Drinking
Water Act Regulations. Like many Public Water Systems, the
SWSC uses an aluminum-based product, an inorganic salt (PC
2800), as a coagulant in our drinking water treatment process.
With the addition of PC 2800 to the West Parish Filters Rapid
Sand filtration plant, our disinfection by-products are below the
Stage 2 Disinfection Byproducts Rule (DBPR) Maximum
Contaminant Level (MCL). Without the enhanced coagulation step
our DBP numbers would most likely climb above the MCL and
push us towards a violation of the Stage 2 DBPR regulation limits.
The proposed Aluminum Criteria Calculator will help us to
determine the allowable aluminum limits using site-specific
dissolved organic carbon (DOC), pH, and hardness, enabling
systems like ours to continue to use aluminum-based coagulants in
our treatment process.
Overall the draft criteria are well-written and organized and the
SWSC believes the proposal should move forward. We offer the
following comments for EPA's consideration:
EPA-HQ-OW-
2017-0260-0028
(Joshua D. Schimmel,
Executive Director,
Springfield Water and
Sewer Commission
(SWSC))
With these new proposed criteria, the SWSC believes it will be able
to more effectively balance protection of the health and safety of
our customers with the protection of aquatic life forms from the
effects of aluminum toxicity.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0029
(Hall & Associates on
behalf of Minnesota
Environmental Science
and Economic Review
Board (MESERB))
Background
The USEPA developed draft aquatic life ambient water quality
criteria for aluminum in freshwater (EPA-822-P-17-001; July
2017). This draft is an update of the 1988 aluminum criteria and
provides EPA's scientific assessment of the ecological effects of
aluminum on aquatic life in freshwater. The draft criteria were
released for public comment on July 28, 2017 (See, 82 FR 35198)
with comments due on or before September 26, 2017.
The current (1988) freshwater aluminum criteria set acute and
chronic impairment thresholds of 750 jug/L and 87 fig/L,
respectively, as total recoverable metal over a pH range of 6 - 9.
The chronic criterion is flagged with the following warning
Thank you for your comment; substantive comments are
addressed in detailed responses.
EPA disagrees with the commenter's assertion that the Water
Effect Ratio applied to the superseded 1988 aluminum
criteria is more appropriate than the 2018 final aluminum
criteria which reflects the current and best available science.
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum. Dissolved, colloidal and
precipitated forms of aluminum are all bioavailable to aquatic
organisms, which supports the criteria as total aluminum.
No edits.
35
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 7: Rciiiirdinii compliments to
Aliiiiiiiniin \\\Q( development
I'.I'A Response
Revision l.ociilion in
20IX Aluminum
Crilcriii Document
concerning the suitability of this criterion:
There are three major reasons whv the use of Water-Effect Ratios
might be appropriate. (1) The value of 87 us/l is based on a
toxicity test with the striped bass in water with pH= 6.5-6.6 and
hardness <10 mg/L. Data in "Aluminum Water-Effect Ratio for the
3MPlant Effluent Discharge, Middleway, West Virginia " (May
1994) indicate that aluminum is substantially less toxic at higher
pH and hardness, but the effects of pH and hardness are not well
quantified at this time. (2) In tests with the brook trout at low pH
and hardness, effects increased with increasing concentrations of
total aluminum even though the concentration of dissolved
aluminum was constant, indicating that total recoverable is a more
appropriate measurement than dissolved, at least when particulate
aluminum is primarily aluminum hydroxide particles. In surface
waters, however, the total recoverable procedure might measure
aluminum associated with clav particles, which might be less toxic
than aluminum associated with aluminum hydroxide. (3) EPA is
aware of field data indicating that many high quality waters in the
U.S. contain more than 87 fig aluminum/L, when either total
recoverable or dissolved is measured. (Emphasis added)
As noted in the footnote with the chronic aluminum criterion, EPA
has long known that pH and hardness influence the toxicity of
aluminum. These draft criteria have been developed to address
these known confounding factors. The revised draft criteria are a
function of pH, hardness, and dissolved organic carbon (DOC)
and represent a dramatic improvement over the current criteria.
However, the revised draft criteria still include significant
uncertainties that warrant site-specific adjustment.
Thus, il .tliimiiiiim ci'ilcna arc based on dissolved
concentrations, toxicity would likely be underestimated, as
colloidal forms and hydroxide precipitates of the metal that
can dissolve under natural conditions and become
biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
total recoverable aluminum (U.S. EPA 1994a,b). This method
is based on acid soluble aluminum where the sample is
acidified to pH<2 and then filtered through a 0.45 |im filter.
This process does dissolve the monomelic and polymeric
forms of aluminum, in addition to colloidal, particulate, and
clay aluminum. However, the EPA Methods 200.7 and 200.8
are the currently approved methods for aluminum.
In the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
This approach would not acidify the sample to pH<2 but
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has recently been
published as Rodriguez, P.H., J.J. Arbildua, G. Villavicencio.
P. Urrestarazu, M. Opa/o, A.S. Cardwell, W. Stubblefield, E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
EPA-HQ-OW-
2017-0260-0034
(James Boswell, Senior
Manager,
Environmental,
Peabody Energy)
Overall Peabody agrees with the EPA's proposed approach to use
multiple linear regression (MLR) to develop the draft criteria and
feels it is a significant improvement over the 1988 criteria and
incorporates the latest science. Peabody has some concerns with
EPA's approach in the draft criteria. Those concerns are focused
on 1) the form of aluminum in the criteria, 2) the range of
hardness, 3) the range of pH, and 4) applicability issues with
selected species in different regions of the U.S.
Thank you for your comment.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development.
As a result, the water chemistry bounds for the 2018 criteria
No edits.
36
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 7: Rciiiudinii compliments lo
Aliiiiiiiniin \\\Q( (lc\olopnicnt
I.PA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
wore llius expanded, Willi details and rationale pi'ovided 111 the
criteria document, and in the responses above.
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum. Dissolved, colloidal and
precipitated forms of aluminum are all bioavailable to aquatic
organisms, which supports the criteria as total aluminum.
Thus, if aluminum criteria are based on dissolved
concentrations, toxicity would likely be underestimated, as
colloidal forms and hydroxide precipitates of the metal that
can dissolve under natural conditions and become
biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
total recoverable aluminum (U.S. EPA 1994a,b). This method
is based on acid soluble aluminum where the sample is
acidified to pH<2 and then filtered through a 0.45 |im filter.
This process does dissolve the monomelic and polymeric
forms of aluminum, in addition to colloidal, particulate, and
clay aluminum. However, the EPA Methods 200.7 and 200.8
are the currently approved methods for aluminum.
In the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
This approach would not acidify the sample to pH<2 but
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has been published as
Rodriguez, P.H., J.J. Arbildua, G. Villavicencio, P.
Urrestarazu, M. Opa/o, A.S. Cardwell, W. Stubblefield, E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
37
-------
( omilKMII
Number
(Oi^;ini/;ilion)
Public C oiii moil I on lopic 7: Rciiiirdinii compliments lo
Aliiiiiiiniin \\\Q( (lc\clopnicn(
I'.I'A Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
Species included in a sensitivity distribution for criteria are
considered surrogates for other taxonomically-related species,
due to genetic conservation of important toxicity response
traits in species. For example, fish in the family Salmonidae,
such as the Atlantic salmon, include many recreationally and
commercially important species, as well as endangered
species, which are have broad relevance across the U.S.
EPA-HQ-OW-
2017-0260-0035
(RichardA. Hyde,
P.E., Executive
Director, Texas
Commission on
Environmental Quality
(TCEQ))
Comments on Proposed Standards
I. General Comments and Overview.
A. The TCEQ supports the development of criteria using site-
specific water chemistry.
It is appropriate to consider the impact of water chemistry on the
toxicity of aluminum in freshwater to aquatic species. The TCEQ
has adopted site-specific toxic criteria for aluminum in fresh water
using Water-Effect Ratio (WER) procedures agreed upon by the
EPA and the TCEQ. These procedures have allowed the TCEQ to
recognize and incorporate the effects of local chemistry on the
bioavailability and toxicity of metals, including aluminum.
Consideration of local water chemistry is particularly important to
develop appropriate criteria for aluminum, due to its interactions
with complexing ions and organic matter in freshwater.
Thank you for your comment. EPA asserts that the 2018 final
aluminum criteria, which reflects the current and best
available science and allows incorporation of local water
chemistry considerations, is more scientifically defensible
than the Water Effect Ratio applied to the superseded 1988
aluminum criteria.
No edits.
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
We are pleased to see that EPA is updating the national freshwater
aquatic life ambient water quality criteria to take into account
water quality parameters that affect Aluminum toxicity and
bioavailability. The current Aluminum criteria, adopted by EPA in
1988, does not appear to be appropriate for receiving waters in the
New England region. The Massachusetts Department of
Environmental Protection (MassDEP) has been in the process of
reviewing their surface water quality standard for Aluminum and
were expected to move forward with proposing changes to their
regulations this fall, as they felt the current criteria to be overly
conservative for many of Massachusetts' waters. These proposed
criteria could impact the state's adoption of new surface water
quality standards.
Thank you for your comment.
No edits.
38
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 7: Rciiiirdinii compliments (o
Aluminum AWQC (lc\clopmcnl
I-IPA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
El'.l-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
Our stales haw identified that the Draft. Iqualic I.ije. Inihienl
Water Quality Criteria for Aluminum 2017 is, overall, an excellent
and valuable document. We agree that the scientific findings are
defensible and accurate. It is understood that the new limitations,
while higher than those from the 1988 criteria, will not increase
the risk to aquatic ecosystems due to the bioavailability of
aluminum when properly derived and applied, and in fact, these
new draft criteria are more reflective of local conditions.
NEIWPCC encourages EPA to move forward and finalize these
Draft Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater.
Tluiuk sou lor sour comment
No edils
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
The State of New Mexico Environment Department (NMED) has
reviewed the Environmental Protection Agency's (EPA's) draft
Aquatic Life Ambient Water Quality Criteria for Aluminum in
Freshwater. The NMED appreciates the work and thoroughness
put forth to revise the 1988 aluminum guidance, which was
instituted almost 30 years ago. Overall, the primary literature
supporting the new guidance appears to be well vetted and
contributes to a more comprehensive understanding of aluminum
bioavailability and toxicity for aquatic organisms in ambient
freshwater systems.
The toxicological nature of aluminum is complex and the scientific
research exploring the various modes of exposure, and conditions
in which aluminum can pose harmful physiological impacts is
expanding, but in many ways, it is still unbound in the scope to
which it needs to be explored. Due to the limited period of time
afforded to the public and government entities that will be
responsible for implementing such guidance, the State of New
Mexico's comments submitted here are limited to a broad overview
of the study, as presented, and some of the foreseen potential
implications of implementing these multi-parameter derived
criteria.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0046
(Jennifer Wigal,
Program Manager,
Water Quality
Standards &
Assessments, Oregon
Department of
Environmental
Quality)
DEQ supports EPA's trend in developing recommended national
water quality criteria that account for the effects of site-specific
water chemistry on toxicity, as this approach improves the
accuracy and protectiveness of criteria.
Thank you for your comment.
No edits.
39
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 7: Rciiiirdinii compliments (o
Aluminum AWQC (lc\clopmcnl
I-IPA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
El 'A-11Q-OW-
2017-0260-0047
(Kathleen M. Roberts,
Executive Director,
North American Metals
Council (NAMC))
The Xorlh American Metals Council (XAMCj In addition lo the
points set forth below, NAMC supports and incorporates by
reference here the positions and views expressed in comments
submitted by the Aluminum Association. We are encouraged and
extremely supportive of the EPA effort to endorse a bioavailability-
based model in deriving the revised aluminum criteria and to
consider site-specific water quality conditions in the development
of those criteria. Our comments below are aimed at insuring
implementation can be done in a manner that is both scientifically
defensible and acceptable to the States.
Thank \ ou lor \ our coimiicnl.
\o cdlls.
EPA-HQ-OW-
2017-0260-0047
(Kathleen M. Roberts,
Executive Director,
North American Metals
Council (NAMC))
MLR Modeling Approach
NAMC supports the EPA proposal to update the ambient water
quality criteria for aluminum as the current approach uses an
outdated approach to deriving criteria and does not reflect today's
scientific advancements. The proposed approach using a MLR
model allows for the incorporation of bioavailability of aluminum
into the criteria dataset, thus providing protection for even the
most sensitive waters of the U.S. without over protecting many
non-sensitive waters. NAMC notes some key areas for further EPA
consideration below.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
On behalf of the Aluminum Ecotoxicity Research Group [Eirik
Nordheim, European Aluminium Association,
nordheim@european-aluminium.eu; William Adams, PhD, Red
Cap Consulting, Adamsw10546@gmail.com; Robert Gensemer,
PhD, GE1 Consultants, Inc., bgensemer@geiconsultants.com;
Robert Santore, PhD, Windward Environmental, LLC.,
RobertS@windwardenv.com; David DeForest, Windward
Environmental, LLC., DavidD@windwardenv.com; Patricio
Rodriguez, PhD, C1MM, phr.consulting@outlook.com; Bill
Stubblefield, PhD, Oregon State University,
bill.stubblefield@oregonstate.edu; Allison Cardwell, Oregon State
University, allison.cardwell@oregonstate.edu], we appreciate the
opportunity to provide comments on the USEPA's 2017 Draft
Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater. Our group has been developing
empirical toxicity test data and bioavailability models for
aluminum for close to a decade. EPA's recent revision of the
aluminum criteria document is extremely timely and reflects the
current state-of-the-science for the evaluation of the potential
effects of metals in the environment. We are encouraged and
extremely supportive of the EPA's efforts to endorse a
bioavailability-based model in deriving the revised aluminum
Thank you for your comment; substantive comments are
addressed in detailed responses.
No edits.
40
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\olopnicn(
I-IPA Response
Kc\ision l.ociiliun in
20IX Aluminum
Criloriii Document
criteria and lo consider site-specific water quality conditions in the
development of that criteria.
As you will see, our comments consist of a series of general overall
comments that apply to the document or to the scientific approach
employed, followed by a series of specific comments that are
defined by page number and section. These are comments which
our group believes to be very important considerations in
assessing the appropriateness and thoroughness of the draft water
quality criteria as it has been written.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xii
Sentence: EPA reviewed these models, published bv DeForest et al
(2017), and verified the results. Thus, the aluminum criteria were
derived using MLR models that incorporate pH, hardness and
DOC as input parameters.
Comment: We appreciate that the Asencv reviewed and supports
the use of the MLR approach. This approach is state of the science
and provides the right level of protection for each water body
based on the site water chemistry. The previous use of one value
for all waters of the US is clearly scientifically outdated.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
First, the District would like to express its strong support for the
EPA's decision to review and update the water quality criteria for
aluminum. In addition, we support the EPA's recommendation to
include appropriate adjustments for site-specific water chemistry
conditions such as pH, hardness, and dissolved organic carbon
(DOC) concentrations that significantly affect the potential toxicity
of aluminum.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEP))
MassDEP is pleased that the U.S Environmental Protection
Agency (EPA) is in the process of updating the aluminum
freshwater aquatic life ambient water quality criteria
recommendation in accordance with §304(a) of the Clean Water
Act (CWA). MassDEP respectfully submits the following comments
on EPA's document entitled Draft Updated Aquatic Life Ambient
Water Quality Criteria for Aluminum in Freshwater, published on
July 28, 2017 [DocketID No. EPA-HQ-OW-2017-0260].
EPA's Draft Updated Aquatic Life Ambient Water Quality Criteria
for Aluminum in Freshwater, published on July 28, 2017, provides
a thoughtful review of the water quality parameters (pH, hardness
and dissolved organic carbon (DOC)) that affect aluminum
bioavailability and toxicity. EPA's current recommended water
quality criteria (750 micrograms per liter (fig/L) acute; 87 jug/L
Thank you for your comment.
No edits.
41
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\olopnicn(
I-IPA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
chronicj [EPA 440 5-86-008, 1988J, which has been adopted by
MassDEP, are absolute numbers that are not adjusted to site-
specific water quality conditions. DOC has been documented to
ameliorate the bioavailability of aluminum and therefore
aluminum toxicity; relatively high concentrations of DOC (up to
12 milligrams per liter (mg/L)) have been measured in
Massachusetts' surface waters, thus reducing the bioavailability of
aluminum in many of our surface waters. Because the current EPA
criteria for aluminum does not incorporate these unique surface
water conditions, MassDEP asserts that EPA's 1988 aluminum
criteria are overly conservative for many of Massachusetts'
waters. MassDEP believes that these new draft criteria are more
reflective of local conditions and encourages EPA to move forward
with finalizing these criteria for aluminum in freshwater.
MassDEP is offering the following specific comments on the draft
guidance for EPA's consideration.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEP))
1. EPA 's Draft Updated Aquatic Life Ambient Water Quality
Criteria for Aluminum in Freshwater includes a user-friendly
Aluminum Criteria Calculator V.1.0 (Aluminum Criteria
Calculator V. 1.0.xlsx) that allows users to enter site-specific
values for pH, total hardness and DOC to calculate the
appropriate recommended freshwater acute and chronic criteria.
MassDEP believes this will be a useful tool for regulators and
permit holders. This calculator incorporates an approach to derive
aluminum criteria in freshwater systems using multiple linear
regression (MLR) models with pH, hardness and DOC as input
parameters.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0056
(Chris Burbage, PhD.,
Environmental
Scientist, Hampton
Roads Sanitation
District (HRSD),
Virginia Beach, VA)
With respect to the EPA request for scientific and technical views,
HRSD offers the following comments for review.
Comment 1: Revision of2008 freshwater aluminum aauatic life
ambient water aualitv criteria.
HRSD approves of the revised 2017 EPA effort updating the
freshwater aluminum A WQC.
The 2017 criteria is the product of additional laboratory toxicity
tests of aluminum on aquatic life published from 1988 to 2015.
Additionally, supplemental toxicity data from 2016 to 2017 was
also used. The original 1988 criteria document included toxicity
data from only 15 total species (representing 14 genera), however
the new 2017 criteria includes a total of 20 species, including an
amphibian (representing 18 genera). The addition of new test
Thank you for your comment.
No edits.
42
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aliiiiiiiniin \\\Q( (lc\olopnicn(
I'.I'A Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
species allows for minimum data requirements (MDRsj to be met
for the calculation of both freshwater Final Acute and Chronic
Values (FA V and FCV).
The fulfillment ofMDRs as described in the 1985 EPA guidance
document (Stephen et al. 1985) provides for scientifically
defensible water quality criteria for aluminum. These newly
derived freshwater FA V and FCV values provide freshwater
aquatic organisms sufficient protection without placing undue
burden on the public.
EPA-HQ-OW-
2017-0260-0056
(Chris Burbage, Ph.D.,
Environmental
Scientist, Hampton
Roads Sanitation
District (HRSD),
Virginia Beach, VA)
Comment 2: Inclusion of ambient water aualitv characteristics
(vH. DOC, hardness) in normalizins toxicity data.
Thank you for your comment.
No edits.
HRSD supports the use of Multiple Linear Regression (MLR)
models for the determination of aluminum water quality criteria.
The 2017 criteria establishes an aluminum freshwater criteria
taking into account the effects of pH, total hardness, and dissolved
organic carbon on the biological uptake potential of aluminum via
gill tissue. MLR models were developed by DeForest et al. (2017)
to characterize the bioavailability of aluminum for freshwater
aquatic organisms based on the above chemical properties.
DeForest et al. (2017) established the relationship between pH,
DOC, hardness and aluminum toxicity through a series of
vertebrate (Pimephales promelas) and invertebrate (Ceriodaphnia
dubia) chronic toxicity tests. These tests were used to evaluate the
ability of MLR models to accurately predict aluminum toxicity
given multiple combinations of model parameters.
The use of these MLR models allows for an accurate assessment of
aluminum toxicity for a given freshwater location that may have
varied chemical (pH, hardness, and DOC) conditions. These
models allow for small scale variations of water quality
parameters while still protecting freshwater organisms.
43
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic 7: Rciiiudinii compliments (o
Aluminum AWQC (lc\olopnicnt
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0056
(Chris Burbage, Ph.D.,
Environmental
Scientist, Hampton
Roads Sanitation
District (HRSD),
Virginia Beach, VA)
Comment 3: Inclusion of calculator for users.
HRSD supports the inclusion of the Aluminum Criteria Calculator
for the assessment of site specific freshwater acute and chronic
water quality criteria.
HRSD is supportive of the development of the criteria "calculator",
however there is concern regarding the appropriate use of data
generated by this tool. The "calculator" and its use of site specific
water quality information (pH hardness, and DOC) in predicting
protective aluminum limits is an improvement over the original
1988 criteria which had fixed values. The calculator allows for a
set range of values to be used in support of this model. If
parameter data is used that falls outside of these acceptability
ranges the calculator issues a warning stating caution in using
produced results for site assessment. HRSD is concerned that the
calculator will be preferentially used for the assessment of
aluminum criteria, and limit the availability of other assessment
tools. For instance, EPA's continued assessment of the use of the
biotic ligand model (BLM) in setting specific aluminum criteria
should not be suspended. Likewise the ability to calculate site-
specific aluminum criteria such as Water Effects Ratio (WER)
studies should not be impacted. The BLM and WER are valuable
tools that should continue to be available for aluminum criteria
assessment in addition to the newly developed "calculator".
Thank you for your comment. Please reference Section 5.3.5
for the rationale as to why the EPA chose to pursue the MLR
models published by DeForest et al. (2018a, b) over the BLM
approach (Santore et al. 2018). EPA asserts that the 2018
final aluminum criteria, which reflects the current and best
available science and allows incorporation of local water
chemistry considerations, is more scientifically defensible
than the Water Effect Ratio applied to the superseded 1988
aluminum criteria.
No edits.
EPA-HQ-OW-
2017-0260-0056
(Chris Burbage, Ph.D.,
Environmental
Scientist, Hampton
Roads Sanitation
District (HRSD),
Virginia Beach, VA)
Comment 4: Continue to provide information regarding reasons
whv specific studies were not used in water aualitv criteria
development.
HRSD supports the inclusion of information that was rejected for
inclusion in the criteria development process. The documentation
of information that was not utilized in the development process
with appropriate explanations as to the reason for its omission
provides a degree of transparency for the public. This is beneficial
for the development of the aluminum criteria but also subsequent
criteria that have yet to be developed. This omitted data with
appropriate explanations demonstrates to the public what types of
data quality are required for inclusion in criteria development. If
the public is supportive of the rationale for inclusion or omission
of specific data then their support of a given criteria will be that
much greater.
Thank you for your comment.
No edits.
44
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\elopnicn(
I'.PA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
EPA-HQ-OW-
2017-0260-0057
(Roger Claffl P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
The American Petroleum Institute (API) represents over 625
companies involved in all aspects of the oil and natural gas
industry (Exploration, Production, Refining, Marketing and
Transportation). We have a substantial interest in federal agency
activity impacting our member companies' operations under the
Clean Water Act (CWA) water quality standards program. API
member companies have facilities in all states and territories of the
United States (U.S.) that generate wastewater, require NPDES
permits to discharge, and may be subject to permit limits based on
aluminum criteria.
API appreciates the opportunity to comment on EPA's Notice of
Availability, Request for Scientific Views: Draft Updated Aquatic
Life Ambient Water Quality Criteria for Aluminum in Freshwater
(82 Fed. Reg. 35198, July 28, 2017, hereafter "Notice") and
associated draft document, Draft Aquatic Life Ambient Water
Quality Criteria for Aluminum 2017 (EPA 822-P-17-001). The
criteria derivation incorporates recent research into the physical
and environmental chemistry of aluminum that drives
bioavailability and thus ecological effects to aquatic life. It is clear
EPA and its collaborators have carefully designed and conducted
high-quality testing programs to populate the toxicity models.
While the proposed multiple linear regression (MLR)-based
criteria are an improvement over the 1988 criteria, there are still
technical and implementation limitations which should be
addressed before the criteria are finalized. Given the age of the
existing criteria and unlikelihood of timely updates there is a
concern that if these issues are not addressed prior to finalizing
the criteria, they will be problematic for decades; API suggests
they should be resolved before the final guidance is issued.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water quality
standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion.
The implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them available
for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0058
(National Council for
Air and Stream
Improvement, Inc.
(NCASI))
The National Council for Air and Stream Improvement, Inc.
(NCASI) respectfully submits the following comments on EPA's
Draft Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater and the associated technical support
document (EPA 2017). NCASI is an independent, non-profit
research institute that focuses on environmental topics of interest
to the forest products industry. Members of NCASI represent
approximately 90% of the pulp and paper production in the United
States. In its capacity as a research organization, NCASI has a
long history of working to contribute to the science needed to
address numerous environmental topics related to the forest
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
No edits.
45
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 7: Rciiiudinii compliments lo
Aluminum AWQC (lc\olopnicn(
I-IPA Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
products industry including effluent regulation, water quality
management, and relationships between human and natural
stressors on aquatic ecosystems. Additionally, in its capacity as a
research organization, NCAS1 has a long history of collaboration
with EPA on the use of sound science needed for the development
and implementation of responsible environmental management
practices. Evidence of this ongoing collaboration is seen in the
selection ofNCASl scientists as participants in numerous EPA
Science Advisory Board and other panels relating to surface water
quality. NCASl's comments on the draft criteria document are
presented below.
The proposed criteria and methods, including use of a multiple
linear regression (MLR) procedure, represent an improvement
over the existing criteria because they are based on the use of
additional test species and science-based knowledge to adjust
criteria values for water quality. Nonetheless, we identify and
describe several concerns for EPA's consideration prior to
adopting revised criteria.
EPA-HQ-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
COMMENTS OF THE UTILITY WATER ACT GROUP ON:
UWAG's purpose is to participate on behalf of its members in
EPA's rulemakings under the CWA and in litigation arising from
those rulemakings.] appreciates this opportunity to comment on
EPA's "Draft Aquatic Life Ambient Water Quality Criteria for
Aluminum, " (EPA-822-P-17-001) (defined here as the "Draft
Criteria "), which was released for public review on July 28, 2017,
82 Fed. Reg. 35,198. The Draft Criteria uses multiple linear
regression (MLR) models to derive site-specific aluminum criteria
based on the pH, hardness, and dissolved organic carbon (DOC)
of the waterbody.
Aluminum is an ubiquitous crustal element of the earth's
lithosphere. As such, the element is present in near-surface strata,
including coal deposits. Many UWAG members own and operate
fossil fuel-fired electric generating facilities, including coal-fired.
The extraction and combustion of coal results in waste and
wastewater streams that may contain aluminum. Therefore, the
development of water quality criteria for aluminum is of interest to
UWAG.
Aquatic life criteria, including for aluminum, should be adequately
protective; they should not, however, be overly conservative such
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
Since the 2017 draft document was released, additional
toxicity tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. As a result, the
water chemistry bounds for the 2018 criteria were expanded,
with details and rationale provided in the criteria document
and summarized below. The criteria calculator can be used to
address waters within a pH range of 5.0 to 10.5. For hardness
values, the criteria calculator allows entry of values between
0.01 and 430 mg/L total hardness; criteria magnitudes will
not increase or decrease by increasing the hardness above 430
mg/L total hardness (as CaC03). For DOC, the criteria
calculator will not extrapolate below the lowest empirical
DOC of 0.08 mg/L and upper limit of the empirical MLR
models will be bounded at a maximum 12.0 mg/L DOC in the
criteria calculator; criteria magnitudes will not increase or
decrease by increasing the DOC above 12.0 mg/L.
No edits.
46
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 7: Rciiiirdinii compliments (o
Aluminum AWQC (lc\clopmcnl
I-IPA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
lhal unnecessary regulatory burdens are imposed on economic
activities important to the states and the nation. Overall, UWAG
believes the scientific basis of the Draft Criteria is relatively
sound, at least for the ranges of the parameters in the underlying
data used in developing the MLR models pH of 6.0 to 8.1;
hardness up to 150 mg/L as CaC03; and DOC of up to 5.0 mg/L.
Outside of those ranges, however, the scientific validity is
questionable.
These comments focus on the technical aspects of the Agency's
derivation of the Draft Criteria to enhance the robustness of the
final updated criteria.
EPA-HQ-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
6. EPA's use of the wood frog data for the freshwater chronic
criterion was appropriate.
UWAG agrees with EPA's use of an amphibian (wood frog)
chronic test result to satisfy the 1985 Guidelines eight-family
minimum data requirement (MDR). If this data point were not
used, the chronic criterion would need to be calculated using an
acute-to-chronic ratio (ACR). Chronic criteria developed using
ACRs, in some cases, have high uncertainty and thus could be
under-protective (Raimondo et al. 2007). Also, the dose-response
pattern of acute exposures to a particular organism may be
different than the dose-response pattern for chronic exposures.
[Cited References]
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0063
(Kevin Oakes, Director
of Wastewater,
Borough of West
Chester, Chester
County, Pennsylvania)
Comments on Proposed Water Oualitv Criteria
A. The Borough of West Chester supports the development of water
quality criteria based on site-specific water chemistry and logical
and scientific approaches.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United
States Steel
Corporation)
United States Steel Corporation (U.S. Steel) is submitting the
following comments to the Environmental Protection Agency
(EPA) in support of the Federal Water Quality Coalition (FWQC)
comments submitted by Barnes & Thornburg LLP on the Notice of
Availability of Request for Scientific Views: Draft Updated Aquatic
Life Ambient Water Quality Criteria for Aluminum in Freshwater.
U.S. Steel is a member of the American Iron and Steel Institute
(AISI), who is represented in the FWQC as indicated in the Barnes
& Thornburg LLP comment letter.
The Draft Criteria will be considered by States in adopting water
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
No edits.
47
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 7: Rciiiirdinii compliments (o
Aluminum AWQC (lc\clopmcnl
I-IPA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
quality standards Jar aluminum, and in issuing ejjluenl limits for
aluminum in discharge permits. U.S. Steel, like the FWQC,
generally supports the Multiple Linear Regression (MLR)
approach which incorporates receiving waterbody quality for
deriving a site-specific aluminum quality criteria. U.S. Steel also
has the same concerns as the FWQC that we believe must be
addressed before the recommended criteria guidance document is
finalized. The following is a summary of these issues:
EPA-HQ-OW-
2017-0260-0068
(Rachel Gleason,
Executive Director,
Pennsylvania Coal
Alliance (PCA))
By way of background, Pennsylvania's aluminum criteria was
approved by US EPA Region 3 in 2001 when the Commonwealth
adopted US EPA's acute criterion, but rejected US EPA's chronic
criterion due to problems with the science on which it was
developed. While we appreciate the US EPA revising the 1988
Criteria and the flexibility that the draft criteria could provide to
operators when treating effluent limitations to meet the aluminum
limits, there are still some major concerns and clarifications that
need addressed by US EPA prior to final publication.
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
No edits.
EPA-HQ-OW-
2017-0260-0071
(Fredric P. Andes,
Coordinator, Federal
Water Quality
Coalition (FWQC))
The Draft Criteria for aluminum are based on a Multiple Linear
Regression (MLR) approach. Using that approach, EPA's
document provides for derivation of site-specific water quality
criteria, based on the levels of pH, hardness, and dissolved
organic carbon (DOC) in that waterbody. The FWQC believes that
this approach represents a substantial scientific improvement over
the methods that EPA has used in the past to develop
recommended aluminum criteria. However, our review of the Draft
Criteria has yielded a number of significant technical and
implementation concerns. We believe that it is critical for EPA to
address these issues before it finalizes the recommended criteria.
Those issues are set forth below.
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Multi-Linear Regression (MLR) modeling approach
The Association supports the EPA's work to update the ambient
water quality criteria for aluminum. The current nationally
recommended criteria date to 1988, and significant additional data
is now available to support their revision. EPA proposes using the
multiple input (pH, hardness, and DOC) MLR methodology as
outlined in the draft criteria document. In particular, the
Association asks that EPA proceed with this work consistent with
its existing 1985 Guidelines for Deriving Numerical National
Water Quality Criteria for the Protection of Aquatic Organisms
and Their Uses "1985 Guidelines"), and that it continue to
collaborate with aluminum toxicology experts such as those
involved with the Aluminium REACH Consortium to reach a final
Thank you for your comment.
No edits.
48
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 7: Rciiiirdinii compliments (o
Aluminum AWQC (lc\clopmcnl
I-IPA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
updated aluminum ambient water quality criteria that accurately
reflects the best available science.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
GEI Consultants, Inc. (GEI), on behalf of the Aluminum
Association, has reviewed the United States Environmental
Protection Agency's (EPA) 2017 Draft Updated Aquatic Life
Ambient Water Quality Criteria for Aluminum in Freshwater. Our
review focused primarily on assessing which toxicity studies were
deemed acceptable by EPA for inclusion in the criteria
development, the rationale for their inclusion, and whether the
results from these studies were used appropriately and in
accordance with the 1985 EPA Guidelines for Deriving Numerical
National Water Quality Criteria for the Protection of Aquatic
Organisms and their Uses. Additionally, we have reviewed whether
the draft criteria document addresses aluminum chemistry and
bioavailability under field conditions as opposed to the simpler
laboratory water quality conditions used in the toxicity tests from
which the draft criteria were derived.
We appreciate the efforts EPA has taken to evaluate the new
toxicity data and bioavailability models for aluminum that have
recently been published. We conclude that EPA's draft aluminum
criteria represent a significant improvement in the scientific
reliability of these criteria compared to EPA's original aquatic life
criteria (EPA 1988). The inclusion of water quality-based criteria
calculations for pH, hardness, and dissolved organic carbon
(DOC) represent significant improvements, and will provide for
much more accurate levels of aquatic life protection than the older
fixed criteria concentrations. Based on our review of the draft EPA
criteria, we provide the following comments regarding several
issues which we believe warrant further explanation or
clarification from EPA.
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
No edits.
EPA-HQ-OW-
2017-0260-0075
(Steven A. Buffone,
CHHM, QEP, GIT,
Supervisor,
Compliance and
Regulatory Affairs,
CONSOL Energy Inc.)
We commend the EPA for reviewing the 1988 A WQC Criterion
and proposing a draft revision that could offer more flexibility to
operators when treating effluent to meet the aluminum limits,
however there are still some concerns and clarifications that
should be addressed by EPA prior to final publication.
Thank you for your comment; substantive comments are
addressed subsequently in detailed responses.
No edits.
49
-------
TOPIC 8: Comments regarding the document in general
( ommcnl
Number
(()ri>
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public (onimcnl on Topic X: Rciiiirdinii (lie (lociiiiionl in
!icncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Criloriii Dociinuiil
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
Additional Comments
WDEQ/WQD commends EPA for the extensive collection and
presentation of toxicity data in the appendices of the draft
document. Though these tables are very useful, WDEQ/WQD has
noticed several discrepancies among the appendices. WDEQ/WQD
did not see any test duration or assessment endpoint information
listed for the studies in Appendix A or B. Further, the studies in
Appendix E, F and I do not include any information on DOC
concentrations. If available, WDEQ/WQD requests that EPA
include this information so each study can be fully evaluated.
Thank you for your comments. As stated in Section 3.1 of the
document: "Most fish and invertebrate data are from acute
toxicity tests that were 96 hours in duration, except the tests
for cladocerans, midges, mysids and certain embryos and
larvae of specific estuarine/marine groups, which were 48
hours in duration." Thus, all studies provided in Appendix A
(FW acute) and Appendix B (SW acute) adhered to the
recommended test duration. Text has been added to clarify
that the assessment endpoint (either EC50 or LC50 depending
on the species) also followed Guidelines recommendations.
The EPA chose not to add the DOC column to Appendices E,
F and I. In Appendix E (Acceptable Toxicity Data of
Aluminum to Freshwater Aquatic Plants) and Appendix F
(Acceptable Toxicity Data of Aluminum to Estuarine/Marine
Aquatic Plants), studies often did not report DOC and there
were not enough data to develop criteria. Appendix I (Other
Data on Effects of Aluminum to Estuarine/Marine Aquatic
Organisms) contains data that are not used in the criteria
derivation because they were not of sufficient quality.
Section 3.1
EPA-HQ-OW-
2017-0260-0025
(Peter T. Goodmann,
Director, Kentucky
Division of Water)
The Kentucky Division of Water appreciates the opportunity to
comment on the Draft Updated Aquatic Life Ambient Water
Quality Criteria for Aluminum in Freshwater (EPA-HQ-OW-2017-
0260).
A review of the material raises several concerns regarding the
draft criteria. The document narrative indicates that the curation
of the recommended acute limit is one hour, however, the table in
the Executive Summary and Table 9 both indicate a duration of
one day. The division urges the EPA to resolve this discrepancy in
amount of time recommended for the acute limit.
Thank you for your comment, the typo was fixed in both
tables and edited to be a" 1-hour average."
Executive Summary
Table 9
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
7. Corrections or clarifications
Tables in the Fact Sheet and the Draft Criteria Document show the
freshwater acute criteria as "1 day, total aluminum "for both the
current and proposed criteria. However, on page xi, the document
states, "The criteria durations are one-hour average for acute and
4-day average for chronic, respectively ... "In addition, the 1988
Criteria Document states that the acute criterion is a one-hour
average concentration.
Thank you for your comment, the typo was fixed in both
tables and edited to be a " 1-hour average."
Executive Summary
Table 9
Fact Sheet
51
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public ( onimcnl oil Topic X: Rciiiirdinii (lie (Inciimcnl in
!icncr;il
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
The 2017 Criteria Document
Page xii of the 2017 criteria document states that the "1988
aluminum freshwater acute criterion was based on dissolved
aluminum concentrations... " and that "This 2017 draft criteria
update is based on total aluminum concentrations... " However,
text on pages 20 to 21 of the 2017 document contradicts the above,
stating that the "1988 AWQC criteria [The phrase, "AWQC
criteria, " literally says "ambient water quality criteria criteria. "
This error occurs at least five times in the draft document. This is a
hazard with overuse of acronyms; they tend to lose meaning to
readers (and authors)] for aluminum were based on acid-soluble
concentrations, and were subsequently expressed in terms of total
recoverable aluminum. The current EPA approved CWA Test
Methods for aluminum in water and wastes by inductively coupled
plasma-atomic emission spectrometry and inductively-coupled
plasma-mass spectrometry measure total recoverable aluminum. "
Pages 3 and 4 of the 2017 document define what the various terms
mean, stating, "the terms filtered, dissolved, unfiltered, and total
and their relationships are defined below. "Dissolved" refers to
constituents that exist in chemical solution in a water sample. The
designation "filtered" pertains to constituents in a water sample
passed through a filter membrane of specified pore diameter, most
commonly 0.45 micrometer or less for inorganic analytes.
Therefore, for interpretation, the filtered samples will be assumed
to be dissolved aluminum. "Total"pertains to the constituents in
an unfiltered, representative water-suspended-sediment sample.
This term is used only when the analytical procedure includes an
acid digestion procedure that ensures measurement of at least 95
percent of the constituent present in both the dissolved and
suspended phases of the sample. Therefore, for interpretation, the
unfiltered samples will be assumed to be total aluminum. "
Thank you for your comment. Additional information has
been added to clarify terminology. In addition, text was
edited to be consistent with identified terms.
Section 2.1
Section 2.6.2
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
NDEP BWQP - Sequential Technical Comments on Draft
Aquatic Life Ambient Water Quality Criteria for Aluminum
(EPA 2017)
1. Pase xii of the Executive Summary states that, "The 1988
aluminum freshwater acute criterion was based on dissolved
aluminum concentrations and data from 8 species... This 2017
draft criteria update is based on total aluminum concentrations... "
The reference to "dissolved aluminum " appears to be an incorrect
statement. Page 14 on the 1988 criterion document states "... it is
Thank you for your comment. Additional information has
been added to clarify terminology used. In addition, text was
edited to be consistent with identified terms.
Section 2.1
Section 2.6.2
52
-------
( <11111110III
Number
(Oi'^iiiii/iilioii)
Public (oniniciK oil Topic X: Rciiiirdinii (lie (lociiiiicnl in
iiCIRTill
I.PA Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
recommended that aquatic lift criteria for aluminum not be
expressed as dissolved aluminum. " The 1988 document also
further states that "... not enough data are available concerning
the toxicity of dissolved aluminum to allow derivation of a
criterion based on dissolved aluminum. "
Instead, the 1988 document appears to define three states of
aluminum in water samples:
field-filtered (i.e., dissolved);
acidified before filtering (i.e., acid soluble, which some
also take as "total");
digested in the lab (i.e., total recoverable)
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
McCauley et al. 1986 - C. dubia (page C-2)
Although not used in the calculation of the SMCV, the MATC
value reported in the table is <1,100 jug/L, while the MATC value
reported in McCauley et al. 1986 is estimated at 1,600 jug/L.
Clarification on the difference is recommended.
An EC2o could only be calculated for the Lake Superior water
test. The UW lab-water test missed the endpoint (no
treatment with insignificant effects). Thus, an EC20 is not
available for this test (neither TRAP model EC20 is
recommended for this test).
No edits.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Line Item Comments
Page ix (Forward)
Sentence: Alternatively, states and authorized tribes mav use
derive numeric criteria based on other scientifically defensible
methods but the criteria must be protective of designated uses.
Comment: Grammatical error "mav use derived". Edit as
appropriate.
Thank you for highlighting this grammatical error, the error
was corrected.
Foreword
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page ix (Forward)
Sentence: Asencv decisions in anv particular situation will be
made by applying the Clean Water Act and EPA regulations on the
basis of specific facts presented and scientific information then
available.
Comment: Grammatical error "then available ". Edit to "when
available ".
Thank you for highlighting this grammatical error, the error
was corrected.
Foreword
53
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (luciimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xiv
Sentence: The 1985Guidelines...
Comment: Grammatical error. Please add a space between 1985
and Guidelines.
Thank \ ou lor liiglilighung llns. granmialieal error, die error
was corrected.
L\eeuli\ e Summary
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 2 (Problem Formulation)
Sentence: aluminiosilicate
Comment: Please check spelling. Should this be aluminosilicate.
not aluminiosilicate?
Thank you for highlighting this spelling error, the error was
corrected.
Section 2.1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 11
Sentence: ... .at vH of 7.6Land 8.05 and...
Comment: Grammatical error. Remove period after 7.61, add
space.
Thank you for highlighting this grammatical error, the error
was corrected.
Section 2.3
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 23
Sentence: an LC50
Comment: Grammatical error. Should be a LC50. not an LC50.
Thank you for highlighting this grammatical error, the error
was corrected.
Section 2.6.2
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 27
Sentence: Recent publications bv Cardwell et al. (2017) and
Gensemer et al. (2017) summarized short-term aluminum chronic
toxicity data...
Comment: The citation of Cardwell et al. (2017) in this sentence is
incorrect in this sentence and should be removed. Gensemer et al.
2017 summarized these data.
Thank you for your suggestion, the citation was corrected.
Section 2.7.1
54
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (luciimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 85, 116, 117
Sentence: References: Call, OSU, Sauer
Comment: Spelling errors in references, change Univeristv to
University.
Thank \ ou lor liiglilighung these spelling errors, die errors
were corrected.
Section "
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Appendix L (Page L-2)
Sentence: Aluminum ACRs could be calculated four freshwater
species, a mussel, a cladoceran, an amphipod and a fish. No
estuarine/marine ACRs could be calculated.
Comment: Grammatical error. Suggest revised sentence:
Aluminum ACRs could be calculated for four.
fCited References]
The ACR appendix was removed from the final document.
No edits
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC'))
3. It is well understood that the toxicity of aluminum is influenced
by the changes in the pH of surface water. To facilitate states'
understanding regarding pH toxicity at various concentrations and
to fully protect aquatic organisms from the effects ofpHfound in
natural surface waters, ADEC requests EPA expand on the tables
found in Appendix K that present criteria for various water
chemistry conditions. The current tables do not sufficiently support
the incremental measurements of pH concentrations (e.g., 6.0, 6.1,
6.2) and the level of variation that ADEC expects to see in Alaskan
surface waters. Small differences in pH result in large differences
in the resulting criteria.
The Aluminum Criteria Calculator allows users to enter up to
500 individual sets of water chemistry conditions at once to
ease facilitation of these incremental pH concentrations.
Please use this macro-enabled Excel file to calculate criteria
magnitudes that are not presented in Appendix K.
Aluminum Criteria
Calculator "Over 20
Scenarios" tab
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC))
Other Issues of Concern
1. ADEC questions why EPA chooses to incorporate
estuarine/marine criteria discussions sections throughout the
document when there is not enough data to develop WQC for
estuaries and marine environments. For clarity, EPA should
publish a separate criteria document when there is enough data to
support criteria development or combine the estuarine/marine
criteria text and data into an appendix.
Criteria document updates typically present all available data
and information (both freshwater and estuarine/marine) for
specific contaminants as recommended by the 1985
Guidelines. Even though estuarine/marine criteria cannot be
recommended with this update, the available information can
be used by different entities (states, tribes, etc.) in other ways.
No edits.
55
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public (onimcnl on Topic X: Rciiiirdinii (lie (lociiiiionl in
!icncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Criloriii Dociinuiil
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
EPA Response to External Peer Review Comments on the Draft
Aquatic Life Ambient Water Quality Criteria for Aluminum - 2017
(July 2017)
Peer-review comments on an earlier draft of the 2017 aluminum
criteria document suggest the reviewers were specialists in aquatic
toxicology rather than aqueous geochemistry. If so, the critical
issue of aluminum solubility in natural waters (i.e., "dissolved"
versus "total") may not have been sufficiently addressed during
the peer-review process. Indeed, a word search of the July 2017
"EPA Response to External Peer Review Comments, "finds no
match for "field-filtered" or "filtered. " Reviewer 4 did, however,
comment that a "Paragraph starting at the bottom of page 2 might
be an appropriate place to mention aluminum solubility and Ksp
(unless a separate section on chemistry is adopted per my
suggestion above). This is an important concept in nature and a
really important concept in the toxicity experiments" and "Page 9
near the top of the page says that at neutral pH aluminum is nearly
insoluble - this should be quantified. The Ksp of aluminum
hydroxide allows clear estimation of the solubility limits of
aluminum."
EPA's response to both of the above comments was that because
"several sources had conflicting Ksp values for Al(OH)3 so we did
not add this information. " The NDEP notes that there is
disagreement of Ksp values for most species, due to the difficulty of
obtaining these values experimentally; however, that is not a valid
reason for ignoring the concept of solubility products entirely.
[Note: the Ksp value is the "solubility product, " which is the
equilibrium constant for a solid dissolving in aqueous solution,
and is typically determined experimentally].
Text was added to the document clarifying the solubility
range for aluminum hydroxide, and solubility values for
aluminum chloride, aluminum nitrate and aluminum sulfate.
Section 2.2
56
-------
( <11111110III
Number
(Oi'^iiiii/iilioii)
Public (oniniciK oil Topic X: Rciiiirdinii (lie (lociiiiicnl in
iiCIRTill
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
El'. l-IKJ-Oll-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
2. Pages 3-4. The 2017 draft document states that, "The terms
filtered, dissolved, unfiltered, and total and their relationships are
defined below. "Dissolved" refers to constituents that exist in
chemical solution in a water sample. The designation "filtered"
pertains to constituents in a water sample passed through a filter
membrane of specified pore diameter, most commonly 0.45
micrometer or less for inorganic analytes. Therefore, for
interpretation, the filtered samples will be assumed to be dissolved
aluminum. "Total"pertains to the constituents in an unfiltered,
representative water-suspended-sediment sample. This term is
used only when the analytical procedure includes an acid digestion
procedure that ensures measurement of at least 95 percent of the
constituent present in both the dissolved and suspended phases of
the sample. Therefore, for interpretation, the unfiltered samples
will be assumed to be total aluminum. "
"Dissolved aluminum " is defined as a sample that is filtered with a
0.45-jum membrane filter. The NDEP is more specific, requiring
that "dissolved" be associated with a water sample that is field-
filtered prior to acidification. (Note that "dissolved" is an
operational definition, based on what portion of the sample passes
through a 0.45-jum filter). Any sample that is acidified prior to
filtering is considered by NDEP to yield "total" metals upon
analysis.
Thank \ou lor \ our comment, le\l allied ;is suggested
Seel io ii 11
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
4. Section 2.2, Pages 7-9. "Environmental Fate and Transport... "
This section attempts to build on the information provided in
Section 2.1. However, Section 2.2 seems to focus on the
geochemical behavior of aluminum in the aqueous environment;
specifically, solubility and speciation of dissolved aluminum.
Figure 2 (page 9) shows the relative abundance of aqueous (i.e.,
dissolved) species through a range of pH values. Unfortunately,
Figure 2 lists "Total Aluminum " on the y-axis; thereby adding to
the confusion inherent in "dissolved" versus "total" discussion.
The title on the y-axis should be changed to something like
"Relative Percent of Dissolved Species. " Figure 2 even seems to
have confused one of the peer reviewers, who stated, "In Figure
2... aluminum in the water column at pH 7.0 is almost all in the
insoluble form of aluminum hydroxide. " NDEP believes this is
incorrect; Figure 2 shows the relative abundance of dissolved
species of aluminum at different pH values.
Passage through a 0.45-jum membrane filter is the operational
Figure 2 is provided in the document to give the reader an
overall perspective of the solubility of aluminum over a wide
pH range. As stated in the document, Figure 2 is taken
verbatim from Zhou et al. (2008) and as described in the
paper, Panel A of the figure illustrates the results of
aluminum speciation of the total added to a saline solution in
the absence of ligands. Thus the "Percent of Total
Aluminum" displayed on the y-axis is relative to the total
added, not percent dissolved relative to percent total. And the
dotted lines indicate solutions that would be supersaturated
with respect to freshly prepared Al(OH)3, or the pH range in
which the calculated concentration of Al(OH)3 exceeds its
solubility. At pH 7, the majority of the aluminum is as
Al(OH)3, and as observed by the authors, the insoluble
Al(OH)3 remained dispersed in solution as a labile, colloidal
suspension (diameter of -400 nm). They also stated that true
equilibration of aluminum solutions with the less soluble,
crystalline form of Al(OH)3 (gibbsite) would take months.
No edits.
57
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public ( onimcnl oil Topic X: Rciiiirdinii (lie (Inciimcnl in
!icncr;il
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
definition for "dissolved" aluminum (and other dissolved metals).
However, the true nature of "dissolved" is calculated based on the
theoretical solubility of aluminum under a range ofpH and
chemical conditions, combined with reaction times needed to
achieve equilibrium. Section 2.2 provides some discussion of how
pH and the presence of complexing ions can affect the solubility of
aluminum in natural waters. Throughout this discussion, the focus
is on dissolved aluminum, and how pH and DOC can affect the
amount of aluminum dissolved in water (which is why the criteria
consider both pH and DOC). From the information provided in
this section, the reader would expect the criteria to be based on
dissolved aluminum; that reader would be wrong.
What is the purpose of considering pH and DOC, when the
proposed 2017 criteria are based on total (i.e., dissolved and
particulate) concentrations of aluminum?
The "Environmental Fate and Transport" section of the
document provides the reader with an overview of the
chemistry of aluminum in the aquatic environment to
compliment the information presented in Section 2.1. It is not
meant to influence how the criteria are derived. The decision
to base the criteria on total aluminum reflects a number of
considerations (analytical procedure, bioavailability, etc.), all
of which potentially impact implementation of the proposed
criteria.
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
6. Pages 12-14. The topic sentence of the last paragraph states
that, "For fish, the gill is the primary site of aluminum toxicity
under either acidic or alkaline conditions (Wilson 2012). " This
follows the amphoteric nature of aluminum, suggesting that the
criteria be based on dissolved aluminum. Likewise, on page 13, the
text notes that the biotic ligand model (BLM), "... estimates the
bioavailable portion of dissolved metals in the water column based
on site-specific water quality parameters such as alkalinity, pH
and dissolved organic carbon... " EPA (2017) opted instead to use
a multiple linear regression model (MLR) because, although such
"... models are less complex than BLM models, they also estimate
the bioavailability of aluminum to aquatic species. " This entire
discussion seems to point to the "bioavailable portion of dissolved
metals. "Again, the document seems confused on the matter of
"total" versus "dissolved. "
Dissolved, colloidal and precipitated forms of aluminum are
all bioavailable to aquatic organisms, which supports the
criteria as total aluminum. Thus, if aluminum criteria are
based on dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
Section 2.1
Section 2.6.2
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
7. Pases 20-21. The 2017 draft document states that, "The 1988
A WOC criteria for aluminum were based on acid-soluble
concentrations, and were subsequently expressed in terms of total
recoverable aluminum. The current EPA approved CWA Test
Methods for aluminum in water and wastes by inductively coupled
plasma-atomic emission spectrometry and inductively-coupled
plasma-mass spectrometry measure total recoverable aluminum
(U.S. EPA 1994a, b). The 1988 criteria considered use of dissolved
aluminum, but instead recommended acid-soluble aluminum... "
The 2017 document needs a thorough review to eliminate such
Thank you for your comment. Additional information has
been added to clarify terminology used by USGS. In addition,
text was edited to be consistent with identified terms.
Section 2.1
Section 2.6.2
58
-------
( <11111110III
Number
(Oi'^iiiii/iilioii)
Public (oniniciK oil Topic X: Rciiiirdinii (lie (lociiiiicnl in
iiCIRTill
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
contradictory statements regarding the nature of the 1988 criteria.
Further, if standard ICP-AES and 1CP-MS analyses actually
measure "total recoverable aluminum " as stated on page 21 of the
2017 document, then it is likely that many states will have only
data for "dissolved" (i.e., field-filtered prior to acidification) and
"total recoverable" (i.e., unfit Itered in the field prior to
acidification and analyzed by ICP-AES or ICP-MS) aluminum.
(Also, please note that "A WQC criteria " in the above quote
literally states, "ambient water quality criteria criteria. ")
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP'))
8. Page 21. The text discusses the relationship between
concentrations described as "dissolved" and "total recoverable, "
versus toxicity. The fifth sentence of the last complete paragraph
on page 21 states:
"Toxicity was only observed when the test solutions were
unfiltered; furthermore, dose-response was only observed using
total aluminum as opposed to measurements of dissolved or
monomeric forms (Gensemer et al. 2017). "
The study cited (Gensemer et al. 2017) is not yet published and
could not be found on-line. Therefore, the data on which this
conclusion is based and the control (or lack thereof) of
confounding variables are unknowns. Could it not be that the
suspended particulates present in an unfiltered sample would have
an adverse effect on the organisms tested? Were all other
parameters accounted for? What is the mechanism by which the
unfiltered portion of the water sample imparts toxicity? Is
Gensemer et al. using "monomeric" as equivalent to "dissolved"
or as a subset of "dissolved" species? Because "dissolved" is an
operational definition only (i.e., that portion of the sample that will
pass through a 0.45-jum membrane filter), was there any attempt to
define particulate sizes that appeared to increase toxicity? Until
these data and this study can be reviewed, the draft criteria cannot
be properly evaluated.
The Gensemer et al. (2018) study was available online pre-
publication at the time of the draft release. It is now published
hardcopy and addresses these questions.
Gensemer, R., J. Gondek, P. Rodriquez, J.J. Arbildua, W.
Stubblefield, A. Cardwell, R. Santore, A. Ryan, W. Adams
and E. Nordheim. 2018. Evaluating the effects of pH,
hardness, and dissolved organic carbon on the toxicity of
aluminum to freshwater aquatic organisms under
circumneutral conditions. Environ. Toxicol. Chem. 37(1): 49-
60.
No edits.
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
10. Section 2.7.1, Pages 28-33. The discussion of the MLR model
focuses on the solubility of aluminum (i.e., dissolved aluminum)
and how it is affected by pH. The other factors (hardness and
DOC) appear to modify the bioavailability of dissolved aluminum
by cation competition (Mg2+, Ca2+) for binding to fish gills or
reduction in toxicity when dissolved aluminum is bound by organic
matter.
As stated previously, the criteria are based on total aluminum
to adequately address the bioavailability of aluminum in the
environment and to also include colloidal and precipitated
forms. Natural field samples are not typically used for
toxicity testing due to the potential for other contaminants to
be present, thereby exerting additional toxic stress on the test
organisms.
No edits.
59
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public ( onimcnl oil Topic X: Rciiiirdinii (lie (Inciimcnl in
!icncr;il
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
The focus of the discussion remains on dissolved aluminum, yet the
new criteria specify use of data for total aluminum, as shown on
Figures 4 and 5, againstpH, which (as noted in the quote above)
affects solubility of aluminum (i.e., dissolved species), and in
Figures 6 and 7, against concentration of DOC and hardness, as
well as pH.
"The negative pH2 term accounts for the fact that AI
bioavailability decreases from pH 6 to pH 7 and then increases
from pH 7 to pH 8, which is expected given the unique solubility
chemistry of aluminum (DeForest et al. 2017). "
The mechanism of toxicity associated with unfiltered samples of
salt solutions prepared and tested in the laboratory is not
adequately discussed in the 2017 criteria document. The 2017
document is internally inconsistent and needs extensive revision
following toxicity testing using samples of field-filtered and
unfiltered waters collected from streams and lakes.
Additional information has been added to clarify terminology
used. In addition, text was edited to be consistent with
identified terms.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Terminology for measured Al
The importance of thoroughly and accurately defining measured Al
concentrations should be reviewed for clarity throughout the
document. The draft criteria incorrectly states that all
concentrations for toxicity tests are expressed as total recoverable
Al. This is not correct in studies reported in Gensemer et al.
(2017), Cardwell et al. (2017), Wang et al. (2017), including the
European Aluminium Association and the Oregon State University
references. These studies reported aluminum concentrations as
"total Al" and not "total recoverable Al". The total Al analytical
methodology used in these studies involved preserving an
unfiltered sample with HN03 to a pH of <2 prior to analysis, and
does not include the additional digestion step used for "total
recoverable ". EPA should review and correct their references to
"total Al" versus "total recoverable Al. " This will also have
implications in criteria/standards implementation.
The commenter is correct, the studies noted should be
described as concentrations for toxicity tests are expressed as
total Al, not total recoverable Al. Gensemer et al. (2018):
Total Al (acidified to pH <2 prior to analysis); Cardwell et al.
(2018): Total Al (acidified to pH <2 prior to analysis); Wang
et al. (2018): Total Al (acidified to pH <2 prior to analysis);
European Aluminium Association (2009): Nominal
concentrations equate to total Al; European Aluminium
Association (2010): Nominal and Total concentrations, don't
specify method for total Al; Oregon State University
(2012a,b,c,d,e,f,g,h & 2013): Total Al (acidified to pH <2
prior to analysis, although reports incorrectly state that
sample collected for total recoverable analysis).
No edits.
60
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (luciimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 3
Sentence: The terms filtered, dissolved, unfiltered, and total and
their relationships are defined below.
Comment: This sentence should be modified (see eeneral
comments on terminology). Sentence should include the term
"acid-soluble" as this terminology is used in the 1988 criteria
document (i.e., acidify the water sample with HN03 topH 1.65-
1.85, followed by filtration through 0.45 jum). Also the
differentiation between "total" and "total recoverable" should be
mentioned.
Thank \ ou lor \ our comment. AiiiiiUonal iiilorniauoii hu;>
been added to clarify terminology.
Section 2.1
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Chemistry of Aluminum
One of the central problems with the extrapolation of laboratory
toxicity data with aluminum to regulatory criteria implementation
in natural waters is the complex chemistry of precipitated or solid-
phases of aluminum. As EPA correctly summarizes on page 22,
natural waters contain mineral particulate forms of aluminum that
may be subject to measurement "uncertainty" when using "total
recoverable " [As per typical analytical methods for total
recoverable metal, the term "total recoverable " should only be
applied to samples that have been acidified by HN03 and HCl,
followed by gentle fluxing (see Table 7 of Cardwell et al. 2017)]
forms of aluminum. While this is an accurate statement, EPA does
not fully explain this uncertainty in terms of how aluminum criteria
should be applied, or even how the toxicity data presented in the
criteria document should be cited or interpreted. These are critical
omissions that would benefit from further explanation as EPA
revises the aluminum criteria. Specific aspects of this concern are
outlined below.
On page 22, EPA states that "All concentrations for
toxicity tests are expressed as total recoverable aluminum
in this document (unless otherwise specified), and not as
the form of the chemical tested. " This is generally not
correctfor most all of the laboratory studies we've
conducted or reviewed, total recoverable metal assays
were not used to express the total concentration of Al in
the test solutions. Rather, most tests used a "total"
aluminum assay which was simply the acidification of
unfiltered test solutions without the additional
digestion/fluxing step used in total recoverable analytical
methods. The simpler "total" aluminum assay is
appropriate for laboratory test solutions (as correctly
Thank you for highlighting these discrepancies; these errors
were corrected. The 1988 AWQC for aluminum were
discussed as acid-soluble concentrations and were
subsequently expressed in terms of total recoverable
aluminum.
Dissolved, colloidal and precipitated forms of aluminum are
all bioavailable to aquatic organisms, which supports the
criteria as total aluminum. Thus, if aluminum criteria are
based on dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
total recoverable aluminum (U.S. EPA 1994a,b). This method
is based on acid soluble aluminum where the sample is
acidified to pH<2 and then filtered through a 0.45 |im filter.
This process does dissolve the monomelic and polymeric
forms of aluminum, in addition to colloidal, particulate, and
clay aluminum. However, the EPA Methods 200.7 and 200.8
are the currently approved methods for aluminum.
In the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
Section 2.6.2
61
-------
( <11111110III
Number
(Or^iini/iilion)
Public ( oninicnt on Topic X: Rciiiirdinii (lie (lociiiiionl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
('rili'rhi Document
pointed out by EPA on page 69) that only contain
dissolved monomeric and precipitated forms (e.g.,
aluminum hydroxides) of aluminum. This simpler total
metal assay should indeed solubilize any precipitated
forms of aluminum, thereby leading to an accurate
measure of all aluminum forms in the test solutions.
Therefore, EPA should not use the terminology of "total
recoverable " aluminum in the context of laboratory tests
unless they are certain the study actually used this more
aggressive analytical method.
On page 69, EPA correctly cites justifications for use of
"total" aluminum concentrations in laboratory tests
based on work presented in Gensemer et al. 2017 and
Santore et al. 2017. However, in this discussion, EPA
incorrectly uses the term "particulate " to describe the
other basic form of aluminum considered in Santore et al.
2017 in addition to dissolved aluminum. In an important
distinction, Santore et al. 2017 uses the term
"precipitated" aluminum to refer to aluminum hydroxides
that precipitate in the test solutions when concentrations
and pH are such that solubility of the dissolved metal is
exceeded. The other papers in this journal series (e.g.,
Cardwell et al. 2017 and Gensemer et al. 2017) also are
careful to use the term "precipitated" aluminum to
distinguish solid phase aluminum that forms specifically
in test solutions following precipitation of the dissolved
(usually acidic) concentrated stock solutions at
circumneutralpH. The term "particulate " can too easily
be confused with mineral particulates in natural waters,
so we suggest that EPA use the term "precipitated"
aluminum in this context.
Overall, we suggest that EPA do more to explain the
"uncertainty" with respect to total or total recoverable
aluminum measurements in natural waters. While EPA
correctly points out on page 69 that total (should be total
recoverable here) concentrations "may overestimate the
potential risks of toxicity... ", further explanation is
warranted to ensure that implementation of these criteria
do not generate too many false positive outcomes (i.e.,
total recoverable aluminum concentrations that exceed
the criteria, but the true bioavailable concentration of
aluminum would not exceed the criteria). A more clear
This approach would nol aadil\ ilic sample lo pll J" bill
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has been published as
Rodriguez, P.H., J.J. Arbildua, G. Villavicencio, P.
Urrestarazu, M. Opa/o, A.S. Cardwell, W. Stubblefield, E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
62
-------
( <11111110III
Number
(Oi'^iiiii/iilioii)
Public ( oniniciK on Topic X: Rciiiirdinii (lie (lociiiiionl in
iicncr;il
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
('rili'rhi Document
understanding of this uncertainly would assist Stales and
Tribes as they seek to develop the best methods or
implementation tools to ensure the criteria are used and
interpreted in the most accurate way possible. For
example, New Mexico uses a coarse (e.g., 10 micron)
filtration step to remove at least some of the non-toxic
mineral phase aluminum as stated in their water quality
standards: "For aluminum, the criteria are based on
analysis of total recoverable aluminum in a sample that is
filtered to minimize mineral phases as specified by the
department. " (see NMAC 20.6.4.900.H)
Although coarse prefiltration presents a possible solution,
analytical methods based on a mild acid-reactive process
would likely provide a more accurate representation of
bioavailable aluminum in waters with significant amounts
of mineral particulates because of the operational nature
of size-based filtration methods. In particular, methods
that use a less strong or aggressive acidification step than
that used in a total recoverable metal assay would likely
provide a more accurate measure of bioavailable
aluminum in natural waters. Such methods might include
the acid soluble test described in the existing national
aluminum criteria (EPA 1988), or even a modified pH 4
extraction method currently under development. We
recognize that these methods are not yet available for
compliance purposes under the Clean Water Act in all
cases. However, until such time as an improved method is
available (e.g., the modified pH 4 method), we suggest
that EPA consider citing the acid-soluble method (EPA
1991; method 200.1) as the recommend method for
implementation as they did in the existing 1988 criteria.
At the very least, we feel a more thorough discussion of
the uncertainties regarding the use of total recoverable
aluminum concentrations in natural waters would great
help end users of these criteria.
63
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I on Topic X: Rciiiirdinii (lie (lociiinonl in
iiOIRTill
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
('rili'rhi Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 4
Sentence: Groundwater concentrations of dissolved aluminum
(filtered using a 0.45 micrometer filter) from the NA WQA database
collected during 1992-2003 are presented in Figure 1, with a 90th
percentile concentration of dissolved aluminum concentrations of
11 fig/L.
Comment: Is 2003 the most recent data collection of groundwater
data? Could this be expanded to more current values?
The figure ib the laical and nios>l up lo dale figure u\ uiluble.
No edib.
EPA-HQ-OW-
2017-0260-0028
(Joshua D. Schimmel,
Executive Director,
Springfield Water and
Sewer Commission
(SWSCJJ
1. The SWSC suggests clarification of the phrase "site-specific
values" by clearly stating that water quality parameters should be
collected from the receiving water.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water quality
standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion.
No edits.
EPA-HQ-OW-
2017-0260-0035
(RichardA. Hyde,
P.E., Executive
Director, Texas
Commission on
Environmental Quality
(TCEQ))
D. The TCEQ recommends EPA be clear and consistent
regarding the speciation of aluminum
The speciation of aluminum in the 1988 criteria document is
referenced inconsistently in EPA's current proposal. EPA should
clarify the speciation, and reference the information consistently.
For example, the following citations in EPA's current proposal
inconsistently reference aluminum speciation of the 1988 criteria:
Page xii: "The 1988 aluminum freshwater acute criterion
was based on dissolved aluminum concentrations..."
Page 20-21: "The 1988 AWQC criteria for aluminum
were based on acid-soluble concentrations, and were
subsequently expressed in terms of total recoverable
aluminum."
Page 21: "The 1988 criteria considered use of dissolved
aluminum, but instead recommended acid soluble
aluminum for several reasons. "
Page 74: Table 9, Summary Overview of 2017 Draft
Aluminum Aquatic Life Criteria Compared to Current
1988 Criteria references aluminum concentrations for
both criteria documents as "total aluminum".
Thank you for your comment. Additional information has
been added to clarify terminology. In addition, text was
edited to be consistent with identified terms.
Section 2.1
Section 2.6.2
64
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public ( onimcnl oil Topic X: Rciiiirdinii (lie (Inciimcnl in
!icncr;il
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
5. Pages 10-11. Sentences within one paragraph (split between
pages 10 and 11 of the draft criteria document) appear to state
opposite results and conclusions. On page 10, the second sentence
of the last paragraph states "Over time as the aluminum from the
stock solution equilibrates with the test water and the pH
increases, the monomeric species of aluminum transform to the
insoluble polymeric hydroxide species, which are more toxic... "
This seems to indicate that precipitates (i.e., insoluble forms) in
older solutions have higher toxicity than the dissolved (i.e.,
soluble) species of aluminum. However, the next two sentences
state that "... soon after test initiation, there is a transformation
period of rapid speciation changes from short-lived transient
amorphous and colloidal forms of aluminum to more stable
crystalline forms (Gensemer et al. 2017). Aged stock solutions
(aluminum solutions that have been given time to form more stable
forms of aluminum) have been shown to be less toxic than those
that are not aged. " Readers will likely imagine that "soon after"
occurs faster than "over time, " and will equate "over time " with
"aged. " One sentence says aged is "more toxic " and one says it is
"less toxic. " Is this poorly stated or misstated? Which qualifier
(i.e., "more" or "less") for toxicity is correct?
Thank you for your comment, text was edited for
clarification.
Section 2.3
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
9. Section 2.7.1, Pase 27. Please define "short-term aluminum
chronic toxicity data. " "Chronic" defines a long-term condition.
What is difference between standard testing for chronic toxicity
and the testing for short-term chronic toxicity? Are the data from
short-term tests of chronic toxicity for aluminum different from
data from standard chronic tests?
As described on page 27 (Section 2.7.1), the "short-term
chronic tests" refer to the 7-day fathead minnow, 7-day C.
dubia and 72-hr algal (Pseudokirchneriella subcapitata) tests.
Standard chronic tests for invertebrates and fish usually span
the life cycle of the species, although 7-day C. dubia and 28-
day early life stage fish tests are routinely used in the
sensitivity distribution for criteria derivation. Algal tests
should be 96 hours as recommended by the Guidelines.
No edits.
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
7. "Hardness" is used interchangeably with "Total Hardness" and
it is difficult to tell without investigating all the subtending
literature if these are being differentiated. Total Hardness is the
parameter used in the 2017 Al guidance calculator. However, New
Mexico's hardness-dependent calculator for compliance with
hardness-dependent numeric criteria, uses dissolved hardness (as
mg CaC03/L). This discrepancy made it difficult to accurately
assess New Mexico's EPA approved hardness-based criteria
against the proposed guidance. Can it be demonstrated that
particulate hardness (solidphase hardness included in the MLR
model inputs) provides protection of aquatic life?
Thank you for your suggestion. To avoid confusion, "total"
was added throughout the document. Currently there is no
data available comparing the dissolved versus particulate
hardness for aluminum.
Throughout the
document, appendices
and the Aluminum
Criteria Calculator.
65
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public Comment oil Topic X: Rciiiirdinii (lie (Inciimcnl in
!icncr;il
I'.I'A Response
Re\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0045
(Lee Lemke, Executive
Vice President,
Georgia Mining
Association (GMA))
3) In the Review Comments, Reviewers 2 and 4 repeatedly
emphasize the enormous disparity between the solubilities of the
aluminum forms used in the Draft Criteria's referenced
experiments compared with aluminosilicate minerals. Reviewer 4
states:
"The current text does explain soluble speciation (i.e.,
complexation) but fails to recognize solid speciation. There is a
large difference between a particle of feldspar or kaolinite from
freshly precipitated aluminum hydroxide" (p. 41, Review
Comments).
Despite the frequency and reasonableness of these comments, the
authors of the Draft Criteria inadequately address these review
comments by simply adding language to the brief treatment of
uncertainty in the Draft Criteria, while offering no suggested
recourse to this problem.
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum.
Dissolved, colloidal and precipitated forms of aluminum are
all bioavailable to aquatic organisms, which supports the
criteria as total aluminum. Thus, if aluminum criteria are
based on dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
total recoverable aluminum (U.S. EPA 1994a,b). This method
is based on acid soluble aluminum where the sample is
acidified to pH<2 and then filtered through a 0.45 |im filter.
This process does dissolve the monomelic and polymeric
forms of aluminum, in addition to colloidal, particulate, and
clay aluminum. However, the EPA Methods 200.7 and 200.8
are the currently approved methods for aluminum.
In the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
This approach would not acidify the sample to pH<2 but
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has recently been
published as Rodriguez, P.H., J.J. Arbildua, G. Villavicencio,
P. Urrestarazu, M. Opa/o. A.S. Cardwell, W. Stubblefield, E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
Section 2.6.2
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Use of analytical methods and terminolosv for measurins
aluminum concentrations
As stated on page 22 of the revised criteria document, natural
waters contain mineral forms ofAl that may not be bioavailable,
therefore aggressive digestions (such as total recoverable
methods) may lead to potential overestimations of bioavailable Al
in natural waters containing suspended solids.
Both the type of analytical method used and the terminology for
measured Al as it relates to the expression of water quality criteria
should be more clearly defined and applied throughout the draft
criteria document. To address the issues of appropriate analytical
methodology for measuring Al in natural waters, our research
group is currently developing methods that will be helpful in
measuring the amount of "bioavailable " Al in natural waters (pH
4 digestion in Table 1). For clarity, we have provided Table 1 to
summarize and further define all of the available methods.
We believe a clearer description of both analytical methods and
analytical terminology would allow users to effectively quantify Al
concentrations in natural waters. A more robust discussion on
measuring Al in natural waters would also provide guidance on
appropriate ways to measure bioavailable Al and avoid measuring
non-toxic mineral phases.
[TABLE 1]
66
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public ( onimcnl oil Topic X: Rciiiirdinii (lie (Inciimcnl in
!icncr;il
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xi (Executive Summary)
Sentence: Multiple linear regression MLR) models were
developed to characterize the bioavailability of aluminum in
aquatic systems based on the effects of pH, hardness and DOC
(DeForest et al. 2017). The authors used 22 chronic tests with the
fathead minnow (Pimephales promelas), and 23 chronic tests with
Ceriodaphnia dubia to evaluate the ability of MLR models to
predict chronic toxicity of aluminum as a function of multiple
combinations of pH, hardness, and DOC conditions.
Comment: The Aeencv failed to mention that the MLR approach
included many studies with green algae as well. While the Agency
does not use these values in their approach to criteria
development, the data provide support for the overall MLR
approach as presented in DeForest et al 2017. Reference to the
algae data would be appropriate.
Since this plant MLR model was not used in the criteria
development it is not needed in the Executive Summary. The
plant MLR model is discussed in Section 2.7.1 and text has
been edited to present the plant MLR model in Section 5.2.
The EPA discussed that, based on existing data, plants are
less sensitive then fish and invertebrates, thus the 2018
aluminum criteria is expected to be protective of aquatic plant
species.
Section 5.2.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xii
Sentence: The 1988 aluminum freshwater acute criterion was
based upon dissolved aluminum concentrations.
Comment: It is not correct that those criteria were based on
dissolved aluminum. Criteria were stated on the basis of "acid-
soluble " measurements.
Thank you for your comment. The sentence has been edited.
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum.
Executive Summary
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xii
Sentence: The MLR eauations applied to the acute toxicity data
were those developed through chronic tests, with the assumption
that the effect of water chemistry on bioavailability remains
consistent.
Comment: This statement should be expanded a little to indicate
that the MLR approach published by DeForest et al. (2017) was
developed using chronic tests and the Agency adopted these to
develop equations for acute testing.
Thank you for your comment. The final 2018 aluminum
criteria document discusses application of the chronic MLR
approach to normalize acute data.
No edits.
67
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (luciimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xiii
Sentence: This 2017 draft criteria update includes new data...
Comment: This should read - This 2017 draft criteria update
includes new acute and chronic data. And, "Minimum Data
Requirements (MDRs) for direct calculation " should read -
Minimum Data Requirements (MDRs) for direct calculation
without the use of an acute to chronic ratio.
Text w ab edited lor daril\.
L\eculi\ e Summary
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xiii, lines 20-22
Comment: It would be auite insightful if the Agency were to
include a MLR calculation of the water chemistry that results in a
CCC value of 87 jug/L, i.e., DOC, 1 mg/L, hardness 50 mg/L and
pH 6.27 (using the MLR published by DeForest et al.)
Since the chronic criterion (CCC) is a function of three water
quality parameters (pH, total hardness and DOC), there are
multiple scenarios where the CCC would be ~87 |ig/L (the
1988 AWQC CCC). For example, in Appendix K, Table K-l
where the DOC=0.1 mg/L, the CCC would be 87 |ig/L when
the pH=6.5 and total hardness=150 mg/L. Another example
would be Table K-8 (DOC=2.5, pH=6.0 and total
hardness=10) where the CCC=81 (ig/L.
No edits.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page xiv (and throughout)
Comment: It would be beneficial for the Aeencv to provide a basis
for the water quality characteristics "example " (pH = 7, hardness
= 100 mg/L, DOC =1 mg/L) that is used throughout the document
as an example of a normalized value. Does the agency believe
these water quality characteristics are of a typical North American
natural water? For background and because it is significantly used
throughout the document, please provide a basis for selecting these
specific values.
The water quality characteristics that the EPA uses a scenario
throughout the document was simply an example scenario. In
other hardness based AWQC documents (i.e., cadmium),
total hardness is usually normalized to a hardness of 100
mg/L as CaC03. The sample DOC and pH was chosen just to
be illustrative of one example scenario. Additional text added
to clear up this confusion and to relate that the sample
scenario is just an example. The calculator allows a wide
range of water quality conditions typical of US waters to be
taken into consideration in deriving criteria.
Executive Summary
(table insert)
Section 2.7.1
Table 9
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 5
Sentence: At the typical ocean pH of 8.0-8.3, aluminum
coordinates with the hydroxide ion, primarily as Al(OH)4.
Comment: Sussest revision: At the typical ocean pH of 8.0-8.3,
aluminum coordinates with the hydroxide ion, primarily as
Al(OH)4, which will precipitate out of solution, for the most part,
which explains the low concentrations in marine waters.
Thank you for your suggestion, text was edited.
Section 2.1
68
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (lociimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
(ritcrhi Document
EI\ i-IfO-()ll-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 6, second last paragraph
Comment: some mention of the impact of soil particles entrained in
the air samples should be mentioned. Total analyses will digest the
particles which are typically high inAl. AirAl concentrations are
highly dependent upon particulate concentrations.
1 'lunik sou for sour suggestion, te\l was allied
Section 2 1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 7
Sentence: In streambed sediment samples collected from locations
in the conterminous U.S. from 1992 to 1996, aluminum
concentrations ranged from 1.4 to 14 fig/g dry weight (Rice 1999).
Comment: Are the units correct- us/s (i.e.. ppm)? More likelv 1.4
- 14%. Soil samples range from 500-142,000, hence, stream
bedded sediments would be quite similar.
Thank you for your suggestion; the correct values are weight
percent, text was edited.
Section 2.1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 9 first sentence;
Sentence:...characteristics are significant because episodic acidic
pulses in streams, for example during winter snow melt, maximize
the solubility of aluminum
Comment: Edit text... characteristics are significant because
episodic acidic pulses in streams, for example during winter
snowmelt, maximize the solubility of aluminum ifpH drops to 5.5
or lower.
Thank you for your suggestion; the text was edited.
Section 2.2
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 10, second paragraph
Sentence: Freeman andEverhart (1971) found that the chronic
toxicity of nominal (unmeasured) aluminum increased as pH
increased from 6.8 to 8.99 in rainbow trout. "
Comment: Does this mean that the toxicity values became smaller?
Aluminum was more toxic at the higher pH when exposed to
same concentration of aluminum (TL50 was 38.9 days at pH
6.8 compared to TL50 of 2.96 days at pH 8.99). Text was
edited for clarity.
Section 2.3
69
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (luciimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 12
Sentence: Bottom-dwelling organisms mav be impacted more bv
aluminum floe in the field than in the laboratory due to the greater
floe layer thickness observed in the field relative to laboratory
exposures.
Comment: This depends upon the water velocity/mixing
zone/movement of water in both the field and lab. Please clarify or
provide citation for this observation.
Thank \ ou lor \ our buggeblion, llie le\l w u;> edited.
Section 2.3
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 12
Sentence: Bioavailabilitv of aluminum is affected bv water
chemistry parameters such as pH, hardness, and DOC.
Comment: Text edit needed. Bioavailabilitv of aluminum is
affected by water chemistry parameters such as pH, hardness, and
DOC and to a lesser extent fluoride.
Thank you for your suggestion; the text was edited
Section 2.3.1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 13
Sentence: Overall, aauatic plants are generally insensitive to
aluminum. Algae productivity and biomass are seldom affected if
the pH is above 3.0. Aluminum and acid toxicity tend to be additive
to some algae when the pH is less than 4.5.
Comment: Gensemer et al. (2017) demonstrated toxicity to the
green algae under varying pH, hardness, and DOC conditions.
Suggest clarification to the statement that algae biomass are
seldom affected if the pH is above 3.0.
Thank you for your suggestion; the text was edited.
Section 2.3
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 13, second last paragraph
Sentence: In contrast, no apparent hardness-toxicitv relationship
was observed for rainbow trout exposed to three different hardness
levels at a controlledpHof8.3 (Gundersen et al. 1994).
Comment: This is consistent with data recently published bv
DeForest et al (2017) and Gensemer et al (2017) demonstrating
that there is a reduced effect of hardness at elevated pH levels.
Thank you for your suggestion; the text was edited
Section 2.3.1
70
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (luciimcnl in
iicncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page and l-l
Sentence: Paragraph starting with "Development of the "biotic
ligand model" (BLM - formerly the "gill model")
Comment: We suggest mentioning and/or discussing how the Al
BLM differs from earlier BLMs with other metals, as the Al BLM
accounts for the dissolved and precipitated fraction.
Thank \ ou lor \ our suggestion, the text w as edited
Section \ v 1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 21
Sentence: The 1988 criteria considered use of dissolved aluminum,
but instead recommended acid-soluble aluminum for several
reasons.
Comment: Correct. Suggest this edit to page xii as well.
Thank you for your suggestion; the text was edited
Executive Summary
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 27
Sentence: These three parameters are thought to be the most
influential for aluminum bioavailability and can be used to explain
the magnitude of differences in the observed toxicity values
(Cardwell et al. 2017).
Comment: The more correct citation for this would be Gensemer et
al. (2017) and Cardwell et al. (2017).
Thank you for your suggestion; the text was edited
Section 2.7.1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 34
Sentence: Throughout this document, unless otherwise stated,
effect concentrations were normalized to pH 7, hardness of 100
mg/L and DOC of 1 mg/L. These specific values were chosen to
represent pH, hardness and DOC levels found in the environment.
Comment: This sentence about the selection of these specific
values as an example is vague. Additional basis for use as an
example would be appreciated.
The water quality characteristics that the EPA uses as a
scenario throughout the document were simply an example
scenario. In other hardness based AWQC documents (i.e.,
cadmium), total hardness is usually normalized to a hardness
of 100 mg/L as CaC03. The sample DOC and pH was chosen
just to be illustrative of one example scenario. Additional text
added to clarify that the sample scenario is just an example.
Section 2.7.1
71
-------
( ommcnl
Number
(Oi^;ini/;ilion)
Public ( o in moil I oil Topic X: Rciiiirdinii (lie (lociiincnl in
iionor;il
I.PA Response
Kc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 49
Sentence: Oregon State University also conducted several chronic
studies for three invertebrate species: an oligochaete, Aeolosoma
sp.; a rotifer, Brachionus calyciflorus; the great pond snail,
Lymnaea stagnalis; and one fish species, an early life cycle test
with the zebrafish (OSU 2012b,c,e, 2013).
Comment: The one fish species (zebrafish) should be under the
vertebrate section and not the invertebrate section.
Thank \ ou lor \ our suggestion, the w rile up lor ilus siud\
was moved to the vertebrate section.
Section V \ 1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 49 (and Appendix C spreadsheet)
Sentence: pond snail 30-dav biomass
Comment: The snail (Lvmnaea stagnalis) study reported bv OSU
and Cardwell et al. (2017) did not calculate a biomass endpoint
(survival and wet weight were calculated and reported). If
additional analysis was conducted by EPA to report a biomass
endpoint, please provide details for clarity.
Biomass was calculated using the reported values in Table 3-
8 (OSU 2012b) by calculating proportion survived by wet
weight. If aluminum reduced survival and growth, the
product of these variables (biomass) was analyzed (when
possible), rather than analyzing them separately.
No edits.
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 50
Sentence: The chronic toxicity of aluminum to fathead minnows
was also evaluated by OSU (2012g). Very similar exposure
methodology and the same dilution water were used as described
above for the amphipod and midge tests (OSU 2012f h), except
that <24-hr old fertilized eggs were used at initiation of the 33-day
test.
Comment: Suggest revision as the onlv similar methodology was
the dilution water and pH control of the water. The methods for
number of replicates, feeding, duration, flow-rate, etc. were all
different from the amphipod and midge. Suggest citation to
Cardwell et al. (2017) which details methodologies for each
species.
Text was edited for clarity.
Section 3.2.1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 50
Sentence: Fry survival was the most sensitive endpoint with an
estimated EC20 of428.6 fig/L, and normalized EC20 of 1,734
fig/L.
Comment: Suggest edit (adding calculated). Fry survival was the
most sensitive endpoint with a calculated EC20 of428.6 fig/L, and
normalized EC20 of1,734 fig/L.
Text was edited for clarity.
Section 3.2.1
72
-------
( <11111110III
Number
(Oi'^iiiii/iilioii)
Public (oniniciK oil Topic X: Rciiiirdinii (lie (lociiiiicnl in
iiCIRTill
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
EPA-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 51
Sentence: The NOEC for survival and growth normalized to a pH
7, hardness of 100 mg/L and DOC of 1.0 mg/L was 9,746 jug/L (the
highest concentration tested), with a chronic value of >9,746 fig/L.
Comment: Suggest removal of "highest concentration tested" as
this was not the highest exposure tested (appears the actual
concentration was 2,000 jug/L).
Text \\a;> edited u;> buggered.
Section 3.2.1
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Page 70
Sentence: Both MLR models and the BLM model are based on the
same toxicity test database.
Comment: Both models include the same toxicity test data, but the
BLM doesn't exclusively use the data (BLM includes data on the
accumulation ofAl on the gills of salmon). This is somewhat
clarified in the next sentence, but we suggest the sentence that both
models are based on the same database should be re-worded.
Thank you for your suggestion; the text was edited.
Section 5.3.5
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
Appendices B, E, G, H
Comment: Suggest EPA provide separate column for DOC
concentrations, as was done in Appendix A and C.
Thank you for your suggestion. The EPA chose not to add
this column. Regarding Appendix B (Acceptable Acute
Toxicity Data of Aluminum to Estuarine/Marine Aquatic
Animals) and Appendix E (Acceptable Toxicity Data of
Aluminum to Freshwater Aquatic Plants), these studies often
did not report DOC and there were not enough data to
develop criteria for estuarine/marine aquatic animals and
plants. Appendix G (Acceptable Bioaccumulation Data of
Aluminum by Aquatic Organisms) data was not used in
criteria derivation. Appendix H (Other Data on Effects of
Aluminum to Freshwater Aquatic Organisms) contains data
that are not used in the criteria derivation because they were
not of sufficient quality.
No edits.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief ofPlanning
Division, Riverside
County Flood Control
and Water
Conservation District)
(M) The draft criteria document states that a total of 7,483 surface
samples were collected and analyzed for dissolved and total
aluminum. The EPA should describe what fraction of the Total
Aluminum measured in these samples was in a form that was likely
to become bioavailable under natural conditions and what fraction
was in the inert, insoluble form previously acknowledged as "not
biologically available." [DAC (2) pg. 69]
The Water Quality Data Portal does not describe what
fractions of these samples are bioavailable, thus, we are
unable to provide this information.
No edits.
73
-------
( <11111110III
Number
(Oriiiini/iilioii)
Public (onimcnl on Topic X: Rciiiirdinii (lie (lociiiiionl in
!icncr;il
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Criloriii Dociinuiil
EPA-HQ-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
5. EPA should be explicit that the criteria are site-specific.
UWAG recommends that EPA clearly state that the criteria, when
finalized, should be applied on a site-specific basis. In several
locations of the Draft Criteria, EPA states that site-specific
measurements of these parameters provides the greatest certainty
ofprotection:
"Like the acute criterion, the freshwater chronic criterion, known
as the Criterion Continuous Concentration (CCC), is also
dependent upon the set of water chemistry conditions at the site."
Draft Criteria at xiii. As the criteria are derived based on site-
specific parameters, they logically are applicable on a site-specific
basis as EPA recognizes.
The criteria can be applied on a site-specific basis, and a state
could choose to apply them on another basis, such as an
ecoregional basis by using water chemistry input data that
would appropriately represent the area selected and the
designated use for those waters.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Other Items
Below are additional areas we have identified within the draft
criteria document which would benefit from correction or
clarification from EPA.
In Section 2.7.1 (p 27) the test conditions of the P.
subcapitata studies evaluated for the MLR as listed may
not be correct.
The range of DOC concentrations tested is given as "0. -
1.9mg/L. " The algae results presented in Gensemer et al.
2017 show the range of DOC concentrations tested were
0.3 to 4.0mg/L.
Thank you for your suggestion; the text was corrected.
Section 2.7.1
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
3. Pase 4. The last paragraph on this pase reports that data
obtained from the "Water Quality Data Portal. " The range of
concentrations reported for dissolved aluminum is from 0.8
micrograms per liter to 20,600 micrograms per liter. The latter
value is provided without qualification, even though it far exceeds
the equilibrium solubility of aluminum in most natural waters. This
is misleading. If the 20.6 milligrams per liter value was from acid
mine drainage, the number would make sense; however, this is not
mentioned. The concluding sentence of this paragraph reports that
the 90th percentile for concentrations of dissolved aluminum in
groundwater is 11 micrograms per liter; this does make sense for
the typical range ofpH values for natural waters, but there is no
mention of the relation to surface waters. The final sentence also
refers the reader to Figure 1, which provides a range of
concentrations for dissolved aluminum in groundwater.
Discussion is needed to put these numbers in context for the
reader.
The Water Quality Data Portal did not provide enough
information to clarify if this is the case. This is the available
data from the Water Quality Data Portal.
No edits.
74
-------
TOPIC 9: Comments regarding the Endangered Species Act
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public C oiiimoil 1 on l opic Rciiiudinii (lie lliidiiii^crcd
Species Ad
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rilcrhi Document
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC'))
2. National consultation for Endangered Species Act
EPA continues to issue revised water quality criteria without
developing the biological evaluations and consultation of the
effects of criteria levels on endangered species as required under
the Endangered Species Act (16 U.S.C. 1531-1544). EPA
consultations done after state adoption of criteria can delay EPA
approval of state criteria for years. ADEC strongly urges EPA to
complete ESA consultation before issuing final criteria. National
ESA consultation prior to publishing final criteria would be most
effective in protecting endangered species and would alleviate
further burden on states and delays in EPA action on state water
quality standards.
The Endangered Species Act does not require EPA to
develop a biological evaluation and consult with the Services
on water quality criteria developed under CWA Section
304(a). Ambient water quality criteria are recommendations
and do not impose legally binding requirements on states to
adopt these specific criteria recommendations, nor do they
bind the Agency to take future federal action with respect to
state standards that are less, more, or equally stringent than
the recommended value. States are not required to adopt the
national recommended criteria. Thus, by developing national
recommended criteria, EPA is not authorizing, funding, or
carrying out an agency action subject to the ESA. In addition,
recommended 304(a) criteria are not reviewable final agency
actions.
EPA's statement in the 2011 Memorandum of Agreement
does not create a binding requirement for the Agency to
engage in ESA consultation. That agreement states that the
"memorandum is intended only to improve the internal
management of EPA and the Services and is not intended to,
and does not, create any right or benefit, substantial or
procedural, enforceable at law or equity by a party against the
United States, its agencies or instrumentalities, its officers or
employees, or any other person". 66 Fed. Reg. 11202, 11217
(Feb. 22, 2001).
Further, endangered species have not been found to be more
toxicologically sensitive than other species, based on
available data. The distribution of any particularly sensitive
listed species which might affect the appropriate local water
quality criteria is location specific; Allowing the most
sensitive location-specific potential concerns to determine
national recommendations, including for locations where
especially sensitive endangered species are not present,
would tend to inappropriately bias those recommendations.
The EPA believes that it is most efficient for states to modify
national criteria recommendations for aluminum, if
necessary, based on the presence of any localized highly
sensitive species-specific concerns or use other scientifically
defensible methods when adopting new or revised water
No edits.
EPA-HQ-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
EPA's duty to complete its Section 7 consultation process prior to
finalizing any recommended criterion is firmly established by the
text of the ESA and by the Memorandum of Agreement that EPA
entered with the Services to clarify the procedures for ESA
compliance in taking action under the CWA. The latter document
states that:
EPA and the Services will conduct a section 7 consultation on the
aquatic life criteria to assess the effect of the criteria on listed
species and designated critical habitat. EPA and the Services will
also conduct a conference regarding species proposed for listing
and proposed designated critical habitat. EPA will consider the
results of this consultation as it implements and refines its criteria
program, including decisions regarding the relative priorities of
revising existing criteria and developing new criteria.
[Memorandum of Agreement Between the Environmental
Protection Agency, Fish and Wildlife Service and National Marine
Fisheries Service Regarding Enhanced Coordination Under the
Clean Water Act and Endangered Species Act at 11 (Jan. 2001)].
EPA asserts that the meaning of water quality criteria in Section
304(a)(1) of the CWA, is "anon-regulatory, scientific assessment
of ecological and human health effects. " [DRAFT AQUATIC LIFE
AMBIENT WATER QUALITY CRITERIA FOR ALUMINUM 2017
(hereafter "DRAFT CRITERIA") at 4. Docket#: EPA-HQ-OW-
2017-0260-0002], However, EPA also correctly notes that these:
If water quality criteria associated with specific surface water uses
are adopted by a state or EPA as water quality standards under
75
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic V: Rciiiirdinii (lie lliidiin^crcd
Species Acl
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
section 303, they become applicable Clean Water Act water quality
standards in ambient waters within that state or authorized tribe.
Water quality criteria adopted in state water quality standards
could have the same numerical values as criteria developed under
section 304. However, in many situations states might want to
adjust water quality criteria developed under section 304 to reflect
local environmental conditions and human exposure patterns. [Id],
The establishment of water quality criteria under Section 304(a)(1)
is an action for purposes of Section 7 because such criteria set the
ceiling for establishment of water quality standards. Even if water
quality criteria are not regulatory per se, like a Forest
Management Plan under the National Forest Management Act or
similar federal agency acts, consequences still flow from the
establishment of the criteria. The federal act of establishing these
criteria has both direct and indirect effects for species, especially
since methodologies are chosen and species get excluded from
consideration now with consequences for how states may proceed
in establishing water quality standards. Additionally, criteria for
toxic pollutants under Section 303(b) are less "optional" than
criteria developed for non-toxic pollutants. This makes the
adoption of criteria for toxics certainly more "regulatory" in
nature.
quality standards under CWA section 303(c). When
appropriate, the EPA intends to consult with the Services
regarding future approvals of new or revised state water
quality standards under Clean Water Act Section 303(c) per
the Endangered Species Act requirements.
EPA-HQ-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
Because of the incredibly endangered status of many freshwater
mussels in the United States, it is simply unacceptable for EPA to
ignore the input of experts in the Fish and Wildlife Service to set a
protective criterion for freshwater mussels. The reality is that the
EPA lacks the capacity and ability to take action that is protective
of endangered species. If EPA finalizes this criterion without
consultations, the Center will take legal action to remedy this
gross deficiency.
EPA also has an independent obligation under Section 7(a)(1), to
"carrying out [its] programs for the conservation of endangered
species and threatened species. "[16 U.S.C. § 1536(a)(1)], By
consulting on national criteria and coordinating with the Services,
EPA can move toward meeting its Section 7(a)(1) obligations.
The Clean Water Act mandates that water quality standards
protect not only fish, but all aquatic organisms and other wildlife
that depend on healthy streams. Section 303(c) requires that such
standards "shall be established taking into consideration their use
In response to concerns raised by the USFWS, and others,
that endangered freshwater mussels may be sensitive to
aluminum, the EPA included recent studies by USGS on
freshwater mussels, the fatmucket mussel (Lampsilis
siliquoidea), in the family Unionidae in the 2018 aluminum
criteria derivation. Freshwater mussels in the family
Unionidae are known to be sensitive to a number of
chemicals, including metals and organic compounds (Wang
etal 2018; U.S. EPA 2013).
While the 96-hr LC50 juvenile test included in the criteria
document failed to elicit an acute 50% response at the highest
concentration tested (6,302 |ig/L total aluminum, or 29,492
|ig/L when normalized), the 28-day biomass normalized
SMCV ranked as the fourth most sensitive senus in the
dataset. The mussel's chronic value is greater than the most
sensitive species, Atlantic salmon, and the freshwater
criterion. Thus, the chronic criterion is expected to be
protective of freshwater mussels and related species. The
No edits.
76
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic V: Rciiiirdinii (lie lliidiin^crcd
Species Acl
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
and value for. . . propagation of fish and wildlife, " among other
things.[33 U.S.C. § 1313(c)(2)(A) (emphasis added); see also id. §
1252(a) (directing states to develop comprehensive programs for
controlling water pollution giving due regard to improvements
necessary to "conserve such waters for the protection and
propagation of fish and aquatic life and wildlife ").] EPA's
regulations require states to develop standards that will "[s]erve
the purposes of the Act, " meaning that they will "provide water
quality for the protection and propagation offish, shellfish and
wildlife, " among other things.[40 C.F.R. § 130.3],
fatmucket mussel tested is not a threatened and/or endangered
species, but the genus Lampsilis contains several listed
species with a wide distribution across the United States.
Additional testing on endangered mussel species, or closely
related surrogates, would be useful to further examine the
potential risk of aluminum exposures to endangered
freshwater mussels.
The EPA believes that it is most efficient for states to modify
national criteria recommendations for aluminum, if
necessary, based on the presence of any localized highly
sensitive species-specific concerns or use other scientifically
defensible methods when adopting new or revised water
quality standards under CWA section 303(c). When
appropriate, the EPA intends to consult with the Services
regarding future approvals of new or revised state water
quality standards under Clean Water Act Section 303(c) per
the Endangered Species Act requirements.
EPA-HQ-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
The Fish and Wildlife Coordination Act ("FWCA ") gives the U.S.
Fish and Wildlife Service ("FWS") broad authority to protect
freshwater wildlife resources through coordination and providing
assistance to all federal agencies regarding actions that may
impact U.S. waters.[16 U.S.C. § 661 et. seq]. To ensure that the
final aluminum water quality criteria is fully protective of all types
of wildlife, EPA should engage the FWS broadly not just as is
clearly legally required by the ESA but also engage other
divisions of the FWS that may have additional expertise and
information that would benefit the EPA.
Congress expected that the EPA would develop water quality
criteria with input from the FWS and other federal agencies. At its
outset, Section 304(a) states "The Administrator, after
consultation with appropriate Federal and State agencies and
other interested persons, shall develop and publish " water quality
criteria.[33 U.S.C. § 1314(a)(1)], Furthermore, Section 511 of the
CWA, affirms that the CWA does not limit or preclude this type of
coordination under the FWCA.[33 U.S.C. § 1371], In passing the
original CWA, the House and Senate proposed different versions of
Section 511. The Senate version would have limited "the
consultation and coordination requirements of the Fish and
Wildlife Coordination Act. . . to the provisions of section 306, the
publication of information under section 304 and the establishment
The EPA disagrees that before making general
recommendations to states regarding future state actions to
adopt aluminum criteria ("national recommendations"), it is
helpful or necessary to first engage in consultation under the
ESA to ensure that any possible subsequent federal action to
approve new or revised state aluminum criteria consistent
with the national recommendations would be protective of
listed species. The national criteria recommendations for
aluminum do not impose legally binding requirements on
states to adopt these specific criteria recommendations, nor
does it bind the Agency to take future federal action with
respect to state standards that are less, more, or equally
stringent than the guidance value. States are not required to
adopt the national recommended criteria. Thus, by
developing national recommended criteria, EPA is not
authorizing, funding, or carrying out an agency action subject
to the ESA. In addition, recommended 304(a) criteria are not
reviewable final agency actions.
The 2018 aluminum Aquatic Life Ambient Water Quality
Criteria provide recommendations for aquatic life. These
criteria recommendations are intended to be protective of
Aquatic Life Designated Uses, not other uses designated by a
state. Aquatic dependent wildlife data, including for birds or
No edits.
77
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C 0111 moil I on lopic V: Kc»;ir(lin» (lie llndiin^oivd
Speck's Ad
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
of guidelines under section 403 but not to the imposition of any
specific effluent limitation on a particular source. "[S. REP. 92-
414, 92nd Cong. (1 972), reprinted in, 1972 U.S.C.C.A.N. 3668, at
3751], The House version did not contain a limitation on the scope
of the Fish and Wildlife Coordination Act, and ultimately, the
Congress adopted a compromise version that did not limit the
scope of the Fish and Wildlife Coordination Act.[S. CONF. REP.
92-1236 (1972)]. Clearly, though, Congress intended that EPA
would involve the FWS in many aspects of the CWA's
implementation.
Coordination under the FWCA should not be burdensome or
formalistic. But the reality is that EPA has consistently and
systemically failed to fully consider the impacts of its proposals on
aquatic wildlife. One of Congress' stated goals in passing the
CWA was to achieve "water quality which provides for the
protection and propagation offish, shellfish, and wildlife. "[33
U.S.C. §1251(a)(2) (emphasis added)]. Despite this clear
statement of a national goal, and despite the repeated inclusion of
wildlife as a top priority for protection under the CWA, EPA has
consistently failed to fully consider aquatic-dependent wildlife in
the development of national criteria.[33 U.S.C. § 1314(a)(1) ("The
Administrator, after consultation with appropriate Federal and
State agencies and other interested persons, shall develop and
publish... from time to time thereafter... criteria for water quality
accurately reflecting the latest scientific knowledge (A) on the kind
and extent of all identifiable effects on health and welfare
including, but not limited to, plankton, fish, shellfish, wildlife, plant
life, shorelines, beaches, esthetics, and recreation which may be
expected from the presence ofpollutants in any body of water,
including groundwater"); 33 U.S.C. § 1314(a)(2) ("The
Administrator, after consultation with appropriate Federal and
State agencies and other interested persons, shall develop and
publish... information... on the factors necessary for the protection
and propagation of shellfish, fish, andwildlife... "); 33 U.S.C. §
1314(a)(5)(A) ("the Administrator, to the extent practicable before
consideration of any request under section 1311(g) of this title and
within six months after December 27, 1977, shall develop and
publish information on the factors necessary for the protection of
public water supplies, and the protection and propagation of a
balanced population of shellfish, fish and wildlife, and to allow
recreational activities, in and on the water. ").] The Center
other taxa, are beyond the scope of the data considered in the
2018 Aquatic Life Ambient Water Quality Criteria.
The references provided were all included in the final
aluminum criteria document with the exception of Naimo
(1995).
78
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic V: Kc»;ir(lin» (lie llndiin^oivd
Speck's Ad
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
recommends that EPA develop water quality criteria that are fully
protective of all types of wildlife, including taxonomic groups that
EPA routinely overlooks and omits from its analysis. Using the
FWCA coordination process as a framework to achieve this would
strengthen the final aluminum criteria.
REFERENCES
Elangovan, R., K.N. White and C.R. McCrohan. 1997.
Bioaccumulation of aluminum in the freshwater snail Lymnaea
stagnalis at neutral pH.
Huebner, J.D., andK.S. Pynnonen. 1992. Viability of glochidia of
two species of Anodonta exposed to low pH and selected metals.
Canadian Journal of Zoology 70: 2348-55.
Kadar, E., J. Salanki, R. Jugdaohsingh, J.J. Powell, C.R.
McCrohan, and K.N. White. 2001. Avoidance responses to
aluminum in the freshwater bivalve Anodonta cygnea. Aquatic
Toxicology 55: 137-148.
Mackie, G.L. and B. W. Kilgour. 1995. Efficacy and role of alum in
removal of zebra mussel veliger larvae from raw water supplies.
Water Research 29(2): 731-744.
Malley, D.E., J.D. Huebner, and K. Donkersloot. 1988. Effects of
ionic composition of blood and tissues of Anodonta grandis
grandis (Bivalvia) of an addition of aluminum and acid to a lake.
Archives of Environmental Contamination and Toxicology 17:479-
491.
Naimo, T.J. 1995. A review of the effects of heavy metals on
freshwater mussels. Ecotoxicology 4(6): 341-362.
Taskinen, J., P. Berg, M. Saarinen-Valta, S. Valila, E. Maenpaa,
K. Myllynen, and J. Pakkala. 2011. Effect of pH, iron and
aluminum on survival of early life history stages of the endangered
freshwater pearl mussel, Margaritifera margaritifera.
Toxicological and Environmental Chemistry 93(9): 1764-1777.
D01: 10.1080/02 772248.2011.610798.
79
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic V: Rciiiirdinii (lie lliidiin^crcd
Species Ad
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
El 'A-11Q-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
Water quality standards under the Clean Water Ad ("CIIA "j must
proted all existing uses in a waterbody, and such "uses" often
include supporting species that are listed as threatened or
endangered pursuant to the Endangered Species Act. [3 3 U.S.C. §
1313], Additionally, under Section 7 of the Endangered Species
Act ("ESA "), and its implementing regulations each federal
agency, in consultation with the U.S. Fish and Wildlife, must
insure that any action authorized, funded, or carried out by the
agency is not likely to (1) jeopardize the continued existence of any
threatened or endangered species or (2) result in the destruction or
adverse modification of the critical habitat of such species. [16
U.S.C. § 1536(a)(2); 50 C.F.R. § 402.14(a)], "Action" is broadly
defined to include actions that may directly or indirectly cause
modifications to the land, water, or air, and actions that are
intended to conserve listed species or their habitat. [50 C.F.R. §
402.02], EPA thus must ensure that any criteria that it
recommends to states for adoption will be fully protective of listed
species.
80
-------
TOPIC 10: Comments regarding exposure routes
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Comment on Topic 10: Rciiiirdinii exposure routes
T.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rileriii Documciil
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
There are also some contradictions relative to the potential
bioavailability of aluminum and whether or not the proposed
criteria is conservative. Section 2.6.2 Measures of Effect includes a
short discussion that application of the aluminum criteria to total
recoverable aluminum may be considered to be conservative as the
total recoverable measurement also includes aluminum bound to
particulates, implying that such bound aluminum is less
bioavailable. However, in section 3.3 Bioaccumulation studies are
presented which show that dietary exposures to aluminum bound
to particulates is bioavailable to grazing aquatic invertebrates. As
water quality criteria are derived to also be protective of these
invertebrates, it would seem that the current proposed criteria
based on total recoverable aluminum measurements are not
conservative, but appropriate for protection of species across the
full range of potential exposure pathways. Therefore, the document
should not overstate the potential for a conservative application of
the criteria though the use of total recoverable aluminum
measurements.
Section 3.3 text discusses that aluminum bound to humic
acids may be bioavailable via grazing. In general, humic
acids do not equate to particulates as suggested by the
comment. Section 3.3 also notes that bioaccumulation and
toxicity via the diet are considered unlikely relative to direct
waterborne aluminum toxicity (Handy 1993; Poston 1991).
This conclusion is also supported by the lack of any
biomagnification within freshwater invertebrates that are
likely to be prey of fish in acidic, aluminum-rich rivers
(Herrmann and Frick 1995; Otto and Svensson 1983; Wren
and Stephenson 1991). The opposite phenomena, trophic
dilution up the food chain, has been suggested (King et al.
1992).
No edits.
81
-------
TOPIC 11: Comments regarding other general issues
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Comment on Topic 11: Rciiiirdinii oilier general issues
T.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rileriii Documcnl
EPA-HQ-OW-
2017-0260-0013
(Ricardo Cantu,
President, OspreyOwl
Environmental, LLC)
1 have modified Method 1669 that I use during my "Clean
Sampling" events. My field blank is set out first, upwind of sample
location, and left open to determine environmental impact of
airborne metals. In all my sampling events I will scout the river to
pick a location that is very representative of river flows, far
removed from bridges and road traffic, with sampling taking place
beyond riverbank tree canopy. What 1 have found for ambient
contamination is very consistent with Japan's findings and in many
instances less. I believe the noted higher levels mentioned in the
draft are a result of contamination carried in by the sampling
team, or improperly cleaned sampling equipment. I believe the
ambient background concentrations of aluminum are overstated
and not representative of actual ambient conditions.
During my sampling projects 1 worked for the City of Springfield,
Massachusetts at both the wastewater and water plants. The main
focus was to test the receiving waters for aluminum, copper and
lead at low flow conditions. This was done over the summer of
2016 when the receiving waters were near, and in a few instances,
below 7Q10 conditions. Background levels were extremely low and
patterns were noted during the few times of rain events. The same
patterns repeat over and over again regardless of the river being
sampled or the watershed location. If the water from the wetlands
is stable, absent of rain and groundwater influx, then the humic,
fiulvic and tannic acids remain in the wetland areas weathering the
rocks and organics that are associated with these wetlands while
building up aluminum concentrations and lowering the wetland
pH. When it rains, the water from the wetlands rises, enters the
main waterbodies, drops the pH in these main waterbodies, which
in turn begins to drop the alkalinity due to the increased buffering
capacity needs, and brings along an increase in both total and
dissolved aluminum.
The West Parish Filters (WPF), Water Treatment Plant had
collected over two years (December of 2012 through February of
2015) of chemical concentration data at their two supply
reservoirs. The Cobble Mountain Reservoir has 23 billion gallons
of storage and the Borden Reservoir has about 2.5 billion gallons
of storage. Both reservoirs are in the same watershed, receive the
same amount of rainfall as they are approximately located V2 mile
from each other and are impacted by the same soils and
Thank you for your comment. Several sentences have been
added to the "Occurrence" section regarding recent common
use of "clean sampling techniques" thereby reducing potential
for any contamination of samples. We expect new methods
for measuring aluminum will be available in the future.
Thank you for submitting the interesting data. However, the
analysis you submitted cannot be used in the criteria
derivation.
The EPA reviewed the study by Lydersen et al. (2002) and
determined that it was not acceptable for criteria derivation.
(Appendix J). The reason the study is deemed unused is that
only one aluminum concentration was tested. However, the
study did show that both Ca and Na reduced fish mortality
(Na reduced mortality more than Ca).
Section 2.1
82
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public ('oninicnl on Topic 11: Rciiiirdinii oilier ^cnoi'iil issues
I'.PA Response
Ke\ision Location in
20IX Aluminum
Crileriii Document
surroundingflora. An aerial vie ir of both reservoirs can be seen al
the following link:
https://www. google. com/maps/@42.1362127,-
72.9225551,3486m/data=!3ml !le3?hl=en
The EPA requested that the WPF collect two-years of chemical
concentration data from both reservoirs, along with several other
locations within their facility, the upper and lower lagoons and
Cook's Brook (the discharge point for the WPF treatment
backwash).
The data from the reservoirs is attached. In my cursory review of
the data I noted that the smaller reservoir has total recoverable
aluminum (TEA) fairly consistently over 100 ug/l. The larger
reservoir consistently had a TEA value of less than 50 ug/l. The
dissolved aluminum even demonstrated a wider difference in range
between the two reservoirs.
I had submitted comments on Thursday 9/21/2017 and did not
include the attachment on the Oslo Study that 1 referenced in my
comments. The acknowledgement # was lkl-8ysj-hulg. Attached
for reference with that document is the Oslo Study.
Abstract: The Effects of Ionic Strength on the Toxicity of
Aluminium to Atlantic salmon (Salmo salar) Under Non-steady
State Chemical Conditions. Please contact the EPA Docket Center,
Public Eeading Eoom to view this document. Address: 1301
Constitution Ave, NWEoom 3334 Washington, DC 20004
Telephone: 202-566-1744 Fax: 202-566- 9744 Email: docket-
customerservice@epa.gov Prepared by Espen Lydersen et al.
Authors: Espen Lydersen et al.
Eeason Eestricted: This attachment is restricted to show metadata
only because it contains copyrighted data.
Publication Eeference: Journal of Limnology 61.1 (2002): 69 - 76
EPA-HQ-OW-
2017-0260-0013
(Ricardo Cantu,
President, OspreyOwl
Environmental, LLC)
In review of the proposed draft Aluminum Criteria it is evident that
much work and review has been done to develop what the current
train of thought believes is the best fit models for determination of
aluminum toxicity. This dynamic approach is much better than the
previous static approach at predicting the toxicity of aluminum to
riverine biota.
Thank you for your comment. The 1988 AWQC for
aluminum were discussed as acid-soluble concentrations and
were subsequently expressed in terms of total recoverable
aluminum.
No edits.
83
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public (onimcnl on Topic 11: Rciiiirdinii oilier ^cnoi'iil issues
I'.PA Response
Ke\ision l.ociilion in
20IX Aluminum
Crileriii Doeiimenl
From the infancy stages of toxicity studies in the 1980s it was
determined that hardness has always played a role in the toxicity
of aluminum. There were other factors that were widely studied,
but to a very limited degree. Many of these studies are outlined in
the greater than 800 references listed on pages 76 through 141 of
the draft.
The 'Gold Book' was the standard reference for acute and chronic
toxicity. Aluminum has a chronic value of 87 ug/l and an acute
value of 750 ug/l within this document. The 'GoldBook' document
did indicate that dissolved aluminum was a better predictor of
actual toxicity than total aluminum. This belief was held for 30
years until the new release of this document. The current document
states, "Toxicity was only observed when the test solutions were
unfiltered; furthermore, dose-response was only observed using
total aluminum as opposed to measurements of dissolved or
monomeric forms (Gensemer et al. 2017). This same effect was
observed in 7-day exposures atpH 7 and 8 with the daphnid
(Ceriodaphnia dubia) where filtered test solutions were less toxic
than unfiltered solutions (Gensemer et al. 2017).... Thus, if
aluminum criteria are based on dissolved concentrations, toxicity
would likely be underestimated, as colloidal forms and hydroxide
precipitates of the metal that can dissolve under natural conditions
and become biologically available would not be measured (GEI
Consultants, Inc. 2010;"
This document uses multiple linear regressions (MLR) models and
did explore biotic ligand models (BLM) to take data results from
varying chemical concentrations during aluminum toxicity
analyses (calcium, sodium, magnesium, chlorides, sulfate etc.) and
fit the impact of these ionic concentrations into three parameters,
pH, hardness and dissolved organic carbon (DOC). It is noted in
the section 5.3 that there are data gaps and uncertainties in the
development of this draft. There is one specific section that makes
a statement offact, yet indicates the ambiguous nature of this
statement because natural waters may contain other species of
aluminum that are not biologically available.
I did develop several questions, but saw that these were brought up
in the Peer Review Comments and noted that the EPA had
responded too many of the questions.
84
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ('oninicnl on Topic 11: Rciiiirdinii oilier iionor;il issues
I'.I'A Response
Revision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
WDEQ/WQD appreciates EPA's thorough review of aluminum
studies and toxicity data for aquatic organisms. Further,
WDEQ/WQD was interested to see how EPA incorporated the
effects of other water quality parameters on aluminum toxicity
through development of multiple linear regression (MLR) models.
Nonetheless, WDEQ/WQD has concerns regarding: the lack of
standardization among toxicity studies selected for criteria
development; the assumptions and procedural exceptions used
during criteria derivation; the limited applicability of MLR
models; as well as unclear or missing information in the criteria
document.
Thank you for your comment. Substantive comments on this
topic were addressed in other sections of this Response to
Comment document. The development of the 2018 final
aluminum criteria followed the procedures described in the
1985 Guidelines, with the advancement of more complex
consideration of water chemistry impacts on aluminum
bioavailability. All studies used in criteria were thoroughly
reviewed for data quality. The applicability of the criteria
across a broader range of US waters was enhanced by the
addition of data and MLR equation incorporate that
additional data. Unclear or missing information noted in
public comments on the 2017 draft was addressed. The
criteria document and all additional data and modeling
included after the 2017 draft document were externally peer
reviewed. EPA asserts the criteria represent the latest and
most scientifically-defensible science.
No edits.
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
In summary, we request that EPA base the updated aluminum
criteria on a wider range of water quality parameters and also
consider the use of filtration to remove natural sources that greatly
increase the aluminum concentrations especially in wet weather.
The potential for aluminum toxicity in surface waters is directly
related to the chemical form of aluminum present, which is highly
dependent on water quality characteristics of the waterway. We
hope that the characteristics typical of many California waterways
are represented and considered during development of the final
recommended standards.
[Attachment A: Natural background concentrations during wet
weather in southern California creeks]
Thank you for your comment.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. As a result, the
water chemistry bounds for the 2018 criteria were expanded,
with details and rationale provided in the criteria document.
The EPA is aware, and has noted in the 2018 aluminum
criteria document, that under natural conditions not all forms
of aluminum would be biologically available to aquatic
species (e.g., clay-bound aluminum). The EPA has also noted
in its 2018 final aluminum criteria document that the EPA
Methods 200.7 and 200.8 are the only currently approved
methods for measuring aluminum in natural waters and
wastes for NPDES permits. The EPA further notes that
research on new analytical methods is ongoing to address
concerns with including aluminum bound to particulate
matter (i.e., clay) in the total recoverable aluminum
concentrations (OSU 2018c). One approach would not acidify
the sample to pH less than 2 but rather to pH 4 (pH 4
extracted method) to better capture the bioavailable fraction
of aluminum (CIMM 2016, OSU 2018c). The method has
recently been published as Rodriguez, P.H., J.J. Arbildua, G.
Villavicencio, P. Urrestarazu, M. Opazo, A.S. Cardwell, W.
No edits.
85
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic 11: Rciiiirdinii other licncrnl issues
I'.I'A Response
Revision l.ociilion in
20IX Aluminum
Crilcriii Document
Slubblclicld. E. Nuidlicnli. and W. Adams. 20 19.
Determination of Bioavailable Aluminum in Natural Waters
in the Presence of Suspended Solids. Environ. Toxicol.
Chem. 29 April 2019. https://doi.org/10.1002/etc.4448. The
expectation is that this approach may better estimate the
bioavailable fraction of aluminum in natural waters. The EPA
is developing implementation guidance on this topic that will
be issued in the future.
EPA-HQ-OW-
2017-0260-0029
(Hall & Associates on
behalf of Minnesota
Environmental Science
and Economic Review
Board (MESERB))
Summary
The draft criteria are a marked improvement over the 1988 aquatic
life ambient quality criteria for aluminum in fresh water. However,
as with the current criteria, the draft criteria include
acknowledged uncertainties that may be resolved using water
effect ratio studies and/or by WET testing with common test
organisms (D. magna, C. dubia, and P. promelas). The draft
criteria should also include a footnote, similar to that provided
with the current aluminum criteria, warning that the calculated
criteria may be inaccurate for pH, hardness, and DOC
concentrations outside the bounds of data used to derive the
criteria. The criteria may also be inaccurate where the aluminum
present is in the form of clays or other materials that are not
bioavailable. This is particularly important for waters with high
turbidity or suspended solids as would be expected in stormwater
runoff. Finally, the criteria should be adjusted where salmonids
are not present.
Thank you for your comment.
The EPA is aware, and has noted in the 2018 aluminum
criteria document, that under natural conditions not all forms
of aluminum would be biologically available to aquatic
species (e.g., clay-bound aluminum). The EPA has also noted
in its 2018 final aluminum criteria document that the EPA
Methods 200.7 and 200.8 are the only currently approved
methods for measuring aluminum in natural waters and
wastes for NPDES permits. A new method has recently been
published as Rodriguez, P.H., J.J. Arbildua, G. Villaviccncio.
P. Urrestarazu, M. Opa/.o. A.S. Cardwell, W. Stubblefield. E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
The 2018 final aluminum criteria discuss the expanded water
chemistry bounds of the criteria, and discuss increased
uncertainty outside of the empirical water chemistry bounds
for the 2018 MLR model's underlying toxicity tests.
The 2018 final aluminum and underlying MLR is reflective
of a larger toxicity and water chemistry database than a
WER, which can depend greatly on the particular "snapshot"
conditions during which the WER tests are conducted.
No edits.
86
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ('oninicnl on Topic 11: Rciiiirdinii oilier iionor;il issues
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0035
(RichardA. Hyde,
P.E., Executive
Director, Texas
Commission on
Environmental Quality
(TCEQ))
Federally-recommended criteria for aluminum were last updated
by the EPA in 1988. The 1988 criteria were developed with a
limited number of toxicity studies, expressed as a fixed value for
waters between 6.5 and 9.0 pH units, and did not account for other
site-specific factors.
Thank you for your comment. The EPA agrees that the 1988
criteria were developed with limited studies and only
addressed pH between 6.5 and 9.0.
No edits.
EPA-HQ-OW-
2017-0260-0041
(John Heggeness and
Mary A. Siders,
Bureau of Water
Quality Planning
(BWQP), Nevada
Division of
Environmental
Protection (NDEP))
Overview of1988 Criteria Document
EPA's 1988 document for aluminum begins with a discussion of
the geochemistry of aluminum in surface water (EPA 1988). The
complexity of its geochemical behavior is attributed to five
characteristics: the amphoteric nature of aluminum, its tendency to
form complexes with anions, the formation of strong complexes
with organic acids, its tendency to form polymers, and its slow
chemical equilibration under certain conditions. These
characteristics are related to the theoretical solubility of aluminum
under different geochemical conditions. From this, it seems that
the focus is clearly on dissolved species, which we approximate by
using an operational definition of "dissolved" as those
components that pass through a 0.45- jim membrane filter.
Conditions ofpH are important specifically because of the greater
solubility of aluminum at both lower and higher pH values; again,
this relates to the theoretical solubility of aluminum across a range
ofpH values. The introduction section of the 1988 criteria
document appears to acknowledge use of "dissolved"
concentrations, stating that, "Hunter et al. (1980) reported that the
toxicity of the test solutions was directly related to the
concentration of aluminum that passed through a 0.45 /im
membrane filter. " This quote, along with the first three pages of
the 1988 document, leads the reader to believe the criteria will be
based on dissolved aluminum (i.e., data from field-filtered
samples); however, the last paragraph of the introduction section
appears to contradict this. The third sentence of the last paragraph
states; "Unless otherwise noted, all concentrations of aluminum in
water reported herein from toxicity and bioconcentration tests are
expected to be essentially equivalent to acid-soluble aluminum
concentrations. " The question becomes, how does the term, "acid-
soluble" relate to the standard definitions of "dissolved" or
"total" (i.e., field-filtered or not)?
Thank you for your comments. The discussion of the 1988
AWQC document in the 2018 final criteria document was
reviewed for clarity and edited where appropriate. The EPA
is developing implementation guidance on this topic that will
be issued in the future.
Section 2.6.2
87
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public (onimcnl on Topic 11: Rciiiirdinii oilier ^cnoi'iil issues
I'.PA Response
Ke\ision l.ociilion in
20IX Aluminum
Crileriii Doeiimenl
The "Implementation " section (pages 10-15) of the 1988 document
attempts to clarify the "filtered/unfiltered" question, stating that,
"Previous aquatic life criteria for metals and metalloids (U.S. EPA
1980) were expressed in terms of the total recoverable
measurement (U.S. EPA 1983a), but newer criteria for metals and
metalloids have been expressed in terms of the acid-soluble
measurement. " The text goes on to explain that, "acid-soluble
measurement does not require filtration of the sample at the time
of collection, as does the dissolved measurement. The only
treatment required at the time of collection is preservation by
acidification to a pH between 1.5 and 2.0, similar to that required
for the total recoverable measurement. " This quote indicates use
of data from samples that are not filtered prior to acidifying the
sample for preservation. By this description, "acid soluble " is
equivalent to what most practitioners would call "total" (i.e.,
unfiltered).
The discussion of "acid-soluble" versus "dissolved" versus
"total" versus "total recoverable" and the timing of filtering and
size of filter (0.1-jum versus 0.45-jum) continues at length in pages
12 through 15 of the 1988 document. In this discussion, it seems
that the acidified sample is then filtered; however, this is a
misrepresentation of the load of dissolved metals in the neutral-pH
stream. For example, page 13 of the 1988 document states:
"The intent of the acid-soluble measurement is to measure the
concentrations of metals and metalloids that are in true solution in
a sample that has been appropriately acidified. Therefore, material
that does not pass through a filter with smaller holes, such as a 0.1
um membrane filter should not be considered acid-soluble even if
it passes through a 0.45 um membrane filter. Optional filtration of
appropriately acidified water samples through 0.1 um membrane
filters should be considered whenever the concentration of
aluminum that passes through a 0.45-um membrane filter in an
acidified water sample exceeds a limit specified in terms of acid-
soluble aluminum."
Based on all the above, it is no wonder that the 2017 criterion
document seems confused as to whether concentrations of
dissolved (i.e., field-filtered) or total (i.e., not field-filtered)
aluminum were used as the basis of the criteria in the 1988
document for aluminum.
88
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic 11: Rciiiirdinii oilier iioiionil issues
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Dociimcnl
El 'A-11Q-OW-
2017-0260-0043
(Blake Beyea,
Standards Unit
Manager, Water
Quality Control
Division, Colorado
Department of Public
Health &
Environment)
The Colorado Water Quality Control Division (division)
appreciates the opportunity to provide comments on the 2017
Draft Aquatic Life Ambient Water Quality Criteria for Aluminum.
The division also appreciates EPA's substantial effort to develop
these draft criteria, as their scientific basis is a significant
improvement over the existing criteria.
The division is responsible for the daily implementation of the
Clean Water Act's water quality programs, including the water
quality standards programs for which states are responsible under
the Clean Water Act. Therefore, the proposed criteria and their
ability to be implemented are of interest to the division.
Thank \ ou lor \ our coimiicnl.
\o cdlls.
EPA-HQ-OW-
2017-0260-0052
(Heidi L. Dunn,
President, Freshwater
Mollusk Conservation
Society (FMCS))
Please accept these comments submitted on behalf of the
Freshwater Mollusk Conservation Society on the Draft Updated
Aquatic Life Ambient Water Quality Criteria for Aluminum in
Freshwater (https://www.gpo.gov/fdsys/pkg/FR-2017-09-
26/pdf/2017-20597.pdf.
We are writing to advocate on behalf of a freshwater standard for
aluminum that is protective of larval and juvenile forms of
freshwater mollusks and of threatened and endangered species in
particular. Freshwater mollusks are the most imperiled group of
organisms in United States with nearly two-thirds of species being
identified as at risk-of extinction. It is thus of utmost importance
for the Environmental Protection Agency to develop water quality
criteria that are protective of these sensitive organisms.
Thank you for your comments. Studies with freshwater
mussels were conducted by USGS (Lampsilis siliquoidea
acute and chronic tests reported by Wang et al. 2018) and are
included in the 2018 aluminum criteria derivation. Additional
responses to mussel comments are included in Topic 15 in
this Response to Comments document.
No edits.
EPA-HQ-OW-
2017-0260-0054
(Anonymous public
comment)
According to this Proposal:
"EPA is updating the aluminum criteria to better reflect the latest
science. Unfortunately, there are not enough data to support the
development of estuarine/marine criteria at this time. "
This critical statement should provoke the recognition of the
reality that the common-sense, responsible action here should be
to concentrate on more research and data before altering criteria,
when obviously the essential data does not exist to address the
many serious questions that are well known to the public and
health professionals and researchers.
The action necessary is research, not formulaic conjecture.
"Unlike the fixed acute and chronic values found in the 1988
document, this draft document provides users the flexibility to
Thank you for your comments. Current science demonstrates
that the toxicity of aluminum to aquatic organisms is
dependent on the water chemistry conditions, thus the criteria
were derived to be sensitive to these key water quality
parameters.
The EPA did not derive criteria for estuarine/marine waters
due to a lack of data, consistent with the comment.
The AWQC document has undergone independent, external
expert peer review and represents the best available science.
The averaging durations for the aluminum criteria are based
on long-standing EPA methodological guidance (1985
Guidelines).
No edits.
89
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public (onimcnl on Topic 11: Rciiiirdinii oilier ^cnoi'iil issues
I'.PA Response
Ke\ision l.ociilion in
20IX Aluminum
Crileriii Doeiimenl
develop site-specific criteria based on a site's water chemistry."
The public does not want flexibility when it comes to the health and
safety of the American people.
"The resulting acute criterion would have an appropriate level of
protection if the one-hour average concentration is not exceeded
more than once every three years on average. If the four-day
average concentration is not exceeded more than once every three
years on average, the chronic criterion is protective.
The flexibility here to have the determination depend on "average "
is not reassuring, because this is a presence that once it's there, it's
there, and entering into the biological chain in any concentration,
no matter the interval of occurrence is not acceptable.
Further, if one is the recipient of an above-average exposure, that
binds with biological functioning, then average is of little
consolation or rationality.
In the absence of data, or further research on the questions that
deeply concern Americans, it is particularly disturbing that the
proposed values in criteria are double the existing standards.
It is stated:
Note: Values will be different under differing water chemistry
conditions as identified in this document.
///////////////////////////////////
This is precisely the reality! Freshwaters are characterized by a
network of tributaries and variable flows. This averaging can be
totally misleading as a discharge into a tributary may have major
impact in its concentration with serious consequences in exposure
along a short segment, but then not register very much on the
average!
"Once final, the criteria will serve as recommendations to states
and tribes by defining the concentration of aluminum in water that
will protect against harmful effects to aquatic life."
To alter criteria in the absence of the necessary data and to impose
a more lenient framework based on "averages" and present it as
90
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public ('oninicnl on Topic 11: Rciiiirdinii oilier iionor;il issues
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
an official recommendation when so much is unknown is
irresponsible.
What we have now is not "final" criteria of any kind.
The EPA is entrusted with the awesome responsibility of protecting
the public from environmental hazards.
To posit this conjecture and knowingly send it out as guidance in
the absence of data and research in range and depth that would
provide assurance of safety is irresponsible.
The American people, if they were fully aware of this, would not be
happy.
EPA-HQ-OW-
2017-0260-0059
(JeffHenderson,
President, Aluminum
Extruders Council)
The Aluminum Extruders Council represents over 100 extruders
and suppliers across the United States. After reviewing all
available information on this issue, we stand in support of the
recommendations of the Aluminum Association. We encourage the
EPA to take those comments under careful consideration as you
deliberate this issue.
Thank you for your comment. The specific comments of the
Aluminum Association were addressed in this Response to
Comment document.
Edits were made
based on the
Aluminum
Association
comments, as
appropriate.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
The US EPA Database
US EPA has spent years developing the Draft Aluminum Criteria.
It is impossible, within the brief time allowed, to adequately assess
each of US EPA's data decisions for inclusion and exclusion of
specific studies. Some of the cited materials have only been
available to the public for a limited time, and they are integral to
the Draft Aluminum Criteria.
Based on a brief comparison, US EPA has included some studies
that West Virginia determined were inappropriate for inclusion,
and vice versa. These small decisions affect the genus mean acute
or chronic values (GMAVs and GMCVs) and therefore are
significant. The following table compares the GMAVs and the
resultant CMC (acute criterion) for the four most sensitive species
in the West Virginia database as compared to the US EPA
database:
[Table 3]
While the CMCs appear comparable, the West Virginia database
was normalized to a hardness of 50 mg/l, whereas the US EPA
database was normalized to a hardness of 100 mg/l. Considering
Thank you for your comments. The AWQC document has
undergone independent, external expert peer review and
represents the best available science.
The commenter is incorrect in stating that the EPA's database
was normalized to a hardness of 100 mg/L. The EPA's
database was normalized based on peer-reviewed multiple
linear regressions (DeForest et al. 2018a, b) accounting for
the variable effects of aluminum across a broad range of total
hardness, dissolved organic carbon and pH conditions. This
approach of including these three water chemistry parameters
in calculating appropriately protective criteria represents the
best available science as indicated in peer-reviewed
publications (e.g., Brix et al. 2017, ET&C). In not
considering all three critical water chemistry parameters
relevant for water chemistry, West Virginia may have come
to conclusions that are different than the EPA's. In fact, peer-
reviewed publications demonstrate that pH and DOC have a
larger overall impact on bioavailability and toxicity of
aluminum than the hardness parameter that is the focus of the
West Virginia analysis. EPA has shared these data and
analyses with West Virginia and discussed available
No edits.
91
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public (onimcnl on Topic 11: Rciiiirdinii oilier ^cnoi'iil issues
I'.PA Response
Ke\ision Location in
20IX Aluminum
Crileriii Doeiimenl
the mitigating effect of hardness, the two numbers are no longer in
the same ballpark.
Likewise, the data decisions affect the ranking of each species. The
D. magna number is significantly higher in the West Virginia
database (which was normalized for hardness) as compared to the
US EPA database (which was normalized using the MLR for
hardness, pH, and DOC for only two species). The normalization
process dramatically shifts the balance between certain species for
toxicity. Ceriodaphnia were the most sensitive genus in the West
Virginia database. Ln the US EPA database, Ceriodaphnia were
not among the four most sensitive genera.
West Virginia and US EPA took a dramatically different approach
with the chronic criterion (CCC). The West Virginia number was
based upon the final acute to chronic ratio, whereas US EPA
constructed a chronic database and calculated the FCVfrom the
GMCVs. Therefore, the comparison is not as simple. However, the
FCVs can be directly compared:
[Table 4]
The US EPA FCV is calculated at a much higher hardness, yet the
value is much lower. We believe this difference is due at least in
part to the inclusion of a recently published mussel study.
However, a more substantial part of the issue may be with the use
of the MLR, as the range of hardness and DOC are very limited in
the EPA database.
Even with the thirty-day extension, US EPA has not allowed
adequate time to evaluate each of its data decisions. This is a
lengthy exercise. Ln communications with WVDEP, US EPA claims
to have been working on the aluminum criteria for roughly four
years. However, US EPA expects the public to assess its work and
to provide meaningful, thorough comments in ninety days.
Our comments are focused on the overall issues with the criteria
development, along with the four most sensitive species identified
in the chronic database. According to the 1985 Guidelines, only
the four GMCVs which have cumulative probabilities closest to
0.05 are selected for calculation of the FCV. When less than 59
GMCVs are available, these will always be the lowest four GMCVs
IIII'OIIIKIIIOIIOIIIOMCIIS k> mussels
The criteria are not presented nor intended to represent
conditions only at a hardness of 100 mg/L. Tables presented
in the criteria document show that the criteria values change
with changing water chemistry and can be calculated for any
water chemistry conditions within the bounds of the model as
specified in the criteria document.
Since the 2017 draft document was released, additional
toxicity tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. The total
hardness of toxicity test waters ranged from 9.8 to 428 mg/L.
The DOC of toxicity test waters ranged from 0.08 to 12.3
mg/L. The pH of toxicity test waters ranged from 6.0-8.7.
Please see the 2018 final aluminum criteria document for a
detailed discussion in Section 2.7.1.
The criteria calculations in the 2018 criteria document and
associated calculator were completed per the 1985 Guidelines
procedures.
92
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public CommcMl on Topic 11: Rciiiirdinii oilier iioiionil issues
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Dociimcnl
(1985 Guidelines, p. 51). The inclusion or exclusion of other
studies will only affect N, a factor used in the FCV calculation.
While we anticipate that the problems are more extensive, the
narrow focus allows for the preparation of timely comments that
demonstrate the issues with the Draft Aluminum Criteria.
EPA-HQ-OW-
2017-0260-0070
(JeffHenderson,
President, Aluminum
Anodizers Council)
The Aluminum Anodizers Council represents over 100 anodizers
and suppliers across the United States. After reviewing all
available information on this issue, we stand in support of the
recommendations of the Aluminum Association. We encourage the
EPA to take those comments under careful consideration as you
deliberate this issue.
Thank you for your comment.
No edits.
EPA-HQ-OW-
2017-0260-0074
(Timothy F. Moore,
Risk Sciences, on
behalf ofLake Elsinore
and Canyon Lake
Nutrient TMDL Task
Force administered by
the Lake Elsinore San
Jacinto Watershed
Authority (LESJWA))
EPA also cites several other field studies where relatively high
aluminum concentrations were associated with reduced richness
and abundance of fish and invertebrate species. [Draft Criteria @
pg. 65-64] However, all of these studies were conducted in lakes
and streams with low pH (<5 s.u.) and very low hardness. Such
conditions are not typical of western waters.
Thank you for your comment. The field studies discussed was
not used in the database for the quantitative criteria
calculation approach.
No edits.
EPA-HQ-OW-
2017-0260-0075
(Steven A. Buffone,
CHHM, QEP, GIT,
Supervisor,
Compliance and
Regulatory Affairs,
CONSOL Energy Inc.)
We recommend that EPA review additional data and studies
available through states, such as West Virginia, and continue to
refine the Criteria so that expanded ranges for both hardness and
DOC are addressed.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. The total
hardness of toxicity test waters ranged from 9.8 to 428 mg/L.
The DOC of toxicity test waters ranged from 0.08 to 12.3
mg/L. The pH of toxicity test waters ranged from 6.0-8.7.
The Multiple Linear Regression equations were updated
based on this new data. As a result, the recommended bounds
of the criteria have expanded. The 2018 aluminum criteria
document provides an extensive discussion of the new,
expanded bounds of the criteria and model in Section 2.7.1.
No edits.
93
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ('oninienl on Topic 11: Rciiiirdinii oilier ^cnoi'iil issues
I'.PA Response
Ke\ision Location in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0067
(Patrick McDonnell,
Secretary,
Pennsylvania
Department of
Environmental
Protection (DEP))
3. Hardness vs. calcium content
Hardness has long been monitored by water companies due to its
tendency to cause mineral deposits in pipes and leave soap scum
on bathtubs. The correlation between hardness and ameliorative
effects on pollutants has long been known, and since this discovery
some work has been done to try to understand what elemental
components of hardness are protective and the mechanism behind
such protection. For example, research by Davies and Hall has
indicated that calcium may be the component in hardness most
responsible for biological protection against some common toxins.
[Trevor D. Davies and Ken J. Hall, "Importance of Calcium in
Modifying the Acute Toxicity of Sodium Sulphate to Hyalella
Azteca and Daphnia Magna," Environmental Toxicology and
Chemistry 26, no. 6 (2007): 1243-1247.] Knowing what
components of hardness are protective and establishing standards
based upon them could ultimately lead to better criteria for aquatic
life protection.
EPA should consider the possible use of calcium and/or
magnesium concentrations to see if they correlate with biological
protection better (or worse) than the more general "hardness"
parameter.
The EPA is aware of studies indicating the importance of
calcium in the effect of hardness on toxicity of chemicals to
aquatic organisms. However, the vast majority of aluminum
toxicity studies available provided only reported total
hardness and not individual Ca and/or Mg concentrations.
The EPA thus based the 2018 final aluminum criteria on total
hardness, a parameter frequently measured by implementing
entities, which also increases the utility of its application.
No edits.
94
-------
TOPIC 12: Comments providing information on external research
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Comment oil l opic 12: Pr
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 12: Pro\ i(liiii£ iiiloniiiilion on
cxlcriiiil rcsciirch
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
the Tit I graphs and 1 add the associated correlating colored
boxes.
[Graphs x 4]
EPA-HQ-OW-
2017-0260-0013
(Ricardo Cantu,
President, OspreyOwl
Environmental, LLC)
Only one Box correlated with increasing TRA and decreasing
DOC (yellow colored box), while both the green and blue boxes
demonstrated increased DOC with increased TRA. If the MLR
model is to hold true DOC would have to decrease with each
increase of TRA. This did not happen in the Cobble Reservoir.
The yellow box and pink box in the alkalinity graph decreased with
increasing TRA and one of the two data points in the green box
decreased with increasing TRA. This is a better predictor than the
DOC, but still did not happen in every case and therefore it would
not satisfy the requirements of a true trending model.
As the pH trends were close it was a bit tougher to follow the
trends. It does seem like in three instances the pH did slightly drop
when there was an increase in TRA.
EPA-HQ-OW-
2017-0260-0013
(Ricardo Cantu,
President, OspreyOwl
Environmental, LLC)
I took the same TRA graph and matched it against sodium and
calcium (below).
[Graphs x 3]
In the four cases where the Cobble Reservoir TRA trended upward
significantly, the sodium and calcium for that month trended
downward. This was identical to the findings of the Oslo Study,
and would be a much better predictor in a MLR model with
apparently more consistency over the use of DOC and Alkalinity.
The DOC in the Borden Reservoir (consistently higher pH) has an
abundance of DOC when compared to the Cobble Reservoir. If
DOC is used more for high aluminum waters this does not bear out
in the data from both reservoirs. The draft proposal does touch
upon DOC, but it doesn't indicate whether or not it should be
higher or lower in high TRA/ASA waters.
I believe future research should focus on the sodium and calcium
aspects ofparameters that are relevant to the toxicity of aluminum.
Also, with such an environment rich example of two reservoirs in
western Massachusetts, with identical flora and fauna, and having
such a vastly different aluminum content, this may be a ripe are for
96
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 12: Pro\ i(liiii£ inToi'miilion on
cxlcriiiil rcsciirch
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
further study.
WPF has these area protected by fencing, the fish are very mature
as no fishing is allowed in either reservoir and conditions are ideal
for a long-term study of the impacts of aluminum when considering
a whole host of impacts. The baseline can be easily established and
an in-vitro real-time study could be completed. U-Mass Amherst
and University of Oslo exchange students could pick up where the
authors of the Oslo Study left off and really get some meaningful
research with little to no doubt regarding data gaps.
EPA-HQ-OW-
2017-0260-0026
(Ricardo Cantu,
President, OspreyOwl
Environmental, LLC)
I had submitted comments on Thursday 9/21/2017 and did not
include the attachment on the Oslo Study that 1 referenced in my
comments. The acknowledgement # was lkl-8ysj-hulg. Attached
for reference with that document is the Oslo Study.
Abstract: The Effects of Ionic Strength on the Toxicity of
Aluminium to Atlantic salmon (Salmo salar) Under Non-steady
State Chemical Conditions. Please contact the EPA Docket Center,
Public Reading Room to view this document. Address: 1301
Constitution Ave, NWRoom 3334 Washington, DC 20004
Telephone: 202-566-1744 Fax: 202-566- 9744 Email: docket-
customerservice@epa.gov Prepared by Espen Lydersen et al.
Authors: Espen Lydersen et al.
Reason Restricted: This attachment is restricted to show metadata
only because it contains copyrighted data.
Publication Reference: Journal of Limnology 61.1 (2002): 69 - 76
Thank you for submitting this study. The EPA reviewed the
study and determined that it was not acceptable for criteria
derivation. (Appendix J). The reason the study is deemed
unused is that only one aluminum concentration was tested.
However, the study did show that both Ca and Na reduced
fish mortality (Na reduced mortality more than Ca).
No edits.
97
-------
TOPIC 13: Comments regarding the lack of marine criteria
( omilKMII
Number
(()ri*;ini/;ilion)
Public C onimoil 1 oil l opic 13: Kcuiirdinu (lie hick of murine
crilcriii
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
Another issue of concern for coastal states is the uncertainty
regarding the potential for aluminum to affect marine aquatic
communities. As presented in the current draft aluminum criteria
document there is insufficient data to develop marine water quality
criteria for aluminum. However, the data included in the current
criteria development document shows that aluminum toxicity to the
most sensitive marine organism studied is measured at aluminum
concentrations one to two orders of magnitude below the
concentrations that are acutely toxic to the most sensitive
freshwater organisms evaluated. EPA should work to close this
data gap quickly. The potential to adjust aluminum water quality
criteria within freshwater portions of rivers and streams that then
flow into marine waters could potentially put those downstream
waters into jeopardy if aluminum is more toxic to marine aquatic
organisms.
EPA was able to obtain some data from Australia on
estuarine/marine toxicity tests for aluminum, and that is
captured in the criteria document, however we still do not
have sufficient data to develop estuarine/marine criteria. EPA
agrees that this remains a data gap.
No edits.
EPA-HQ-OW-
2017-0260-0055
(Anonymous public
comment)
I do not think the EPA should update the aquatic criteria for "site
specific" due to recent technological advances. The criteria for
aquatic life should be updated entirely for the safety of sea life.
Many fish are being affected by pollution even though they are
within the water. EPA please protect our sea life not for site
specific but for the entirety.
Thank you for your comment. Current science indicates that a
location's water chemistry greatly affects aluminum
bioavailability and toxicity. Thus EPA's 2018 final aluminum
criteria reflects this information. EPA was able to obtain
some data from Australia on estuarine/marine toxicity tests
for aluminum, and that is captured in the criteria document,
however we still do not have sufficient data to develop
estuarine/marine criteria. EPA agrees that this remains a data
gap.
No edits.
98
-------
TOPIC 14: Comments regarding the MLR (multiple linear regression) models
( omilKMII
Number
(()ri*;ini/;ilion)
Public C onimoil 1 on lopic 14: Kc^iirdin^ (lie MI.R (multiple
lincnr regression) models
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( riloriii Document
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
There is no temperature element to the MLR model despite
significant literature that suggest temperature modulates
aluminum toxicity. For instance, Stubblefield et al. (2012) studied
several different aquatic species at pH 6. They found that pH,
dissolved organic matter, and temperature had the largest
influence on aluminum toxicity with calcium, sodium and fluoride
having only having a minor influence. Other studies have found
similar relationships between aluminum toxicity and temperature
for brown trout and Atlantic salmon. The impacts of water
temperature are noted in the literature review section of the
document, but no justification was provided as to why this
parameter was not a part of the model. The model should include
temperature, or at the very least acknowledge the influence
temperature plays in the bioavailability of aluminum to aquatic
organisms, and explain why it was not incorporated into the
guidance.
We are unable to locate the reference cited; the citation is not
provided in your comments.
Temperature was not considered because of the lack of
experimental data that could be used to develop an additional
parameter in the MLR.
No edits.
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC))
The MLR models developed by DeForest et al., which EPA uses to
normalize aluminum criteria were developed with chronic toxicity
data from two animals species, one invertebrate (C. dubia; a
sensitive species) and one fish (fathead minnow; a moderately
sensitive species). If EPA recognizes the uncertainties and
limitations of the model, EPA should consider additional studies
that minimize the uncertainties and thus bolster the models'
protectiveness across a more diverse range of species before
moving forward with issuing final criteria.
Thank you for your comment. We consider the fathead
minnow to be representative of other vertebrates, and that the
cladoceran is representative for other invertebrates.
In particular, the mechanisms of aluminum toxicity to fish
based on bioavailability of aluminum are expected to be
similar across freshwater species due to similarity in gill
microenvironment among fathead minnows and other species
(e.g., salmonids). It is well known that the solubility of
aluminum decreases as pH is elevated in acidic water
(ambient surface or gill microenvironment). Aluminum
toxicity subsequently increases because aluminum
polymerizes and accumulates on the gill surface. Thus,
because of the similarity in the gill microenvironment among
freshwater fishes in soft water, it is not expected that
aluminum toxicity would be expressed differently in
salmonids, for example, as compared to the fathead minnow.
The EPA also used the invertebrate Ceriodaphnia dubia as a
surrogate for other invertebrates. The use of surrogate species
to predict effects in other organisms is a standard practice in
ecological risk assessment because toxicity data are typically
limited. C. dubia and P. promelas were used as surrogates to
test the effects of water chemistry on aluminum
bioavailability and toxicity, not for the purposes of
establishing the relative sensitivity of genera, which is
No edits.
EPA-HQ-OW-
2017-0260-0029
(Hall & Associates on
behalf of Minnesota
Environmental Science
and Economic Review
Board (MESERB))
The database and the procedures used to normalize the data for
criteria development are also subject to uncertainty as discussed in
the Draft.
There are additional uncertainties, beyond those described above,
associated with the normalization of aluminum toxicity data using
the MLR models developed by DeForest et al. (2017). The models
were developed with chronic toxicity data from two animal species,
one invertebrate (C. dubia; a sensitive species) and one fish
(fathead minnow; a moderately sensitive species). Incorporating
additional species in the model development would improve the
representativeness of all species, and further validate the MLR
model use across species. Though the pH, hardness, and DOC do
explain the majority of differences seen in the toxicity data between
the two species, there are two MLR models developed (invertebrate
C. dubia model and vertebrate P. promelas model), which better
99
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
delineate the differences in their uptake of aluminum. Because the
arthropod phylum is highly diverse, there is uncertainty in the
application of the C. dubia model across other invertebrate taxa.
(Draft at 71)
The MLR models are used to normalize the toxicity test result data
to specific conditions ofpH, hardness and DOC for evaluating the
CMC and CCC. As noted above, these regressions were developed
for a single invertebrate (C. dubia) and a fish (fathead minnow),
with the MLR models applied to all invertebrates and vertebrates,
respectively. Given the diversity of the invertebrates, this approach
lends itself to a high degree of uncertainty. Moreover, in
developing the MLR model for C. dubia, the Draft notes that a
negative pH2 term was added to account for the fact that aluminum
bioavailability decreases from pH 6-7 and then increases from
pH 7-8. (Draft at 29).
c;ipiural mi llie sciisiiiv ii\ distribution lor llie criierm.
In the 2018 final aluminum criteria, the EPA used separate
MLRs for fish and invertebrates to best capture the effects of
water chemistry on toxicity for the taxa and differences in
trends across water chemistry; Section 2.7.1 discusses the pH,
hardness and DOC normalization approach the EPA took in
the 2018 aluminum criteria document. Appendix L of the
2018 criteria document discusses the comparison of the MLR
models used to normalize the toxicity data and compares the
results of the fish and invertebrate and pooled taxa MLR
approaches in detail. In addition, the ranges of pH for the
toxicity tests was broadened above pH of 8.
The EPA used the best available science to generate a
scientifically sound updated 2018 aluminum criteria
document and described uncertainties in the criteria
document.
EPA-HQ-OW-
2017-0260-0029
(Hall & Associates on
behalf of Minnesota
Environmental Science
and Economic Review
Board (MESERB))
A review of the text justifying this relationship (aluminum toxicity
lowest at normal pH (approximately 7.0) with toxicity increasing
as the pH increases or decreases from normal) only identifies
studies using rainbow trout. (Draft at 11) Consequently, it is not
apparent that the MLR model presented for C. dubia is
appropriate, as it would seem more relevant for salmonids. The
graphic illustrating the chronic toxicity data for C. dubia and the
MLR model fit to these data (Figure 4, Draft at 30) does not
clearly show increasing toxicity as pH varies above and below 7.5.
This is due to the fact that toxicity was not evaluated at a pH of 7.5
(which would show if toxicity is further reduced at this point) and
no measurements were made at pH > 8.1 to verify that aluminum
toxicity continues to increase at higher pHfor this organism.
Whole effluent toxicity tests using aluminum sensitive organisms
(D. magna for acute tests and C. daphnia for chronic tests) are
warranted to resolve this uncertainty.
EPA-HQ-OW-
2017-0260-0030
(Nelson Brooke,
Riverkeeper et ah,
Black Warrior
Riverkeeper)
Finally, we agree with, and would like to reiterate certain
comments submitted by David Waterstreet on behalf of the
Wyoming Department of Environmental Quality, who notes:
The invertebrate and vertebrate MLR models were derived based
solely on chronic toxicity data for the cladoceran, Ceriodaphnia
dubia, and the fathead minnow, Pimephales promelas,
respectively.
After reviewing the MLRs, WDEQ/WQD's initial concern is the
100
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Re^iirriin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Re\ision l.ociilion in
20IX Aluminum
Crileriii Document
applicability of such species-specific models to broader
invertebrate and vertebrate taxonomic groups. Further, neither C.
dubia nor P. promelas are among the most sensitive taxa used to
derive criteria values at a normalizedpH of 7, hardness of 100
mg/L and DOC of 1 mg/L. EPA does not present any information
on how the MLRs would be representative of other species and/or
genera and acknowledges that including other species would
improve model representativeness, notably for the invertebrate
MLR due to arthropod diversity. Therefore, prior to finalizing the
models and criteria, WDEQ/WQD recommends that EPA explore
how other taxa may respond to varying levels of pH, hardness and
DOC. Without such an analysis, there remains fundamental
uncertainty regarding the applicability of the recommended
criteria to other taxa.
While we understand that C. dubia and P. promelas are common
indicator species used for determining toxicity, we agree that the
criteria should be evaluated for toxicity across a much broader,
more representative range of taxa.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The EPA should perform additional studies necessary to expand
the range of hardness and DOC used to derive the MLR model in
order to encompass the higher natural hardness concentrations
(>300 mg/L) commonly observed in the arid southwest and the
higher natural DOC concentrations commonly observed in
stormwater runoff (> 10 mg/L ). Pending completion of such
studies, the Tables in Appendix K should be revised to delete the
recommended values for hardness concentrations greater than 150
mg/L because these "bounded estimates" are speculative and not
supported by any actual evidence in the given range.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEP))
MassDEP has conducted independent laboratory studies in
cooperation with USGS to investigate the influence of natural
dissolved organic matter on aluminum toxicity for low hardness
waters using an aluminum-sensitive test species.
The USGS investigated the influence of dissolved organic matter
on aluminum toxicity to the species, Ceriodaphnia dubia, by
conducting a series of 7-day/3-brood chronic tests, with endpoints
of survival and reproduction. Test waters consisted of serial
dilutions of two low hardness natural waters collected from sites in
Massachusetts (Beaver Brook at South Royalston, USGS
01163900; and Unnamed Tributary 2, Whitehall Res, NR,
Woodville, USGS 010974573), which had DOC concentrations of
Thank you for your comment. Since we do not have access to
the data, the results cannot be considered at this time.
No edits.
101
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
6 and 10 mg L, well above the 5 mg L DOC maximum in the MLR
model. Tests were conducted at hardness levels of 20 and 35 mg/L,
with hardness-adjusted site waters mixed with hardness-adjusted,
low-DOC lab water (diluted well water; DOC <0.4 mg/L) to
produce test waters containing 100 percent, 50 percent, and 25
percent of the original DOC concentration. Toxicity tests were
conducted in an incubator with a controlled C02 atmosphere to
maintain pH close to the target range of 5.8-6.2. Results of these
tests are being used to estimate chronic effect concentrations for
C. dubia (e.g., EC50 for 50percent reduction in reproduction),
expressed as total (unfiltered) aluminum concentrations. Results
from the tests are being finalized. Publication of the results is
expected in March 2018. MassDEP requests that these data be
considered by EPA and that the MLR model be adjusted to extend
the upper range of DOC.
EPA-HQ-OW-
2017-0260-0058
(National Council for
Air and Stream
Improvement, Inc.
(NCASI))
Lt is helpful that a peer review panel has provided EPA with input
on the modeling approaches that EPA has considered, including
the MLR model. However, the Deforest et al. (2017) article has
only been recently made available in an early published, electronic
form. As a result, there is considerable additional scientific
information that may be generated that could support or refine the
information contained in this article, or could lead to substantial
changes in model parameters and use. In addition, EPA has
decided to select the MLR statistical modeling approach over a
more mechanistic biotic ligand model (BLM) approach. The basis
for EPA's decision should be clearly described. We note that a
BLM approach is used in EPA's revised copper aquatic life water
quality criteria (EPA 2007, EPA 2016). One advantage of
mechanistic models is that they can better capture causal
mechanisms and may therefore better predict toxicity in previously
unmeasured conditions when adequate data are available (EPA
2009). However, we acknowledge that the MLR approach may
represent an adequately predictive method for many situations that
is simpler to implement (e.g., fewer model inputs, and therefore
potentially easier to provide appropriate values for all input
parameters), while still providing practical improvements over the
existing criteria. We also note that an aluminum BLM may still be
used as an optional alternative, scientifically valid approach. EPA
should continue to assess the science and relative merits of the
MLR and BLM approaches for aluminum to ensure that important
differences are considered in future revisions of aluminum water
quality criteria.
Please reference Section 5.3.5 for the rationale for why the
EPA chose to pursue the MLR models published by DeForest
et al. (2018a, b) over the BLM approach (Santore et al. 2018).
No edits.
102
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United
States Steel
Corporation)
EPA should compare the MLR approach to other methods such as
the Biotic Ligand Model to confirm the reliability of the input
variables MLR results.
The EPA asked external expert peer reviewers to investigate
the performance of the Aluminum BLM compared to MLR
models that incorporated only pH and total hardness. Please
refer to the 2018 aluminum criteria web page
(httos://www.era.eov/wac/aauatic-life-criteria-
aluminum#2018) or EPA docket for this information. Please
also reference Section 5.3.5 for the rationale for why the EPA
chose to pursue the MLR models published by DeForest et al.
(2018a, b) over the BLM approach (Santore et al. 2018).
No edits.
EPA-HQ-OW-
2017-0260-0071
(Fredric P. Andes,
Coordinator, Federal
Water Quality
Coalition (FWQC))
While we generally support use of the MLR models, it is important
to recognize that there are some uncertainties involved
(particularly in derivation of acute criteria). Also, because the
method focuses on a few specific variables, it may not be as fully
reflective of the water quality variables that drive aluminum
toxicity as other approaches that utilize more variables. Therefore,
we believe that EPA should consider developing a comparison of
the MLR approach and other methods, such as the Biotic Ligand
Model, to help confirm the reliability of the MLR results.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Given that significant acute toxicity data exist under a range of
water quality conditions, is it possible for EPA to provide an
analysis of the applicability of the chronic MLR, perhaps even
using the BLM as an independent means of checking the accuracy
of acute normalization outcomes using the chronic MLR? Given
that a number of studies used to calculate the ACR were
unbounded, we recognize that there is some uncertainty with the
ACR presented in the draft criteria document. We recommend that
EPA review whether any additional studies would be acceptable
for refinement of the ACR. Additionally, it would be beneficial if
EPA could provide a discussion on whether a "reverse "
application of an ACR would be acceptable approach for deriving
acute criteria as an alternative to the application of the chronic
MLR to normalize acute data. Another option might be to use this
reverse application of the ACR as a bounding calculation to
confirm the accuracy of acute criteria calculations derived using
application of the chronic MLR to acute data.
[Cited References]
The 2018 final criteria document notes, in Section 5.3.5, that
both the MLR (DeForest et al. 2018a, b) models and the
BLM model (Santore et al. 2018) include the mostly the same
toxicity test data, with the BLM including additional data on
the accumulation of aluminum on the gills of Atlantic salmon
(Santore et al 2018). The MLR approach empirically curve-
fits log-log pH, total hardness and DOC relationships (with
interaction terms) to the empirical data. The BLM uses a
mechanistic model based on an underlying theory of how
water chemistry input parameters affect aluminum toxicity,
although it still has empirically derived factors.
EPA agrees that the use of the chronic MLR to normalize
acute toxicity data is an area of uncertainty. It is discussed in
the document in the Effects Characterization Section 5,
specifically in Subsection 5.3.
Chronic data were used in the MLR model used to reflect the
effects of pH, DOC and hardness on aluminum
bioavailability and toxicity to normalize the sensitivity
distribution data. Application to acute toxicity data assumes
that the same relationship with aluminum bioavailability and
aquatic toxicity are present under shorter, acute exposures,
which is postulated to be an appropriate assumption to make
given available data. This uncertainty associated with the
No edits.
103
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
model is a liilui'c research area dial could be further
investigated.
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
WDEQ/WQD has other concerns regarding MLR model
development. Both the invertebrate and vertebrate models were
developed using chronic toxicity data. To account for acute toxicity
in the models, EPA assumes that the effect of water chemistry on
aluminum bioavailability remains constant across exposure
duration. Though EPA checked acute toxicity data against the
MLR models, WDEQ/WQD questions whether this assumption is
completely valid. WDEQ/WQD recommends that EPA evaluate
MLR models based on acute toxicity data and compare these to
current models. WDEQ/WQD also questions the appropriateness
of using different assessment endpoints for each model, i.e., mean
biomass endpoints for the fathead minnow and reproduction
endpoints for the cladoceran. Similar to the concerns identified for
criteria development, WDEQ/WQD believes that different
endpoints represent different levels of organismal toxicity. Again,
WDEQ/WQD recommends that EPA standardize data when
possible and discuss the potential uncertainties that may arise
when data are not standardized.
Not all toxicity studies measure the same effects. Therefore,
the EPA chooses the most sensitive endpoint based on
growth, survival or reproduction, consistent with the 1985
Guidelines. Note: biomass is chosen over growth when
available.
No edits.
104
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
C'ri(eri;i Document
EPA-HQ-OW-
2017-0260-0035
(RichardA. Hyde,
P.E., Executive
Director, Texas
Commission on
Environmental Quality
(TCEQ))
Justification is needed to address the applicability of the MLR
model, which was developed using results of chronic tests, to the
development of acute toxic criteria. In the proposal, EPA states the
"MLR equations applied to the acute toxicity data were developed
through chronic tests, with the assumption that the effect of water
chemistry on bioavailability remains the same. "Achieving a high
degree of confidence in the results of acute and chronic toxicity
tests is inherently difficult, due to the large amount of variability
that may be introduced while conducting the test, including but not
limited to: (1) source and condition of test organisms, (2) known
quality and condition of test waters, (3) control of laboratory
conditions to conduct the test, (4) instrument calibration, and (5)
training of laboratory staff. Incorporating the results of acute
toxicity tests into the MLR model, including any evaluation of the
differences in bioavailability, is needed due to the high potential
for uncertainty already inherent in toxicity tests, and since
exposure scenarios and endpoints are not consistent among acute
and chronic tests. Information such as results of validations tests,
or detailed information regarding the assumptions in the model
may also be beneficial. Additionally, the EPA should elaborate on
the use of "acute studies [that] did not report a definitive LC50
(i.e., yielded greater than values) because the highest
concentration did not cause more than 50% mortality. "
Information such as extent of censored data, and a rationale
explaining the relative impact to the toxicity dataset should be
provided to describe this technical limitation. Use of the censored
results may not be appropriate, if the amount of censored data
comprising the dataset is substantial.
The EPA discusses the use of the chronic toxicity data
evaluating the effects of water chemistry to acute data in the
2018 final aluminum criteria document; this approach reflects
the extrapolation of the effects of water chemistry across test
durations, reflecting the same assumptions in principle
accepted in the 2007 Copper BLM-based criteria. The
approach is the most scientifically-defensible approach at this
time, based on available data.
The toxicity data that were used in the development of the
MLR models did not include censored data.
Censored toxicity values were only included for a few species
in the species sensitivity distribution; the inclusion was
intended to provide the most complete data set to represent
the range of taxa present in the environment. Use of "greater
than" values follows the "decision rule" as described in the
final aluminum criteria document (Section 3.1), as follows:
"greater than" (>) low chronic values and "less than" (<) high
chronic values were not used in the calculation of the SMCV;
but "less than" (<) low chronic values and a "greater than"
(>) high chronic values were included in the SMCV. This
approach was also followed for acute SMAV calculations.
The methodology is based on the finding that "greater than"
values for concentrations of low magnitude, and "less than"
values for concentrations of high magnitude do not generally
add significant information to the toxicity analysis. In the
2018 Final Aluminum Criteria document in Section 3.1, all
Species Mean Acute Value (SMAV) calculations were re-
evaluated to verify that they adhere to the decision rule. This
approach to the use of "greater than" values was initially
described in the 2013 Aquatic Life Ambient Water Quality
Criteria for Ammonia in Freshwater and has continued to be
applied in subsequent criteria.
No edits.
EPA-HQ-OW-
2017-0260-0047
(Kathleen M. Roberts,
Executive Director,
North American Metals
Council (NAMC))
NAMC notes that EPA applied the MLR based on the chronic
dataset to normalize the acute dataset in the development of the
acute MLR model. It is not clear whether this is the best approach
to deriving acute criteria. NAMC requests EPA to compare this
approach with the use of an acute to chronic ratio (ACR) used in
reverse, i.e., use the MLR outputs divided by the ACR. EPA could
also use the Biotic Ligand Model (BLM) to develop acute criteria
for purposes of comparison and determining the best approach to
generate the final acute value.
105
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crileriii Document
El 'A-11Q-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
EPA 's proposed criteria document uses a MLR based on chronic
data to normalize the acute dataset. The justification behind this
decision is lacking and more discussion on the validity of this
application is needed. On page 37, EPA simply states that this was
done, but provides no discussion or justification regarding whether
or not that is a valid application of the MLR. Although such
extrapolations are common from acute-to-chronic datasets (using
acute-to-chronic ratios), we are not aware of any precedence for
essentially doing this in "reverse ". We suggest additional
discussion from EPA on whether this is a valid application of the
MLR.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
The Association notes that EPA applied the MLR based on the
chronic dataset to normalize the acute dataset in the development
of the acute MLR model. Acute-to-chronic dataset extrapolations
are common, but not the reverse. Because this is a novel
application of a chronic bioavailability model to acute data, the
Association believes that additional explanation and/or
verification steps are essential to confirm the validity of this
approach. Below are several options that EPA should explore
toward this end.
Use an acute-chronic ratio (ACR) in reverse to derive
acute concentration predictions from MLR-normalized
chronic criteria concentrations. There is precedent for
this approach as used in the development of the copper
BLM to derive chronic data from the acute BLM. As part
of this approach, EPA should conduct validation of the
draft criteria ACR as there are a significant number of
unbounded acute values in the dataset provided.
Use the Biotic Ligand Model (BLM) to develop acute
criteria for purposes of comparison with the proposed
acute MLR model and use that comparison in determining
the best approach to generating final acute values.
More information on these options can be found in the attached
GEL letter report.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
We recommend that EPA further explain or justify using a MLR
based on chronic data to normalize the acute dataset. On page 37,
EPA simply states that this was done, but provides no discussion
regarding whether or not that is a valid application of the MLR.
While such extrapolations are commonly done from acute to
chronic criteria using acute-to-chronic ratios (ACRs), we are not
aware that this has been done, effectively, "in reverse. " Perhaps
106
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA can also consider comparing acute MLR calculations against
acute BLM calculations as a sensitivity analysis to further justify
the accuracy or protectiveness of applying the chronic MLR to the
acute data for purposes of calculating acute criteria.
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
WDEQ/WQD is concerned with EPA's use of exceptions for the
1985 Guidelines. In addition to the example above, EPA also opted
to use data for the fathead minnow that did not meet early life
stage (ELS) requirements when developing the vertebrate MLR
model (model discussed more below). Ln this instance, EPA states
that the fathead toxicity values are comparable to acceptable ELS
tests defined in the 1985 Guidelines, therefore their use in MLR
model development is considered appropriate. WDEQ/WQD
understands that appropriate data are not always available,
however guidelines are in place to ensure consistency and
defensibility. If exceptions are allowed, WDEQ/WQD requests that
the 1985 Guidelines be revised to include the MDR exceptions or
that EPA describe the exceptions in a formal, standalone document
so they may be considered by other entities when developing water
quality criteria.
The 1985 Guidelines has a "best available science" clause
that allows the EPA to pursue different avenues for criteria
derivation, if they are scientifically defensible. The fathead
minnow data identified by the commenter meets all
appropriate data quality requirements to be used for criteria
derivation, except that the exposure duration is not long
enough (7 days versus 28 days). The data are therefore only
used to develop bioavailability models for aluminum toxicity
and are not included in Appendix C. Appendix C is
Acceptable Chronic Toxicity Data of Aluminum to
Freshwater Aquatic Animals and includes the data used in the
species sensitivity distribution that has 28-day duration.
These studies are vital for explaining the magnitude of
differences seen in aluminum toxicity when water chemistry
conditions vary between studies. However, the 7-day fathead
minnow values were not included as core chronic data in the
sensitivity distribution used to derive the criterion for
aluminum because the exposure duration is too short
compared to the other tests used in the sensitivity distribution,
thus making relative sensitivity difficult to determine.
The aluminum criteria document and the MLR models
underlying the criteria were all subjected to independent
external peer reviewed, with positive feedback.
No edits.
EPA-HQ-OW-
2017-0260-0057
(Roger Claff, P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
The MLR may be simpler than the biotic ligand model (BLM) to
comprehend and apply but may not be as fully descriptive of the
water quality variables driving aluminum ecotoxicity. A
comparison such as depicted in the attached (Figure 1, for copper)
would be beneficial to ensure the aluminum MLR provides reliable
results in most cases, and to understand and provide guidance for
the cases where there is discrepancy between the MLR and the
BLM. It is possible such a comparison was made during criteria
development and assessment, but if so the comparison is not
presented in the document. EPA has not provided access to the
database of values used in the MLR/BLM, preventing commenters
from making this comparison independently.
Current research on modeling indicates that the MLR and
Biotic Ligand models have comparable performance in
predicting aquatic toxicity for several chemicals, as long as
both models are well-constructed and are supported with
sufficient data. For example, Brix et al. (2017) concluded that
the MLR and BLM models' performance for copper were
comparable across a wide range of water chemistries and
species (Environ. Sci. Technol., 2017, 51(9): 5182-5192).
However, the aluminum BLM has not been updated with the
new available data and has not been finalized.
The EPA asked external expert peer reviewers to investigate
the performance of the Aluminum BLM compared to MLR
models that incorporated only pH and total hardness. Please
No edits.
107
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Kc^iirdin^ llic MI.U (multiple
linciir regression) models
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
refer to the 2018 aluminum criteria web page
(httos://www.era.eov/wac/aauatic-life-criteria-
aluminum#2018) or EPA docket for this information. Please
also reference Section 5.3.5 for the rationale for why the EPA
chose to pursue the MLR models published by DeForest et al.
(2018a, b) over the BLM approach (Santore et al. 2018).
EPA-HQ-OW-
2017-0260-0046
(Jennifer Wigal,
Program Manager,
Water Quality
Standards &
Assessments, Oregon
Department of
Environmental
Quality)
In general, EPA chose assumptions that lead to more conservative
outcomes in favor of more accurate outcomes in multiple steps of
the criteria development process. At a minimum, EPA should
evaluate, quantify, and report the uncertainty in criteria values
that result from application of conservative assumptions they
applied to multiple steps in the criteria development process, and
provide the evidence that led them to conclude that conservative
assumptions are necessary to protect aquatic life uses.
Primarily, EPA used assumed Dissolved Organic Carbon (DOC)
values to normalize the toxicity endpoints (i.e. EC20's) in the
underlying toxicity studies when these were not reported in the
studies themselves (p.37). It is unclear whether the estimated DOC
values are meant to be accurate or conservative. It is also unclear
how the use of assumed DOC concentrations in the toxicity studies
impacts the accuracy of the resulting criteria. The effect of DOC
on aluminum bioavailability is foundational to the models. The
normalization of the toxicity data the criteria is based on should
favor accuracy, rather than be conservative, if assumptions are
necessary.
EPA also assumes the DeForest et al. linear models that were
developed for Pimephales promelas, and Ceriodaphnia dubia are
generally applicable to all vertebrates and invertebrates,
respectively. In addition, EPA also assumes that the DeForest et
al. linear models, which were developed using only chronic
toxicity endpoints, also adequately describe the response of acute
toxicity to changes in DOC, hardness, andpH. EPA has identified
these assumptions in Section 5.3 as a data gap.
We understand that EPA compiled the best data available, and in
some cases the ideal data is limited. However, DEQ is concerned
by the number of assumptions made in the criteria, which include
assumptions in the underlying toxicity data, assumptions upon
which the sensitivity of different species are normalized by the
models, and the expansion of the range ofparameters beyond
The DOC concentrations in the MLR equations used to
normalize the toxicity data were all measured.
The default DOC values used in the final 2018 aluminum
criteria document, when measured concentrations were not
reported by the external study authors for species in the
sensitivity distribution, are the same as those found in
Appendix C of the 2007 freshwater copper criteria document.
These default DOC values were based on a scientific analysis
of the different water types used in the studies. Authors of
published studies were contacted, the USGS and the EPA
databases were consulted, and city officials at drinking water
plants were contacted to verily the default DOCs used. Please
refer to Appendix C for more details
(lUtDs://\Yw\Y.cDa.ao\/\Yac/aaiiatic-lirc-critcria-coDDcr). since
these estimated values are meant to be as accurate as possible
given the analysis.
The best available data are being used at this time, and the
EPA chooses to be clear and transparent with all assumptions.
We consider the fathead minnow to be representative for
other vertebrates, and that the cladoceran is representative for
other invertebrates.
In particular, the mechanisms of aluminum toxicity to fish
based on bioavailability of aluminum are expected to be
similar across freshwater species due to similarity in gill
microenvironment among fathead minnows and other species
(e.g., salmonids). It is well known that the solubility of
aluminum decreases as pH is elevated in acidic water
(ambient surface or gill microenvironment). Aluminum
toxicity subsequently increases because aluminum
polymerizes and accumulates on the gill surface. Thus,
because of the similarity in the gill microenvironment among
freshwater fishes in soft water, there is no reason to expect
aluminum toxicity to be expressed differently in salmonids,
No edits.
108
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
which the model is validated. Given that there are multiple
assumptions made, starting with the most fundamental data, DEQ
questions the level of confidence EPA holds in there being enough
accurate information to formulate a valid criteria at this time.
We encourage EPA to consider refining the criteria by conducting
additional studies to expand the underlying toxicity relationships,
especially to address aluminum toxicity relative to DOC, and
normalization models for additional species that may have a
different toxic response from P. promelas and C. dubia.
for example, as compared to the fathead minnow. The EPA
also used the invertebrate Ceriodaphnia dubia as a surrogate
for other invertebrates. The use of surrogate species to predict
effects in other organisms is a standard practice in ecological
risk assessment because toxicity data are typically limited.
EPA-HQ-OW-
2017-0260-0058
(National Council for
Air and Stream
Improvement, Inc.
(NCASI))
MLR model output is most sensitive to changes in DOC
concentration (DeForest et al. 2017). Unlike pH, DOC is often not
measured or measured with limited frequency, and unlike hardness
there are no more easily obtained, measured parameters such as
specific conductivity that correlate satisfactorily with DOC to
provide adequate predictions of site specific DOC measurements
[https://www.oregon.gov/deq/RulesandRegulations/Documents/BL
M-TSD.pdf], Because DOC is the most important input parameter
affecting aluminum aquatic toxicity, it is suggested that EPA add
language recommending that DOC values be measured rather than
estimated when generating site specific aluminum water quality
criteria.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water quality
standards under CWA section 303(c). The implementation
documents that the EPA is developing are intended to provide
assistance to states and authorized tribes that adopt into the
water quality standards a criterion based on or similar to the
EPA's recommended criterion. The implementation
documents are also intended to provide assistance to other
stakeholders and the public. The EPA recognizes that there
are several aspects of the recommended criterion that will
benefit from technical support documents to enhance
implementation of state and tribal criteria and is planning to
develop such documents and make them available for public
comment.
No edits.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
The draft criteria document uses 0.5 mg/L DOC (per the
recommendations in the 2007 EPA copper criteria) for
reconstituted laboratory waters where DOC was not measured. In
the database used to develop the MLR, however, other values were
used where DOC was not measured (e.g., 0.3 mg/L DOC was used
for McCauley et al. 1983). Given the sensitivity of the MLR to
DOC, the default value used has a very significant effect on the
resulting criteria after normalization, even for a seemingly small
reduction in the default from 0.3 from 0.5 mg/L, As stated in
Appendix C of the 2007 copper criteria document, "[t]he
recommended default TOC (DOC) value for laboratory prepared
reconstituted water is 0.5 mg carbon/L (note: some newer
laboratory water systems can achieve a TOC of less than 0.5
mg/L). " The draft criteria would benefit from discussion on the
selection of 0.5 mg/L DOC, as opposed to lower values used to
develop the MLR, for the purposes of normalizing water chemistry.
Our recommendation is that EPA consider using 0.3 mg/L as a
default value for unmeasured DOC values.
The 2007 freshwater copper criteria document's Appendix C
recommendations note, "For tests with reconstituted, city tap,
or well water, default DOC values can be applied if the
author does not report a measured value. The recommended
default TOC (DOC) value for laboratory prepared
reconstituted water is 0.5 mg carbon/L (note: some newer
laboratory water systems can achieve a TOC of less than 0.5
mg/L). The recommended default value for laboratory-
prepared reconstituted water is based on the arithmetic mean
of recent measurements of DOC in reconstituted water
prepared at two Federal (U.S. EPA Cincinnati, OH, and
USGS Yankton, SD) and two consulting (Commonwealth
Biomonitoring and GLEC) laboratories (range 0.1 to 1
mg/L)."
Based on this analysis and to be consistent with other
published AWQC recommendations, the default DOC value
of 0.5 mg/L, for reconstituted water will stay the same. When
No edits.
109
-------
( omilKMII
Number
(Oi^;ini/;ilion)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
die aullioi' reports a value, llial value will be used.
Additionally, if the author reports a less than value, half that
value will be used.
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC))
ADEC is concerned that the multiple linear regression (MLR)
model will not adequately address the complexity of water
chemistry found in Alaskan waters. Diverse geologic, topographic
and hydrologic features, including the presence of permafrost, will
affect the fate and transport of aluminum. The current version of
the MLR model development bounds the upper limits of dissolved
organic carbon (DOC) at 5 mg/L; Alaska has surface waters that
naturally exceed this concentration. ADEC requests EPA expand
on the existing model's upper limits to take into account waters
with greater physicochemical ranges.
Thank you for your comment. Since the draft document was
released, additional toxicity tests were conducted with
Ceriodaphnia dubia and Pimephales promelas thereby
expanding the water chemistry empirical data used for model
development. The total hardness of toxicity test waters
ranged from 9.8 to 428 mg/L. The DOC of toxicity test
waters ranged from 0.08 to 12.3 mg/L. The pH of toxicity test
waters ranged from 6.0-8.7. As a result, the recommended
bounds have changed. The criteria calculator can be used to
address waters within a pH range of 5.0 to 10.5. For hardness
values, the criteria calculator allows entry of values between
0.01 and 430 mg/L total hardness; criteria magnitudes will
not increase or decrease by increasing the hardness above 430
mg/L total hardness (as CaC03). For DOC, the criteria
calculator will not extrapolate below the lowest empirical
DOC of 0.08 mg/L and upper limit of the empirical MLR
models will be bounded at a maximum 12.0 mg/L DOC in the
criteria calculator; criteria magnitudes will not increase or
decrease by increasing the DOC above 12.0 mg/L.
No edits.
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed-
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
The draft 2017 aluminum criteria also accounts for the influence of
other water quality parameters on aluminum toxicity. Using the
results from previous studies, EPA developed MLR models that
normalize aluminum toxicity data for invertebrate and vertebrate
taxa as a function of ambient measurements ofpH, hardness as
calcium carbonate (CaC03), and dissolved organic carbon
(DOC). The invertebrate and vertebrate MLR models were derived
based solely on chronic toxicity data for the cladoceran,
Ceriodaphnia dubia, and the fathead minnow, Pimephales
promelas, respectively.
After reviewing the MLRs, WDEQ/WQD's initial concern is the
applicability of such species-specific models to broader
invertebrate and vertebrate taxonomic groups. Further, neither C.
dubia nor P. promelas are among the most sensitive taxa used to
derive criteria values at a normalized pH of 7, hardness of 100
mg/L and DOC of 1 mg/L. EPA does not present any information
on how the MLRs would be representative of other species and/or
genera and acknowledges that including other species would
improve model representativeness, notably for the invertebrate
Thank you for your comments. It is common when evaluating
effects on organisms to use surrogate species to represent
untested species. Surrogate species are typically used as
indicators of how other species will respond. The EPA does
note the uncertainty surrounding this approach.
The mechanisms of A1 toxicity to fish based on
bioavailability of aluminum are expected to be similar across
freshwater species due to similarity in gill microenvironment
among fathead minnows and other species (e.g., salmonids).
It is well known that the solubility of aluminum decreases as
pH is elevated in acidic water (ambient surface or gill
microenvironment). Aluminum toxicity subsequently
increases because aluminum polymerizes and accumulates on
the gill surface. Thus, because of the similarity in the gill
microenvironment among freshwater fishes in soft water,
there is no reason to expect aluminum toxicity to be
expressed differently in salmonids, for example, as compared
to the fathead minnow. The EPA also used the invertebrate
Ceriodaphnia dubia as a surrogate for other invertebrates.
No edits.
110
-------
(o 111 I1H-II1
Nil in her
(Oiiiiiiii/iilion)
Public C oiii moil I on lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crileriii Document
MLR due to arthropod diversity. Therefore, prior to finalizing the
models and criteria, WDEQ/WQD recommends that EPA explore
how other taxa may respond to varying levels of pH, hardness and
DOC. Without such an analysis, there remains fundamental
uncertainty regarding the applicability of the recommended
criteria to other taxa.
The use of surrogate species to predict effects in other
organisms is a standard practice in ecological risk assessment
because toxicity data are typically limited.
Further, the EPA submitted the document to independent,
external peer review, with a favorable outcome.
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
WDEQ/WQD's final concern with model development are the
limited ranges for input values. EPA developed the MLR models
within the ranges of 5.0 - 9.0 SU, 9.8 - 127 mg/L and 0.08-5.0
mg/L for pH, CaC03 and DOC, respectively. EPA cautions model
users to avoid using higher or lower input values since these may
yield limited or extrapolated criteria values. WDEQ/WQD
questions the applicability of the MLRs to Wyoming surface waters
and requests that EPA elaborate on how the models/criteria are to
be used if ambient measures ofpH, CaC03 and DOC fall outside
of the input ranges.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. For the 2018
final criteria, the total hardness of toxicity test waters ranged
from 9.8 to 428 mg/L. The DOC of toxicity test waters
ranged from 0.08 to 12.3 mg/L. The pH of toxicity test waters
ranged from 6.0-8.7. As a result, the recommended bounds
have changed.
The criteria calculator can be used to address waters within a
pH range of 5.0 to 10.5. For hardness values, the criteria
calculator allows entry of values between 0.01 and 430 mg/L
total hardness; criteria magnitudes will not increase or
decrease by increasing the hardness above 430 mg/L total
hardness (as CaC03). For DOC, the criteria calculator will
not extrapolate below the lowest empirical DOC of 0.08
mg/L and upper limit of the empirical MLR models will be
bounded at a maximum 12.0 mg/L DOC in the criteria
calculator; criteria magnitudes will not increase or decrease
by increasing the DOC above 12.0 mg/L.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development.
As a result, the water chemistry bounds for the 2018 criteria
were thus expanded, with details and rationale provided in the
criteria document and summarized below. The criteria
calculator can be used to address waters within a pH range of
5.0 to 10.5. For hardness values, the criteria calculator allows
entry of values between 0.01 and 430 mg/L total hardness;
criteria magnitudes will not increase or decrease by
increasing the hardness above 430 mg/L total hardness (as
CaC03). For DOC, the criteria calculator will not extrapolate
No edits.
EPA-HQ-OW-
2017-0260-0021
&
EPA-HQ-OW-
2017-0260-0022
(Daryll Joyner,
Administrator, Water
Quality Standards
Program, Florida
Department of
Environmental
Protection (DEP))
While the pH range covered by the proposes Al criteria covers the
general range expected in natural freshwaters, the range of DOC
addressed by the proposed criteria is very limited and well below
the levels typically found in the majority of Florida freshwaters.
More than 90 percent of Florida's lakes and streams have DOC
concentrations about the 5 mg/L upper limit used in the proposed
criteria. Similarly, approximately 35 percent of Florida's streams
have hardness levels above the 150 mg/L upper limit for the
proposed criteria. The limited ranges of DOC and hardness
incorporated into the proposed criteria would result in Al criteria
that are more stringent than required for the protection of many
Florida freshwaters. Therefore, DEP recommends that EPA
conduct the necessary studies to expand the range of hardness and
especially DOC covered prior to finalizing the proposed criteria.
EPA-HQ-OW-
2017-0260-0023
(Stan Dempsey Jr.,
CMA President,
Colorado Mining
Association (CMA))
The Multiple Linear Regression (MLR) model is a significant
improvement over EPA's current recommended criteria because it
accounts for changes in pH, hardness, and DOC and the effect that
they have on toxicity. For hardness and DOC, the EPA limited the
criteria to the range of hardness and DOC that was used in the
MLR studies. However, for pH EPA attempts to expand the range
beyond what was used in the MLR studies. The MLR studies did
not include a pH range below 6.0 or above 8.1. Applying the model
beyond these boundaries is unacceptable. EPA needs to apply pH
limitations similar to how hardness and DOC were handled in the
criteria calculation. If the pH of water is beyond the range, then
the criteria should be calculated with a pH level equal to the upper
Ill
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
C'ri(eri;i Document
or lower extent of the range. For example, ifpH of the water is 9.5,
the criteria should be calculated with a pH of 8.1, which is equal
to the upper extent of the MLR model.
below the lowest empirical DOC of 0.08 mg/L and upper
limit of the empirical MLR models will be bounded at a
maximum 12.0 mg/L DOC in the criteria calculator; criteria
magnitudes will not increase or decrease by increasing the
DOC above 12.0 mg/L.
EPA-HQ-OW-
2017-0260-0025
(Peter T. Goodmann,
Director, Kentucky
Division of Water)
The pH for the draft recommended aluminum criteria is bound in
the 5.0 to 9.0 pH range, however, some waters, especially in areas
with historical resource extraction activities, will experience pH
outside of this range. The draft does not indicate how the
recommended criteria apply when the stream pH is outside of the
range. The division believes that further clarity or guidance is
needed for these conditions.
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
CASQA requests that the final criteria extend the range for the
DOC parameter. DOC is one of the most sensitive parameters of
the criteria calculation methodology. As DOC increases, the
bioavailability of aluminum decreases, resulting in lower criteria
for the waterway being evaluated. As proposed in the Aluminum
Notice, the maximum DOC is 5.0 mg/L. However, many waterways
in California have significantly higher concentrations of DOC. An
assessment of natural (i.e., un-impacted) streams in Southern
California found natural background concentrations of DOC
above 5 mg/L (flow-weighted) in ten of the 14 streams during wet
weather. [Stein, E. and V. Yoon. 2007. Assessment of Water
Quality Concentrations and Loads from Natural Landscapes.
Southern California Coastal Water Research Project Technical
Report 500, Appendix Vlll. February.] Three of the streams had
DOC concentrations above 20 mg/L. (See Attachment A). A study
of the Los Angeles River found that all dry weather and wet
weather samples from the main stem and tributary sites exceeded 5
mg/L. [Larry Walker Associates. 2014. Final Report Copper
Water-Effect Ratio Study to Support Implementation of the Los
Angeles River and Tributaries Metals TMDL. April.] In the
proposed criteria, the Multiple Linear Regression (MLR) criteria
outputs are bounded at a maximum DOC of 5.0 mg/L because the
available toxicity data did not extend beyond 5 mg/L. Securing the
additional toxicity data will require additional time, however, it
will allow a more accurate assessment of bioavailability and
decrease the potential for California waterways being erroneously
identified as impaired by aluminum.
We also note that a peer reviewer indicated that more data would
be needed to calibrate the model, especially for higher DOC
values, before using the model for regulatory purposes. This data
is needed to represent commonly encountered natural
112
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
Crilerhi Document
environmental conditions. As currently proposed, the DOC range
is not representative of California waterways.
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
Increasing hardness generally has the effect of decreasing the
toxicity of metals and this is true for aluminum. Aluminum is
substantially less toxic at higher levels of hardness. Similar to
DOC, the MLR criteria outputs for hardness are bounded such that
it will not be possible to accurately assess site-specific conditions
for many California waterways. The maximum total hardness used
as input in the MLR is 150 mg/L as CaC03. This is because the
toxicity input data for developing the model ranged from 9.8 to 127
mg/L. Limiting the hardness used in the MLR to 150 mg/L results
in toxicity being overestimated for values above 150 mg/L. Many
waterways in California have hardness values above the maximum
assessed in developing the proposed criteria. For example, median
values for dry weather hardness in the Los Angeles River are
shown below and significantly exceed the 150 mg/L cap. [Ibid.
Excerpted from Table 2-6 (source: City of Los Angeles WMP).
Also see Table AS.]. Similarly, average hardness in the Santa Ana
River averages between 220-350 mg/L. [Santa Ana Watershed
Project Authority (SAWPA). 2012. 2011 Annual Report of Santa
Ana River Water Quality. August]
Other approved standards have been based on higher limits for
hardness. Colorado's revised water quality standards were
approved by EPA Region 8 in 2011. These standards provide an
example of the effect of hardness values above the 150 mg/L cap
used in the Aluminum Notice. [Colorado Dept. of Public Health
and Environment - Water Quality Control Commission. Regulation
No. 31: Basic Standards and Methodologies for Surface Water (5
CCR 1002-31). Effective March 1, 2017. See Table IV: Table
Value Standards for Selected Hardnesses.
https://www.colorado.gov/pacific/sites/default/files/31 2017-
03.pdf Note: Table III - Metal parameters indicates that the
aluminum criterion is based on total recoverable. Table IV,
however, incorrectly includes the following in parenthesis in the
title: concentration in ug/L, dissolved. Use of total recoverable is
correct for aluminum in Colorado based on the discussion on page
196] The Colorado criteria apply to total recoverable aluminum,
but unlike the 1988 EPA criteria, they are adjusted for hardness.
[The previous Colorado standards included the EPA 1988 acute
and chronic recommended criteria of 750 jug/L and 87 jug/L for
total aluminum, respectively, except that the 87 jug/L chronic
113
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
Crilerhi Document
criterion did not apply when the pi I urn > 7 and hardness > 50
ppm.] They are not adjusted for DOC. The hardness cap used in
Colorado is 220 mg/L rather than the 150 mg/L used in the
Aluminum Notice. The following table shows the Colorado criteria
for selected hardness values.
[TABLE]
Excerpted from Table IV - Table Value Standards for Selected
Hardnesses; the upper cap on the calculations is a hardness of 220
mg/l; where pH is less than 7.0 in the receiving water after mixing,
either the 87 jug/l chronic total recoverable aluminum criterion or
the criterion resulting from the chronic hardness-dependent
equation will apply, whichever is more stringent.
As seen in the table above, increasing the maximum of the
hardness range to 220, as was done in Colorado, significantly
increases the acute and chronic criteria due to the decrease in
bioavailability. For aluminum, the equations are valid only for
dissolved hardness concentrations of0-220 mg/L. For dissolved
hardness concentrations above 220 mg/L, the aluminum criteria
for 220 mg/L apply.
The new standards reflected in the table above were approved by
EPA Region 8 in 2011. [U.S. EPA-Region VIII (Carol L.
Campbell, Asst. Regional Administrator; Office of Ecosystems
Protection and Remediation). Letter to Peter Butler, Chairman of
the Colorado Water Quality Control Commission Approving the
2010 Revisions to the Basic Standards and Methodologies for
Surface Water. August 4, 2011.
https://www.colorado.gov/pacific/sites/default/files/2011EPAaug4.
31.pdf] EPA stated:
Although the revised table value standards for aluminum are
substantially different from CWA§ 304(a) recommendations for
aluminum [i.e., 1988 criteria], EPA agrees that the revised
standards are scientifically defensible and protective of aquatic
life.
In the approval letter, EPA included a comparison of acute toxicity
data with the 1988 EPA acute criterion of 750 jug/L and the new
Colorado hardness dependent criterion.
114
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ke^iirriin^ llic MI.U (multiple
linoiir regression) models
I.PA Response
Ke\ision l.ociilion in
20IX Aluminum
Crileriii Document
[FIGURE]
Excerpted from EPA 8 August 4, 2011 letter approving revised
water quality standards in Colorado.
As seen in the figure, the Colorado hardness-adjusted acute
criteria are protective at the higher levels of hardness based on the
available toxicity data.
More testing will be needed to establish new aluminum criteria
capable of assessing higher levels of hardness together with DOC.
Nevertheless, the results will allow a more accurate assessment of
risk in waterways with relatively high levels of hardness, such as
those in California.
EPA-HQ-OW-
2017-0260-0029
(Hall & Associates on
behalf of Minnesota
Environmental Science
and Economic Review
Board (MESERB))
The draft criteria were developed using multiple linear regression
(MLR) models to predict the toxicity of aluminum as a function of
multiple combinations of pH, hardness, and DOC conditions.
(Draft at xi) based on 22 chronic tests with fathead minnows and
23 chronic tests with Ceriodaphnia dubia. The resulting MLR
criteria are bounded at a maximum of 150 mg/L hardness and 5.0
mg/L DOC, to reflect the bounds of the underlying model data,
whereas the pH covers the range of 5.0 to 9.0.
It should be noted that the MLR criteria outputs are bounded at a
maximum of 150 mg/L total hardness, as CaC03, and DOC of 5.0
ms/L, because the available toxicity data did not extend bevond
these maxima (input data ransed from 9.8 to 127 ms/L for
hardness and 0.08 to 5 mg/L for DOC). The user can input values
for areas with hardness greater than 150 mg/L and DOC of 5
mg/L, but the criteria output for these parameters will be limited at
the bounds stated due to underlvins data limitations. The pH ranee
of the model is from 5.0 to 9.0. extending bevond the ranse of
empirical data used for model development (pH 6.0 to 8.1). This is
provided to be protective of a broader range of natural waters;
however, values estimated outside of the ranse of the data are
more uncertain. (Draft at xii) (Emphasis added)
The criteria values outside of the model input data range are more
stringent than those within the model input range under the same
hardness and DOC conditions and have greater uncertainty.
(Draft at 5 7)
115
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Kc^iirdin^ llic MI.U (multiple
liuciir regression) models
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
As discussed, factors that tend to mitigate the toxicity of aluminum
(hardness and DOC) were bounded based on the range of
concentrations present in the toxicity database used to develop the
criteria. The pH range, however, was extended beyond the
empirical data range. The rationale presented for extending the pH
range is to be protective, but as noted, results in predictions that
are more uncertain because they lie outside the range that was
evaluated. Given this acknowledgement of uncertainty, the
proposed criteria should also include a footnote indicating that use
of a water effect ratio may be appropriate where the ambient pH,
hardness, or DOC falls outside the testing boundaries used for the
development of the revised criteria.
EPA-HQ-OW-
2017-0260-0032
(Phillip M. Gonet,
President, Illinois Coal
Association (ICA))
The Multiple Linear Regression (MLR) predicts the toxicity of
aluminum based on the level for pH, hardness, and DOC. The
Draft Criteria points out that the studies used to develop the MLR
had a pH range of 6.0 - 8.1 standard units. The criteria should be
limited to this pH range and should not be extrapolated beyond it.
The reliability of MLR models results is uncertain above a pH of
8.1 or below a pH of 6.0.
EPA-HQ-OW-
2017-0260-0034
(James Boswell, Senior
Manager,
Environmental,
Peabocfy Energy)
The draft criteria have an upper bound for hardness of 150 mg/L.
Peabody agrees that this is appropriate based on the empirical
data that was incorporated into the MLR model. However, this
limitation reduces the models representativeness in regions with
high hardness. Coal mining facilities are located in sedimentary
rock deposits which often have high hardness levels under natural
conditions. The coal mining process of blasting and mixing the
overburden strata increases the hardness further, to levels in
excess of 150 mg/L. As a reference, the sites provided in Table 1
showed hardness ranges of21-824 mg/L in New Mexico, 36 -
3,838 mg/L in Arizona, and 130 - 818 mg/L in Colorado. The
hardness limit significantly underestimates the hardness ranges
seen in the environment, including the undisturbed environment
characterized by these concentration ranges. Generally speaking,
for most metals criteria increased hardness is associated with
reduced toxicity to aquatic organisms. As such, limiting the
criteria to a hardness of cap of 150 mg/L likely does not account
for this phenomenon at higher hardness levels. EPA needs to
examine opportunities to expand the hardness range of the criteria.
For example, aluminum criteria that were developed in the western
states of New Mexico and Colorado had an upper bound for
hardness of220 mg/L, based on data from Kimball (1978) that is
116
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Re^iirriin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Re\ision l.ociilion in
20IX Aluminum
Crileriii Document
also cited in I he draft criteria. EPA should determine if these or
other data can be incorporated into the criteria development that
could expand upon the range of hardness.
EPA-HQ-OW-
2017-0260-0034
(James Boswell, Senior
Manager,
Environmental,
Peabody Energy)
The draft criteria sates that the MLR model was based on
empirical data with a pH range of 6.0 to 8.1 standard units. The
draft criteria go on to expand the pH range of the model from 5.0
to 9.0 standard units to "be protective of a broader range of
natural waters". Peabody notes that the resulting aluminum
criteria reduce exponentially at pH levels less than 6 and greater
than 8. This questions whether it is appropriate to expand the
model beyond the data boundaries that it was originally based on.
This is particularly suspect considering the significant reducing
effect that the higher and lower pH levels have on the resulting
criteria. EPA should limit the applicability of the criteria to the
bounds for pH (6.0-8.1) just as it did for hardness and DOC,
where the criteria remains constant at pH values above and below
those bounds. The extrapolation that EPA is currently proposing
above and below this pH range is not scientifically valid.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development.
As a result, the water chemistry bounds for the 2018 criteria
were thus expanded, with details and rationale provided in the
criteria document and summarized below. The criteria
calculator can be used to address waters within a pH range of
5.0 to 10.5. For hardness values, the criteria calculator allows
entry of values between 0.01 and 430 mg/L total hardness;
criteria magnitudes will not increase or decrease by
increasing the hardness above 430 mg/L total hardness (as
CaC03). For DOC, the criteria calculator will not extrapolate
below the lowest empirical DOC of 0.08 mg/L and upper
limit of the empirical MLR models will be bounded at a
maximum 12.0 mg/L DOC in the criteria calculator; criteria
magnitudes will not increase or decrease by increasing the
DOC above 12.0 mg/L.
The pH of toxicity test waters ranged from 6.0-8.7. The EPA
has determined that for pH it is reasonable to allow the user
to extrapolate beyond these values for criteria derivations.
The criteria calculator can be used to address all waters
within a pH range of 5.0 to 10.5. Thus, criteria values for pH
input values beyond the range of the underlying empirical pH
data used for model development (pH 6.0 to 8.7) can be
generated using the criteria calculator. (This is also reflected
in the criteria lookup tables in Appendix K of the 2018 Final
Aluminum AWQC document). The EPA took this approach
for pH so that the recommended criteria can be provided for,
and thus are protective of, a broader range of U.S. natural
waters. Extrapolated criteria values outside of the empirical
pH data tend to be more protective of the aquatic
environment (i.e., lower criteria values) in situations where
pH plays a critical role in aluminum toxicity. However,
criteria values generated outside of the range of the pH
conditions of the toxicity tests underlying the MLR models
are more uncertain than values within the pH conditions of
No edits.
EPA-HQ-OW-
2017-0260-0035
(RichardA. Hyde,
P.E., Executive
Director, Texas
Commission on
Environmental Quality
(TCEQ))
B. The TCEQ recommends expanding the range of possible
measurement inputs in the proposed Multiple Linear Regression
(MLR) model which has limited applicability in Texas waters.
The MLR model as proposed by the EPA is not reflective of water
chemistry observed in western surface waters, such as Texas. As
currently proposed, the MLR model criteria outputs are
constrained by total hardness of 150 mg/L as CaC03, and
Dissolved Organic Carbon (DOC) at 5.0 mg/L. These constraints
limit the utility and applicability of the model in Texas, where total
hardness values and DOC may exceed 1,525 and 270 mg/L,
respectively. The EPA should adjust the model as needed to
increase its applicability, or provide options for states to allow
local water chemistry of surface waters to be incorporated.
Adjustment may result in changes to the EPA's proposal.
EPA-HQ-OW-
2017-0260-0036
(Barry N. Burnell,
Water Quality Division
Administrator, State of
Idaho Department of
Environmental Quality
(DEQ))
DEQ is concerned with the upper bounds of 5 mg/L for dissolved
organic carbon and 150 mg/L as CaC03for hardness. A hardness
of 150 mg/L as CaC03 represents the 73rd percentile of hardness
collected from stream and river sites sampled throughout Idaho in
the summer of 2016 [DEQ (Idaho Department of Environmental
Quality). 2017. Statewide Monitoring for Inputs to the Copper
Biotic LigandModel. Boise, ID: DEQ.
http://www.dea.idaho.sov. media. 60180618/58-0102-1502-
statewide-monitorine-inputs-copper-biotic-lisand-model-
0817.pdf], EPA should consider expanding the model's bounds for
117
-------
( omilKMII
Number
(Oi^;ini/;ilion)
Public C oiii moil I oil lopic 14: Re^iirriin^ llic MI.R (multiple
linciir regression) models
I.PA Response
Re\ision l.oeiilion in
20IX Aluminum
Crileriii Document
both hardness and DOC.
llic Ml U losis. and llius should bo considered
carefully and used with caution.
The total hardness of toxicity test waters underlying the MLR
models ranged from 9.8 to 428 mg/L. Since a decrease in
total hardness tends to increase aluminum toxicity, the EPA
has determined it is reasonable to extrapolate on the lower
bound of the hardness data to enable generation of lower
criteria at low hardnesses beyond the limit of the empirical
data. Thus, hardness input values in the criteria calculator can
be entered that are less than 9.8 mg/L down to a limit of 0.01
mg/L. This is consistent with existing EPA approaches to low
end hardness (U.S. EPA 2002). However, criteria values are
bounded at the approximate upper limit of the empirical MLR
models' underlying hardness data, at a maximum of 430
mg/L total hardness (as CaC03). The user can input hardness
values into the criteria calculator that are greater than 430
mg/L for total hardness, but the criteria magnitude will reach
its maximum value at 430 mg/L total hardness (as CaC03),
and criteria magnitudes will not increase or decrease by
increasing the hardness above 430 mg/L total hardness (as
CaC03). This is also consistent with existing EPA guidance
on high end hardness "caps" (U.S. EPA 2002). (These total
hardness bound approaches are also reflected in the criteria
lookup tables in Appendix K of the 2018 Final Aluminum
AWQC document.) The EPA took this approach so that the
recommended criteria can be provided for, and will be
protective of, a broader range of U.S. natural waters. Criteria
values generated beyond the lower bound of the hardness
conditions of the toxicity tests underlying the MLR models
are more uncertain than values within the hardness bounds of
the MLR toxicity test data.
The DOC of toxicity test waters ranged from 0.08 to 12.3
mg/L. Since most natural waters contain some DOC, the
lower bound of the empirical toxicity test data (0.08 mg/L) is
the lowest value that can be entered into the criteria
calculator; thus, no extrapolation below the lowest empirical
DOC of 0.08 mg/L is provided. The criteria values generated
with the criteria calculator are bounded at the upper limit of
the empirical MLR models' underlying DOC data, at a
maximum 12.0 mg/L DOC. The user can input DOC values
EPA-HQ-OW-
2017-0260-0037
(Anonymous public
comment)
EPA extended the pH range of the proposed aluminum calculator
(5-9) beyond the range of reliable data (-6-8) to be more
protective of water bodies. EPA did not extend the range of DOC
above 5 mg/L. Discharges to high DOC water may be held to a
more conservative standard than is necessary. EPA should address
the fairness of extending the range of only one parameter (pH)
beyond reliable scientific data. Why should there not be a similar
extrapolation to higher DOC values (>5 mg/L)?
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
The maximum dissolved organic carbon (DOC) limit for the new
calculator is 5 mg/L. Some water bodies have significantly higher
DOC and therefore potentially significantly higher toxicity limits.
Lt would be an undue hardship for a permittee discharging to a
water body with high DOC to be required to meet an unreasonably
low Aluminum limit just because the scale of the model maxes out
at 5 mg/L for DOC. The model should be expanded to account for
higher DOC concentrations observed in New England waters.
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
EPA should provide updated guidance for performing calculations
and/or studies to determine higher regulatory Aluminum toxicity
limits when water bodies are not within the calculator's limits for
pH, hardness, and DOC.
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
The user-friendly Aluminum Criteria Calculator appears to be a
useful tool, but some of the parameters, particularly Dissolved
Organic Carbon (DOC), do not fully encompass all ambient
conditions in our member states. We request that the model
parameters be expanded to reflect the full range of observed
concentrations in our states' surface waters, for example
Massachusetts's values for DOC tend to fall between 3 and 12
mg/L, with median values of 6.0 mg/L. Based on the model's upper
boundary of 5 mg/L DOC, it does not adequately represent the
range of conditions in all waters. Appendix K offers a broader
range of input values for pH and Hardness. The addition of these
into the calculator, along with expanded parameters for DOC
would be of valuable.
EPA-HQ-OW-
2017-0260-0042
(Bruce A. Stevens,
President, Indiana
Coal Council, Inc.
(ICC))
The Multiple Linear Regression (MLR) accounts for changing
toxicity based on the pH, hardness, and DOC of the water column.
The Draft Criteria points out that the studies used to develop the
MLR had a pH range of 6.0 -8.1 standard units. But the EPA goes
on to expand the pH range beyond what was used in the MLR
studies. The MLR was not validated above a pH of 8.1 or below a
118
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
pH of 6.0. There is significant uncertainty in model predictions
above a pH of 8.1. For pH levels below 6.0, singling out toxic
effects from aluminum is complicated by the toxic effects of the
acidity of the water. EPA should not expand the model beyond the
pH range of 6.0 - 8.1.
greater than 12.0 mg/L into the calculator, but the criteria
magnitude will reach its maximum value at 12.0 mg/L DOC,
and criteria magnitudes will not increase or decrease by
increasing the DOC above 12.0 mg/L. This is also reflected in
the criteria lookup tables in Appendix K of the 2018 Final
Aluminum AWQC document. This is consistent with the
existing approach for hardness (U.S. EPA 2002) to provide
for protection of aquatic organisms through the use of
protective, conservative values when water chemistry
conditions are beyond the upper limits of the empirical
toxicity test data.
Please work with your local EPA Region and Headquarters'
staff to regarding any refinements sought for situations where
water chemistry for a particular water falls outside the bounds
of the model.
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
The bounds of inputs to populate the model were 9.8 mg/L to 127
mg/L for total hardness, 0.08 mg/L to 5.0 mg/L for dissolved
organic carbon (DOC) and 6.0 to 8.1 for pH. However, the extent
in which the guideline applies to, for pH and hardness, extends
beyond the range of empirical data used for model development.
The guidance should not assert or assume knowledge of toxic
responses beyond the range of empirical data used for model
development. Toxicological studies have been clear that expanding
the boundaries of a study beyond that in which the toxicology
supports is not defensible. Linear regression as it pertains to
toxicological responses can vary drastically beyond the scope of
the study parameters and it would be inappropriate to simply
extend the bounds of applicability without sufficient demonstration.
The State of New Mexico does not have adequate information,
from what was provided by EPA in the proposed guidance, to
ascertain the reasoning or defensibility for extending the reaches
of the criteria beyond the scope of the study, and without such
demonstration cannot support this assertion.
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
The 1988 guidance expressly excluded waters with pH values
below 6.5 or above 9.0. This proposed guidance does not provide
any additional input on the limits to which aluminum toxicity
impacts aquatic life in waters with pH values below 6.0 and above
8.1. The linear regression model, as proposed extends beyond the
scope of the data to include pH values ranging from 5.0 to 9.0;
however, this assertion is not defensible as it is known that
toxicology in these outlying pH ranges changes drastically from
the circumneutral zone.
The findings in older primary literature regarding the influence of
alkaline pH on toxicity seems mixed (Gundersen et al., 1994,
Gensemer & Playle, 1999), yet the MLR model becomes more
protective as one progresses from circumneutral pH to the more
alkaline range. A review of the literature regarding aluminum
toxicity at alkaline pH suggests equivocal effects at best, but trend
toward less toxic aluminum forms as waters become more alkaline.
Colorado (prior to adopting hardness-based criteria) and North
Dakota (currently) incorporate(d) EPA's 1988 guidance with the
119
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Kc^iirdin^ llic MI.U (multiple
linciir regression) models
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
caveat that the chronic criteria would not apply at high pll, or
with appreciable water hardness, due to the low toxicity of
aluminum at this pH range. The State of New Mexico would like
the guidance to include defensible aluminum criteria for waters
with pH values below 6.0 and above 8.1.
EPA-HQ-OW-
2017-0260-0046
(Jennifer Wigal,
Program Manager,
Water Quality
Standards &
Assessments, Oregon
Department of
Environmental
Quality)
The multi-linear regression model by DeForest et al. (2017), upon
which the criteria are based, was validated to a pH range of 6.5 -
8.0, hardness of 9.8 - 127 mg/L, and DOC of 0.1 -5 mg/L. In the
criteria, EPA extrapolates the range ofpH inputs from 5.0-9.0,
more than doubling the validated pH range. We also note EPA
chose not to extrapolate hardness and DOC to higher ranges, even
though these would more accurately reflect aluminum
bioavailability in a more natural range of water quality conditions
because this would make the resulting criteria less stringent.
DEQ is concerned that EPA recommends using the calculator to
generate criteria values for conditions outside the pH range
validated for the DeForest et al. model, particularly because
conditions outside the validated model limits (i.e. changing from a
pH of 6.5 to 5) have a very large impact on the resulting criteria
values for aluminum that should be justified. For example, using
statewide median concentrations in Oregon for DOC (1.8 mg/L)
and hardness (35 mg/L) as reference values, the criteria values
change dramatically with pH. At pH 6.5, the lower bound of the
DeForest validation, the CCC under these median conditions is
310 jug/L. At pH 5, the range to which EPA extrapolated the
model, the CCC is 6.1 jug/L. This change brings the criterion far
below typical natural background levels of aluminum found in
Oregon waters.
Criteria values of this low magnitude are driven by pH values not
validated by either the DeForest et al. model nor represented in
the underlying toxicity data. EPA did not cite the evidence that led
them to suggest criteria values in the extrapolated range of pH are
necessary to protect the use, nor how certain they are in the
accuracy of these criteria values. EPA should provide this
evidence or limit the model to pH ranges supported by the data
and model.
In addition, the pH water quality standard in Oregon is 6.5 to 9.0.
In western Oregon, it is not uncommon to see naturally occurring
pH levels as low as 6 due to rainfall. If pH drops below 6, it is
120
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crileriii Document
most lil
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Re^iirriin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Re\ision l.ociilion in
20IX Aluminum
Crileriii Document
extrapolate pll values beyond the range of the available data, but
there is no explanation or basis for this decision or the accuracy
and protectiveness of the approach. Based upon the limitations of
the empirical data and the effect of the approach on the site-
specific criteria, we recommend that USEPA limit the pH input
values as they have done for hardness and DOC.
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
The Draft Criteria itself notes that studies used to develop the MLR
had a pH range of 6.0 - 8.1, yet the Draft Criteria expands the
range to 5.0- 9.0. There is an insufficient basis apply the criteria to
waters with a pH outside the range of 6.0 - 8.1.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEP))
The MLR models were developed using data that encompass a pH
range of 6-8.1, DOC range of 0.08-5 mg/L and hardness range of
9.8-127 mg/L (as CaCO). MLR criteria outputs are bounded at a
maximum of 150 mg/L total hardness and a DOC of 5.0 mg/L
because available toxicity data did not extend beyond these
maxima. The user can input values of hardness greater than 150
mg/L and DOC greater than 5 mg/L, but criteria output will be
limited to these bounds due to underlying data limitations during
model development. The hardness and pH values that were
selected for this analysis appear to be representative of the surface
water quality conditions in Massachusetts; however, the upper
boundary of 5 mg/L DOC does not adequately represent the range
of conditions in Massachusetts' waters. MassDEP supports the
incorporation of pH, hardness and DOC into the model, but has
concerns about the maximum range of DOC.
To evaluate the impact of the potential criteria, MassDEP
reviewed available data for hardness, pH, and organic carbon in
Massachusetts streams from the U.S. Geological Survey (USGS)
National Water Information System (NWLS) database
(https://water data, usgs.gov/nwis).
A total of6,462 samples had been analyzed for one or more ofpH,
hardness, DOC, or total organic carbon (TOC). USGS averaged
the values for each parameter at each site yielding the following
number of data points per parameter: 556for hardness, 765for
pH, 158 for DOC, and 103 for TOC.
Hardness varies across Massachusetts. The median value of
hardness (as CaC03) across all Massachusetts sites was 34 mg/L
and most values (80 percent) were between 12 and 99 mg/L.
122
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilerhi Document
Hardness irm consistently high in the Ilousatonic and Hudson
basins in western Massachusetts; most values were greater than 60
mg/L as CaC03, with values greater than 120 mg/L as CaC03 at
about one-third of sites in the Housatonic basin. Relatively high
values of hardness were also common in the Connecticut River
Basin in west-central Massachusetts. Hardness was relatively low
in the Millers, Nashua, Chicopee, Quinebaug, and French basins.
Average pH at most sites in Massachusetts ranged between 6.1 and
7.4, and the median value was 6.8. Regional patterns in pH were
similar to those of hardness, in that relatively high values (greater
than 6.8) were more consistently present in the western and west-
central parts of the Commonwealth. Relatively low values (less
than 6.8) were present in north-central and southeastern
Massachusetts. Values in the northeast were mixed.
DOC and TOC data were available for many fewer sites than
hardness or pH, and most data were from sites in the eastern half
of Massachusetts. Most values ranged between 3 and 12 mg/L,
with median values of 6.0 mg/L for DOC and 6.6 mg/L for TOC.
While we are pleased to see that the MLR takes into account pH,
hardness and DOC, we believe that the upper limit for DOC does
not reflect the range of DOC concentrations in Massachusetts. We
request that the model be expanded to reflect the range of DOC
concentrations observed in Massachusetts surface waters.
EPA-HQ-OW-
2017-0260-0053
(Abdul Alkhatib,
Director,
Massachusetts Water
Works Association
(MWWA))
1 have reviewed EPA's proposal and believe that the changes
proposed are beneficial and should move forward, however, we do
offer the following comments for EPA's consideration before the
new criteria is finalized:
1. Maximum DOC in Multiple Linear Regression Models: The
maximum dissolved organic carbon (DOC) limit for the new
calculator is 5 mg/L. Some water bodies have significantly higher
DOC and therefore potentially significantly higher toxicity limits.
It would be an undue hardship for a permittee discharging to a
water body with high DOC to be required to meet an unreasonably
low Aluminum limit just because the scale of the model maxes out
at 5 mg/L for DOC. The model should be expanded to account for
higher DOC concentrations observed in New England waters.
123
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crileriii Document
El 'A-11Q-OW-
2017-0260-0057
(Roger Claff, P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
Although the influences of hardness and dissolved organic carbon
(DOC) are considered in the MLR, the effect ranges are truncated
at concentrations of those two variables that are relatively low in
relation to natural surface waters in many parts of the U.S. For
hardness and DOC, values greater than 150 mg/L and 5 mg/L,
respectively, would be expected to provide even greater limitations
on aluminum bioavailability and result in higher but more
accurate criteria values. EPA's 2016 Draft Technical Support
Document, Recommended Estimates for Missing Water Quality
Parameters for Application in EPA's Biotic Ligand Model (EPA
820-R-l5-106), indicates a considerable proportion ofU. S.
surface waters exceeds the 1 Oth-percentile calcium concentration
(60 mg/L) corresponding to 150 mg/L hardness (as CaC03); this
distribution suggests many water bodies would be subject to overly
stringent aluminum criteria if the criterion ceiling is established
based on the assessed distribution as specified in the guidance
document. A similar concern exists for DOC for which the
conservative defaults (1 Oth or 20th percentile), and thus many
local water bodies, exceed 5 mg/L. While the proposed approach
would likely yield protective criteria, valuable resources would be
diverted to establish alternative site-specific criteria that would
likely be at higher aluminum concentrations than those produced
by default assumptions. Those resources could be better used
where real problems exist, and areas where real problems exist
may only be identified if the implemented criteria model is
accurate across the wider range of chemistries that exist across the
U.S.
EPA-HQ-OW-
2017-0260-0057
(Roger Claff, P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
As EPA observes, the MLR was developed from chronic toxicity
studies conducted using testpH values that ranged between 6.0
and 8.1 (DeForest et al. 2017); however, EPA has chosen to apply
the MLR to normalize toxicity data and, hence, generate aluminum
criteria concentrations for waters outside this range both at low
pH (i.e., between 5.0 and 6), and high (i.e., from 8.1 to 9.0) pH.
While application of any model outside the test conditions used to
derive or validate the model carries a high amount of scientific
uncertainty, this is particularly the case for empirical models such
as the MLR for which the equations can truly only be considered
valid within the tested range. Whereas a mechanistic model such
as the BLM may provide additional confidence when extrapolating
outside the test conditions, extrapolating empirical models is much
more problematic. In the draft criteria, EPA choses to not
extrapolate use of the MLR for data normalization and criteria
124
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Kc^iirdin^ llic MI.U (multiple
linciir regression) models
I.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
calculations outside the tested hardness and DOC presented in
DeForest et al. (2017) for this very reason-that the empirical
model should not be used for conditions not considered in
development of the model. API recommends that EPA treat pH the
same as hardness and DOC with respect to use of the MLR, and
not use the MLR for any calculations outside the water quality
conditions used to develop the model. This would effectively "cap"
MLR pH input values to being no less than 6.0, or no greater than
8.1. This is particularly important because EPA's application of
the MLR in the draft aluminum criteria generates criteria
concentrations that are significantly more stringent at pH values
both above and below the tested range. Unless EPA can provide
additional scientific confidence supporting the accuracy of criteria
calculations outside the tested pH range, significant regulatory
problems and costs could be incurred when little confidence exists
that such problems are actually real.
EPA-HQ-OW-
2017-0260-0060
(Katie Kistler,
Environmental
Manager of air
Programs, AK Steel
Corporation)
As explained through EPA's various support documents, the draft
criteria were developed using multiple linear regression (MLR)
models to predict the toxicity of aluminum as a function ofpH,
hardness, and dissolved organic carbon (DOC). That is, the
criteria are intended to be implemented and calculated based upon
site specific pH, hardness and DOC. However, the criteria
calculations are bounded at a maximum of 150 mg/L hardness, a
maximum of 5.0 mg/L DOC and pH range of 5.0 to 9.0 s. u. The
upper bounds for hardness and DOC were chosen because they
generally reflect the range of conditions encountered during the
toxicity tests upon which the criteria are based. However, we
believe that the upper bound hardness values does not reflect
conditions in many receiving streams.
EPA explained that hardness and DOC generally reduce aluminum
toxicity. The hardness of receiving streams for the majority ofAK
Steel facilities substantially exceeds 150 mg/L, and may exceed 5.0
mg/L DOC. We have found no statements in the proposed
language of the actual governing criteria restricting the
applicability of the proposed criteria to these ranges, nor have we
found any language in the actual proposed criteria cautioning the
user against utilizing the criteria when site-specific conditions are
outside of these ranges. In fact, the calculator developed by EPA to
provide quick calculation of the criteria values allows the input of
values outside of the hardness and DOC ranges, while limiting the
values used in the calculation to a hardness of 150 mg/L and a
125
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ke^iirriin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
Crileriii Document
DOC concentration of 5 nig L.
Many NPDES permit holders, in addition to AK Steel, discharge to
receiving streams with hardness values of greater than 150 mg/L
and DOC concentrations greater than 5.0 mg/L. Use of the
proposed criteria for these receiving streams would therefore
likely result in ambient criteria that would be substantially more
stringent than necessary, or at least, the level of protection
provided by the draft criteria would be uncertain.
Given the prevalence of receiving streams outside of the criteria
parameter bounds, AK Steel suggests that EPA withdraw postpone
the proposed criteria and include a broader range of hardness and
DOC values into the supporting toxicity tests, MLR models and the
resulting criteria.
EPA-HQ-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
UWAG agrees that the draft criteria should be based on the
measurable water quality parameters that affect aluminum
bioavailability and toxicity: pH, hardness, and DOC. And over the
range of the underlying pH, hardness, and DOC data, the criteria
derived using the proposed MLR models are appropriate. UWAG
is concerned, however, over the use of the proposed MLR models
to derive criteria for sites with parameters outside of the range of
the underlying data. The scientific validity of doing so is
questionable, and EPA has not provided information to support its
validity. If the MLR models are to be used for sites with
parameters outside of the range of the underlying data, the ability
to extrapolate where appropriate outside of the range must be
allowed. This is of particular concern for hardness.
Based on an analysis of the toxicity tests used to develop the acute
and chronic MLR models, EPA sets an upper (maximum) total
hardness value of 150 mg/L as CaC03for deriving aluminum
criteria. Not allowing hardness values greater than 150 mg/L to be
used for site-specific application of the criteria is problematic as,
in some regions (e.g., the arid southwest), background (ambient)
hardness values are considerably higher than 200 mg/L. In
addition, many process wastewater discharges have hardness
values much higher than 150 mg/L. Even as EPA acknowledges
that such situations do occur, the Agency provides no room for
extrapolation:
"... the user can apply the model in areas with hardness values
126
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
Crilerhi Document
greater than 150 mg L and DOC of 5 mg L, but the model output
for these parameters will be limited at the bounds stated due to
underlying data limitations."
Draft Criteria atxiii. While UWAG understands the technical
reason for this, use of the MLR models as proposed to derive
criteria for receiving streams and wastewaters with hardness
values greater than 150 mg/L is scientifically questionable and
inappropriately overprotective. EPA should allow for the ability to
extrapolate to parameters outside of the range of the underlying
toxicity data or limit the use of the MLR models to the underlying
range.
For example, the Arid West Water Quality Research Project
(AWWQRP; 2006) updated the EPA 1988 aluminum aquatic life
criteria (U.S. EPA, 1988) toxicity database and found a significant
positive relationship with acute effect measurement and test media
water hardness for species having the most toxicity endpoints
reported (r2 = 0.76; P < 0.03). The pooled slope of the acute
endpoint and water hardness equation was 0.833. Using this slope
and a CMC value of1,289 jug/L total aluminum, the authors
extrapolated protective acute criteria at water hardness values
between 25 - 400 mg/L (Table 3-8 of report).
UWAG also notes that, of the 119 acceptable acute tests listed in
the draft criteria document (Appendix A), only 23 of these tests
used a water hardness concentration greater than 100 mg/L.
Optimally, EPA could conduct some additional toxicity tests at
hardness values > 150 mg/L (e.g., within the range 150 - 400
mg/L) before the final criteria document is issued. Alternatively,
the Agency could evaluate the acute effect-water hardness
relationship and determine if the slope of tests having water
hardness values >100 mg/L differed from tests where lower
hardness values were used. If the acute endpoint values in tests
with water hardness values >100 mg/L have a similar pattern
relative to tests with lower hardness values, EPA should extend the
regression slope for hardness values > 150 mg/L.
127
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
Crilerhi Document
El 'A-11Q-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
EPA does extend the range for pll in the MLR models beyond the
underlying data. Although the underlying data used to develop the
MLR models are limited to a pH range of 6.0 to 8.1, EPA intends
to apply the models to waters with a pH range of 5.0 to 9.0, albeit
recognizing that criteria derived outside of the 6.0 to 8.1 range are
more uncertain. Research has shown, however, that the MLR
models are not inaccurate for pH from 8.1 to 9. (See Aluminum
Industry comments.) As discussed above, UWAG supports
extrapolation beyond the range of the underlying data, but only
where it has been shown to be appropriate. Lt has not been shown
to be appropriate for pH. For that reason, applicability of the MLR
models should be limited to the pH range of the underlying data.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
US EPA's handling of pH in the Draft Aluminum Criteria warrants
re-evaluation. US EPA goes to considerable effort to develop
aluminum criteria for waters that would be considered impaired
for pH. The national recommended criteria for pH are limited to
6.5 to 9.0 due to the impact of low pH on aquatic life.
Mount (1973) performed bioassays on the fathead minnow,
Pimephales promelas, for a 13-month, one generation time period
to determine chronic vH effects .... At the two lowest vH values
(4.5 and 5.2) behavior was abnormal and the fish were deformed.
At vH values less than 6.6. ess production and hatchabilitv were
reduced when compared with the control. It was concluded that a
pH of 6.6 was marginal for vital life functions.
Based on present evidence, a pH range of 6.5 to 9.0 appears to
provide adequate protection for the life offreshwater fish and
bottom dwelling invertebrates fish food organisms outside of this
range, fish suffer adverse physiological effects increasing in
severity as the degree of deviation increases until lethal levels are
reached. (Quality Criteria for Water, 1986). Despite the well-
documented effect of low pH on fish, the chronic aluminum
database is based on largely studies that are outside the
acceptable pH range. The studies for all four of the species utilized
to calculate the FCVwere conducted atpH<6.5. The database
includes twenty-nine studies. More than half of the reported EC2o
values were for studies conducted at pH<6.5. These studies had to
be adjusted upward based on the data for C. dubia to pH 7.0.
Two studies for C. dubia are included in the chronic database. The
normalized chronic value for the study conducted at pH 7.70 was
128
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Kc^iirdin^ llic MI.U (multiple
linciir regression) models
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
3,569 fjg / based on a biomass endpoint. The normalized chronic
value for the study conducted at pH 6.20 was 1,734 jug/l based on
a survival endpoint. The study with a more sensitive endpoint
resulted in a much higher chronic value when compared to the
study conducted at a lower pH. The data suggests that the impact
of pH as an independent stressor may affect the normalized values
within the chronic database.
US EPA proposes to extend the pH range of the criteria (pH 5.0 to
9.0) well beyond the range of the empirical data used for model
development (pH 6.0 to 8.1). On the converse, US EPA proposes to
limit the range for DOC and hardness to precise boundaries of the
empirical data for the MLR. This disparity is irrational. Moreover,
the application of MLR outside the empirical data range yields
questionable results. "The criteria values outside of the model
input data range are more stringent than those within the model
input range under the same hardness and DOC conditions and
have greater uncertainty." (Draft Aluminum Criteria, p. 57). US
EPA cannot corroborate the results of the MLR outside the
empirical data range, and the effects at pH>8.1 are particularly
suspect. While the WVCA believes the Draft Aluminum Criteria
should be entirely redeveloped, the extension of the criteria beyond
pH 6.0 to 8.1 is particularly egregious. The criteria for pH>8.1
should be "capped" similar to the method employed for hardness
and DOC.
US EPA cites recent studies which suggest that dissolved and
suspended aluminum species (particularly insoluble hydroxides)
are toxic to aquatic life. However, most of the toxicity studies were
conducted at low pH to maximize the dissolved aluminum
concentrations. Because of the important independent effect ofpH,
US EPA should obtain additional studies at circumneutral pH, at
least for the four most sensitive species utilized for criteria
calculations.
EPA-HQ-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United
States Steel
Corporation)
Maximum hardness and DOC values should not be capped at 150
mg/L and 5.0 mg/L, respectively. Natural background levels and
stormwater can exceed these values.
129
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crileriii Document
El 'A-11Q-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United
States Steel
Corporation)
Thepll range over which the criteria apply should be limited to
those values used to develop the MLR model and should not
extrapolate beyond those values. Doing so introduces additional
uncertainty.
EPA-HQ-OW-
2017-0260-0069
(Julia Young, Water
Quality Standards
Coordinator, Kansas
Department of Health
and Environment
(KDHE))
3) The draft aluminum criteria has an upper bound for CaC03 of
150 mg/L and DOC of 5 mg/L. The majority of waterbodies
throughout Kansas have CaC03 levels that exceed 150 mg/L and
DOC levels that exceed 5 mg/L. It is understood that EPA bounded
hardness and DOC at these limits because the available toxicity
data did not extend beyond the maxima, however, EPA should
consider expanding the bounds for both hardness and DOC to be
more realistic to measured stream concentrations.
EPA-HQ-OW-
2017-0260-0071
(Fredric P. Andes,
Coordinator, Federal
Water Quality
Coalition (FWQC))
The draft approach specifies a maximum hardness value of 150
mg/l, and a maximum DOC value of 5.0 mg/l. There are many
waterbodies around the country that exceed those values due to
natural levels. Also, many effluents - including stormwater
discharges, cooling tower blowdown and utility water - will exceed
those values. This situation is especially problematic when the
receiving water is effluent-dominated during critical low-flow
conditions. In all of these circumstances, application of the EPA
maximum values will yield aluminum criteria that are unduly
conservative, without any technical basis.
EPA-HQ-OW-
2017-0260-0071
(Fredric P. Andes,
Coordinator, Federal
Water Quality
Coalition (FWQC))
EPA has expanded the pH range over which the criteria apply. The
current EPA recommended criteria apply from 6.5 to 9.0, while the
Draft Criteria extend that range down to pH of 5.0. The MLR
models were developed using data with pH values no lower than
6.0. Moreover, there are questions about accuracy of the MLR
models at pH above 8.1. We are concerned that extrapolation of
the models to lower or higher pH values, beyond the scope of the
scientific studies concerning the models, carries substantial
uncertainty. EPA needs to provide a technical basis for the pH
range used in the Draft Criteria.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
The Association notes that the EPA capped model input hardness
values to no higher than 150 mg/L and model input DOC values no
higher than 5 mg/L based on the uncertainty of modeling
predictability above those thresholds. The Association believes that
a similar restricted approach to model input pH should also be
pursued as detailed below. Moreover, the Association notes that
the application of the MLR approach in establishing criteria for
130
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
waters with hardness levels above 150 mg L, DOC over 5 mg L, or
pH lower than 6 or higher than 8.1 may exceed the appropriate
use of the MLR, and result in overly stringent site-specific criteria.
The Association thus urges the agency to further consider and
explain the use of the MLR for such waters.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
According to EPA's proposal and information, the MLR is fully
validated between pH 6 and 8.1, which captures the range of most
waterbody chemistry. In the draft criteria, EPA extrapolates model
performance up to waterbody pH 9 and based on this
extrapolation, above pH 8.1 the model predicts increasing toxicity
(i.e., lower criteria concentrations) up to pH 9.0. However,
because the model is not validated above pH 8.1, there is
significant uncertainty in the model's predictions when modeling
in the range of pH 8.1 to pH 9.0. In fact, it may not even be
mechanistically correct for aluminum concentrations above pH 8.1
to exhibit increasing toxicity as above pH 8.5 the speciation of
dissolved aluminum changes considerably to more strongly favor
the aluminate anion, with a likely concurrent reduction in toxicity
due to its lesser binding potential on fish gill surfaces. Several
recent studies referenced in the GEL review attached to these
comments support that understanding. Given this significant
uncertainty and the relative lack of acceptable acute and chronic
toxicity data at higher pH's, the Association requests that EPA cap
the model utilization to no higher than pH 8.1 unless/until a more
thorough understanding of aluminum toxicity and model validation
above pH 8.1 is available. Under this capping scenario, if a
waterbody pH were found to be at a level greater than pH 8.1, a
pH of 8.1 would be entered into the model and the resulting model
output would be used to set the aluminum water quality criteria
limit for that waterbody.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
According to EPA's proposal and information, the MLR is fully
validated between pH 6 and 8.1 pH and the Nationally
Recommended Water Quality Criteria for pH is in the range ofpH
6.5 to pH 9. However, EPA extrapolates the MLR output down to
waterbody pH 5 in the draft criteria. As more fully explained in the
attached GEL comments, the Association believes that there
continues to exist uncertainty in the MLR operation below its
validated range although there is recognition that below pH 6 the
dissolved ionic and monomeric forms of aluminum increase which
generally leads to an increase in aquatic toxicity. However, the
amount toxicity increases as pH decreases below 6 is not yet
incorporated into the MLR, so the accuracy of the draft criteria
131
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
calculations in this range arc unknown. Given this uncertainly in
model performance below pH 6, the Association requests that the
EPA set a floor ofpH 6 for model usage with the recognition that
this still provides an expansion of modeling applicability below the
1988 Nationally Recommended Water Quality Criteria lower level
ofpH 6.5. If EPA desires to use the MLR model down to pH 5, it
must perform an MLR model validation down to that pH level and
then expand the use of the model down to pH 5 using a data
validated model rather than using an untested extrapolation.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
The multiple linear regression (MLR) models used as the basis for
normalizing toxicity tests were developed using data only using pH
values ranging from 6.0 to 8.1. Little justification for extrapolating
the MLR outside this range is provided in Section 4.1. For example
(see pages 56-57), EPA only points out that they are indeed
extrapolating the MLR equations outside the tested pH range, and
that criteria concentrations calculated outside this range (both low
and high ends of the range) "are more stringent than those within
the model input range under the same hardness and DOC
conditions and have greater uncertainty. " We suggest that at a
minimum, EPA provides additional scientific justification for either
the accuracy or protectiveness of these criteria, not just to state
that they are subject to greater uncertainty. Most importantly, have
any studies used to derive or validate the BLMprovide at least
some support to extrapolating the MLR equations outside this pH
range? For example, gill complexation data used to develop the
BLM provided by NIVA (Norway) provide information on effects
on salmonids at pH values less than 6. We suggest the BLM be
used to utilized to evaluate the accuracy of chronic criteria values
at pH values below 6 and above 8 to see if there is consistency in
responses with the MLR. Consistency would support further use of
the MLR in these ranges, whereas any significant discrepancies
may argue instead for a different approach than use of the MLR as
currently proposed.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
And while the MLRs developed for aluminum show strong
agreement between observed and predicted toxicity between pH 6
to 8.1, for an effects-based model like the MLR, extrapolation to
test conditions beyond the original calibration parameters adds
significant uncertainty. The use of the MLR to normalize data
below pH 6.0 would not account for the change in the speciation of
aluminum, nor the change in the mode of toxicity as pH decreases
to more of an ionoregulatory mechanism. Therefore, the inclusion
and MLR-normalization of toxicity data pH < 6 or > 8.1 is
132
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I.PA Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
questionable and would benefit from further justification by EPA.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Considering this uncertainty below pH 6, we recommend that EPA
provide either additional analysis to demonstrate that the MLR is
valid below this range, or that EPA consider whether the
aluminum criteria equations should be "capped" to use values no
lower than pH 6. Such capping would be consistent with limits
EPA is already recommending for hardness and DOC, and is a
familiar approach similar to hardness equation caps often used for
other metals (e.g., hardness no greater than 400 mg/L). Additional
justification, particularly including comparisons to BLM criteria
predictions at pH values below 6, would provide users of the model
additional confidence that the MLR can accurately predict toxicity
over this range ofpH. Even if should instead EPA choose to simply
"cap " the criteria to values no lower than pH 6, it would still
expand the pH range over which the criteria are applied compared
to the current criteria (i.e., no lower than pH 6.5).
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Furthermore, we believe it may not be mechanistically correct for
aluminum criteria concentrations above pH 8.1 to be more
stringent than concentrations derived at lower pH. While the
effects of the aluminate anion (Al(0H)4~, which dominates
aluminum speciation as pH increases beyond 8.5) on aluminum
toxicity are poorly known, recent studies suggest that aluminate
will not bind strongly to the fish gill and, hence, not contribute to
aluminum toxicity to a significant degree. Examples include Poleo
and Hytterod (2003), who concluded that the toxicity of the
aluminate ion to Atlantic salmon was low at pH 9.5 (lower than the
corresponding toxicity of cationicAl hydroxides) and Winter et al.
(2005), who showed that aluminum accumulation on the gills of
rainbow trout was lower at high pH (pH 10) owing to poor binding
of the aluminate ion to the positively charged gill surface.
Therefore, we suggest that EPA reconsider extrapolating the MLR
above pH 8.1 because of the strong likelihood thatAl is less toxic
at this pH owing to the limited bioavailability of aluminate. Given
the relative lack of acceptable acute or chronic aluminum toxicity
data at the high end of this pH range, we recommend that EPA
consider "capping" the pH values to which the MLR would apply
to no greater than 8.1. Such a cap would set the MLR pH input
value to 8.1 for any pH greater than 8.1 up to a pH of 9.0.
133
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.R (multiple
linoiir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilerhi Document
El 'A-11Q-OW-
2017-0260-0074
(Timothy F. Moore,
Risk Sciences, on
behalf ofLake Elsinore
and Canyon Lake
Nutrient TMDL Task
Force administered by
the Lake Elsinore San
Jacinto Watershed
Authority (LESJWA))
Although the draft water quality criteria for aluminum includes a
method to make adjustments for site-specific water chemistry
factors, the proposed procedure "caps" these adjustments based on
a maximum hardness of 150 mg/L and a maximum DOC
concentration of 5 mg/L. EPA was reluctant to extrapolate beyond
the range of input values used to develop the MLR model.
However, such caps are artificially low compared to the hardness
and DOC levels commonly measured in the arid southwest. For
example, Canyon Lake has an average hardness of300 mg/L and
an average DOC concentration of 7 mg/L; both well above the
range used to develop the MLR model.
Using the highest ("capped") hardness and DOC values shown in
Table K-8 of the draft criteria document, the maximum
recommended chronic criteria (CCC) for aluminum in Canyon
Lake is only 2,000 ug/L. However, prior to commencing the alum
application program, the Task Force conducted a number of site-
specific chronic toxicity tests using EPA's published Water Effects
Ratio procedure to determine the "safe dose." [U.S. EPA. Interim
Guidance on Determination and Use of Water-Effect Ratios for
Metals. EPA-823-B-94-001 (Feb., 1994)] These tests confirmed
that adding 40 mg/L of alum to samples of Canyon Lake water had
no adverse effect on Fathead minnow survival or srowth or
Ceriodaphnia dubia survival or reproduction. [Since alum is
comprised of 9% aluminum (by weight), 40 mg/L of alum is
equivalent to 3,600 ug/L of aluminum.] Thus, in some cases, it
appears that the "capped" MLR formula significantly
underestimates the appropriate aluminum criteria.
The draft criteria document acknowledges the need for additional
data to accurately characterize the effects of higher hardness and
DOC on the potential for aluminum toxicity. [Draft Criteria @ pg.
71] And, as noted above, the MLR model should also be expanded
to include the binding properties of phosphorus. However,
collecting the data needed to improve the MLR model will take
many years and another 2 or 3 decades may go by before EPA
elects to update the aluminum criteria again. Until then, there is a
better alternative available.
The Task Force recommends that EPA revise the aluminum
criteria document to explicitly authorize and encourage the use of
the existing Water Effects Ratio (WER) methods. The draft criteria
134
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Kc^iirriin^ llic MI.U (multiple
linciir regression) models
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crileriii Document
document makes no men I ion of I he If'Hi procedure. This omission
may be misinterpreted to imply that it cannot or should not be used
for aluminum.
Given the narrow range of hardness and DOC concentrations
evaluated in the laboratory experiments that EPA considered when
developing the MLR model, the WER procedure can be used to
develop more appropriate site-specific standards or permit limits
for aluminum when ambient water chemistry conditions fall
outside the normal range of the MLR model. In addition, the WER
procedure provides a method for evaluating how the numerous
other site-specific factors that are not yet included in the MLR
model (e.g. phosphorus binding) because EPA lacked sufficient
data to accurately estimate the parameter coefficients.
In sum, the proposed water quality criteria for aluminum does not
make appropriate adjustments for hardness concentrations greater
than 150 mg/L or for dissolved organic carbon (DOC)
concentrations greater than 5 mg/L. In addition, the draft criteria
makes no adjustment whatsoever for the mitigating effects of
phosphorus on the potential for aluminum toxicity. Collectively,
these limitations and omissions may make it far more difficult to
authorize the use of alum in future NPDES permits unless EPA
also endorses additional tools such as the WER procedure.
EPA-HQ-OW-
2017-0260-0075
(Steven A. Buffone,
CHHM, QEP, GIT,
Supervisor,
Compliance and
Regulatory Affairs,
CONSOL Energy Inc.)
The 2017 draft criteria is more complex than the 1988 A WQC
Criteria, with the addition of species, the derivation of data
through normalization by application of a multiple linear
regression model, and addressing the influence of numerous
receiving water quality parameters including pH, DOC, and
hardness.
It is unclear how the criteria would be applied to discharges when
pH, DOC, and hardness concentrations fall outside the limited
thresholds defined by EPA's proposed regression model. For
instance, in Pennsylvania, treatment to pH values above the model
default limit of 9.0 s.u. is often required to facilitate manganese
precipitation needed to comply with permitted effluent limits of 1
mg/L or less. In these cases, water treatment processes raise the
pH to as high as 10.0 s.u.; however, the proposed EPA calculator
does not account for pH levels above 9.0. Similarly, the hardness
of permitted effluents is routinely above the maximum 150 mg/L as
CaC03 limit included in the criteria calculator as a result of
135
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 14: Rc^iinlin^ llic MI.R (multiple
linciir regression) models
I'.I'A Response
Rc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
conventional chemical treatment processes, which incorporate
hydrated lime for neutralization and removal of metals. Certainly
the states have much additional data available for these
parameters that could be used to expand the limited thresholds
included in EPA's model.
CONSOL recommends that EPA clarify in the draft
recommendation that in cases where pH is above the
default limit of 9.0 that the draft criterion can still be
utilized, as is the case for hardness and dissolved organic
carbon (DOC) outside of the default limits of the criterion
calculator.
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC'))
In the draft AWQC for aluminum, EPA states that aluminum
solubility increases in lower temperatures and in the presence of
complexing ligands (both inorganic and organic) (EPA 2017).
Given the lower average temperatures naturally present in Alaskan
surface water and potentially higher DOC concentrations, ADEC
would like further clarification as to why EPA did not consider the
effects of temperature when considering model development.
Temperature was not considered because of the lack of
experimental data that could be used to develop an additional
parameter in the MLR.
No edits.
EPA-HQ-OW-
2017-0260-0063
(Kevin Oakes, Director
of Wastewater,
Borough of West
Chester, Chester
County, Pennsylvania)
The USEPA developed Water Quality Criteria (WQC) for
aluminum in 1988 based on a limited number of toxicity studies,
which was expressed as a fixed value for waters between 6.5 and
9.0 pH units, and did not account for other site-specific factors.
These WQC were adopted by the Pennsylvania Department of
Environmental Protection.
On July 28, 2017, the EPA published in the Federal Register
Request for Scientific Views: Draft Updated Aquatic Life Ambient
Water Quality Criteria for Aluminum in freshwater. The EPA is
seeking public comment on the proposed draft WQC for aluminum,
which were updated to reflect the "latest science knowledge ".
Many studies have concluded that aluminum can accumulate on
the surface of fish gill, leading to respiratory dysfunction, and
possibly death.
For years, researchers have been using bioavailability to measure
the element in the environment that is available to enter living
organisms, such as fish and other aquatic lives. The bioavailability
of aluminum is dependent on the chemical properties of water that
includes total hardness, pH and dissolved organic carbon (DOC),
those compounds can affect the toxicity of aluminum by affecting
the bioavailability of aluminum in the water to fish and other
136
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
aquatic lives.
However, despite the abundant scientific evidence and well
established theories, it seems the effects of water temperature on
aluminum toxicity was not considered by the USEPA in developing
the new WQC for aluminum.
In addition, we also like to recommend for the USEPA's
consideration to develop seasonal WQC for aluminum for cold
seasons when aluminum toxicity is lower.
This proposal is similar to the USEPA's existing policies that
apply less stringent limits for cold months (and more stringent
seasonal NPDES limits for summer months) for ammonia and
nutrients that include total nitrogen and total phosphorus, etc.
EPA-HQ-OW-
2017-0260-0058
(National Council for
Air and Stream
Improvement, Inc.
(NCASI))
EPA should more rigorously evaluate and communicate MLR
model prediction and criteria uncertainties throughout the
document
There are repeated references throughout the document (perhaps
as a carryover from the DeForest et al. 2017 article) to model
predictions that are within a "factor of two "for some percentage
of the test data. This characterization is overused, and reflects an
incomplete, potentially misleading and non-transparent approach
to characterizing the performance of the MLR model. An example
of the problem occurs in Panel A of Figure 4 in DeForest et al.
(2017). This plot of two sets ofpredicted and observed values from
two different MLR models shows that in both cases there is a high
percentage of observed values that fall within a factor of two, yet
the quality of the fit for one model is substantially better based on
visual and other numeric/statistical measures. EPA should rely
less on the "factor of two " characterization throughout the
document, replacing it with additional, more rigorous and relevant
information on predictive performance as outlined, for example, in
EPA (2009) and McLaughlin (2015).
Relatedly, in the evaluation of the C. dubia MLR model (page 29),
the document states "No clear pattern was observed in the
residuals over a wide range of water chemistry conditions or
relative to single independent variables (DeForest et al. 2017). "
The statement "No clear pattern... " is a broad generalization that
The EPA was clear and transparent regarding the
performance of the MLR model predictions in the draft
aluminum criteria. Various performance metrics were
described (i.e., R2, AIC, BIC, visual performance, factor of
two and residuals). While the factor of two is described in
several places, this is not the only metric described (please
refer to Section 2.7.1). The EPA disagrees that there is an
upward trend in Figure S4, Panel F. Furthermore, the residual
trends or lack of trends are the conclusions of the authors
(DeForest et al. 2018a).
The final aluminum criteria document is clear and transparent
regarding the performance of the MLR model predictions. In
the 2018 final aluminum criteria, the EPA used separate
MLRs for fish and invertebrates to best capture the effects of
water chemistry on toxicity for the taxa and differences in
trends across water chemistry; Section 2.7.1 discusses the pH,
hardness and DOC normalization approach the EPA took in
the 2018 aluminum criteria document. Appendix L of the
2018 criteria document discusses the comparison of the MLR
models used to normalize the toxicity data and compares the
results of the fish and invertebrate and pooled taxa MLR
approaches in detail.
Thank you for finding the typographical error for the adapted
Figures, these items were fixed.
Section 2.7.1
Figure 4
Figure 5
Figure 6
Figure 7
Appendix L
137
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
docs noI communicate the presence of some potentially important
details in the actual fit of the models. In fact, there are some clear
patterns, as shown in Fig. S4, panel F, where an upward trend
exists, or in panels C, D, and E of the same figure (which is
incorrectly labeled "D") which show decreasing variance with
increasing DOC. A similar statement is made in regard to the P.
Promelas model (page 32). "No clear pattern " in the model
residuals indicates equal model prediction accuracy across all
ranges of input variables. For both taxa, it is important to
acknowledge that the data set available to evaluate the assertion of
"no clear patterns" is relatively small. For example, in panel A of
Figure 7, only two pH values are available to evaluate model fit
for DOC of 2.7 mg/L and hardness of 122 mg/L, with no data at
pH values above 7. With uncertainties as large as those shown by
the plotted error bars, one must conclude that there is substantial
opportunity for additional validation of this model, and EPA
should be clear about this fact. The other models shown represent
similar data-limited situations.
Additionally, it appears that the reference to Figure S7 in the title
of the document's "Figure 7" is incorrect. Figure S7 is a 6-panel
figure of genus sensitivity distributions that appears to not be
related to Figure 7. Therefore, it is not clear how Figure 7 would
be "adapted" from the referenced figure. A similar comment
applies to Figure 6 as well.
It appears that all of the available C. dubia and P. promelas data
were used to create the statistical model, and that no cross-
validation of the model was conducted. For example, a portion of
the dataset could be excluded from that used for model
development, and used instead for subsequent evaluation (see EPA
2009). If this is the case, the quality of the model fit may be
overestimated. Furthermore, it is important to keep in mind that in
the face of the data limitations for models of these two species, the
draft criteria are based on the application of these models to other
species where no data are available to evaluate the quality of the
predictions (EPA acknowledges this on p. 71).
In Table 3 on page 42, SMAVs and GMAVs are presented without
including counts or standard deviations of the data from which the
averages are derived. This limits the transparency of the science
used to derive water quality criteria. Furthermore, on page 68, the
As stated in the 1985 Guidelines (pages 29 and 31), "For each
species for which at least one acute value is available, the
Species Mean Acute Value (SMAV) should be calculated..."
and "For each genus for which one or more SMAVs are
available, the Genus Mean Acute Value (GMAV) should be
calculated..." Thus, one toxicity test result is sufficient to
generate a SMAV/GMAV for the particular species/genus.
The uncertainty associated with this approach is described in
the document.
138
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
document stales "There are a number of cases in the acute
database where only one acute test is used to determine the SMA V
and subsequently the GMA V is based on the one acute test. In this
situation, there is a level of uncertainty associated with the GMAV
based on the one test result since it does not incorporate the range
of values that would be available if multiple studies were
available. " Of course, it is also true that "there is a level of
uncertainty" when more than one SMAVis available (see
McLaughlin 2011). The wording used here seems to reflect
inadequate attention on the part of EPA to the basic description of
scientific uncertainties that exist in their recommended criteria,
and the opportunities for additional scientific study to reduce them.
EPA should revise the presentation ofSMAVs and GMAVs in
Table 3 to include information on the number of tests and the
standard deviations of the toxicity data used to derive the draft
aluminum criteria.
In conclusion, EPA's proposed criteria represent an improvement
over the existing criteria, and should yield benefits in the effort to
protect aquatic life from the adverse effects of aluminum.
However, we encourage EPA to incorporate these comments in
order to ensure that the strengths and limitations of the MLR
approach are fully transparent, that model outcomes are
implemented appropriately, and that continued important
advances in the understanding of aluminum toxicity and modeling
in support of aquatic life water quality criteria are encouraged.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
The MLR is used twice - both as the basis of the data
normalization and for the final criteria calculation. US EPA
applied the MLR developed based on chronic toxicity to the acute
data, with the large leap offaith assumption that the effect of water
chemistry on bioavailability remains consistent across exposure
durations and for lethal and sublethal endpoints. US EPA is
attempting to quickly pass through a new approach without
allowing States and affected stakeholders adequate time to fully
critique the new approach. The data manipulation seriously affects
the original toxicity determinations as well as calculation of the
acute and chronic criteria based on pH, hardness, and DOC.
To calculate the normalized values, one must assume that the MLR
prepared from data for three species is applicable to all other
species in the acute and chronic databases. One must make a
greater leap of faith that the effects of pH, hardness, and DOC are
The MLR models (i.e., the normalization equations described
in Section 2.7.1) are used to normalize all of the freshwater
acute and chronic toxicity data to common water chemistry
conditions. Those normalized values are then ranked
according to GMAV/GMCVs and criteria are calculated
according to the method described in the 1985 Guidelines.
The MLR models, therefore, are only used once for each
criteria calculation, for normalizing the toxicity data). This
procedure is repeated for all criteria calculations when the
chosen water chemistry conditions are different (i.e., pH, total
hardness and DOC). The criteria calculator, following the
statistical approach outlined in the 1985 Guidelines, generates
the criteria magnitude values for each set of water chemistry
conditions. These values are also provided in summary tables
in Appendix K of the 2018 criteria document. Please refer to
Section 2.7.1 which elaborates on these normalization trends.
No edits.
139
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision Location in
20IX Aluminum
Crilerhi Document
consistent for both chronic and acute exposures. If all Jish and
invertebrates behaved the same way, we would have no diversity in
our ecosystem. We know this is not accurate. We also know that
certain fish are more sensitive than others, and the same holds true
for invertebrates.
This is evident in the studies for the two most sensitive species in
the chronic database. McKee reported that "significant
reductions" in RNA content and DNA content in Atlantic salmon
exposed to aluminum. (McKee, p. 3). Cleveland found no clear
impact on RNA content or DNA content in the highest test
exposures for brook trout.
(Cleveland, Table 8). This is significant in the use of the MLR for
P. promelas to normalize the chronic values for all fish prior to
calculation of the FCV. The MLR assumes that all fish species will
respond equally to aluminum and to the impacts of hardness,
DOC, and pH. If the effect of aluminum on two sensitive fish within
the salmonid family differs, it seems very unlikely that the
sensitivity data for fathead minnows is applicable to all fish.
Considering their diversity in behavior and function, it is even
more unlikely that all invertebrates are represented by the data for
C. dubia.
Even for C. dubia, which were used to develop the MLR, the
normalized results vary dramatically. All thirteen reported EC2qS
for C. Dubia are for the reproduction endpoint, but the normalized
chronic values range from 563.4 jug/l to 2,719 jug/l. More than half
of the studies were conducted concurrently. (Gensemer 2017). If
the MLR were appropriate, the normalized chronic values for C.
dubia should be nearly identical. Before proceeding with the Draft
Aluminum Criteria, US EPA must build a scientific demonstration
that the MLRs are appropriate for normalizing the chronic data. It
is beyond reason that the chronic MLR can be utilized to normalize
the acute database. A separate MLR must be developed based on
acute exposures.
The MLR yields criteria that do not demonstrate the expected
relationship to hardness and pH. "[B]oth C. dubia and P.
promelas EC20s generally increase with each independent
variable (DOC, pH, and hardness) regardless of the levels of the
other two variables. " (Deforest, p. 7). Therefore, one would expect
The EPA discusses the use of the chronic toxicity data
evaluating the effects of water chemistry to acute data in the
2018 final aluminum criteria document; this approach reflects
the extrapolation of the effects of water chemistry across test
durations, reflecting the same assumptions in principle
accepted in the 2007 Copper BLM-based criteria. The
approach is the most scientifically-defensible approach at this
time, based on available data.
RNA and DNA content are not used in the Aluminum criteria
calculations; criteria are based on survival, growth and
reproduction.
The underlying basis of the 2018 final aluminum criteria is
that water chemistry, specifically pH, hardness and DOC,
affect bioavailability, and hence toxicity of aluminum, as
reflected in the MLR normalizations underlying the criteria.
As indicated in the 2018 final criteria document, increasing
hardness generally increases criteria values, up to the
hardness bounds of the model; in the 2018 final criteria
document, at DOC=1.0 mg/1 and pH 7.5, the calculated
chronic criterion is 580 jxg/1 at a hardness of 25 mg/1, but is
660 mg/1 at a hardness of 150 mg/1.
EPA has clearly described the trends in criteria across water
chemistry conditions, through graphical representations and
criteria tables presented in the document. The commenter is
directed to those. In general, increasing DOC and hardness
tend to decrease bioavailability, resulting in increased
protective criteria values, while low and high pHs tend to
increase aluminum bioavailability, resulting in decreased
protective criteria values.
The 1988 national recommended aluminum chronic criteria
was 87 jxg/1, not 750 jxg/1 as the commenter incorrectly
suggests. The EPA recommend 750 jxg/1 as the acute (one
hour) criteria in 1988.
140
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic 14: Ko^iirdin^ llic MI.U (multiple
linoiir regression) models
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
Crilerhi Document
the criteria to increase with each of these parameters. Based on
the MLR, this does not occur. At D0C=1.0 mg/l andpH 7.5, the
calculated chronic criterion is 900 jug/l at a hardness of 25 mg/l,
but is 540 mg/l at a hardness of 150 mg/l. In other words, hardness
is predicted to make aluminum more toxic at one of the most
common pH values for natural waters. The calculated chronic
criterion, which drives calculation of effluent limitations for
NPDES permittees, is LOWER than the current US EPA
recommended chronic criterion of 750 jug/l. The same downward
trend exists for hardness at pH 7.0. Based on footnotes to the chart
presented in Appendix K, Table K-4, the four most sensitive species
at both pH 7.0 and 7.5 in all scenarios are fish and invertebrates.
Why does the criterion decrease with hardness?
We are unable in the allotted time to re-create the MLR to
investigate these surprising trends. Considering the species
rankings in the chronic database, we would anticipate chronic
criteria that would increase with pH, hardness, and DOC. This is
not represented in Appendix K. In fact, the criteria are often
inversely related to hardness and pH in the circumneutral to
alkaline range. It appears that US EPA has spent so much time
focusing on waters with impaired pH that no effort has been made
to ensure the criteria are sensible for unimpaired waters.
While we appreciate US EPA's efforts to improve the aluminum
criteria through the development of the MLR, something has
clearly gone wrong. We ask US EPA to reconsider the application
of the MLR, as many healthy waters will be listed as impaired
based on the calculated chronic criteria.
141
-------
TOPIC 15: Comments regarding mussel toxicity data
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Com nun 1 on Topic 15: Ke^inlin^ mussel lo\icil> (l;il;i
I'.I'A Response
Kc\ision Location in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0039
(Brett Hart I,
Government Policy
Director, Center for
Biological Diversity)
... there are indications from the record that Section 7
consultations would be beneficial here. For example, freshwater
mollusks are the most imperiled group of organisms in United
States with nearly two-thirds of species being identified as at risk-
of extinction. Mussels are particularly sensitive to contamination
from dissolved metals (Naimo 1995). Aluminum can be lethal to
mollusks and is added to water by some water treatment facilities
to kill the young larvae of invasive mussels (Mackie and Kilgour
1995).
In response to concerns expressed by the USFWS and others
that endangered freshwater mussel sensitive to aluminum
needs to be taken into consideration in deriving the criteria,
EPA identified studies by USGS (Wang et al. 2018) on
freshwater mussels in the family Unionidae, a family of
freshwater mussels found to be sensitive to some toxicants,
These new data on aluminum toxicity to the fatmucket
mussel (Lampsilis siliquoidea) are included in the final
aluminum criteria.
While the 96-hr LC50 juvenile test failed to elicit an acute
50% response at the highest concentration tested (6,302 |ig/L
total aluminum, or 29,492 |ig/L when normalized), the 28-
day biomass normalized SMCV ranked as the fourth most
sensitive genus in the chronic dataset. The mussel's chronic
value is greater than the most sensitive species, Atlantic
salmon, and the freshwater criterion. Thus, the chronic
criterion is expected to be protective of this and related
mussel species. The fatmucket tested is not a threatened
and/or endangered species, but the genus Lampsilis contains
several listed species with a wide distribution across the
United States. Additional testing on endangered mussel
species, or closely related surrogates, would be useful to
further examine the potential risk of aluminum exposures to
endangered freshwater mussels.
The studies the commenters noted were reviewed by the EPA
and their information considered.
Regarding Taskinen et al. (2011): The study was unused
because the river water used for dilution water was not
characterized.
Huebner and Pynnonen (1992) data were deemed "unused"
for criteria numeric calculations. Anodonta anatina and
Anodonta cygnea are not native to North America, nor do
they have naturally reproducing populations, but there are
species of the Anodonta genus present in the United States. In
the Huebner and Pynnonen (1992) data the glochidia 24.-hr
EC50 conducted at pH 4.5 was approximately 18,000 |ig/L.).
Other data in these taxa by Pynnonen (1990) and Kadar et al.
No edits.
EPA-HQ-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
Several published studies indicate that native freshwater mollusks
can be harmed by aluminum pollution. Huebner and Pynnonen
(1992) found that exposure to increased aluminum decreased the
viability ofglochidia of the unionids Anodonta anatine and
Anodonta cygnea. Malley et al. (1988) added aluminum sulfate to
an experimental lake in Ontario to test the effects on adult mussels
of the addition of aluminum and increasing acid levels in soft
water, and found that Anodonta grandis grandis experienced blood
and tissue ionic changes indicative of stress and exhibited
aluminum accumulation in tissues. The authors concluded that in
increasingly acidic conditions with high levels of aluminum, adult
mussels could experience significant damage to their shells.
EPA-HQ-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
In the Ahtavanjoki River in Finland, Taskinen et al. (2011)
reported that the endangered freshwater pearl mussel
Margaritifera margaritifera experienced low reproductive success
attributable to high concentrations of aluminum and iron
accompanied with periods of low pH. Though the adult mussels
appeared to be tolerant to periods of water quality variation and
were able to produce glochidia, the early life cycle stages of
mussels in the river were not successfully recruited into the
population due to metal exposure. In laboratory experiments on
mussels collected from the river, exposure to high but
environmentally realistic levels of aluminum was toxic to free
glochidia with most individuals dying within 72 hours.
Importantly, the survival of control glochidia was significantly
higher than that of any group ofglochidia that were exposed to
aluminum at any level. The authors also found that the survival of
juvenile mussels was lower in groups exposed to aluminum than in
the control group.
142
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I on Topic 15: Kc^iirdin^ mussel l«i\icil> (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0039
(Brett Hartl,
Government Policy
Director, Center for
Biological Diversity)
Though aluminum is most readily uptaken by mollusks in acidic
waters, Elangovan et al. (1997) found that the freshwater snail
Lymnaea stagnalis accumulated significant levels of aluminum in
neutral water in its soft tissues, gut, digestive gland and kidneys.
Kadar et al. (2001) examined the effect of aluminum on the
filtering behavior of the mussel Anodonta cygnea in neutral water
at environmentally relevant concentrations and found that mussels
closed their shells and avoided filtering at the higher
concentration. Interestingly, the mussels exposed to the lower dose
accumulated more aluminum in their tissues because they did not
reduce filtering time in response to exposure as did the mussels
exposed to the higher dose. They found that the mussels
accumulated most of the aluminum in their kidneys and digestive
glands. Their study provides evidence for the bioavailability and
toxicity of aluminum to mussels at neutral pH.
(2001) are discussed in the criteria document Effects
Characterization section 5.4.3 (EPA also reviewed additional
studies on mussels in developing the criteria document.
Malley et al. (1988): Data deemed "Unused".
Anodonta grandis grandis is native to North America, and the
study did show blood and tissue ionic changes due to both pH
and increased aluminum, but the pH and aluminum levels
were variable during the exposure making it difficult to
determine specific effect concentrations for pH and
aluminum.
Elangovan et al. (1997): Study deemed "Unused" (steady
state not reached inbioaccumulation study).
Kadar et al. (2001): Study was not used in criteria calculation
but discussed in Effects characterization. (Anodonta cygnea is
not a North American species).
EPA-HQ-OW-
2017-0260-0052
(Heidi L. Dunn,
President, Freshwater
Mollusk Conservation
Society (.FMCS'))
Mussels are particularly sensitive to contamination from dissolved
metals (Naimo 1995). Aluminum can be lethal to mollusks and is
added to water by some water treatment facilities to kill the young
larvae of invasive mussels (Mackie and Kilgour 1995).
Several published studies indicate that native freshwater mollusks
can be harmed by aluminum pollution. Wang et al. (2017) recently
reported acute and chronic toxicity of aluminum to juvenile
Lampsilis siliquoidea. Based on chronic toxicity results, the mussel
ranks as the 4th most sensitive species tested to date. Huebner and
Pynnonen (1992) found that exposure to increased aluminum
decreased the viability ofglochidia of the unionids Anodonta
anatina and Anodonta cygnea.
Malley et al. (1988) added aluminum sulfate to an experimental
lake in Ontario to test the effects on adult mussels of the addition
of aluminum and increasing acid levels in soft water, and found
that Anodonta grandis grandis experienced blood and tissue ionic
changes indicative of stress and exhibited aluminum accumulation
in tissues. The authors concluded that in increasingly acidic
conditions with high levels of aluminum, adult mussels could
experience significant damage to their shells.
In the Ahtavanjoki River in Finland, Taskinen et al. (2011)
reported that the endangered freshwater pearl mussel
Margaritifera margaritifera experienced low reproductive success
attributable to high concentrations of aluminum and iron
143
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 15: Kc^iirdin^ mussel l«i\icil> (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
accompanied with periods oflowpll. Though the adult mussels
appeared to be tolerant to periods of water quality variation and
were able to produce glochidia, the early life cycle stages of
mussels in the river were not successfully recruited into the
population due to metal exposure. In laboratory experiments on
mussels collected from the river, exposure to high but
environmentally realistic levels of aluminum was toxic to free
glochidia with most individuals dying within 72 hours.
Importantly, the survival of control glochidia was significantly
higher than that of any group ofglochidia that were exposed to
aluminum at any level. The authors also found that the survival of
juvenile mussels was lower in groups exposed to aluminum than in
the control group.
Though aluminum is most readily uptaken by mollusks in acidic
waters, Elangovan et al. (1997) found that the freshwater snail
Lymnaea stagnalis accumulated significant levels of aluminum in
neutral water in its soft tissues, gut, digestive gland and kidneys.
Kadar et al. (2001) examined the effect of aluminum on the
filtering behavior of the mussel Anodonta cygnea in neutral water
at environmentally relevant concentrations and found that mussels
closed their shells and avoided filtering at the higher
concentration. Interestingly, the mussels exposed to the lower dose
accumulated more aluminum in their tissues because they did not
reduce filtering time in response to exposure as did the mussels
exposed to the higher dose. They found that the mussels
accumulated most of the aluminum in their kidneys and digestive
glands. Their study provides evidence for the bioavailability and
toxicity of aluminum to mussels at neutral pH.
In light of these studies demonstrating that aluminum can be
harmful to mussels and snails in freshwaters, we urge you to
implement criteria that are protective of all life stages mollusks.
Inclusion of mussel chronic toxicity data in recalculation of the
aluminum chronic criterion would help ensure that mollusks are
protected.
[Cited References]
144
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I on Topic 15: Kc^iirdin^ mussel l«i\icil> (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0064
(Scott G. Mandirola,
Director, West
Virginia Department of
Environmental
Protection (WVDEP))
Finally, it should be noted that commenters have not been given
sufficient time to examine the freshwater mussel studies recently
published and long-awaited for the development of this Draft
Aluminum Criteria. It is difficult to comment on the Draft
Aluminum Criteria when presented with research that is either still
in peer-review or only very recently published.
Again, WVDEP thanks EPA for the opportunity to comment on the
Draft Aluminum Criterion, as it is quite important and particularly
relevant to the state of West Virginia. WVDEP continually seeks to
appropriately protect its aquatic environment, including protection
from aluminum. WVDEP appreciates the monumental effort EPA
has undergone for the last many years to develop this criterion.
However, West Virginia would like to see additional consideration
for the issues discussed herein to make this criterion a fully-
protective nationally-recommended standard.
The most recent study, Wang, N., C.D. Ivey, E.L. Brunson,
D. Cleveland, C.G. Ingersoll, W.A. Stubblefield and A.S.
Cardwell, was published in January of 2018. Acute and
chronic toxicity of aluminum to a unionid mussel (Lampsilis
siliquoidea) and an amphipod (Hyalella azteca) in water-only
exposures. Environ. Toxicol. Chem. 37(1): 61-69.
No edits.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
Mussel Studv
A study of the impacts of water column concentrations of
aluminum on growth to juvenile mussels does not reliably reflect
the exposure mechanism of immature mussels. Growth is more
likely to be influenced by sediment and interstitial water (IW)
concentrations of metals:
"The development of the juvenile mussel sediment/IW test is
important in determining site toxicity because it focuses on the
environment that they inhabit. Yeager et al. (1994) found that
juvenile mussels pedal-feed in the substrate, exposed mostly to
the sediments and IW, with little exposure to the water column."
(Simon, p. 13)(Emphasis added). In the unionid mussel study,
significant differences occurred between the measured
concentration of total aluminum in composite samples collected
from the lower portion of the water column or at the bottoms of the
test beakers compared to the upper portion of the water column. It
is very likely that the reported toxicity was based on the
concentrations in the lower portion of the water column, which by
the end of the 28-day test were nearly double the concentration for
the 1,200 jug/l exposure. The difference between the water column
concentration and the bottom of the beaker in the lowest exposure
concentration was even greater. Therefore, considering that IW is
a more likely exposure mechanism, the aluminum concentrations
from the bottom of the beaker are more representative of the actual
exposure concentrations for the unionid mussels. If the
The authors (Wang et al. 2016, 2018) follow ASTM protocol
(ASTM E2455-06) to use the average concentration from the
water column to calculate the EC20s. Based on Figure 1, in
the nominal 300 |ig/L treatment, the water column value is
200 versus 400 in the bottom portion of the beaker. Note: the
EC2o reported for biomass is 169 |ig/L.
The EPA agrees that exposure via sediment may be an
important exposure pathway for juvenile and immature
mussels. However, the aquatic life ambient water aluminum
criteria use toxicity studies with exposure to aluminum in the
water column.
No edits.
145
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 15: Kc^iirdin^ mussel l«i\icil> (l;il;i
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
concentration from the bottom of the beaker were used instead of
the water column concentration, then it is likely that the fatmucket
unionid mussel would no longer be among the four most sensitive
species in the chronic database.
146
-------
TOPIC 16: Comments regarding plant toxicity data
( omilKMII
Number
(()ri*;ini/;ilion)
Public Com nun 1 on Topic H>: Rciiiirdinii pkinl lo\ici(\
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rilcriii Document
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
Additionally, as part of the evaluation of impacts to marine
systems, EPA identified that certain marine aquatic plants are
sensitive to aluminum. The limited data included in the document
shows that a marine grass species (Halophila stipulacea) that is
not native to the United States, is impacted by levels of aluminum
below the calculated freshwater criteria. Restoration of eel grass
within marine waters is a concern to states in New England. As
part of developing a marine water quality criteria for aluminum
additional information is needed to determine if native marine
plant species are also sensitive to aluminum, potentially disrupting
eel grass restoration activities.
Marine aluminum toxicity data are severely limited and
therefore no estuarine/marine criteria can be recommended at
this time.
No edits.
EPA-HQ-OW-
2017-0260-0025
(Peter T. Goodmann,
Director, Kentucky
Division of Water)
Finally, the toxicity data for green algae, for which the most data
is available, appears to indicate that these plants are sensitive to
aluminum. The Hornstrom et al. 1995four-day toxicity studies
(Appendix E of the supplemental materials), at pH 6.8 and
hardness 14.9, indicate a LOEC of200 jug/L and 100 jug/L for
Monoraphidium dybowskii andMonoraphidium griffithii,
respectively. Section 5.2 acknowledges that aluminum effect
concentrations for freshwater algae ranged from 50 jug/L to 6,447
jug/L, with most values below 1,000 jug/L. Table 7 shows that the
four most sensitive aquatic animal genera for chronic toxicity have
GMCVs of508.5 jug/L to 1,102 jug/L. This appears to indicate that
green algae are at least sensitive, if not more sensitive, to
aluminum toxicity as aquatic animals, and should be explained
more thoroughly in Section 5.2 and 5.3.
Thank you for your comment. Additional text has been added
to the document. Plant data, and the associated water
chemistry data necessary to normalize the plant toxicity test
results for comparison with other taxa, were very limited. We
reconsidered inclusion of the Gensemer at al 2017 algae data
but did not include this plant toxicity tests because the tests
were not of 96-hour duration. However, the information is
included in Appendix H.
Some aquatic plants have similar sensitivity to aquatic
animals, thus the calculated criteria are expected to also
protect these species.
Section 5.2
EPA-HQ-OW-
2017-0260-0048
(William Stubblefield,
Professor,
Environmental and
Molecular Toxicology,
Oregon State
University on behalf of
Aluminum Ecotoxicity
Research Group)
The draft criteria document does not reflect the extant algae data
reported in Gensemer et al. 2017 due to the 72-hr duration
(Appendix H: pages H4-H15), which is shorter than the USEPA
algae test duration of 96-hrs. We suggest EPA reconsider the
inclusion of this dataset to Appendix E as the 72-hr test duration is
the standard OECD methodology for chronic algae tests.
Additionally, this dataset is extensive under varying pH, hardness,
and DOC conditions; is used in the MLR equations in DeForest et
al. (2017); and provides valuable insight into the toxicity of Al to
freshwater algae.
147
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic l(>: Kc^iirdin^ pliinl lo\icil> (l;il;i
I'.I'A Response
Revision l.ociilion in
20IX Aluminum
Crilcriii Document
El 'A-11Q-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Algae data from Gensemer el a/. 2017 were not included in
Appendix E because of the test duration.
These studies tested the effects of pH, hardness, and DOC on algae
growth. While these tests used a 72-hour test duration which is not
consistent with the 96-hr requirement from the 1985 Guidelines,
we suggest that 72-hr still represents a valid chronic exposure
period given their rapid cell division rates and population level
response that was measured. So while these data would not
ultimately be used in criteria calculations, we suggest that EPA
consider their inclusion in Appendix E since they represent a
significant and valuable database regarding the effects of water
quality on aluminum toxicity to algae.
148
-------
TOPIC 17: Comment regarding Multi-Sector General Permit (MSGP)
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public C oiiimoil 1 on lopic I"7: Rciiiudinii Mulli-Soclor Cciicnil
IVrmil
I'.I'A Response
Kc\ision Location in
20IX Aluminum
( riloriii Document
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
Request to update the Multi-Sector General Permit (MSGP) with
the new criteria
We request that EPA, in a separate regulatory action, update the
MSGP with the new aluminum criteria. The MSGP currently uses
the acute exposure criterion as a benchmark for aluminum
discharged from industrial facilities. An update to the MSGP will
allow states that have borrowed this benchmark to update their
own benchmarks or action levels for industrial stormwater
permits. This action is necessary for California to update the
unnecessarily low action level in the statewide Industrial General
Permit.
Thank you for your comment.
No edits.
149
-------
TOPIC 18: Comments regarding implementation issues
( ommcnl
Number
(()ri*iini/:ilioii)
Public ( ci in moil I oil lopic IS: Kcuiirriiii^ implcmciil;ilioii
issues
1"PA Response
Kc\ ision Locution in
20IX Aluminum
( rilcriii Document
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
Anti-backsliding: Since the Clean Water Act contains anti-
backsliding provisions, we wonder how EPA will handle cases
where the new criteria results in a higher applicable criteria? It is
our position that if new methodologies are available for
calculating criteria, then permittees should be given the
opportunity to apply the new methodology and their permit should
be modified to adopt the new criteria, even if it results in criteria
that may be considered move "favorable " than the old criteria.
Permits must rely on the best available science and should not be
bound by anti-backsliding provisions if new information is
available.
The intended protection goal of the 2018 final aluminum
criteria remains the same as that of the 1988 criteria,
protection of approximately 95% of genera in an ecosystem
to support protection of an aquatic life designated use. The
differences in the criteria values reflect an expanded toxicity
database and an improved incorporation of the effects of
water chemistry on bioavailability and toxicity in the 2018
final criteria.
The EPA's criteria provide recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). The
implementation documents that the EPA is developing are
intended to provide assistance to states and authorized tribes
that adopt into the water quality standards a criterion based
on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0053
(Abdul Alkhatib,
Director,
Massachusetts Water
Works Association
(MWWA))
Anti-backsliding: Since the Clean Water Act contains anti-
backsliding provisions, the regulated community wonders how
EPA will handle cases where the new criteria results in a higher
applicable criteria? It is our association's position that if new
methodologies are available for calculating criteria, then
permittees should be given the opportunity to apply the new
methodology and their permit should be modified to adopt the new
criteria, even if it results in criteria that may be considered move
"favorable" than the old criteria. Permits must rely on the best
available science and should not be bound by anti-backsliding
provisions if new information is available.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Implementation of the new criteria
Extensive new science developed since the existing 1988
guidelines were finalized has contributed to significant additional
understanding of the chemical behavior of aluminum in water and
this science is the foundation of EPA's proposed aluminum water
quality criteria revisions. In order to fully recognize the value of
that new science during both the development and implementation
of the new criteria, it must be applied broadly to both new and
existing permitted discharges of aluminum, even in cases where
the application of better science may increase aluminum discharge
limits when compared to existing limits. This is consistent with the
concept of providing an exemption to the general prohibition
against permit backsliding as found in CWA 402(o)(2)(B) for
situations "where information is available which was not
available at the time of permit issuance. "
150
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0057
(Roger Claffl P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
Several states already use the BLM and WER models for
establishing site-specific standards. EPA should explicitly state
that the BLM approach will be accepted in state standards
provided the technical requirements of the BLM are appropriately
met.
The Aluminum AWQC are recommendations. States may
choose other scientifically defensible methods to develop
aluminum criteria.
Current research on modeling indicates that the MLR and
Biotic Ligand models have comparable performance in
predicting aquatic toxicity for several chemicals, as long as
both models are well-constructed and are supported with
sufficient data. For example, Brix et al. (2017) concluded that
the MLR and BLM models' performance for copper were
comparable across a wide range of water chemistries and
species (Environ. Sci. Technol., 2017, 51(9): 5182-5192).It
should be noted that the MLR approach requires less data to
implement and is more transparent to the public and users
than the BLM.
The MLR (and the BLM model described above) is reflective
of a substantially larger toxicity database than a WER, which
can depend greatly on the particular "snapshot" conditions
during which the WER tests are conducted.
No edits.
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC'))
Implementation Issues
1. Data collection for the inputs to the proposed aluminum
criteria model
Like many recent national water quality criteria, the draft
aluminum criteria will require implementation of permit specific
criteria requiring new data collection efforts for the inputs to the
model (i.e., pH, hardness and dissolved organic carbon). The
analysis of the data may also pose significant implementation
challenges based on DEC's experience with the biotic ligand
model for copper. Implementation questions include: What will be
considered sufficient data? How do we identify "critical
conditions"? What percentiles should be entered as inputs or are
instantaneous criteria to be implemented as variable permit
limits? How do you calculate the criterion when little or no data is
available for the inputs? Many waters in Alaska have not been
monitored by any agency or permittee, so Alaska cannot rely on
"available data" from independent sources. Such challenges may
affect the timeline for adoption of this proposed criteria compared
to traditional fixed or hardness-based toxics criteria. Because of
the unresolved implementation issues, the timeframe for criteria
adoption will have to be prioritized based state needs through the
state Triennial Review process rather than national program
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). As noted in the
criteria document, the EPA decided to use an empirical MLR
approach in the aluminum criteria update rather than a BLM
model due to: 1) the relative simplicity and transparency of
the model, 2) the relative similarity to the available BLM
model outputs, and 3) the decreased number of input data on
water chemistry needed to derive criteria at different sites.
The EPA is also separately compiling an updated national
database of water chemistry conditions relevant to the MLR
model: hardness, pH and DOC, and will make that data
available to in the future to support states and stakeholders
needs for model input data, when their own data are not
available.
The implementation documents that the EPA is developing
are also intended to provide assistance to states and
authorized tribes that adopt into the water quality standards a
criterion based on or similar to the EPA's recommended
criterion. The implementation documents are also intended to
No edits.
151
-------
( ommcnl
Number
(Or^iini/iilion)
Public ( o in moil 1 on Topic IS: Kcuiirdinu implcmcnl;ilion
issues
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
priorities.
pro\ iiie ubbibluiice lo oilier stakeholders and ihe public. 1 lie
EPA recognizes that there are several aspects of the
recommended criterion that will benefit from technical
support documents to enhance implementation of state and
tribal criteria and is planning to develop such documents and
make them available for public comment.
EPA-HQ-OW-
2017-0260-0036
(Barry N. Burnett,
Water Quality Division
Administrator, State of
Idaho Department of
Environmental Quality
(DEO))
Many states, including Idaho, have very limited DOC data
available. In the absence of sufficient DOC data, states will be
unable to estimate protective aluminum criteria in waters where
data are unavailable. EPA should provide states with options on
how to implement these criteria when data for calculating the
MLR are absent.
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
Our member states are also concerned they will not have the
resources to determine aluminum limitations for areas that do not
currently collect pH, DOC, and/or hardness data. With the
ongoing reductions in funding, it will be difficult to implement
additional monitoring programs to obtain this data. For areas
where these parameters are tested in effluent rather than in the
water body, it will be difficult to determine aluminum limitations.
We urge EPA to address this concern with site specific data, not
regional default data.
EPA-HQ-OW-
2017-0260-0043
(Blake Beyea,
Standards Unit
Manager, Water
Quality Control
Division, Colorado
Department of Public
Health &
Environment)
The draft criteria require the user to input hardness, pH, and
dissolved organic carbon (DOC) data to calculate protective
criteria for a given site. The division was unable to determine what
summary statistic for each parameter should be used when more
than one data point is available for a given site. For instance,
when data from multiple samples are available, should average
hardness, pH, and DOC be used to calculate the criteria? Or,
would a percentile, median, etc. be more appropriate? It would be
helpful if EPA provided more clarity regarding implementation of
the criteria to ensure criteria are calculated appropriately and
protectively. If employing a summary statistic of input parameter
data from multiple samples is not appropriate, how would EPA
recommend implementing the resultant multiple final criteria
values from multiple dates or sample sites?
EPA-HQ-OW-
2017-0260-0043
(Blake Beyea,
Standards Unit
Manager, Water
Quality Control
Division, Colorado
Department of Public
Health &
Environment)
Does EPA have recommendations for minimum data requirements
for the input parameters (i.e., hardness, pH, and DOC)? When
possible, it is important to ensure the data used to calculate the
criteria adequately capture any variability that may occur in a
site's water chemistry.
152
-------
( oiiimeiil
Nil in her
(()r»;ini/;ilion)
Public ( o in moil 1 on Topic IS: Kcuiirdinu implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
El A-IIQ-OW-
2017-0260-0058
(National Council for
Air and Stream
Improvement, Inc.
(NCASI))
EPA should explicitly recommend that MLR model input data be
collected during the same sampling event(s) whenever possible.
EPA proposes using MLR models to characterize the toxicity of
total aluminum in freshwater aquatic systems as a function of pH,
hardness, and dissolved organic carbon (DOC). These models are
thought to capture the major independent variables affecting
aluminum toxicity. It is recommended that EPA add language
stating that it is most appropriate to use model input data obtained
from sampling events in which the full suite of model inputs
(hardness, pH, DOC) are collected simultaneously. These input
parameters are spatially and temporally variable and do not
necessarily vary in the same ways or to the same extent. For
example, export of DOC from most aquatic systems is driven by
hydrological processes (Schlesinger and Melack, 1981). It is
commonly observed that DOC concentrations in streams peak
during periods of high flow, which may typically occur during
times of snow melt runoff in early spring, and then decline rapidly
(Lewis and Grant 1979; Boyer et al. 1997; Sebestyen et al. 2008).
The temporal pattern of DOC in snowmelt dominated systems is
thought to be due to flushing ofpore water from the upper soil
horizons as the water table rises (Hornberger et al., 1994) and this
flushing phenomenon often exhausts the terrestrial DOC pool for
the year (Boyer et al., 1997). In contrast with DOC, hardness may
peak in concentration during the late summer months to early fall
months when steam flow is at its lowest. The USGS has
comprehensive statistics on stream flow by state within their
National Streamflow Statistics Program
[https://water.usgs.gov/osw/programs/nss/NSSpubs Rural.html#or
], and that program's results consistently show the lowest stream
flow during the months of July, August and September. The use of
datasets comprised of temporally- and spatially-linked water
quality endpoints helps ensure that consistent and reasonable
combinations of data inputs are used for the MLR models.
EPA-HQ-OW-
2017-0260-0060
(Katie Kistler,
Environmental
Manager of air
Programs, AK Steel
Corporation)
Method of Criteria Avvlicabilitv
The parameters upon which the draft criteria are based (pH,
hardness and DOC) are known to vary within the same receiving
stream.
Through our review of the draft criteria documents, we have not
located guidance or proposed governing language regarding
selection of these values when calculating the site specific criteria.
153
-------
( oiiimeiil
Nil in her
(()r»;ini/;ilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
Selection of the input parameter values will be of critical
importance to NPDES permit holders; in many cases of equal
importance to the criteria themselves. AK Steel also notes that
while hardness concentrations and pH ranges of receiving streams
are generally known, receiving stream concentrations of dissolved
organic carbon (DOC) are much less available, making it difficult
for NPDES permittees to fully evaluate the possible impact of the
revised criteria at this time.
Based upon our review, it is unclear to AK Steel whether EPA
intends for the draft criteria to be utilized with site-specific Water
Effect Ratio studies. Such studies may account for higher site
specific hardness and DOC concentrations than those upon which
the criteria are based, or may account forms of'particulate'
aluminum that may be less bioavailable than forms involved with
the toxicity tests upon which the criteria are based (e.g., aluminum
bound by clays).
Without such information, stakeholders are unable to provide
proper feedback. AK Steel requests that EPA withdraw or postpose
the proposed criteria until it coordinates with NPDES permitting
authorities on developing guidance for criteria implementation
and until any such information and guidance receives public
review and comment.
EPA-HQ-OW-
2017-0260-0069
(Julia Young, Water
Quality Standards
Coordinator, Kansas
Department of Health
and Environment
(KDHE))
4) Many states do not collect sample data for DOC, but do collect
TOC data. When developing the implementation guidelines for the
aluminum criterion it is recommended that EPA address the use of
conversion factors.
EPA-HQ-OW-
2017-0260-0020
(Jon Tack, Chief,
Water Quality Bureau,
Iowa Department of
Natural Resources
(DNR))
4. Implementation
The implementation procedure; the draft criteria need to address
the implementation issue and clearly state that States have the
discretion on how to implement the criteria. In the meantime, Iowa
has questions on aluminum criteria implementation :
(1) Are default criteria values (or input parameters) necessary? If
the answer is yes; please explain why.
(2) If default criteria values are necessary, do default criteria
values (or input parameters) require EPA approval? If the answer
154
-------
( ommcnl
Number
(Or^iini/iilion)
Public ( o in moil 1 on Topic IS: Kcuiirdinu implcmcnl;ilion
issues
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
is yes, please explain why.
EPA-HQ-OW-
2017-0260-0014
(David Waterstreet,
Manager, Watershed
Protection Program,
Water Quality
Division, Wyoming
Department of
Environmental Quality
(WDEQ/WQD))
Finally, WDEQ/WQD requests that EPA develop implementation
materials to accompany the criteria document. These materials
will allow WDEQ/WQD to fully evaluate the proposed criteria and
determine whether they are applicable to Wyoming surface waters.
EPA-HQ-OW-
2017-0260-0021
&
EPA-HQ-OW-
2017-0260-0022
(Daryll Joyner,
Administrator, Water
Quality Standards
Program, Florida
Department of
Environmental
Protection (DEP))
3. The documentation currently available for the proposed Al
criteria includes no guidance concerning EPA's recommended
implementation of the criteria. While EPA indicates
implementation guidance will be provided after the criteria are
finalized, States cannot conduct a complete evaluation of the
proposed Al criteria without this information. Details concerning
the planned implementation of a water quality criterion are a key
factor in understanding the protectiveness of any water quality
criterion, as well as the implications associated with adopting the
criteria. Therefore, DEP recommends that EPA complete and
provide their implementation guidance prior to finalizing the
proposed criteria.
EPA-HQ-OW-
2017-0260-0035
(RichardA. Hyde,
P.E., Executive
Director, Texas
Commission on
Environmental Quality
(TCEQ))
II. Lack of Guidance for Incorporation of the Criteria into Water
Quality Standards Programs of the Clean Water Act.
A. The TCEQ recommends that EPA coordinate with the states
and tribes to develop guidance, and should postpone the adoption
of the criteria until all the necessary information, including the
guidance, receives public review and comment.
The proposed criterion lacks guidance for the development of state
water quality standards. Guidance is needed to assist states in the
development of water quality standards. The following key areas
need to be addressed in the guidance:
Data needed to run the MLR model, such as DOC, may
be limited in state surface water quality datasets. EPA
should provide guidance to reliably estimate needed
parameters when data are limited. The EPA has
developed similar draft guidance to estimate parameters
for use in the biotic ligand model (BLM) for copper,
which may also be appropriate for aluminum. EPA
should clarify if methods described in Draft Technical
Support Document: Recommended Estimates for Missing
155
-------
( oiiimeiil
Nil in her
(()r»;ini/;ilion)
Public ( o in moil 1 on Topic IS: Kcuiirdinu implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
11 ater Quality Parameters far BL\ 1 are appropriate.
States, including Texas, have relied-upon procedures
such as WERs to modify EPA's 1988 aluminum criteria to
ensure site-specific conditions affecting the
bioavailability and toxicity of aluminum are
incorporated. Guidance is needed to clarify how to
address potentially-conflicting results between WERs and
EPA's proposal, to assist states when considering the
proposed criteria for adoption.
Given the complex nature of the proposal and the significant
change to the approach, the EPA should postpone finalizing the
proposed criteria and coordinate with states and tribes regarding
the expectations for inclusion in triennial reviews. Informational
material should be provided for review prior to finalization of the
criteria. Without this additional information, stakeholders cannot
completely evaluate the proposal and will miss the opportunity to
provide proper feedback.
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
Implementation: We understand that EPA does not have any
implementation guidance available at this point, but we strongly
suggest that the guidance be developed and ready upon
finalization of the criteria.
EPA-HQ-OW-
2017-0260-0057
(Roger Claff P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
In closing, API appreciates EPA's efforts to improve water quality
criteria derivation methodology through better consideration of
the effects of associated water chemistry on bioavailability. As our
comments suggest, however, additional analyses are needed to
improve the model to be more broadly applicable and avoid the
likelihood of misspent effort in implementing criteria in the
significant proportion of waters not represented by the model.
The water chemistry bounds for the 2018 criteria were
expanded, with details and rationale provided in the criteria
document in Section 2.7.1.
Text, tables and MLR
equations edited to
incorporate new
toxicity data
throughout the
document.
156
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I.PA Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0068
(Rachel Gleason,
Executive Director,
Pennsylvania Coal
Alliance (PCA))
In reviewing the Draft Aluminum Criteria, it is unclear how the
criteria would be applied to discharges that have a pH above the
default limit of 9.0. This occurs in Pennsylvania due to the
stringent treatment requirements for manganese at limit of 1 mg/L
on discharges. The calculator does not account for pH levels
above 9.0, however in order to precipitate manganese pH is often
raised to as much as 10.0.
PCA recommends that US EPA clarify in the draft
recommendation that in cases where pH is above the
default limit of 9.0 that the draft criterion can still be
utilized, as is the case for hardness and dissolved organic
carbon (DOC) outside of the default limits of the criterion
calculator.
We also feel that it's important that the states retain their
primacy and be allowed to develop their own criteria or
adopt the recommended criteria or portions of it as they
feel is appropriate for their unique regional variations.
Since the draft document was released, additional toxicity
tests were conducted with Ceriodaphnia dubia and
Pimephales promelas thereby expanding the water chemistry
empirical data used for model development. The bounds for
pH of the toxicity tests underlying the models ranged from
6.0-8.7. The EPA is allowing the user to extrapolate beyond
the pH values used to generate the MLR models. The criteria
calculator can be used to address all waters within a pH range
of 5.0 to 10.5. For additional discussion see Section 4 in the
2018 criteria document.
States can adopt the recommended criteria of other
scientifically-defensible criteria. The 2018 aluminum criteria
recommendations do enable inclusion of unique regional
variations in water chemistry.
No edits.
EPA-HQ-OW-
2017-0260-0045
(Lee Lemke, Executive
Vice President,
Georgia Mining
Association (GMA))
A unique characteristic of Georgia's mining industry is the
significant presence of the kaolin mining sector. Given the
chemical composition ofkaolinite, implementation of the Draft
Criteria will impose a serious and unnecessary regulatory burden
on the kaolin mining industry in Georgia. The cost to comply with
an aluminum standard based on the Draft Criteria in its current
form will lead to significant economic hardships for kaolin mining
companies with real potential for job loss and associated adverse
effects for companies providing support to this industry.
[Figure 1]
KAOLINITE
Kaolinite is a widespread aluminosilicate clay mineral in soils.
Kaolinite is particularly prevalent in warm, moist climates, such
as the southeastern United States (Figure 1). Kaolinite
(Al2Si205(0H4)) is composed of a tetrahedral and an octahedral
sheet, which constitute a single layer in a triclinic unit cell. This
structure renders kaolinite particularly resistant to weathering and
transformation, and provides relatively few adsorption sites
compared to many other clay minerals (Birkeland 1999).
"When two kaolinite sheets are superposed, the 0-present on the
upper surface and the H+ of the lower surface develop a strong
hydrogen bond OH between them, conferring with the van der
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum.
Dissolved, colloidal and precipitated forms of aluminum are
all bioavailable to aquatic organisms, which supports the
criteria as total aluminum. Thus, if aluminum criteria are
based on dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
total recoverable aluminum (U.S. EPA 1994a,b). This
method is based on acid soluble aluminum where the sample
is acidified to pH<2 and then filtered through a 0.45 |im
filter. This process does dissolve the monomeric and
polymeric forms of aluminum, in addition to colloidal,
particulate, and clay aluminum. However, the EPA Methods
200.7 and 200.8 are the currently approved methods for
aluminum.
Section 2.6.2
157
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I.PA Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
Waals bonds a great stability to the stack of sheets against the
action of water. " (El Brahmi and Zoukaghe 2016, p.69-70).
Kaolinite forms as a result of specific weathering reactions
(hydrolysis of feldspathic minerals) in the seasonal tropics and
subtropics. The resulting kaolinite mineral is an end product of
weathering in all but circumequatorial climates (Hugget 2011),
and is considered to be insoluble in water and otherwise inert
(Dixon 1977; Bloom 2004). Kaolinite is thus non-toxic and will not
become toxic at pH ranges of natural waters. Furthermore, among
clay minerals, kaolinite has a very low cation exchange capacity
(Birkeland 1999), and is therefore not a carrier ofbioavailable
aluminum species.
In the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
This approach would not acidify the sample to pH<2 but
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has been published as
Rodriguez, P.H., J.J. Arbildua, G. Villavicencio, P.
Urrestarazu, M. Opazo, A.S. Cardwell, W. Stubblefield, E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
EPA-HQ-OW-
2017-0260-0045
(Lee Lemke, Executive
Vice President,
Georgia Mining
Association (GMA))
The Draft Criteria relies on many experimental studies evaluating
the potential toxicity of aluminum using highly soluble metal salts
(e.g., aluminum chloride, aluminum sulfate, and aluminum
nitrate). These forms of aluminum are experimentally efficient to
use, as they fully dissolve in water. However, these forms are not
representative of many naturally-occurring forms of aluminum.
The aluminum in these highly soluble compounds becomes
biologically available on very short time frames (seconds to
hours). In contrast, the average kaolinite particle (or crystal) is
insoluble and accordingly will remain stable in the environment at
pH = 5.0 for 6,000,000+ years (Bloom 2004). Thus, the
experimental design of the referenced studies, and the conclusions
drawn from those studies, are completely inapplicable to
aluminosilicate minerals, including kaolinite, whose aluminum
atoms are physically bound within the mineral lattice. Soluble
aluminum salts are therefore not appropriate proxies for
kaolinite and other soil minerals that are insoluble and nontoxic.
The availability of aluminum species is indicated by a chemical
concept known as the solubility product constant (Ksp), where the
more readily soluble the material is in water, the higher the Ksp.
(Note: Ksp values of less than 10~4 are considered to be insoluble
(Bailor et al., 1978).) The enormous difference in the availability
of the aluminum in the salts used in the EPA's experiments and the
aluminum in kaolinite is illustrated by the fact that the Ksp values
differ by up to forty orders of magnitude (Table 1).
158
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
This guidance, from its title identifies the application of the
aluminum standard to be applicable only under ambient
conditions. The State of New Mexico would like guidance for
criteria which would be applicable under non-ambient
conditions.
The EPA's aquatic life criteria provide recommendations for
states and authorized tribes to consider in their adoption of
water quality standards under CWA section 303(c). States
may use these criteria to assess ambient waters and in
development of permit limits for discharges.
The intended protection goal of the 2018 final aluminum
criteria remains the same as that of the 1988 criteria,
protection of approximately 95% of genera in an ecosystem
to support protection of an aquatic life designated use. The
differences in the criteria values reflect an expanded toxicity
database and an improved incorporation of the effects of
water chemistry on bioavailability and toxicity in the 2018
final criteria.
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
As we reviewed the proposed criteria, the following question was
raised and should be addressed in Guidance: How are the criteria
translated into a discharge limit for a permit? Water quality can
change seasonally and therefore permittees wonder which samples
will be used for establishing the discharge limits.
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
Some of our states see the potential for deriving National Pollutant
Discharge Elimination System (NPDES) permit limits for
aluminum from multiple linear regression analysis. We request
that EPA provide detailed guidance on the data collection
necessary to support reasonable potential analysis for NPDES
permits and the site-specific adjustment of aluminum criteria.
Further, how will anti-backsliding requirements be applied to the
development of site specific criteria requests where data could
allow for an increase in aluminum concentrations? Further
guidance from EPA is requested to address such situations.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEP))
Implementation: NPDES Permits - Data Collection & Anti-
backsliding - National Pollutant Discharge Elimination System
(NPDES) permits for wastewater treatment facilities (WWTFs) and
Drinking Water Treatment Facilities (DWTFs) may include
effluent limits for total aluminum. EPA Region 1, who is the lead
permitting authority in Massachusetts because we do not have
NPDES delegation, is including aluminum limits at some WWTFs
and DWTFs. Many of these facilities are likely to need costly
retrofits and/or alternative treatment coagulants in order to meet
those limits.
To develop permit limits, permit writers conduct an analysis to
determine if there is a "reasonable potential" that water quality
standards will be violated. Permit limits must comply with existing
water quality standards and the determination of the final limits
must also include an "anti-backsliding" analysis to maintain the
integrity of receiving waters. Anti-backsliding statutory and
regulatory provisions prohibit restrictions on effluent discharge in
an existing permit that are less stringent than the restrictions
established in previous permits at the same facility, except under
159
-------
( ommcnl
Number
(Or^iini/iilion)
Public ( o in moil 1 on Topic IS: Kcuiirdinu implcmcnl;ilion
issues
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
specific circumstances.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEP))
6. Once finalized, the aluminum MLR could also be used to
conduct reasonable potential analysis and, if necessary, derive
NPDES permit limits for aluminum. If a permit holder wishes to
pursue the development of site-specific criteria using the
aluminum MLR, data collection will be necessary. The ideal time
to begin data collection is two to three years before the permit
expires. Massachusetts intends to work with the regulated
community to conduct reasonable potential analyses and, if
necessary, collect data site-specifically to support adjusting the
aluminum criteria to local conditions in the Commonwealth. We
request that EPA provide explicit guidance on the data collection
necessary to support reasonable potential analysis for NPDES
permits and the site-specific adjustment of aluminum criteria. In
addition, MassDEP requests guidance on how the anti-backsliding
provisions will be implemented in cases where the new criteria
model results in a higher applicable aluminum criteria.
EPA-HQ-OW-
2017-0260-0053
(Abdul Alkhatib,
Director,
Massachusetts Water
Works Association
(MWWA))
Implementation: I understand that EPA does not have any
implementation guidance available at this point, but we strongly
suggest that the guidance be developed and ready upon
finalization of the criteria.
EPA should provide updated guidance for performing calculations
and/or studies to determine higher regulatory Aluminum toxicity
limits when water bodies are not within the calculator's limits for
pH, hardness, and DOC.
As our organization reviewed the proposed criteria, the following
question was raised and should be addressed in Guidance: How
are the criteria translated into a discharge limit for a permit?
Water quality can change seasonally and therefore permittees
wonder which samples will be used for establishing the discharge
limits.
It is important for EPA to define the "site "for sampling the water
quality parameters that are input into the model. I understand that
the samples for the water quality parameters (hardness, TOC,
DOC, pH) should be done in the receiving waters and not from the
discharge, but EPA should make that explicit in the final
document.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
The water chemistry bounds for the 2018 criteria were
expanded, with details and rationale provided in Section 4.0
in the criteria document.
No edits.
160
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United
States Steel
Corporation)
States should be allowed to use the dissolved form of acid-soluble
aluminum with site-specific dissolved-particulate studies to
determine a particular facility's permit limits.
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum.
Dissolved, colloidal and precipitated forms of aluminum are
all bioavailable to aquatic organisms, which supports the
criteria as total aluminum. Thus, if aluminum criteria are
based on dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
total recoverable aluminum (U.S. EPA 1994a,b). This
method is based on acid soluble aluminum where the sample
is acidified to pH<2 and then filtered through a 0.45 |im
filter. This process does dissolve the monomeric and
polymeric forms of aluminum, in addition to colloidal,
particulate, and clay aluminum. However, the EPA Methods
200.7 and 200.8 are the currently approved methods for
aluminum.
If a state used a dissolved form of aluminum with site
specific dissolved particulate studies for permit limits, it
would not address the colloidal and precipitated forms of
aluminum that are bioavailable to organisms. In addition, it is
unclear how the method for using the aluminum criteria
calculator would need to change to address this approach.
In the 2018 Final aluminum criteria document, the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
This approach would not acidify the sample to pH<2 but
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has been published as
Rodriguez, P.H., J.J. Arbildua, G. Villavicencio, P.
Urrestarazu, M. Opazo, A.S. Cardwell, W. Stubblefield, E.
No edits.
EPA-HQ-OW-
2017-0260-0043
(Blake Beyea,
Standards Unit
Manager, Water
Quality Control
Division, Colorado
Department of Public
Health &
Environment)
The division recommends including a discussion about seasonal
variability in the next version of the criteria document. For
example, during high-flow, snowmelt conditions, DOC often
increases while hardness decreases; these types of changes in
water chemistry could result in the need for more stringent criteria
during part of the year to ensure protection of aquatic life. While
the division understands that this type of variability would likely
need to be addressed on a site-specific basis, it is important to
acknowledge that seasonal conditions may cause changes in water
chemistry and potentially the bioavailability of aluminum to
aquatic life.
161
-------
( oiiimeiil
Nil in her
(Or^iini/iilioii)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
Nul'dllullll. alld W. Adallls. 2019. Delei'llllllaUull ul
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. Concerns
about different conditions in regard to flow and changes in
water chemistry due to seasonal variation will be discussed.
The EPA recognizes that there are several aspects of the
recommended criterion that will benefit from technical
support documents to enhance implementation of state and
tribal criteria and is planning to develop such documents and
make them available for public comment.
EPA-HQ-OW-
2017-0260-0023
(Stan Dempsey Jr.,
CMA President,
Colorado Mining
Association (CMA))
EPA should provide some guidance as to whether the MLR can be
modified to be state, region, or species specific. When States
implement these criteria recommendations in the water quality
standards, States often begin with the final EPA criteria and then
modify it to be more applicable to the aquatic species within the
state waters. States should be given some guidance as to whether
modification of the MLR would follow the same approach.
Thank you for your suggestion. Please work with your local
EPA Region and the EPA Headquarters' staff to develop site-
specific criteria values (i.e., add/delete species/genera), if
appropriate.
No edits.
EPA-HQ-OW-
2017-0260-0029
(Hall & Associates on
behalf of Minnesota
Environmental Science
and Economic Review
Board (MESERB))
A review of the criteria for various water chemistry conditions
(Appendix K of the Draft), includes summaries of ranking for the
four most sensitive genera, used to derive total aluminum acute
and chronic criteria values. The CMCs and CCCs presented in
Appendix K show that within the pH range of 6.0 - 8.0, criteria
values are very dependent upon the presence of salmonids (Salvo,
Salvelinus, Oncorhynchus). This suggests that either there should
be separate cold water and warm water criteria for aluminum, or
criteria should be reevaluated for waters that do not support
salmonids.
Species included in a sensitivity distribution for criteria are
considered surrogates for other taxonomically-related
species, due to genetic conservation of important toxicity
response traits in species. Fish in the family Salmonidae,
such as the Atlantic salmon, include many recreationally and
commercially important species, as well as endangered
species, which are have broad relevance across the U.S.
Further, regarding comments from the Minnesota
Environmental Science and Economic Review Board
regarding the utility of the aluminum criteria due to the
inclusion of salmonids in the sensitive genera, the Minnesota
Department of Natural Resources' website
(httDs://w\\w.dnr.state, mn. Lis/fishina/trout streams/trout sdc
No edits.
162
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0032
(Phillip M. Gonet,
President, Illinois Coal
Association (ICA))
Selected Species
The Draft Criteria states that the most sensitive species in the MLR
model was the Atlantic salmon. EPA should not be using a species
with such limited range to define nationwide criteria.
Furthermore, EPA should provide guidance to states that explains
how the criteria can be modified to be state-specific, such that
each state can determine what species list is most appropriate for
their waterways.
cies.html) specifics that three soccics of trout are found in
southeast Minnesota, brook trout (Salvelinus fontinalis),
brown trout (Salmo trutta) and rainbow trout (Oncorhynchus
mykiss), thus three salmonid genera are present in
southeastern MN. The DNR site also notes that trout lakes
are found primarily in northeastern MN and are extremely
popular with recreational fishers
(h11ds://www.dnr.statc.mn.us/fishina/trout lakes/index.html).
EPA-HQ-OW-
2017-0260-0034
(James Boswell, Senior
Manager,
Environmental,
Peabody Energy)
Avvlicabilitv of Species: The draft criteria document notes that the
fish genus Salmo, represented by the Atlantic salmon, was the most
sensitive genus. EPA should include some discussion of how this
species impacts the resulting criteria and what potential options
for modification there are for states where a recalculation
procedure may be appropriate. States often use a recalculation to
modify criteria based on species present within a state or region.
Peabody expects that aluminum will be no different than other
metals, so EPA should provide some discussion of the options
states have for modifying this criteria to a state-specific value or
species' subset value. For example, many of Peabody's western
operations are located in areas with limited aquatic life and no or
very few fish as a result of limited streamflow. Such areas would
be a prime target for application of the standard based on a subset
of species. Basing a standard on species that are absent in a
region will again result in unnecessary costs to states and industry
studying aluminum levels and implementing reduction measures
when it is not necessary to protect the aquatic life that is present.
[TABLE 1]
and that lake trout (Salvelinus namaycush) and rainbow trout
are also found in Lake Superior. Thus, inclusion of salmonids
is broadly useful for aluminum criteria development relevant
to at least a number of areas in Minnesota.
Regarding the utility of including salmonids in the sensitivity
distribution for aluminum in Illinois, the IL DNR notes that
brook trout live in streams in the northern one-fourth of the
state and in Lake Michigan, and that both brown trout and
rainbow trout are stocked in IL in Lake Michigan and other
lakes and streams for recreational fishing.
(littDs://\\w\\.dnr.illinois. eov/education/Paees/W AFSalmon.a
spx)
Regarding the utility of including salmonids in the sensitivity
distribution for aluminum in Indiana, the Indiana DNR, notes
that brook, brown, lake and rainbow trout are found in the
northern area of the state near the Great Lakes region, with
brook trout and lake trout native to the Great Lakes area;
rainbow and brown trout introduced to Indiana.
(httos://www.in.eov/dnr/fishwild/files/fw-trout.i3df)
Due to the complexity of the final 2018 aluminum criteria,
please work with your local EPA Region and the EPA
Headquarters' staff to develop site-specific criteria values,
including species recalculation procedures, as appropriate.
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
It is also not clear how multiple samples from a site should be used
in the Aluminum Criteria Calculator. For sites with multiple input
datasets (i.e., pH, hardness, and DOC collected at different times),
would the approach for the calculator be similar to the approach
for the Biotic LigandModel (BLM) for copper? The criteria
document should also include a discussion regarding the
applicable geographic extent of any site-specific water quality
criteria, particularly in light of downstream protection provisions
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
No edits.
163
-------
( oiiimeiil
Nil in her
(()r»;ini/;ilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
within Water Quality Standards. For example, it would be
necessary to evaluate changes in water quality throughout a
watershed to determine if there is a potential for aluminum to
become more bioavailable based on water chemistry changes
further downstream in the watershed.
implementation documents arc also intended lo pro\ ide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
EPA-HQ-OW-
2017-0260-0042
(Bruce A. Stevens,
President, Indiana
Coal Council, Inc.
(ICC))
When States implement the Federal criteria recommendations into
State water quality standards, they are typically allowed to modify
the criteria to be specific to the species that inhabit the state. EPA
should include some discussion of how this can be accomplished
with the MLR model. This is particularly true because the most
sensitive species in the MLR model is the Atlantic salmon, which is
extremely limited in its range. States should be able to modify the
MLR to account for a more representative species list and EPA
should provide some discussion of this process.
Species included in a sensitivity distribution for criteria are
considered surrogates for other taxonomically-related
species, due to genetic conservation of important toxicity
response traits in species. Fish in the family Salmonidae,
such as the Atlantic salmon, include many recreationally and
commercially important species, as well as endangered
species, which are have broad relevance across the U.S.
Please work with your local EPA Region and the EPA
Headquarters' staff to develop site-specific criteria values, if
appropriate.
No edits.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
8. The District understands that the EPA's 304( a) water quality
criteria are merely "guidance" and impose no direct binding
obligation on dischargers until states act to adopt these criteria as
water quality standards. [DAC @ pg. iv] However, states are
required to review and update their water quality standards every
three years and, where a 304(a) criteria has been developed, are
expected to adopt a standard based on that criteria or using some
other scientifically-defensible approach. Based on years of prior
precedent, it is evident that few states have the resources or
expertise to develop their own water quality criteria and, instead,
elect to rely on the EPA's 304(a) guidance to establish new water
quality standards for most pollutants. In fact, more recently some
states simply use the EPA's 304(a) guidance to "translate" existing
narrative water quality standards when preparing triannual
303(d) water quality assessment or deriving numeric effluent
limits. For this reason the District is deeply concerned that the
proposed criteria for Total Aluminum will be applied without
regard for the many nuanced "uncertainties" the EPA calls out in
the draft document. [DAC @pg. 69] This is made more likely by
the EPA's decision to discount these uncertainties by describing
its approach as "conservative. " Such a claim leaves a false
impression that the "conservative " approach somehow addresses
the residual scientific uncertainties, when it does no such thing.
Thank you for your comment.
EPA is confident that the criteria developed and externally
peer reviewed represent the latest science and are protective
of aquatic life designated uses. There are uncertainties in all
scientific analyses and for transparency EPA included a
discussion of uncertainties in data available and in
extrapolation of criteria beyond the bounds of the empirical
model data. However, the overall database for aluminum in
freshwater is robust and the criteria developed represent the
latest science.
No edits.
164
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
(K) The EPA should consider developing a separate water quality
criteria recommendation for warm water and cold water
ecosystems. Past experience has shown that stakeholders
throughout the country must pay consultants to perform the exact
same recalculation procedure to adjust for highly sensitive cold
water species (like trout and salmon) that are not present in warm
water streams. It would save considerable cost if the EPA were to
do this calculation itself and publish the results as an acceptable
warm water alternative so that state authorities could consider this
difference from the outset rather than having to undertake the
burdensome rule-making procedure required to adopt site-specific
standards on case-by-case basis.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
Species included in a sensitivity distribution for criteria are
considered surrogates for other taxonomically-related
species, due to genetic conservation of important toxicity
response traits in species. Fish in the family Salmonidae,
such as the Atlantic salmon, include many recreationally and
commercially important species, as well as endangered
species, which are have broad relevance across the U.S.
Please work with your local EPA Region and the EPA
Headquarters' staff to develop site-specific criteria values, if
appropriate.
No edits.
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
The Draft Criteria uses the Atlantic salmon as the most sensitive
species in the Multiple Linear Regression ("MLR") model. This
species, which has a limited range and does not exist in many of
the states that will be potentially impacted by the criteria, is not
appropriate for use in establishing national recommended criteria.
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
When considering the input values used for the derivation of the
acute water quality criterion for aluminum and the 1-hour in three
year duration and frequency for the criterion, there may need to be
a consideration of the impacts of storm water on receiving water
chemistry. In water bodies that are significantly affected by
stormwater, water chemistry may change for one or more hours as
a result of storm events. These changes may need to be considered
as part of the evaluation of the appropriate values for pH,
hardness, and DOC to be used in criteria derivation.
165
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
Given the high concentration of aluminum in natural soils and the
high concentration of soil particles entrained in storm water and
stream flows during periods of wet weather, it is absolutely certain
that very high concentrations of Total Aluminum will be reported
for samples collected under such conditions. This data will be used
to conclude that a huge number of lakes and rivers are "impaired"
by excessive aluminum and should be added to the state 303(d)
lists. Recent experience indicates that this will most likely occur by
claiming that the measured concentrations of Total Aluminum
violate the state's narrative toxicity standard. [See, for example,
Line of Evidence #8321 for Decision ID #65478 in California's
2014-2016Integrated 303(d) Report where EPA's 1988 aluminum
criteria was used to translate a narrative toxicity standard.] These
listings will trigger a follow-up requirement to develop TMDLs
with Load and Waste load Allocations which will also be translated
directly from the EPA's 304(a) criteria. Therefore, it is incumbent
on the EPA to make certain that the proposed criteria for Total
Aluminum include detailed guidance to ensure that it is interpreted
and implemented in a manner that is consistent with the numerous
caveats and assumptions scattered throughout the draft document.
To that end, the District offers the following recommendations:
The EPA is aware, and has noted in the 2018 aluminum
criteria document, that under natural conditions not all forms
of aluminum would be biologically available to aquatic
species (e.g., clay-bound aluminum). The EPA has also noted
in its 2018 final aluminum criteria document that the EPA
Methods 200.7 and 200.8 are the only currently approved
methods for measuring aluminum in natural waters and
wastes for NPDES permits. The EPA further notes in the
2018 criteria document that research on new analytical
methods is ongoing to address concerns with including
aluminum bound to particulate matter (i.e., clay) in the total
recoverable aluminum concentrations (OSU 2018c).
environment.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
(F) The final criteria should recommend that the Criteria
Maximum Concentration (CMC), not the Criteria Continuous
Concentration (CCC), are more appropriate for evaluating the
true potential for instream toxicity to occur during wet weather
events. In addition, the EPA should warn that the 1-hour exposure
assumption normally used for priority pollutant metals might not
be appropriate for non-conventional pollutants such as aluminum.
The EPA determined that the one-hour average assumption
for the CMC is appropriate. This position is supported by the
1985 Guidelines. More specifically, page 5 of the 1985
Guidelines states that"For the CMC the averaging period
should again be substantially less than the lengths of the tests
it is based on, i.e., substantially less than 48 to 96 hours. One
hour is probably an appropriate averaging period because
high concentrations of some materials can cause death in one
to three hours. Even when organisms do not die within the
first hour or so, it is not known how many might have died
due to delayed effects of this short of an exposure. Thus, it is
not appropriate to allow concentrations above the CMC to
exist for as long as one hour. The durations of the averaging
periods in national criteria have been made short enough to
restrict allowable fluctuations in the concentration of the
pollutant in the receiving water and to restrict the length of
time that the concentration in the receiving water can be
continuously above a criterion concentration." Page 6 of the
1985 Guidelines further states that"the one-hour average
should never exceed the CMC." The duration of a criterion is
based on scientific considerations of toxicological activity;
No edits.
166
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ iiiiplemoiiUilioii
issues
I'.I'A Response
Ke\ isiou Locution in
20IX Aluminum
Crileriii Document
status on the existing priority pollutant list is not considered.
EPA-HQ-OW-
2017-0260-0063
(Kevin Oakes, Director
of Wastewater,
Borough of West
Chester, Chester
County, Pennsylvania)
B. The Borough of West Chester believes seasonal water quality
criteria for aluminum should be developed to reflect the water
environment in the receiving stream in cold seasons.
In general, the period between November 1st and March 31st is
considered as "non-or low growth season "for aquatic lives. In
fact, both USEPA and PaDEP have traditionally applied higher
seasonal NPDES limits for ammonia, total nitrogen (TN) and total
phosphorus (TP) in treated wastewater during that period.
The seasonal WQC can be beneficial to the environment because
the demands for chemicals and energy will be lower to treat
wastewater prior to discharging treated wastewater to receiving
surface water with lower temperatures. It should be noted that
when the chemical application rates become lower, the treated
wastewater may contain lower amount of residual chemicals that
do not react effectively with wastewater at lower temperatures.
These facts are particularly true for the states in
northeastern US, where surface water temperature can be lower
than 10 degrees Celsius (50 degrees Fahrenheit) for months in a
year as shown on Figure No. 1.
[Figure 1]
As stated above, without seasonal limits for cold months, much
more chemicals (alum for coagulation and settling, soda ash for
pH adjustment, polymers for thickening and dewatering, etc.),
energy and associated sampling and testing are needed to achieve
the same year-round limits in cold months, because the rates for
chemical, biological and bio-chemical reactions are slower. As
such, the cost to treat wastewater and to remove and dispose of
additional sludge becomes much higher than summer months when
water temperature is higher.
Please also note that the production, transport and the use of more
chemicals and energy can produce more negative impacts on the
environment. Because those processes will produce more air,
water and solids pollutants and wastes.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
Temperature was not considered because of the lack of
sufficient experimental data that could be used to develop an
additional parameter in the MLR. Further, the EPA is aware
of existing scientific information that indicates that
temperature effects to toxicants may simply reflect time to
observed effect, but not necessarily a lesser sensitivity to the
magnitude of exposure (i.e., the concentration).
No edits.
167
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ iiiiplemoiiUilioii
issues
I'.I'A Response
Ke\ isiou Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0063
(Kevin Oakes, Director
of Wastewater,
Borough of West
Chester, Chester
County, Pennsylvania)
The effects of low water temperature on aluminum toxicity are
discussed below:
1. At lower water temperature, the solubility of aluminum becomes
lower, which lowers the toxicity of total aluminum.
2. At lower water temperature, the chemical/biochemical reaction
rates and the metabolic rates also become lower, which means
lower rates of consumption and utilization by aquatic lives and
therefore, lower overall toxicity of total aluminum.
3. In accordance with Arrhenius' equation, the typical chemical
reaction rate doubles for every 10 degree Celsius increase in
temperature.
Figure No. 1 above shows water temperature records (obtained
from USGS Gaging Station 01480178) between January 2013 and
July 2017, for the East Branch Brandywine Creek in southeast
Pennsylvania. In general, the duration of water temperature below
50 degrees F ranges between four (4) and five (5) months (or
longer) in a year.
4. The results and conclusion of many aluminum toxicity studies on
fishes at low water temperature (such as the research work by
Antonio Poleo of the Department of Biology, University of Oslo,
etc.) are well documented and published. However, those results
and findings were not used by the USEPA to develop the proposed
aluminum water quality criteria for various reasons (i.e. the
quality of data, sampling and testing protocols do not agree with
the standards established by the USEPA, or the article text are in
foreign language, etc.).
Nevertheless, we have attached for USEPA's review, a 2002
article published in the Journal of Limnology, entitled "Seasonal
Variation in Mortality of Brown Trout (Salmo trutta) in an Acidic
Aluminum-rich Lake " by Espen Lydersen, et al. of Norwegian
Institute for Water Research. The authors of this article conducted
extensive experiments in their research, their results also show the
toxicity of aluminum becomes lower when water temperatures are
lower.
In light of the findings of many published aluminum toxicity
Temperature was not considered because of the lack of
sufficient experimental data that could be used to develop an
additional parameter in the MLR. Further, the EPA is aware
of existing scientific information that indicates that
temperature effects to toxicants may reflect time to observed
effect, not necessarily a lesser sensitivity to the magnitude of
exposure (i.e., the concentration).
No edits.
168
-------
( oiiimeiil
Nil in her
(()r»;ini/;ilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
related scientific articles, it is our opinion that the L'SEPA
research should include temperature as the fourth parameter as a
part of the MLR analysis or for future WQC development.
Further, we believe the development and implementation of
seasonal limits (or equivalent) for aluminum can be beneficial to
the environment as discussed above.
EPA-HQ-OW-
2017-0260-0057
(Roger Claffl P.E.,
Senior Scientific
Advisor, American
Petroleum Institute
(API))
EPA's description of aluminum environmental loadings, chemistry,
and fate indicates most aluminum influx to aquatic environments
comes from atmospheric sources via both wet and dry deposition
and subsequent runoff from land. A total maximum daily load
(TMDL) derived from the aluminum loadings within a watershed
would allow little allocation for permitted discharges;
understating acceptable aluminum concentrations in waters
receiving permitted discharges because the MLR model is
incomplete may result in costly and potentially ineffective
mitigation efforts by dischargers.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c).
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
Use of Water Effect Ratios
The appropriate use of the Water Effect Ratio (WER) procedure
should be discussed in the Criteria Document. For some
waterways, a WER would demonstrate that the form of aluminum
typically found in stormwater (e.g. aluminum clays naturally
bound to silicates, oxides and calcites) is not bioavailable or toxic.
The states and permittees need clarification of if and when a WER
would be appropriate.
Research on new analytical methods is ongoing to address
concerns with aluminum bound to particulate matter (i.e.,
clay) from natural waters being included in the total
recoverable aluminum concentrations. This approach would
not acidify the sample to pH<2 but rather to a higher pH to
better capture the bioavailable fraction of aluminum. The
EPA briefly discusses this new research in the final 2018
aluminum criteria document.
The 2018 final aluminum criteria reflects the current science
and a larger database than a Water Effect Ratio applied to the
superseded 1988 aluminum criteria. The Water Effect Ratio
depends greatly on the particular "snapshot" conditions
during the WER tests.
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
No edits.
169
-------
( ommcnl
Number
(Or^iini/iilion)
Public ( o in moil 1 on Topic IS: Kcuiirdinu implcmcnl;ilion
issues
I'.I'A Response
Rc\ ision Locution in
20IX Aluminum
Crilcriii Document
as>s>is>laiice lo oilier ilakeholdeii and llie public. Hie LPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The final criteria should recommend that states consider using the
Water Effects Ratio (WER) procedure to determine if measured
concentrations of Total Aluminum are actually in a bioavailable
form and particularly where the ambient hardness and DOC fall
outside the range of values used to develop the EPA's multiple
regression model. It should be noted that the WER relies on the
same test methods used to develop the recommended criteria and,
when correctly applied, are not "less protective. "
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). The
implementation documents that the EPA is developing are
intended to provide assistance to states and authorized tribes
that adopt into the water quality standards a criterion based
on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
The 2018 final aluminum criteria is reflective of a larger
toxicity and water chemistry database than a WER, which
can depend greatly on the limited and particular "snapshot"
conditions during the WER tests.
No edits.
EPA-HQ-OW-
2017-0260-0021
&
EPA-HQ-OW-
2017-0260-0022
(Daryll Joyner,
Administrator, Water
Quality Standards
Program, Florida
Department of
Environmental
Protection (DEP))
The spreadsheet criterion calculator is very helpful in
understanding the proposed criteria and how they were developed.
However, as currently structured, it would be very cumbersome to
adopt the calculator into a State's water quality standards rule or
to implement it in a statewide 303(d) assessment. DEP
recommends that EPA simplify the expression and calculation of
the criteria into single equations for the acute and chronic criteria,
ifpossible.
The two MLR models (equations) are used to normalize the
aluminum toxicity data for fish and invertebrates, and the
criteria are calculated through the criteria calculator, not
through direct use of the equations. It would be appropriate to
reference Appendix K or the Aluminum Criteria Calculator in
a State's water quality standards rule.
No edits.
EPA-HQ-OW-
2017-0260-0025
(Peter T. Goodmann,
Director, Kentucky
Division of Water)
The EPA recommends numeric criteria for aluminum at pH = 7,
total hardness = 100 mg/L as CaC03, and DOC = 1 mg/L.
However, the recommended criteria vary as these three
constituents change. States may find it impractical, or may even be
prohibited by state administrative regulatory requirements to
codify a model as a state water quality standard.
The water quality characteristics that the EPA uses as a
scenario throughout the document were simply selected as an
example scenario. It would be appropriate to reference
Appendix K or the Aluminum Criteria Calculator in a State's
water quality standards rule.
No edits.
170
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0036
(Barry N. Burnett,
Water Quality Division
Administrator, State of
Idaho Department of
Environmental Quality
(DEQ))
DEQ supports the use of a multiple linear regression model (MLR)
to derive site and time-specific criteria. This approach has several
advantages over black-box models (e.g., the biotic ligand model)
for state implementations, including providing the public and
regulated community the ability to calculate criteria based on site
conditions without the need for specialized software. While the
lookup tables and the Aluminum Criteria Calculator V.1.0 are
useful tools, DEQ believes that the criteria statement should
include, as an alternative option, the actual MLR equation used to
derive aluminum criteria.
Thank you for your comment. The two MLR models
(equations) are used to normalize the aluminum toxicity data
for fish and invertebrates, and the criteria are calculated
through algorithms in the criteria calculator that reflect the
1985 Guidelines methods for criteria calculation from a
sensitivity distribution of genera, not through direct use of
the equations. It would be appropriate to reference Appendix
K or the Aluminum Criteria Calculator in a State's water
quality standards rule.
No edits.
EPA-HQ-OW-
2017-0260-0040
(Susan J. Sullivan,
Executive Director,
New England
Interstate Water
Pollution Control
Commission
(NEIWPCC))
Relating to the Aluminum Criteria Calculator, NEIWPCC requests
that EPA clarify how the criteria will be expressed in surface
water standards. It would also be helpful to understand how the
criteria calculator can be used as part of a surface water quality
standard. For example, additional clarity is needed for the
following:
It is not clear what (or how many) inputs are needed to
the Aluminum Criteria Calculator in order to adjust the
criteria to local conditions. For example, the draft
document does not provide a recommended number of
samples needed to account for the effects of seasonality,
diurnal water quality changes and/or site-specific
variability on the input parameters for the MLR models
used to derive aluminum criteria.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). The
implementation documents that the EPA is developing are
intended to provide assistance to states and authorized tribes
that adopt into the water quality standards a criterion based
on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
8. The proposed guidance for aluminum is one that implements a
multi-parameter approach (pH, hardness, DOC and aluminum).
There are several areas to consider as it pertains to this type of
guidance.
a. Implementation of such a guidance must be carefully explored
as the impacts to development of numerical criteria in National
Pollutant Discharge Elimination System (NPDES) permits which
already require in-depth calculations for development, will now
have to evaluate based on multiple water quality parameters, and
the regulated community may have to increase monitoring for
these additional parameters.
b. Additional resources for State and EPA regulatory staff may be
incurred to develop appropriate criteria in NPDES permits that
reflect the protective limits of aluminum as well as a need for
guidance for implementation of NPDES permits. In order to
incorporate the new criteria into NPDES permits, reasonable
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). As noted in the
criteria document, the EPA decided to use an empirical MLR
approach in the aluminum criteria update rather than a BLM
model due to: 1) the relative simplicity and transparency of
the model, 2) the relative similarity to the available BLM
model outputs, and 3) the decreased number of input data on
water chemistry needed to derive criteria at different sites.
The EPA is also separately compiling an updated national
database of water chemistry conditions relevant to the MLR
model: hardness, pH and DOC, and will make that data
available to in the future to support states and stakeholders
needs for model input data, when their own data are not
available.
The implementation documents that the EPA is also
No edits.
171
-------
( oiiimeiil
Nil in her
(()r»;ini/;ilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(liiii£ implcmcnl;ilion
issues
I.PA Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
potential calculations will need modification to account for the
multi-parameter approach to aluminum.
c. Consideration must be given to the implementation of water
quality standards across boundaries. The State of New Mexico
poses a unique scenario as it has shared waters with four states, a
downstream country and 23 Tribes; all of which have the authority
to adopt and impose water quality standards of their own to which
the Waters of the State must consider and adhere to as they cross
jurisdictional boundaries. The ability to coordinate maintaining
water quality standards using this multi-parameter process will be
challenging given the numerous entities involved.
d. The multi-parameter approach employs the use of dissolved
organic carbon (DOC). Currently, the laboratory to which the
State uses for analytical analysis is not equipped to analyze DOC.
If the State were to consider implementing the new guidance, it
would require additional resources to retain a laboratory capable
of conducting DOC analysis. This issue is also applicable to
permittees and organizations that collect data which is submitted
to the State for use in assessment. Implementation guidance will
need to address issues wherein resources are not readily
available to states and tribes (and their labs) to use the criteria
calculator.
developing arc iiilcinlcil lo pio\ kle assistance lo siales and
authorized tribes that adopt into the water quality standards a
criterion based on or similar to the EPA's recommended
criterion. The implementation documents are also intended to
provide assistance to other stakeholders and the public. The
EPA recognizes that there are several aspects of the
recommended criterion that will benefit from technical
support documents to enhance implementation of state and
tribal criteria and is planning to develop such documents and
make them available for public comment.
EPA-HQ-OW-
2017-0260-0046
(Jennifer Wigal,
Program Manager,
Water Quality
Standards &
Assessments, Oregon
Department of
Environmental
Quality)
The criteria recommendation document states that "when any of
the water quality parameters selected is 'outside model inputs', the
Aluminum criteria calculator ... provides the user a warning to use
discretion in applying criteria in those conditions." (p. 72). This
indicates that states have discretion whether or not to apply
criteria values generated in the extrapolated range of the input
parameters as regulatory criteria. EPA should clearly state that
they are reserving for states the discretion to apply, or not to
apply, criteria values generated using inputs outside the validated
range as regulatory criteria. The current reference to "generated
warnings" is ambiguous.
In other words, EPA is allowing states to decide the bounds of the
criteria calculator they wish to use beyond the calibrated range of
the model (6.5 to 8.5). Only criteria values that are valid for
regulatory use should be included as part of the recommended
criteria. A calculator for research purposes with extended non-
regulatory parameter ranges should instead be provided as a
Thank you for your comments. Since the draft document was
released in 2017, additional toxicity tests were conducted
with Ceriodaphnia dubia and Pimephales promelas thereby
expanding the water chemistry empirical data used for model
development. As a result, the recommended bounds have
changed.
The water chemistry bounds for the 2018 criteria were thus
expanded, with details and rationale provided in the criteria
document. These bounding approaches for hardness, DOC
and pH are reflected in the criteria lookup tables. These
approaches were taken so that the recommended criteria can
be provided for, and will be protective of, a broader range of
U.S. natural waters. Recommended extrapolated criteria
values outside of the empirical data tend to be lower values
and will be more protective of the aquatic environment.
Criteria values estimated outside of the range of the empirical
data are more uncertain. The calculator provides warning
Aluminum Criteria
Calculator
172
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
supplement. DEQ agrees that in no case should input values
outside the extrapolated ranges (i.e. outside pH 5.0- 9.0) be used
to generate regulatory criteria.
when outside the bounds of extrapolated data and future
implementation guidance addresses this point also.
EPA-HQ-OW-
2017-0260-0051
(Douglas E. Fine,
Assistant
Commissioner for
Water Resources,
Massachusetts
Department of
Environmental
Protection
(MassDEPJJ
4. Implementation: Standards Adoption - User Guidance for
Criteria Calculator - EPA created a user-friendlv Aluminum
Criteria Calculator V.1.0 (Aluminum Criteria Calculator
V.l.O.xlsx). EPA should explain in the guidance document how this
criteria calculator can be adopted into a surface water quality
standard. It is requested that EPA clarify the preferred approach
for how the criteria will be expressed in surface water standards
(e.g., as an equation, calculator, look up table, regional criteria
based on monitoring). An example of the preferred approach will
be helpful to states as they move to adopt these criteria.
It is not clear what inputs are needed to Aluminum Criteria
Calculator V.1.0 in order to adjust the criteria to local conditions.
For example, the draft document does not provide a recommended
number of samples needed to account for the effects of seasonality
and/or site-specific variability on the input parameters for the
MLR models used to derive aluminum criteria.
It is also not clear how multiple samples from a site should be used
in the Aluminum Criteria Calculator. For sites with multiple input
datasets (i.e., pH, hardness and DOC collected at different times),
would the approach for the calculator be similar to the approach
for the BLMfor copper? Should users enter multiple input datasets
from one site into the calculator to generate multiple CMC and
CCC criteria and then derive final CMC and CCC criteria using
appropriate summary statistics?
Another tab was added to the Aluminum Criteria Calculator
that provides instructions. It would be appropriate to
reference Appendix K or the Aluminum Criteria Calculator in
a State's water quality standards rule.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). The
implementation documents that the EPA is developing are
intended to provide assistance to states and authorized tribes
that adopt into the water quality standards a criterion based
on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
Aluminum Criteria
Calculator "Read Me"
tab
EPA-HQ-OW-
2017-0260-0028
(Joshua D. Schimmel,
Executive Director,
Springfield Water and
Sewer Commission
(SWSCJJ
The SWSC recommends including guidance on how the criteria
should be applied to setting a discharge limit in a NPDESpermit.
Ever-changing water quality values of DOC, hardness, and pH
will result in an ever-changing allowable limit. How will samples
be used to determine permit limits?
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). The
implementation documents that the EPA is developing are
intended to provide assistance to states and authorized tribes
that adopt into the water quality standards a criterion based
on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
No edits.
EPA-HQ-OW-
2017-0260-0028
(Joshua D. Schimmel,
Executive Director,
Springfield Water and
Sewer Commission
(SWSCJJ
If the new criteria is approved they should be included in
individual permits issued under the NPDES program for water
treatment facilities.
173
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
available for public comment.
EPA-HQ-OW-
2017-0260-0030
(Nelson Brooke,
Riverkeeper et ah,
Black Warrior
Riverkeeper)
In the absence of site-specific, historic data for pH, DOC, and
hardness at sites of interest, which is a particular problem here in
Alabama where there is dearth of reliable water quality data,
regulators, consultants, or non-profits aiming to enforce water
quality criteria will be forced to sample and analyze each of the
four necessary variables just to get a measurement for the site-
specific criteria, rather than simply comparing the analysis for
aluminum to the predetermined water quality criteria under the
current recommendation. The increase in necessary analysis will
result in a significant increase in the time spent collecting samples
in addition to drastically increasing the cost of analysis for
regulators and other interested parties. While this increase in cost
may not be a significant problem for many well-funded federal or
state agencies, it could place an undue burden on organizations
with limited budgets, such as ours, and some state agencies, like
the Alabama Department of Environmental Management, that are
woefully underfunded.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). As noted in the
criteria document, the EPA decided to use an empirical MLR
approach in the aluminum criteria update rather than a BLM
model due to: 1) the relative simplicity and transparency of
the model, 2) the relative similarity to the available BLM
model outputs, and 3) the decreased number of input data on
water chemistry needed to derive criteria at different sites.
The EPA is separately compiling an updated national
database of water chemistry conditions relevant to the MLR
model: hardness, pH and DOC, and will make that data
available to in the future to support states and stakeholders
needs for model input data, when their own data are not
available.
The implementation documents that the EPA is developing
are intended to provide assistance to states and authorized
tribes that adopt into the water quality standards a criterion
based on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0030
(Nelson Brooke,
Riverkeeper et ah,
Black Warrior
Riverkeeper)
Furthermore, there is an inherent standard of error in any
measurement of water quality parameter. While a single
measurement for aluminum would carry but a single standard of
error, requiring four separate measurements, each with its own
standard of error, to calculate a single, site-specific criteria, will
mean that the error inherent in each measurement is compounded
in the final criteria calculation. This means that the standard of
error in the site-specific criteria will be much greater than that of
a single measurement resulting in a criteria calculation that
carries a high degree of uncertainty.
The multiple measurements (Al, pH, total hardness and
DOC) are necessary to account for the bioavailability and
toxicity of aluminum, based on the current science. A single
criterion value for aluminum is no longer thought to represent
the best available science, and would result in over-protection
in some cases, and under-protection in others, depending on
the water chemistry at a location.
No edits.
174
-------
( oiiimeiil
Nil in her
(Oiliiiiii/iilion)
Public ( o in moil 1 on Topic IS: Kcu;ir(lini£ implcmcnl;ilion
issues
I'.I'A Response
Kc\ ision Locution in
20IX Aluminum
Crileriii Document
EPA-HQ-OW-
2017-0260-0038
(Jennifer Pederson,
Executive Director,
Massachusetts Water
Works Association et
al.)
We think it is important for EPA to define the "site" for sampling
the water quality parameters that are input into the model. We
understand that the samples for the water quality parameters
(hardness, TOC, DOC, pH) should be done in the receiving waters
and not from the discharge, but EPA should make that explicit in
the final document.
The 2018 aluminum criteria are Ambient Water Quality
Criteria, as is explicitly stated in the document's title. The
EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water
quality standards under CWA section 303(c). The
implementation documents that the EPA is developing are
intended to provide assistance to states and authorized tribes
that adopt into the water quality standards a criterion based
on or similar to the EPA's recommended criterion. The
implementation documents are also intended to provide
assistance to other stakeholders and the public. The EPA
recognizes that there are several aspects of the recommended
criterion that will benefit from technical support documents
to enhance implementation of state and tribal criteria and is
planning to develop such documents and make them
available for public comment.
No edits.
EPA-HQ-OW-
2017-0260-0036
(Barry N. Burnell,
Water Quality Division
Administrator, State of
Idaho Department of
Environmental Quality
(DEQ))
DEQ also requests that EPA develop guidance on implementation
of the MLR to assist states that may be considering adoption of the
revised aluminum criteria. Although time and site-specific criteria
are not new, recent adoptions of these types of criteria have
highlighted several implementations issues, such as identifying
critical conditions, determining minimum data requirements, how
to reconcile variable criteria for permitting purposes, and site
delineation. EPA should provide clear guidance and options for
states seeking to adopt these criteria, rather than requiring states
to solve these issues as a prerequisite to adoption of the EPA
recommended criteria.
175
-------
TOPIC 19: Comments regarding implementation issues with measuring aluminum
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public C oiiimoil 1 on l opic 1'): Rciiiirdinii iniplcnicn(;i(ion
issues willi iiiciisurinii iiluiniiiiiiii
I'.I'A Response
Kc\isiou Location in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0013
(Ricardo Cantu,
President, OspreyOwl
Environmental, LLC)
I have been sampling rivers in New Hampshire and Massachusetts
(using clean sampling techniques) since 2009. The majority of the
sampling focused on aluminum, with minor sampling for copper
and lead. My initial sampling efforts focused on adoption and use
of Method 1669.
The EPA had proposed in the 2007 NPDES draft for the City of
Manchester an aluminum limit of 87 ug/l indicating the Merrimack
River was impaired for aluminum. A one-year study was
undertaken by the City of Manchester using the "Clean Sampling"
techniques. Even with the use of these techniques, samples had to
be occasionally corrected for ambient environmental conditions. A
final report was issued with the data collected over a year's time.
The findings indicated that the Merrimack River, in the location of
Manchester, was not impaired for aluminum and the proposed
limit was dropped in the subsequent NPDES issuance.
The next permit draft issued in 2012 for Manchester included
copper and lead as permit parameters. Another "Clean Sampling"
event was undertaken and subsequently lead was dropped from the
permit with Copper being raised over eight fold from the proposed
WQ limit.
Clean sampling has made a big difference in findings for all metals
parameters measured in collected samples. The draft Aluminum
Criteria has the following information regarding ambient
aluminum during sampling events. "Average total aluminum
concentrations in the atmosphere were observed to range from
0.005 to 0.18 ng/m3 (Hoffman et al. 1969; Potzl 1970; Sorenson et
al. 1974). These concentrations are dependent on the location,
weather conditions and industrial activity in the area with most of
the airborne aluminum present in the form of small suspended
particles of soil (dust) (ATSDR 2008).
Average total aluminum precipitation concentrations reported in
the rural area (107.2 jug/L, range of28.8-222.7 fig/L) were higher
than observed in the urban area (83.9 fig/L, range 35.8-125.4
fig/L). Samples of wet deposition collected in semi-rural Dexter,
Michigan had an average mean total aluminum concentration of
57 jug/L (Landis andKeeler 1997).
Thank you for your comment.
No edits.
176
-------
(o 111 I1H-II1
Nil in her
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Remind inii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
permittees even though the exceedances are well within the range
of natural aluminum concentrations and are very unlikely to have
an adverse effect on the waterways.
In California, the EPA recommended aluminum criteria have also
been used to evaluate or translate narrative toxicity objectives in
the Water Quality Plans into numeric effluent limits in some
NPDESpermits. [Example: California Regional Water Quality
Control Board, Central Valley Region. 2010. Waste Discharge
Requirements for the Sacramento Regional County Sanitation
District, Sacramento Regional Wastewater Treatment Plant, Elk
Grove (NPDES No. NPDES No. CA0077682). When reissued in
2016 as Order R-5-2016-0020, the aluminum effluent limits were
removed based on the reasonable potential analysis (RPA), which
was partially based on the acute criterion of 750 fig/L.] This has
not yet occurred for stormwater permits, however, numeric effluent
requirements are increasingly being used in stormwater permits in
conformance with EPA guidance that emphasizes measurable
permit requirements and, where feasible, numeric effluent
limitations. [EPA Memorandum to Water Division Directors.
2014. Revisions to the November 22, 2002 Memorandum
"Establishing Total Maximum Daily Load (TMDL) Wasteload
Allocations (WLAs) for Storm Water Sources and NPDES Permit
Requirements Based on Those WLAs". November 26.]
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
An additional concern is that stormwater permittees in the future
may need to implement Wasteload Allocations (WLA) for
aluminum as specified in TMDLs. Currently, nine waterways in
California are included on the 303(d) list for impairment by
aluminum. In TMDLs, the WLAs are often established as numeric
effluent limits at the EPA criterion value. Because of the
ubiquitous nature of aluminum, these WLAs may be very difficult
to achieve.
EPA-HQ-OW-
2017-0260-0027
(Jill Bicknell, Chair,
California Stormwater
Quality Association
(CASQA))
New Mexico, with approval from EPA, has implemented aluminum
standards that use filtration to remove or minimize the "mineral
phases"present in the sample. This approach will reduce the
natural particulates and other soil residue that normally carry a
significant portion of the aluminum but that do not present risk to
aquatic organisms. This is appropriate because aluminum toxicity
is associated with dissolved aluminum with the exception of low
pH waterways.
The New Mexico hardness-adjusted acute criteria are the same as
180
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
containing these non-biologically available forms of aluminum will
overestimate the toxicity of aluminum in these waters. As with the
existing aluminum criterion, the proposed criterion should include
a warning that water effect ratios may be appropriate for assessing
the actual toxicity of aluminum in these waters.
EPA-HQ-OW-
2017-0260-0031
(Robert P.
Baumgartner,
Regulatory Affairs
Department Assistant
Director, Clean Water
Services (Districtj)
Aluminum concentrations in natural surface waters can be
strongly affected by the minerals in the river bed, natural upland
runoff, dust from the atmosphere and natural groundwater inflows.
As stated in the Request for Scientific Views, aluminum is the third
most abundant element and the most abundant metal in the Earth's
crust. Although abundant in the crust, aluminum is rarely present
in a dissolved form in surface waters that are not acidic. In the
crust, aluminum exists predominantly in aluminosilicate minerals,
including feldspar and mica, which weather into a wide variety of
clay minerals. These clay minerals are abundant in soils and in the
aquatic environment in particles on the order of a few micrometers
(fim) to small fractions of a jxm in size. In all these mineral forms,
aluminum is strongly bound in the mineral's crystal structure. In
these solid phases, aluminum is not as biologically available (if at
all) as it is in the dissolved, ionic forms or the chemically
precipitated forms that are used in assessing aluminum toxicity in
the laboratory. It is invalid to apply criteria based on dissolved,
biologically available aluminum to waters where the aluminum is
often present in an entirely different form.
The small size and ubiquity of clay particles in the environment
complicate the measurement of the bioavailable aluminum present
in surface waters. Conventional sample filtration (using a 1-jum or
0.45-jum filter) may reduce but will not eliminate suspended clays
present in colloidal form which includes most clay particles.
Performing a "total" aluminum analysis by acid digestion of a
water sample containing suspended clays, even if filtered using
conventional methods, will overestimate the amount of aluminum
that is actually in solution. Using conventional filtration to define
"dissolved aluminum " as defined in the EPA rule or using total
aluminum analyses of surface waters will thus lead to listing of
water bodies as water quality limited for aluminum, when in fact
the bulk of aluminum in many streams may not be in a toxic form.
Listing streams as water quality limited imposes significant costs
on dischargers to those streams. Such costs include effluent
treatment and product substitution (e.g. discontinuing the use of
alum that aids in nutrient removal and replacing with other
184
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
The following are our concerns for the Draft Criteria.
The EPA has proposed the criteria with the total recoverable form
of aluminum. This analytic method measures both toxic and
nontoxic forms of aluminum in the sample because it uses an acid
digestion that will incorporate the nontoxic aluminum contained in
the sediment particles. Illinois is dominated by agriculture and as
a result, many streams receive runoff during precipitation events
that is high in suspended sediment. Measuring the aluminum that
is bound in these clays and silts will overestimate the toxicity of the
water. A moderate amount of suspended sediment in a sample can
cause a sienificant increase in aluminum above the criteria. EPA
should use an alternative form of aluminum in the criteria. In
determining what form of aluminum is more appropriate, EPA
should consider the level of difficulty associated with its sampling
and analysis. For example, use of dissolved method is already
widely accepted by laboratories and industry.
EPA-HQ-OW-
2017-0260-0032
(Phillip M. Gonet,
President, Illinois Coal
Association (ICA))
EPA should also include some discussion of how the analytic
method can be implemented into NPDESpermits. Permittees
should not be required to translate the aluminum to total form for
incorporation into the NPDES permit. For aluminum, the
requirement for translator studies should be made consistent with
current scientific views that the bioavailable form of aluminum is
the primary concern, and not the form associated with the
suspended sediment.
EPA-HQ-OW-
2017-0260-0034
(James Boswell, Senior
Manager,
Environmental,
Peabocfy Energy)
The EPA draft criteria are based on the total form of aluminum.
This will present issues with samples containing high suspended
sediment loads, which is the majority of samples collected during
or following precipitation or snowmelt events. EPA indicates that
use of the dissolved fraction alone is likely to underestimate the
potential for toxic effects of aluminum. However, use of the total
recoverable or total fraction will significantly overestimate the
bioavailable fraction of aluminum, particularly for samples that
contain elevated sediment concentrations. Many regions of the
U.S. show elevated sediment loads during intense precipitation or
snowmelt events. As noted in the draft criteria, many of these
sediments contain aluminum in their structure, which will be
measured using the total or total recoverable analytic methods.
For example, the arid west contains many highly erosive
environments and streams convey significant sediment loads
following storm events. Peabody mining operations have collected
The 1988 AWQC for aluminum were discussed as acid-
soluble concentrations and were subsequently expressed in
terms of total recoverable aluminum.
Dissolved, colloidal and precipitated forms of aluminum are
all bioavailable to aquatic organisms, which supports the
criteria as total aluminum. Thus, if aluminum criteria are
based on dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
The current EPA approved CWA Test Method (Methods
200.7 and 200.8) for aluminum in water and wastes by
inductively coupled plasma-atomic emission spectrometry
and inductively-coupled plasma-mass spectrometry measures
No edits.
186
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i 1
189
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
"lolal" aluminum for criteria seems unlikely to be a workable
approach for assessing toxicity of aluminum to aquatic organisms
in natural waters.
EPA-HQ-OW-
2017-0260-0042
(Bruce A. Stevens,
President, Indiana
Coal Council, Inc.
(ICC))
The EPA has proposed the criteria as the total recoverable form of
aluminum. Use of a total recoverable analysis measures both toxic
and nontoxic forms of aluminum in the sample and is an overly-
conservative method of applying the criteria. As stated in the Draft
Criteria document, aluminum is an abundant element in the earth's
crust. This includes the clays and silts that are easily entrained in
runoff to streams and rivers following precipitation events. This is
particularly true for Indiana, which is dominated by agricultural
land use. Use of a total recoverable method will inherently
measure the aluminum that is bound in these clays and silts,
overestimating the actual toxicity of the water. This will subject
NPDES permittees to potential for false-positives in their effluents,
cause inaccurate 303(d) listings of streams, and lead to a waste of
both state and industry resources. EPA should use an alternative
analytic method that onlv measures the aluminum that is
bioloeicallv available to aauatic life. It is ICC's preference that
EPA instead use the dissolved form because of its already wide
acceptance and use by laboratories, industries, and the scientific
community.
EPA should also include some discussion of how this alternative
analytic method can be implemented into NPDES permits.
Requiring permittees and states to translate all parameters to a
total form for inclusion in an NPDES permit is the result of a dated
regulatory requirement. These translator studies are often an
unnecessary waste of resources on all parties involved. This needs
to be brought in line with current scientific views on protection of
aquatic life. EPA should include a reasonable alternative that
allows application of the revised analytic method directly into
NPDES permits.
The EPA presents the Draft Criteria as a more scientifically valid
approach and significant improvement over the current 1988
criteria. The EPA also argues that this is a significant relaxation of
the 1988 criteria, which was set at 87 and 750 jug/L. However, the
fact is that many states do not actively implement the 1988 criteria.
One of the primary reasons for this is that the 1988 criteria was
based on total aluminum. Again, this overestimates the actual
toxicity and is impossible to implement at a statewide level because
192
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
of the permitting and assessment issues it creates. In order for EPA
to tout any revised criteria as an "improvement", EPA must first
address these primary issues.
EPA-HQ-OW-
2017-0260-0045
(Lee Lemke, Executive
Vice President,
Georgia Mining
Association (GMA))
The Georgia Mining Association (GMA) appreciates the
opportunity to provide comment on the Draft Criteria published by
the EPA in July 2017. We have carefully reviewed the document
and agree that recent research findings on aluminum toxicity and
chemistry warrant an update to EPA's 1988 recommended criteria
for aluminum. However, we have significant concerns about the
assumptions made and conclusions reached in the Draft Criteria.
As a result, we urge EPA to revise the proposed Draft Criteria to
recognize and accommodate the natural conditions of Georgia and
the southeast, which include high levels of inert, biologically
unavailable aluminum. In the Draft Criteria, EPA explicitly states:
"... natural water samples may also contain other species of
aluminum that are not biologically available (i.e., suspended
particles, clays and aluminosilicate minerals) (Wilson 2012;
Santore et al. 2017). This creates uncertainty because the total
aluminum concentrations measured in natural waters may
overestimate the potential risks of toxicity to aquatic organisms"
(p. 69, Draft Criteria)
In addition, the biological literature establishes that kaolinite is
inert and nontoxic to freshwater biota (e.g., Goldes et al. 1988;
Tao et al. 1999; Tao et al. 2002; Beck et al. 2015). However, the
Draft Criteria, which focuses on total aluminum, makes no
adjustment for forms of aluminum that are well understood to be
biologically unavailable and nontoxic. Accordingly, it is critical
that EPA address the significant overestimation of the bioavailable
aluminum resulting from its suggested analytical methods in the
Draft Criteria.
In these comments, we provide an overview of GMA and the
mining industry in Georgia, describe the aluminosilicate mineral
kaolinite, provide specific comments regarding the Draft Criteria
with suggestions for revisions, and end with concluding remarks.
EPA-HQ-OW-
2017-0260-0045
(Lee Lemke, Executive
Vice President,
Georgia Mining
Association (GMA))
Given these particular properties of kaolinite and its importance in
Georgia and the southeastern United States, we respectfully submit
the following comments on the Draft Criteria:
1) The Draft Criteria requires measurement of total recoverable
193
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-11Q-OH-
2017-0260-0047
(Kathleen M. Roberts,
Executive Director,
North American Metals
Council (NAMC))
Test Methods Jor Aluminum in Water Ll'A notes in the Jruji
criteria that natural waters contain mineral particulate forms of
aluminum that are subject to measurement uncertainty when using
"total recoverable " measurements of aluminum. This is a critical
issue and further guidance from EPA is needed in managing this
uncertainty in interpreting the toxicity data and applying
aluminum criteria. The extrapolation of laboratory toxicity data
for aluminum to regulatory criteria implementation in natural
waters has long been problematic due to the complex chemistry of
precipitated and solid phases of aluminum. The methods for total
recoverable aluminum use a strong acid digestion resulting in an
overestimation of the potential risks of toxicity to aquatic
organisms due to the overly aggressive digestion procedure, which
captures the mineral phases of aluminum that are normally non-
bioavailable in the environment. This analytical procedure has
resulted in numerous waters across the U.S. being listed as
impaired for aluminum when in actuality the aluminum comes
from the solids and is non-toxic. NAMC is concerned that absent
further EPA explanation and guidance in this area, there will be
numerous false positive outcomes in the implementation of the
criteria generated because the total recoverable aluminum
concentrations will exceed the criteria, whereas the true
bioavailable concentration of aluminum would not exceed the
criteria. Below, we provide further details on the magnitude of the
issue of aluminum derived from suspended solids using a strong
acid digestion, as well as our recommendation on an analytical
approach to solve this issue. The multi-linear regression (MLR)
approach is a step in the right direction and will alleviate some of
the problems due to the fact that the criterion will go up for neutral
and alkaline waters. Slightly acidic natural waters, which are high
in suspended solids, however, will not be able to meet the new
proposed criteria.
EPA-HQ-OW-
2017-0260-0047
(Kathleen M. Roberts,
Executive Director,
North American Metals
Council (NAMC))
A review of data available in the United States Geological Survey
(USGS) - National Water Information System (NW1S) for total
aluminum, dissolved aluminum, and total suspended solids (TSS)
to evaluate the relationships across U.S. waters shows little or no
relationship between dissolved aluminum versus TSS. A strong
relationship exists, however, between total aluminum and TSS
based on 22,607 samples. (See Figure 1 below.) If one draws a line
across the Figure at 87 fig/L, it is clear that more than 85% of the
national surface waters would not be able to meet the current
chronic water quality standard (87 fig/L). A standard of400-500
197
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
test method would help inform the ultimate implementation of the
revised criteria and facilitate its implementation in the future.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
Our primary concern is that, if adopted in their current form, the
proposed criteria may cause some state authorities to wrongly
conclude that many waterbodies are "impaired" based on Total
Aluminum concentrations measured in samples of stormwater
runoff. This, in turn, will likely result in a large number of
inappropriate 303(d) listings, unnecessary TMDLs, and inaccurate
permit violations. The basis for our concern and our
recommendations for avoiding this unintended outcome are
described below.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The draft document acknowledges that aluminum is found in most
rocks, clays, soils, and sediment, and that these natural sources
are the dominant source for aluminum entering aquatic
ecosystems. [DAC @pg. 2] Specifically, the EPA cites
authoritative sources stating that, due to its abundance in the
earth's crust, soil concentrations of aluminum average
approximately 71,000 mg/kg. [DAC @pg. 6] This is critically
important because storm water runoff contains naturally high
concentrations of total suspended solids (TSS) such as clay soils
and entrained sediment.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The proposed criteria is expressed as Total Aluminum. [DAC @
pg. xii] Total Aluminum is measured using an unfiltered sample
and an acid digestion procedure that is intended to quantify at
least 95% of the aluminum, including any suspended sediment, in
the sample. [DAC @pg. 3 & 4] The District's long-term water
quality monitoring data indicates that the average TSS
concentration in stormwater runoff is approximately 300 mg/L and
is about ten times higher than normally seen in samples collected
during dry weather conditions. Consequently, it is reasonable to
expect that a random sample of stormwater runoff with 300 mg/L
of TSS could contain more than 20,000 ug/L of Total Aluminum.
Such concentrations are more than four times higher than the
maximum estimated acute criteria (4,300 ug/L) and ten times
higher than the maximum estimated chronic criteria (2,000 ug/L)
described in draft criteria document and would constitute an
"exceedance" were such criteria used to evaluate compliance with
water quality standards. [DAC@pg. K-5 (see Tables K-7 andK-
8)]
199
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
El 'A-11Q-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The draft document stales that aluminum can become sorted to
clay particles or complexed to DOC and later be converted to
reactive (toxic) form. [DAC@pg. 3] The EPA explains that
measuring only dissolved aluminum fails to consider such
conversions and, as a result, would likely underestimate the
potential for aquatic toxicity. [DAC @ pg. 22] Measuring Total
Aluminum will account for the "colloidal forms and hydroxide
precipitates of the metal that can dissolve under natural conditions
and become biologically available." [DAC @pg. 22] However, by
breaking the strong molecular bonds in the soil particle itself, the
laboratory-based acid digestion procedure will also significantly
overestimate the concentration of Total Aluminum that can become
bioavailable under the natural stream conditions (e .g. pH 5-9)
for which the proposed criteria were intended to apply.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
4. While toxic species of aluminum may sorb to soil particles, the
draft document explicitly acknowledges that particles themselves
contain svecies of aluminum that are not biolosicallv available.
[DAC@pg. 6] In natural soils, aluminum is molecularly bound
with silica, oxygen, and other minerals to form inert and insoluble
molecules. These mineral molecules have a very low Product
Solubility Coefficient (i.e. Ksp<0.01 mol/L) and cannot be
dissolved by freshwater in the range ofpH commonly found in the
natural environment. A strong acid (pH<3) is required to break
such bonds and convert any aluminum in the soil particle itself to a
more bioavailable and toxic form. The laboratory procedure used
to measure Total Aluminum does precisely that by using nitric acid
(pH~2) to "digest" the sample. This laboratory procedure not only
releases all of the aluminum "sorbed" to the soil particle, it also
dissolves all of the aluminum in the soil particle. As such, the
laboratory digestion procedure creates a condition that does not
occur naturally and greatly overestimates the potential for aquatic
toxicity by assuming all of the aluminum that is safely bound
within such particles "might" become bioavailable. Outside of
exposure to acid mine drainage, it is difficult to imagine how such
extreme pH conditions could occur naturally. This is particularly
true in the arid southwest where soils have more pH buffering
capacity than is typically seen in the eastern U.S. [DAC@pg. 9]
200
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
El 'A-11Q-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
All of the studies that Ilie EPA evaluated to develop the
recommended criteria relied solely on toxicity data from
laboratory studies using soluble aluminum salt compounds (e.g.,
aluminum chloride, aluminum nitrate, and aluminum sulfate).
[DAC @pg. 20] These man-made compounds are designed to
dissolve in water and convert the aluminum to a highly-
bioavailable form (see Table 1 below). This is reasonable as one
must start with a toxic form of aluminum in order to evaluate how
pH, hardness, and DOC affect that toxicity. However, the chemical
behavior of these soluble salts is not representative or predictive of
how insoluble aluminum-bearing minerals react when exposed to
water with a relatively neutral pH. By expressing the proposed
water quality criteria as Total Aluminum, the EPA has ignored this
important distinction and improperly assumed that the aluminum
found in common granitic soils is as likely to become toxic under
natural conditions as the aluminum in soluble salts. The External
Peer Reviewers expressly warned the EPA against making or
applying such simplistic assumptions. [EPA Response to External
Peer Review Comments. July, 2017 (see Reviewer #4@pgs. 41 &
43).J
[Table 1]
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The draft document states that pH, hardness and DOC "are
thought to be the most influential for aluminum bioavailability and
can be used to explain the magnitude of differences in the observed
toxicity values. " [DAC@ pg. 21] In fact, it also acknowledges that
the specific mineral form of aluminum is another significant factor.
However, the EPA warns that: "although many factors might affect
the results of toxicity tests of aluminum to aquatic organisms,
water quality criteria can quantitatively take into account only
factors for which enough data are available to show that the factor
similarly affects the result of tests with a variety of species. "
[DAC@pg. 27] The problem with this approach is that
laboratories deliberately choose to use only soluble salt
compounds to evaluate the potential toxicity of aluminum.
Commonly occurring insoluble mineral compounds with high
concentrations of Total Aluminum (e.g., kaolinite, feldspar,
gibbsite, bauxite, etc.) are never tested because there is no
expectation that these compounds can or will become toxic in
water. Nevertheless, the EPA relies on the absence of such studies
to support the proposition that there isn't enough data to make
appropriate adjustments to account for aluminum solubility in the
201
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
proposed aluminum criteria. In this instance, the absence of data is
conclusive evidence of the existing scientific consensus that some
insoluble forms of Total Aluminum are not and cannot cause
toxicity under natural conditions.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The EPA describes its decision to propose a Total Aluminum
criteria as "conservative" because it includes all forms of
aluminum. [DAC @pg. 21] While such an approach might be
considered appropriate when dealing with soluble aluminum salts
or colloidal aluminum sorbed to insoluble mineral particles, it
grossly misrepresents the ecological risk posed by the particles
themselves. Measuring the total recoverable aluminum
concentration present when exposing test organisms to soluble
salts is not a reasonable proxy for estimating the potential
bioavailability of insoluble mineral compounds. Deliberately
selecting an approach that severely over-estimates the potential for
toxicity is no more accurate or acceptable than one that under-
estimates it. The External Peer Reviewers warn that the "document
is written as if aluminum is like other metal contaminants and
aluminum is very different. " The External Peer Reviewers go on to
recommend that the EPA revise the document to consider
aluminum solubility (Ksp) and recognize solid speciation. [EPA
Response to External Peer Review Comments. July, 2017 (see
Reviewer #4@pg. 41)] However, the EPA ignored this
recommendation and made no substantive changes to the proposed
water quality criteria despite conceding that: "research on
analytical methods is on-going to address concerns with aluminum
bound to particulate matter (i.e. clay) from natural waters being
included in the total recoverable aluminum concentrations." [DAC
@ pg. 21] The lack of proper analytical methods to adequately
distinguish between bioavailable and non-bioavailable forms of
aluminum does not justify continued reliance on a false assumption
that all forms of aluminum share an equal potential to cause
aquatic toxicity.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
As presently written, it would be easy for readers who are not well
trained in the nuances of aluminum chemistry to misunderstand
exactly what the EPA means when it says that: "aluminum sorbed
to clay particles or complexed to DOC and later be converted to
reactive form " or that "colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and become
biologically available." [DAC @pg. 22] Therefore, the EPA
should explicitly state that while the colloidal aluminum and
hydroxide precipitates bound to clay particles may become
202
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
bioavailable, the aluminum that is molecularly-bound within such
particles cannot become toxic under natural conditions and the
final criteria is not intended to apply to natural aluminum-bearing
minerals with very low Ksp values.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The Total Aluminum criteria should not apply to natural mineral
soils until well-controlled laboratory experiments have been
conducted to confirm the presence of a valid dose-response
relationship for these insoluble forms of aluminum. [DAC@ pg. 21
(referencing Gensemer et al, 2017).] The EPA took great care to
avoid predicting the potential toxicity of aluminum outside the
range of hardness (<150 mg/L) and DOC (<5 mg/L) that had been
evaluated in well-designed laboratory experiments. For the same
reason, the EPA should avoid extrapolating outside the range of
aluminum solubility that has been tested and caution others
against doing so as well.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The final criteria should state, explicitly, that the standard acid-
digestion method will overestimate the amount of Total Aluminum
that is likely to become bioavailable or toxic if the sample also is
contaminated by the presence of natural soils and sediments (i.e.,
measurable TSS).
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The final criteria should describe other analytical procedures that
are presently available (such as the weak-acid digestion method)
to estimate the concentration of Total Aluminum that is sorbed to
clay particles without also including the concentration of
aluminum within the insoluble particles themselves.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The final criteria should describe and endorse the techniques some
states (e.g., New Mexico) use to minimize the risk of over-
estimating the concentration of bioavailable aluminum by
analyzing parallel samples that have been passed through a course
filter to reduce TSS associated with stormwater runoff.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The EPA should advise states that when high concentrations of
Total Aluminum appear to be closely associated with elevated TSS
levels in a given sample, it may be appropriate to employ the
EPA's methods for evaluating sediment-based toxicity to assess the
true potential for that aluminum to become bioavailable.
203
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
through a 0.45 fim membrane filler after the sample has been
acidified to a pH between 1.5 and 2.0 with nitric acid. " 1988
Criteria at 10-11. While dissolved aluminum is filtered, it is not
acidified before being filtered. And unlike total recoverable
aluminum measurement, acid-soluble aluminum measurement does
not require digestion. It is that digestion step of the approved Part
136 analytical method for total recoverable aluminum that
solubilizes the non-toxic forms of aluminum (e.g., aluminum bound
to clays).
IJWAG acknowledges the benefit of expressing the criteria as total
(i.e., total recoverable) aluminum since the aqueous chemistry of
this element is complex and there are no EPA-approved methods to
measure the various species of aluminum. Expressing the criteria
in the total form allows ease in terms of monitoring, reporting, and
calculating WQBELs, where these are necessary. However, there
must be a translator to address that not all of the total recoverable
aluminum is in fact toxic. Otherwise, the resulting criteria will be
unnecessarily over-protective with no measureable benefit for the
additional treatment cost. EPA concedes this in acknowledging the
need to develop a method that will "address concerns with
aluminum bound to particulate matter (i.e., clay) from natural
waters being included in the total recoverable aluminum
concentrations. " Draft Criteria at 21. That is exactly what
measuring the acid-soluble fraction as described in the 1988
Criteria is designed to do. Additionally, UWAG requests that EPA
express the criteria as total recoverable concentrations, not simply
total, though the Agency appears to use the terms synonymously in
the Draft Criteria.
EPA-HQ-OW-
2017-0260-0062
(John St. Clair,
Rosebud Mining
Company)
The Criteria for Aluminum stresses the dependence of the
bioavailability of aluminum to living organisms based on the
chemical properties of water. Dissolved aluminum is bioavailable
and toxic to aquatic life, yet the proposed recommended levels of
aluminum are based on total aluminum. In often cases, total
aluminum can also measure suspended clay sediment which is not
toxic to aquatic life as long as suspended solids limits are applied.
This can be seen in mine drainage from active underground coal
mines that typically contain suspended clay sediment associated
with the underclays of coal seams. While the water may contain
elevated aluminum levels, the aluminum is bound to other elements
of the clay particles and is not bioavailable to aquatic life.
Therefore, the chronic and acute criterion should be based on
206
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
impairment and needlessly scheduled for 1'MDL development.
[This, of course, assumes that US EPA plans to treat all States
equally and unilaterally impose the requirement to adopt the
revised national criteria for aluminum.]
EPA-HQ-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United-
States Steel
Corporation)
Expression of the aluminum criteria as acid-soluble aluminum and
not total recoverable aluminum. Species that are not acid-soluble
are not known to have toxic effects on aquatic life per EPA's past
practice in developing the criteria.
EPA-HQ-OW-
2017-0260-0067
(Patrick McDonnell,
Secretary,
Pennsylvania
Department of
Environmental
Protection (DEP))
DEP has concerns regarding the use of a "total" rather than
"dissolved" standard for aluminum and the resulting impairment
issues raised with spikes in total aluminum concentrations caused
by soil-laden samples that are often collected after storm events.
Additionally, we would like to comment that instead of using
"hardness" in its criteria, EPA consider using calcium
concentration, in the hopes of correlating more precisely with the
element of hardness which may be responsible for its protective
effect on the metal toxicity.
EPA-HQ-OW-
2017-0260-0067
(Patrick McDonnell,
Secretary,
Pennsylvania
Department of
Environmental
Protection (DEP))
Analyzing samples for non-dissolved aluminum requires collection
of unfiltered samples, which, depending on how recently
precipitation has occurred, may contain significantly varying
quantities of suspended soil. Such soil-laden samples are then
subjected to "digestion"per EPA method 200.7, which has been
shown to extract aluminum from clays (See the 2016 work of He
and Ziemkiewicz and the references cited therein). [Y. Thomas He
and Paul F. Ziemkiewicz, "Bias in Determining Aluminum
Concentrations: Comparison of Digestion Methods and
Implications on AlManagement, " Chemosphere 159 (September
2016): 570-76, doi: 10.1016/j. chemosphere.2016.06.052] Our
scientists are observing surges in total aluminum to values above
the EPA's impairment threshold after rain-related events where
large amounts of earth are stirred up into the water column.
However, such high flow events do not coincide with the adverse
effects to stream biology that would be expected with toxic metals
concentrations. This supports the theory that the sampling and
extraction methods result in the reporting of aluminum fractions
that are not readily bioavailable; over-representing the
bioavailable fraction of aluminum in the sample.
Considering the forgoing, if the EPA's "total aluminum" criteria
212
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiiciisuriiiii ;iIn ini nil in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
In fact, because aluminum is ubiquitous in Kansas soils, KDHE
utilizes "total recoverable aluminum " to characterize if a stream
sample is from a base flow or runoff condition. During runoff,
metals concentrations can be elevated, but tightly bound to the
sediment and not reactive biologically. In the absence of actual
measured flow data, KDHE has determined there is a high
probability that runoff was occurring during sampling if the
aluminum concentration 1 mg/L or greater.
EPA-HQ-OW-
2017-0260-0071
(Fredric P. Andes,
Coordinator, Federal
Water Quality
Coalition (FWQC))
USE OF TOTAL RECOVERABLE ALUMINUM FORM: The Draft
Criteria are expressed as total recoverable aluminum. That is a
substantial change from EPA's past practice, which was to use the
acid-soluble aluminum form. We believe that using total values
disregards what we know about aluminum toxicity, and results in
regulation of aluminum presence that has no effect on aquatic life.
Instead, EPA should express the criteria as acid-soluble aluminum.
The Agency should also allow States to use the dissolved form
instead, with site-specific dissolved-particulate studies to
determine the appropriate permit limits.
EPA-HQ-OW-
2017-0260-0072
(Paul Bedore, M.S.,
Senior Scientist,
Robertson-Bryan, Inc.
(RBI on behalf of Port
of Stockton, San
Joaquin County,
California)
We appreciate the opportunity to provide comments on the report
"Draft Aquatic Life Ambient Water Quality Criteria for Aluminum
2017" (hereinafter "Draft Report") and the draft aluminum
criteria provided therein (USEPA, 2017a). Our firm, Robertson-
Bryan, Inc. (RBI), has been instrumental in the development and
adoption of refined water quality objectives in the Central Valley
of California. RBI has developed technical reports with supporting
scientific literature and data to support the rule process of refining
water quality objectives, considering all beneficial uses of the
water body, and has developed site-specific and refined region-
wide objectives for temperature, pH, and turbidity, and site-
specific objectives for trihalomethane compounds, that have been
adopted by California's state and regional water boards and
approved by USEPA. Comments provided herein on the Draft
Report were prepared on behalf of the Port of Stockton (Port), a
Phase 1 Municipal Separate Storm Sewer System (MS4) located in
San Joaquin County, California. The Port's MS4 discharges into
the lower segment of the San Joaquin River, which drains a
watershed of approximately 15,600 square miles.
Comments provided herein address the need for the Draft Report
to evaluate and account for non-bioavailable aluminum species,
since such species are common in aquatic environments. This issue
214
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
is of particular relevance to the Port because the San Joaquin
River, and presumably many other waterways throughout the
nation, is seasonally affected by high wet season rainfall that
washes particulate aluminum species from watershed soils into the
river. Particulate aluminum mobilized from terrestrial sources is
largely composed of recalcitrant forms of aluminum that are not
bioavailable. If adopted as is, monitoring for compliance with the
draft aluminum criteria using the total recoverable aluminum
methodology will measure all forms of aluminum in a sample -
bioavailable and non-bioavailable, dissolved and particulate - and
thus will provide an indication that aluminum levels in the San
Joaquin River seasonally exceed the draft aluminum criteria. Yet
the aluminum species responsible for the exceedance (soil- and
geologically derived aluminum particulates) are not bioavailable,
and do not warrant considering aquatic life beneficial uses of the
San Joaquin River as impaired with regards to aluminum.
Accounting for the prevalence, yet lack of bioavailability, of
particulate aluminum species (largely aluminosilicate minerals),
will avoid unintended regulatory consequences, such as
unnecessarily listing waterways on the federal Clean Water Act
Section 303(d) list that requires the development of a control
program to address the designated impairment.
EPA-HQ-OW-
2017-0260-0072
(Paul Bedore, M.S.,
Senior Scientist,
Robertson-Bryan, Inc.
(RBI on behalf of Port
of Stockton, San
Joaquin County,
California)
Comment 1. The particulate fraction of natural water samples
overwhelminslv consists of aluminum-bear ins silicate minerals
(i.e.. aluminosilicates). and sreater discussion and evaluation of
this form of aluminum in the Draft Report is warranted due to their
pervasiveness.
Section 2.6.2 of the Draft Report states that the appropriate
analytical technique for assessing compliance with the draft
aluminum criteria is total recoverable aluminum. This conclusion
was made on the basis that the toxicity of a sample would likely be
underestimated were dissolved aluminum used to assess
compliance, since colloidal forms and hydroxide precipitates of
aluminum can dissolve under certain conditions and become
bioavailable. Measurement of total recoverable aluminum will
quantify both particulate and dissolved aluminum species in a
water sample, including mineral forms of aluminum such as
aluminum oxide/hydroxides and aluminosilicates. The Draft
Report indicates in Section 2.6.2 that "Applying the aluminum
criteria to total recoverable aluminum may be considered
conservative because it includes monomeric (both organic and
215
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
El 'A-11Q-OW-
2017-0260-0072
(Paul Bedore, M.S.,
Senior Scientist,
Robertson-Bryan, Inc.
(RBI on behalf of Port
of Stockton, San
Joaquin County,
California)
Comment 2: The draft aluminum criteria or compliance testing
should account for the fact that aluminosilicates are larselv not
bioavailable relative to aluminum-containins chemical species that
currently form the basis of the draft aluminum criteria.
Aluminum in aluminosilicates is not directly bioavailable, meaning
the most abundant aluminum-containing particulate species
occurring in natural aquatic environments is not the
toxico logically relevant form of aluminum. There are numerous
species of aluminosilicates that can be present in the environment.
USEPA's ECOTOX database contains acute toxicity test results for
two aluminosilicates (bentonite and kaolinite) and two aquatic
organisms (Rainbow Trout (Oncorhynchus mykiss) and Daphnia
pulex), which help illustrate this point (USEPA, 2017b). Survival
LC50s for these species were available for the two
aluminosilicates on a dry weight basis (i.e., jug aluminosilicate per
liter), and the aluminum content of the respective aluminosilicates
was used to convert the LC50s to an "as aluminum " basis (i.e., jug
aluminum per liter). As shown in Table 1, the LC50s (in jug
aluminum per liter) for the two aluminosilicates are > 200,000
fig/L.
Aluminum toxicity data utilized by USEPA to update the draft
aluminum criteria were based on toxicity studies with aluminum
salts that are readily dissolvable in water, and such aluminum
species are of significantly greater bioavailability than
aluminosilicates. USEPA accompanied the Draft Report with the
MS Excel file "Aluminum Criteria Calculator vl .0 Macro"
(USEPA, 2017c), from which the Genus Mean Acute Value
(GMAV) for Oncorhynchus spp. and Daphnia spp. could be
calculated for the aluminum toxicity dataset used to develop the
draft aluminum criteria (i.e., studies using aluminum salts). The
GMAVs for Oncorhynchus spp. and Daphnia spp. for aluminum
salts are compared to the LC50s for Rainbow Trout and D. pulex
with bentonite and kaolinite, respectively, in Table 1. The Rainbow
Trout LC50 for bentonite (200,000 fig/L) is 41 times higher than
the GMAVfor Oncorhynchus spp. with aluminum salts (4,860
fig/L). Further, the D. pulex LC50 for kaolinite (>235,000 fig/L) is
>66 times higher than the GMAVfor Daphnia spp. with aluminum
salts (3,519 fig/L). Toxicity testing data for the aluminosilicates
bentonite and kaolinite illustrate the fact that aluminosilicates
represent a form of aluminum in water of significantly different
217
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic IV: Rciiiirdinii implcmcnl;ilion
issues willi iiic;isii riMLt ;iIn in i mi in
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
chemical form, reactivity, and bioavailability relative to the
aluminum species that form the basis of the draft aluminum
criteria - aluminum salts and the sparingly soluble aluminum
hydroxide particulates or colloids with which dissolved aluminum
species are in equilibrium.
Since aluminosilicates are one of the most environmentally
prevalent aluminum species in natural waterways, their
bioavailability is significantly lower than bioavailability of
aluminum species used to revise the draft aluminum criteria, and
they are measured in the total recoverable aluminum analysis, it is
appropriate to account for their presence and bioavailability when
setting federal water quality criteria. Various means to account for
the non-bioavailable fraction should be considered, and the most
appropriate approach should be incorporated into the Draft
Report and/or draft aluminum criteria. Means to account for the
bioavailability of aluminum in a water sample could include, but
are not limited, to the following.
Since by far, the most prevalent forms ofparticulate
aluminum in the environment are aluminosilicates
(Filella, 2007), and not sparingly soluble aluminum
particulates that may be present in lab-based experiments
using aluminum salts, the Draft Report could allow
compliance monitoring using the dissolved aluminum
fraction of a water sample.
Extensive research has gone into developing
methodologies for estimating bioavailable aluminum in
natural water samples (see Berger et al. 2008 and
citations therein), and the literature, as well as experts in
aluminum minerology and chemistry, could be consulted
in the development of a methodology for estimating the
bioavailable aluminum fraction of a water sample for the
purposes of assessing compliance with the draft
aluminum criteria.
EPA-HQ-OW-
2017-0260-0073
(Curt Wells, Director
of Regulatory Affairs,
The Aluminum
Association)
Test methods for aluminum in water
The extrapolation of laboratory toxicity data for aluminum to
regulatory criteria implementation in natural waters has long been
problematic due to the complex chemistry of precipitated and
solid-phases of aluminum. EPA notes in the draft criteria that
natural waters contain mineral particulate forms of aluminum that
are subject to measurement uncertainty when using 'total
recoverable' measurements of aluminum. This is a critical issue.
218
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public C 0111 moil I on lopic IV: Rciiiirdinii i in p lomoii 1 ;i 1 i
-------
TOPIC 20: Comments regarding EPA policies
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Com nun 1 on Topic 20: Rciiiirdinii EPA policies
EPA Response
Ke\ision l.ociilion in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0064
(Scott G. Mandirola,
Director, West
Virginia Department of
Environmental
Protection (WVDEP))
West Virginia is also concerned that EPA is adopting yet another
aquatic life criterion while its 1985 Guidelines for Deriving Water
Quality Criteria for the Protection of Aquatic Life and Their Uses
is outdated and not based on the latest data evaluation methods
(USEPA Guidelines 1985) [USEPA. 1985. Guidelines for deriving
numerical national water quality criteria for the protection of
aquatic organisms and their uses. United States Environmental
Protection Agency, Washington, D.C. NTISNo. PB85-227049. 98
pages.
http://water.epa.gOv/scitech/swguidance/standards/criteria/aqlife/u
pload/85guidelines.pdf] As EPA states on its website, the 1985
Guidelines are in need of revision. In fact, EPA states:
"the existing Guidelines for Deriving Water Quality Criteria for
the Protection of Aquatic Life and Their Uses have not been
updated since 1985. Although based on science of that time, the
past 30 years have witnessed substantial scientific advancement in
aquatic toxicology, aquatic biology, fate, transport, and effects
modeling, and ecological risk assessment. Such advancements,
coupled with increasing complexity of water quality impairment
issues requires criteria derivation approaches beyond the existing
Guidelines methods" [https://www.epa.gov/wqc/aguatic-life-
criteria-and-methods-toxics#guide]
Indeed, if the 1985 Guidelines are still considered valid, then EPA
should have used them to develop this Draft Aluminum Criteria; if
they are not, new guidelines should be developed before any new
criteria are recommended. In fact, EPA developed the current
Draft Aluminum Criteria outside the boundaries established in the
1985 Guidelines, including the use of a multiple linear regression
model, the use of which was never addressed in the 1985
Guidelines.
The 1985 Guidelines remain valid, and were followed in the
development of the aluminum criteria. The 1985 Guidelines
document contains a "best available science" clause that
allows the EPA to pursue different avenues for criteria
derivation, if they are scientifically defensible. EPA plans to
update the 1985 Guidelines including robust public
engagement and peer review.
The 1985 Guidelines did specifically discuss the concept of
toxicity data normalization and criteria development based on
water chemistry. In Section VII.B on page 22 of the 1985
Guidelines states: "When enough data are available to show
that chronic toxicity to at least one species is related to a
water quality characteristic, the relationship should be taken
into account.. .If two more factors affect toxicity, multiple
regression analysis should be used." Additionally, 1985
Guidelines Section VII. C states "Because the best
documented relationship is that between hardness and acute
toxicity of metals in fresh water and a log-log relationship fits
these data, geometric means and natural logarithms of both
toxicity and water quality are used. For relationships based on
other water quality characteristics, such as pH, temperature,
or salinity, no transformation or a different transformation
might fit the data better..."
The MLR approach to normalizing aluminum toxicity data
was published twice in peer-reviewed journals by
internationally-recognized experts in the field of metal
bioavailability and toxicity (DeForest et al 2018 a,b).
The final 2018 aluminum criteria document has also
separately undergone independent, external expert peer
review and represents the best available science.
The MLR models underlying the criteria were all also
subjected to independent external peer reviewed, with
positive feedback.
The 2018 final aluminum criteria represent the best available
science using the most current bioavailability and toxicity
information on aluminum. The final aluminum criteria take
into consideration the impact of water chemistry, including
No edits.
223
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 20: Rciiiudinii l.l'A policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Criloriii Document
all key factors (pH. DOC and hardness) on toxicity.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
The 1985 Guidelines
More than thirty years has passed since US EPA prepared the
1985 Guidelines for Deriving Numerical National Water Quality
Criteria for the Protection of Aquatic Organisms and Their Uses
(the 1985 Guidelines). In this time, US EPA has made no progress
in developing criteria that are representative of the actual
chemistry in surface waters. Scientists are far too busy conducting
worst-case scenario studies in laboratories rather than evaluating
what is truly occurring in surface waters. Regulators appear
immune to the social, economic, and environmental cost of the
over-regulation of naturally occurring substances and the
subsequent overtreatment of surface waters to comply with
imaginary numbers developed in the laboratory.
When West Virginia adopted its hardness-based criteria, US EPA
reviewed the proposed revision for compliance with the 1985
Guidelines. US EPA does not hold itself to these same standards.
The Draft Aluminum Criteria make use of a newly published
multiple linear regression (MRL) model. The approach used to
develop the MLR is not addressed in the 1985 Guidelines.
Inadequate time has been provided to review the validity of this
complex approach. Whereas the MLR may be sufficient for
publication in a scientific journal, it has not been established as a
regulatory tool. If US EPA seeks deference to this approach from
States, then the 1985 Guidelines should be revised to provide the
basic parameters for preparation of a satisfactory MLR so that
stakeholders can evaluate its validity for criteria development.
States and regulated entities are being forced to accept criteria
developed by mechanisms for which no technical boundaries have
been established. Specific comments on the MLR are addressed in
a separate section herein.
Likewise, US EPA disregards the requirement for eight difference
taxonomic groups for development of the chronic criterion. The
1985 Guidelines set forth minimum data requirements for
development of aquatic life criteria. The acute dataset is typically
more robust than the available chronic data, and therefore the
criteria are often prepared by utilizing an acute to chronic ratio
for calculation of the chronic criterion. Only seven taxonomic
groups are represented in the chronic database for the Draft
The 1985 Guidelines remain valid and were followed in the
development of the aluminum criteria. Further, the final
aluminum criteria take into consideration the impact of water
chemistry on toxicity.
The 1985 Guidelines did specifically discuss the concept of
toxicity data normalization and criteria development based on
water chemistry. In Section VII.B on page 22 of the 1985
Guidelines states: "When enough data are available to show
that chronic toxicity to at least one species is related to a
water quality characteristic, the relationship should be taken
into account.. .If two more factors affect toxicity, multiple
regression analysis should be used." Additionally, 1985
Guidelines Section VII. C states "Because the best
documented relationship is that between hardness and acute
toxicity of metals in fresh water and a log-log relationship fits
these data, geometric means and natural logarithms of both
toxicity and water quality are used. For relationships based on
other water quality characteristics, such as pH, temperature,
or salinity, no transformation or a different transformation
might fit the data better..."
The MLR approach to normalizing aluminum toxicity data
was published twice in peer-reviewed journals by
internationally-recognized experts in the field of metal
bioavailability and toxicity (DeForest et al 2018 a,b). The
final 2018 aluminum criteria document has also separately
undergone independent, external expert peer review and
represents the best available science.
The EPA disagrees that there has been no progress in
developing criteria that are representative of the actual
chemistry in surface waters. The final 2018 criteria are able
to address a substantial percentage of the waters found in the
U.S. Very low pH waters (less than pH 5) and high pH waters
(pH greater than 10.5) are examples of waters that are not
directly represented because the EPA determined not to
extrapolate to those pHs.
The EPA disagrees with the comment about the limited size
No edits.
224
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I on Topic 20: Rciiiirdinii I.PA policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
Aluminum Criteria. Instead of utilizing the robust acute database
and the acute to chronic ratio, EPA proceeded with the
development of the chronic criterion directly from the incomplete
chronic database.
US EPA dismisses the limited size of the chronic database as
meaningless, using a tree frog study that does not qualify for
inclusion to round out the required eight taxonomic groups. The
chronic database relies on only twelve genus mean chronic values
(GMCVs), compared to eighteen GMAVs in the acute database.
The robustness of the database affects the calculated final chronic
value (FCV). The following are the criteria from the US EPA
database, assuming only an increase to N (the total number of
GMCVs):
[Table 2]
Therefore, EPA's use of a limited chronic database directly affects
the Draft Aluminum Criteria.
If US EPA believes the 1985 Guidelines are arcane, then they
require revision. This should be done prior to development of new
criteria, not afterward (or worse, never). We are years past the
development of the copper biotic ligand model, and the 1985
Guidelines have not been revised to address this novel approach.
US EPA cannot continue to change criteria development without
amending the applicable guidelines.
of the chronic database and the EPA determined that the data
fulfilled the guideline requirements. The CCC was calculated
using the eight family MDR approach as recommended by
the Guidelines. There is less uncertainty associated with this
approach than there is using acute to chronic rations to
estimate chronic data.
225
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I on Topic 20: Rciiiirdinii I.PA policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0012
(Nancy Sonafrank,
Program Manager,
Alaska Department of
Environmental
Conservation (ADEC'))
The Alaska Department of Environmental Conservation (ADEC)
has reviewed the U.S. Environmental Protection Agency's (EPA)
Draft Aquatic Life Ambient Water Quality Criteria (AWQC) for
Aluminum in Freshwater provided to the states for comment by
September 26, 2017. ADEC appreciates the opportunity to
comment on these draft recommendations.
Under Clean Water Act (CWA) section 303(c), EPA requires states
to regularly review and update CWA 304(a) criteria based on EPA
recommendations. While this is a good goal and has the potential
to help maintain strong state and national water quality standards
programs, it is premature to require states to adopt further
revisions to criteria until EPA acts on the tremendous backlog of
revised water quality standards already adopted by states and
provides updated implementation tools (e.g., variances for new
criteria may require a tool for determining substantial and
widespread economic and social impacts that may result from
implementation of the new criteria).
ADEC recognizes EPA's efforts to compile and review the data
quantifying the toxicity of aluminum to aquatic organisms while
assessing the basis for a criterion that will protect population
assemblages offish, amphibians, aquatic invertebrates and plants.
States are required to review their WQS on a triennial basis.
For parameters for which the EPA has issued new or revised
304(a) criteria recommendations, the WQS regulation at
131.20(a) requires that "if a State does not adopt new or
revised criteria for parameters for which the EPA has
published new or updated CWA section 304(a) criteria
recommendations, then the State shall provide an explanation
when it submits the results of its triennial review to the
Regional Administrator consistent with CWA section
303(c)(1) and the requirements of paragraph (c) of this
section."
The regulation does not, however, require states to adopt
revised criteria based on the EPA's latest recommendation.
No edits.
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
Perform More Rigorous Peer Review
KCA is concerned that EPA has adopted an overly literal
definition of "peer review " by having its Draft Criteria reviewed
only by employees of the agency. Any Draft Criteria should be
subject to more rigorous peer review that includes the views of
scientists who work outside the agency, and particularly those with
an understanding of the real world implications and
implementation difficulties raised by the Draft Criteria.
The commenter is incorrect. In additional to internal EPA
expert peer review, the aluminum aquatic life was reviewed
by independent external experts in the field of aquatic
toxicology, as is the case for all EPA aquatic life water
quality criteria.
First, the 2017 Draft criteria document was reviewed by five
independent external peer reviewers, and their comments and
the EPA's associated responses are publicly available. The
2017 underlying bioavailability modeling approach was also
independently, externally peer reviewed. These external
expert peer reviews of the 2017 draft criteria document can
be found at: htft>s://www.era.eov/wac/2017-draft-aauatic-
hfe-criteria-ahiminum-freshwater.
Following the 2017 public comment period and criteria
revisions, the 2018 criteria basis underwent 3 additional
external expert peer reviews. The two new toxicity studies
included in the 2018 MLR models were externally peer
No edits.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
US EPA's Peer Reviewers
US EPA's peer review of the Draft Aluminum Criteria follows the
most literal interpretation possible of that term: All seventeen
individuals listed as reviewers of the Draft Aluminum Criteria
are direct employees of US EPA. Peer review also reauires a
certain level of expertise in the topic, which in this case is the
bioavailability and toxicity of aluminum to aquatic life.
The peer reviewers are generally high-level officials within US
EPA. Direct experience in water quality criteria development does
not appear to be a prerequisite for peer review, nor does expertise
226
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I on Topic 20: Rciiiirdinii I.PA policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
in aluminum toxicity. To provide any real substantive impact on
the Draft Aluminum Criteria, a peer reviewer would be required to
study the underlying toxicity tests and whether they qualify for
inclusion in the Draft Aluminum Criteria. The peer reviewer would
then be required to verify the conversion of the data from those
toxicity studies to normalized hardness, pH, and dissolved organic
carbon (DOC) concentrations. The details of the linear regression
model would require verification. This level of review would to
ensure that US EPA has not made a fundamental technical error in
its approach. The work is both time-consuming and of critical
importance.
The peer reviewers were allowed roughly three weeks to answer
very specific questions regarding the content and approach
presented in the Draft Aluminum Criteria. Instead of providing a
detailed critique, the peer review comments (at least as presented
in the US EPA summary) are largely gratuitous, intra-agency
back-pats or meaningless general discussions without any review
or comment on US EPA's fundamental data decisions.
Most importantly, perhaps, the peer reviewers should have asked
whether the criteria make logical sense: Evaluating whether the
calculated criteria present numbers that are reasonable, given
what we know about water chemistry and overall concentrations of
aluminum in waters within the United States. Based on the peer
review summary (once again, compiled by US EPA), the
appearance is this largely did not occur. Notably, "Reviewer 4"
offered some excellent comments regarding the difference in
solubility and behavior among minerals and the kinetics of land
scale processes. A few comments mention aluminum speciation
and behavior and appear to question the counterintuitive results of
the linear regression model across certain pH and hardness
concentrations. These comments were obviously not adequately
considered, as reflected by the very publication of this draft federal
criterion.
As set forth in the following sections, US EPA appears
disinterested in developing aluminum criteria that can be
reasonably implemented by the States who will be forced to adopt
them once finalized. An obligatory, superficial review by US EPA
employees with no vested interest in the impacts of the Draft
Aluminum Criteria is a DIRECT INSULT to the many States who
review by external scientists who are experts in the fields of
aquatic toxicity and metal bioavailability. The 2018 MLR
models were also externally peer reviewed by experts in the
field of aquatic toxicity. These three additional external
expert peer reviews can be found at:
httos://www.era.eov/wac/aauatic-life-criteria-
aluminum#2018.
The names and affiliations of the external peer reviewers are
available in the external peer review reports posted on the
EPA webpage. These include peer reviewers of the 2017 draft
aluminum criteria, peer reviewers of the two additional
toxicity studies and peer reviewers of the 2018 MLR models.
Prior to, to external expert peer review, EPA conducted
internal peer review of the criteria document. The names of
the internal peer reviewers included in the criteria document.
227
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 20: Kc^iirdin^ l.l'A policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
have struggled with adopt ion and implementation of the Ambient
Aquatic Life Water Quality Criteria for Aluminum published by US
EPA in 1988 (the "1988 Criteria"). Many of these states, like West
Virginia, have labored for decades to develop environmentally
protective standards that reflect specific conditions only to have
them linger at US EPA awaiting federal approval (West Virginia's
current proposal has been pending review since October 2015).
For West Virginia and other states to now be confronted with a
clearly hurried federal proposal is nothing short of offensive.
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
Pursuant to the Notice of Availability published by the U.S.
Environmental Protection Agency ("EPA ") in the Federal Register
on July 28, 2017 at 82 FR 35,198, the Kentucky Coal Association
("KCA ") submits the following comments on EPA's Draft Updated
Aquatic Life Ambient Water Quality Criteria for Aluminum in
Freshwater (the "Draft Criteria").
KCA represents over 80% of Kentucky's coal production, and its
membership includes over 120 additional companies that support
the coal mining industry. As of 2016, Kentucky was the fourth-
largest coal producing state in America.
KCA's coal-producing members operate in two geographically
distinct coal basins: the Central Appalachian basin in the eastern
part of the state, and the Illinois Basin in the western part of
Kentucky. As such, KCA and its members are acutely aware of the
drawbacks of "one size fits all" approaches to environmental
regulation that fail to take into account regional differences in
geological, chemical, and hydrological conditions. The Draft
Criteria represent such a misguided attempt at national regulation
in an area where a state-by-state approach is more appropriate.
Accordingly, KCA requests that EPA withdraw the Draft Criteria
and leave decisions related to aluminum in freshwater to state
permitting authorities, who have more expertise with respect to
local conditions and the real-world impacts (or lack thereof of
aluminum in surface waters.
The aluminum criteria are recommendations for the states and
authorized tribes and are not water quality standards. The
EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water quality
standards under CWA section 303(c). States may also adopt
other scientifically defensible criteria that are scientifically
defensible and protect the designated use. EPA has not
identified aluminum as a priority pollutant, and therefore
states are not required to develop state water quality
standards for aluminum.
The 2018 final aluminum criteria specifically allow for states
to adjust the criteria to their local water chemistry conditions,
and are thus inherently not a "one size fits all approach."
No edits.
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
The Clean Water Act expressly provides that states should have the
primary role of establishing water quality standards and adopting
water quality criteria that apply within their borders. Aluminum in
particular is a parameter where the primary role of states should
be respected. In any document adopting a final recommended
criteria for aluminum, EPA must emphasize that the criteria are
merely recommended, and make clear that states are free to, and
228
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 20: Kc^iirdin^ l.l'A policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
indeed should be encouraged to, adopt stale-specific criteria - or
no criteria at all - where aluminum criteria are not necessary to
protect designated uses. Kentucky provides an important example.
As explained by Kentucky's regulators during a prior triennial
review, due to the naturally high aluminum concentrations in soils
throughout Kentucky, streams that exceed the existing national
recommended criteria nonetheless contain healthy and
reproducing biological communities. Indeed, "some of the
commonwealth's highest quality waterbodies often exceed the
criteria. " [Kentucky Energy and Environment Cabinet, Division of
Water, Statement of Consideration Relating to 401 KAR 10:031
(2012).] There is simply no basis for adoption of stringent criteria
- and the staggering compliance costs that accompany them -
where streams that exceed the criteria are some of the highest
quality in the state and support flourishing aquatic life.
The comments already filed to date by state regulatory authorities
demonstrate that a national approach to aluminum regulation is
simply inappropriate. States regulators from areas as diverse as
Alaska, Kentucky, Texas, Wyoming and others have all raised, in
various respects, questions or concerns about the appropriateness
of applying a national standard to state-specific waters. In light of
the current EPA's repeated statements that it will respect
federalism and the important role of its state partners, EPA must
make clear in any adopted criteria that states are free to deviate
from it to address local conditions.
EPA-HQ-OW-
2017-0260-0044
(Shelly Lemon, Chief,
Surface Water Quality
Bureau, New Mexico
Environment
Department)
The State of New Mexico has already adopted hardness-dependent
criteria for aluminum. The State of New Mexico implores the U.S.
EPA to engage and, at the very least, consult with States and
Tribes in future development of water quality criteria early in the
development process.
Thank you for your comment. EPA routinely engages with
states and tribes on criteria development through the
Association of Clean Water Agencies (ACWA).
No edits.
EPA-HQ-OW-
2017-0260-0049
(Stuart E. McKibbin,
Chief of Planning
Division, Riverside
County Flood Control
and Water
Conservation District)
The EPA should describe how the new 304(a) criteria for
aluminum should be applied in states where it has promulgated
federal water quality standards for other trace metals under the
National Toxics Rule or the California Toxics Rule.
The EPA's criterion provides recommendations for states and
authorized tribes to consider in their adoption of water quality
standards under CWA section 303(c). Other implementation
documents that the EPA is developing related to these
aluminum criteria are intended to provide assistance to states
and authorized tribes that adopt into the water quality
standards a criterion based on or similar to the EPA's
recommended criterion. The implementation documents are
also intended to provide assistance to other stakeholders and
the public. The EPA recognizes that there are several aspects
No edits.
229
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 20: Rciiiirdinii I.PA policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
or die recommended criterion llial will benefit 1'i'oni technical
support documents to enhance implementation of state and
tribal criteria and is planning to develop such documents and
make them available for public comment.
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
The Abundance of Aluminum In the Natural Environment
Despite recognizing that aluminum is one of the most abundant
metals in the Earth's crust, EPA fails to acknowledge the
implications of this finding. Simply put, if most of the earth's crust
contains aluminum, it is illogical to adopt regulatory policies
aimed at the near-total elimination of aluminum from surface
waters. As the Kentucky experience demonstrates, many of the
highest quality waters in the country contain aluminum
concentrations that would exceed the Draft Criteria. The reason
that we can have abundant aluminum and healthy streams is that
most aluminum is not bioavailable. But despite apparently
recognizing that much aluminum in the real world is not
bioavailable, EPA has nonetheless relied on aluminum toxicity
tests that do not use natural geological materials in order to set the
standard. This is a clear case of "science" that is divorced from
real world application on the ground and therefore sets a standard
that is both overly costly and unnecessary for environmental
protection.
Along the same lines, while EPA recites that it has reviewed
thousands of aluminum samples in its research, it appears to have
ignored hundreds of thousands of aluminum samples taken by
those who must comply with existing aluminum criteria. For
example, the thousands of samples reported by West Virginia coal
miners each month, and the experience of those miners in
addressing aluminum toxicity at their operations, appear not to
have been accounted for in any meaningful way in the Draft
Criteria. The experience of those coal operators, if taken into
account, should cause EPA to stop this effort in its tracks. This is
because our members who operate in West Virginia report that the
efforts undertaken to treat discharges to reduce aluminum content
result in more environmental harm than benefit. Without taking
into account the unintended consequences of establishing overly
stringent water quality criteria, EPA is ignoring sound science.
The EPA's aluminum criteria reflect the best available
science, based on bioavailable aluminum. Dissolved,
colloidal and precipitated forms of aluminum are all
bioavailable to aquatic organisms, which supports the criteria
as total aluminum.
In the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations. In
the 2018 Final aluminum criteria document the EPA has
noted that external research on new analytical methods is
ongoing to address concerns with aluminum bound to
particulate matter (i.e., clay) from natural waters being
included in the total recoverable aluminum concentrations.
This approach would not acidify the sample to pH<2 but
rather to a higher pH to better capture the bioavailable
fraction of aluminum. The method has been published as
Rodriguez, P.H., J.J. Arbildua, G. Villavicencio, P.
Urrestarazu, M. Opa/o, A.S. Cardwell, W. Stubblefield, E.
Nordheim, and W. Adams. 2019. Determination of
Bioavailable Aluminum in Natural Waters in the Presence of
Suspended Solids. Environ. Toxicol. Chem. 29 April 2019.
https://doi.org/10.1002/etc.4448. The expectation is that this
approach may better estimate the bioavailable fraction of
aluminum in natural waters.
To the best of our knowledge, the EPA has reviewed all
aluminum toxicity studies available at this time. The public
comment period was open to allow for submission of any
additional research that the public may identify.
No edits.
230
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I on Topic 20: Rciiiirdinii I.PA policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
States should be allowed to retain existing dissolved criteria.
Many states currently express aluminum aquatic life criteria as
dissolved criteria and/or allow aluminum translator studies to be
conducted on a site-specific basis (U.S. EPA 1996). UWAG
recommends that EPA add language specifying that the final
criteria, if still expressed as total recoverable concentrations,
should not supersede previously state-adopted (and EPA
approved) dissolved aquatic life criteria for aluminum.
The EPA's aluminum criteria are based on total recoverable
aluminum toxicity in laboratory studies and the criteria
document reflects consideration of what is bioavailable to
aquatic organisms in the natural environment as described in
the latest high quality peer reviewed literature. Dissolved,
colloidal and precipitated forms of aluminum may all be
bioavailable to aquatic organisms, which supports the criteria
as total aluminum. If aluminum criteria are based on
dissolved concentrations, toxicity would likely be
underestimated, as colloidal forms and hydroxide precipitates
of the metal that can dissolve under natural conditions and
become biologically available would not be measured.
Current research and methods development for aluminum
analytical methods are expected to improve quantification of
the bioavailable aluminum.
For parameters for which the EPA has issued new or revised
304(a) criteria recommendations, the WQS regulation at
131.20(a) requires that "if a State does not adopt new or
revised criteria for parameters for which the EPA has
published new or updated CWA section 304(a) criteria
recommendations, then the State shall provide an explanation
when it submits the results of its triennial review to the
Regional Administrator consistent with CWA section
303(c)(1) and the requirements of paragraph (c) of this
section." The regulation does not, however, require states to
adopted revised criteria based on the EPA's latest
recommendation.
The EPA's final 304(a) recommendation does not constitute a
federal promulgation for states. Until and unless a state
adopts a revision to its own aluminum criteria the state's
previously adopted and the EPA-approved criteria are
applicable for CWA purposes.
No edits.
EPA-HQ-OW-
2017-0260-0064
(Scott G. Mandirola,
Director, West
Virginia Department of
Environmental
Protection (WVDEP))
The West Virginia Department of Environmental Protection
(WVDEP) thanks the Environmental Protection Agency (EPA) for
the opportunity to comment on its Draft Aquatic Life Ambient
Water Quality Criteria for Aluminum 2017 (Draft Aluminum
Criteria). WVDEP offers the following comments on the Draft
Aluminum Criteria.
West Virginia first adopted EPA's 1988 recommended aquatic life
The EPA's response to specific state packages submitted to
the EPA is outside the scope of this response to comments.
No edits.
231
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C cim moil I on Topic 20: Rciiiirdinii I.PA policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
water qualify criteria for aluminum in 1993. AJicr adopting the
recommended criteria, West Virginia made several changes to it
over the years as additional data on aluminum toxicity
accumulated. In 1997, West Virginia revised its aluminum criteria
by removing the chronic portion of the standard, which at that time
had been 87 fig/l; this revision was subsequently disapproved by
EPA. By 2005, the West Virginia Environmental Quality Board
who then managed the state's Water Quality Standards, had re-
inserted the chronic criterion at 87 fig/1, and specified the criteria
as being for dissolved aluminum. In 2008, after WVDEP had taken
over management of standards, West Virginia adopted a standard
for dissolved aluminum which included 750 fig/l acute and chronic
exposure for warm waters, and 750 fig/l acute exposure and 87
fig/l chronic exposure for trout waters. This dissolved aluminum
water quality standard was subsequently approved by EPA. Then,
in 2015, after studying research which indicated aluminum to be
dependent upon hardness, West Virginia revised its dissolved
aluminum criteria by amending Legislative Rule 47 CSR 2,
Requirements Governing Water Quality Standards [Current
version of 47 CSR 2 found here:
http://dep.wv.gov/WWE/Programs/wqs/Documents/47CSR2%2007
0816. pdf], submitting the rule to EPA for review and approval on
October 26, 2015. This change, which remains in 47 CSR 2 and
awaits EPA approval or disapproval, would implement a
hardness-based criterion only for the pH range of 6.5 to 9.0. Above
and below this pH, WV's previous aluminum criterion is still in
place. EPA commented on West Virginia's aluminum criteria
revision in February 2016, citing EPA's ongoing effort to revise
the existing criteria recommendations for aluminum,
[http://dep. wv.gov/WWE/Programs/wqs/Documents/EPA % 2 ODocu
ments/EPA%20Comments%20on%20WV%20Se%20and%20Al%2
02-23-16.pdf]" and again in March 2016 to share limited
preliminary results from the mussel study they had requested West
Virginia wait for completion of in 2013. However, West Virginia
has received no official response regarding approval or
disapproval of the aluminum criteria revision WVDEP submitted
to EPA 24 months ago, although this determination is required of
EPA pursuant to Section 303(c) of the Clean Water Act.
232
-------
( oiiimeiil
Nil in hoi'
(Oiliiiiii/iilion)
Public ( cim moil I oil Topic 20: Kc^iirdin^ l.l'A policies
I'.PA Response
Kc\ision l.ociilion in
20IX Aluminum
Crilcriii Document
EPA-HQ-OW-
2017-0260-0075
(Steven A. Buffone,
CHHM, QEP, GIT,
Supervisor,
Compliance and
Regulatory Affairs,
CONSOL Energy Inc.)
We also feel that it's important that the states retain their primacy
and be allowed to develop their own criteria or adopt the
recommended criteria or portions of it as they feel is appropriate
for their unique regional variations.
The aluminum AWQC are recommendations for the states
and authorized tribes and are not water quality standards.
States may adopt the 304(a) criteria into their water quality
standards and can also adopt other criteria if they are
scientifically defensible and protective of use, as determined
by the state.
No edits.
233
-------
TOPIC 21: Comments regarding the regulatory burden of aluminum criteria
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public ('onimcnl oil l opic 21: Keuiirriinu (lie re^uhilon
burden «il*;ilniiiiiniin crilcriii
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0030
(Nelson Brooke,
Riverkeeper et al.,
Black Warrior
Riverkeeper)
Thank you for the opportunity to provide comments on the above-
referenced recommendation regarding the application of ambient
water quality criteria for aluminum. We write on behalf of Black
Warrior Riverkeeper ("Riverkeeper"), a nonprofit organization
dedicated to protecting and restoring the Black Warrior River and
its tributaries. Currently, several streams in the Black Warrior
River watershed are impaired due to excessive concentrations of
aluminum, with TMDLs having already been approved. Several
other streams and river segments are also at risk for aluminum
impairment due to excessive contributions from existing and/or
abandoned facilities. We are concerned with the proposed
recommendation because of the potential implications it could
have for state regulators, as well as organizations such as
Riverkeeper, which have a vested interest in protecting water
quality, and must rely on EPA guidance in enforcing specific water
quality criteria.
Riverkeeper enthusiastically supports the idea of basing regulatory
decision-making on the application of the best scientific literature
available and expanded data sets. We believe that the proposed
guidance is a well-intended effort to do just that. However we fear
that in its haste to update the science supporting the proposed
recommendation to include the multiple linear regression (MLR)
model, EPA has actually done little more than unnecessarily
complicate the calculation of aluminum criteria in a manner that
will increase costs to regulators (and non-profit organizations,
such as ours), and increase the uncertainty of the site-specific
criteria by compounding the standard of error in each calculated
measurement (Al, pH, DOC, and Hardness), resulting in criteria
calculations that are less protective of water quality in most cases.
The EPA based its 2018 aluminum criteria on publicly
available, peer-reviewed science, with the underlying model
and toxicity data largely developed by external leaders in the
field.
No edits.
EPA-HQ-OW-
2017-0260-0031
(Robert P.
Baumgartner,
Regulatory Affairs
Department Assistant
Director, Clean Water
Services (District))
Clean Water Services (District) appreciates the opportunity to
comment on the U.S. Environmental Protection Agency's proposed
aquatic life water quality criteria for aluminum. The District is a
county service district, located in Washington County, Oregon,
providing sanitary sewer service, stormwater management and
environmental restoration for more than 560,000 residents and the
businesses and industries that support the local and global
economy. The District holds an integrated watershed-based
NPDESpermit covering the sanitary sewer conveyance system,
four wastewater treatment plants and the municipal separate storm
sewer system serving urbanized Washington County. Adoption of
The EPA based its 2018 aluminum criteria on publicly
available, peer-reviewed science, with the underlying model
and toxicity data largely developed by external leaders in the
field. The aluminum AWQC are recommendations for the
states and authorized tribes and are not water quality
standards. States may adopt the 304(a) criteria into their
water quality standards and can also adopt other criteria if
they are scientifically defensible and protective of use, as
determined by the state.
No edits.
234
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 21: Ko^iirdin^ llic rciiiihilon
burden of ;iIn ini mi in crilcrhi
I'.I'A Response
Kc\ision Location in
20IX Aluminum
Criloriii Document
the draft aluminum criteria in its current form would significantly
impact the District and the communities it serves with little to no
added benefit to water quality.
EPA-HQ-OW-
2017-0260-0065
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
WVCA is concerned with the overall approach by US EPA in
preparing, peer reviewing, and publishing the Draft Aluminum
Criteria. As set forth more fully below, the effect of US EPA's
insular view of aluminum chemistry creates a circumstance where
the capital and compliance costs of the Draft Aluminum Criteria
will be staggering, while providing little or no actual
environmental protection or enhancement.
The EPA based its 2018 aluminum criteria on publicly
available, peer-reviewed science, with the underlying model
and toxicity data largely developed by external leaders in the
field. The aluminum criteria and underlying basis underwent
5 independent, external expert peer reviews; these reports and
EPA's responses are available on EPA's website
rhttos://www.eoa.sov/wac/aauatic-life-criteria-aluminum).
The aluminum AWQC are recommendations for the states
and authorized tribes and are not water quality standards.
States may adopt the 304(a) criteria into their water quality
standards and can also adopt other criteria if they are
scientifically defensible and protective of use, as determined
by the state.
No edits.
235
-------
TOPIC 22: Comments regarding a request for an extension on the comment period
(o 1111111-111
Nil in her
(Oi'Uiini/iilioii)
Public Comment on l opic 22: Keu;ir(lini£ ;i rc(|iics( lor ;in
extension on the comment period
I'.I'A Response
Ke\ision l.ociilion in
20IX Aluminum
( rileriii Document
EPA-HQ-OW-
2017-0260-0050
(J. Tyler White,
President, Kentucky
Coal Association
(KCA))
Inadequate Time for Public Input
The public has not been given sufficient time or opportunity to
review the use of the MLR, which forms the underpinning of much
of the Draft Criteria but was only published recently. More
generally, KCA and other impacted members of the public cannot
be expected to reasonably comment on the numerous technical
considerations that have formed the Draft Criteria in the time
provided. Given that states have struggled for years to apply the
1988 criteria, in many cases without resolution, it is unreasonable
for EPA to assume that the voluminous information in the new
criteria can be fully analyzed and addressed in the short timeframe
for public comment.
A 60-day review period is typical for a water quality criteria
document. Since an additional 30 days was added onto the
review period (total of 90 days), the EPA believes that a
sufficient amount has been allotted for document review by
the public.
No edits.
EPA-HQ-OW-
2017-0260-0061
(Penny Shamblin,
Hunton & Williams
LLP on behalf of
Utility Water Act
Group (UWAG))
EPA must afford the public access to kev technical papers that
All these studies have been published and are available. All
studies are identified in the bibliography of the criteria
document.
No edits.
have vet to be published.
As an initial matter, EPA must ensure all the information relied on
to establish the Draft Criteria is available for public review and
comment. It has not done so in this case. The Draft Criteria cites
and is based on MLR models developed by D.K. DeForest and
others. Their work, however, currently is unpublished. DeForest,
D.K., K.V. Brix, L.M. Tear and W.J. Adams. 2017 (Manuscript).
Multiple Linear Regression (MLR) models for predicting chronic
aluminum toxicity to freshwater aquatic organisms and developing
water quality guidelines. Environ. Toxicol. Chem. (submitted); see
also Brix, K. V., D.K. DeForest, L. Tear, M. Grosell and W.J.
Adams. 2017 (Manuscript). Use of multiple linear regression
models for setting water quality criteria for copper: A
complimentary approach to the biotic ligand model. Environ.
Toxicol. Chem.
For stakeholders to be able to meaningfully review and comment
on the Draft Criteria, it is essential that the underlying information
on which it is based be available to all during the review
period. [UWAG understands that members of the Society of
Environmental Toxicity and Chemistry (SETAC) have access to the
DeForest et al. (2017) unpublished paper. Others, however, do
not.] The ability to review the DeForest et al. (2017) paper is
especially crucial as it contains the empirical aluminum toxicity
data used to develop the MLR models, which in turn are used to
develop the acute and chronic criteria. EPA should make these
papers available and extend the comment period to afford
236
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 22: Kc^iirdin^ ;i rc(|iicsl I'or ;ni
extension on (lie comment period
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crileriii Document
stakeholders the opportunity to consider that information in
commenting on the Draft Criteria.
EPA-HQ-OW-
2017-0260-0066
(David Smiga,
Assistant General
Counsel-
Environmental, United-
States Steel
Corporation)
The MLR model development papers have not yet been published.
The papers should be published to allow stakeholders the ability to
review these development documents, along with a comment
period of sufficient time to appropriately comment on the scientific
approach and validity.
All these studies have been published and are available. All
studies are identified in the bibliography of the criteria
document.
A 60-day review period is typical for a water quality criteria
document. Since an additional 30 days was added onto the
review period (total of 90 days), the EPA believes that a
sufficient amount has been allotted for document review by
the public.
No edits.
EPA-HQ-OW-
2017-0260-0071
(Fredric P. Andes,
Coordinator, Federal
Water Quality
Coalition (FWQC))
The papers that document the scientific basis for the MLR models
have not yet been published. It is important that stakeholders be
able to review these papers before EPA moves ahead to issue
recommended criteria based on those models. The Agency should
make the papers available, and provide extended time for submittal
of comments concerning the papers and their impact on the
scientific approach embodied in the Draft Criteria.
EPA-HQ-OW-
2017-0260-0006
(State of New Mexico
Environment
Department)
This comment is a request for an extension of the public comment
period for the draft "Updated Aquatic Life Ambient Water Quality
Criteria for Aluminum in Freshwater".
Although this document has been anticipated for some time, the
implications this document will have on water quality standards
across the nation, once approved, warrants a thorough
assessment. Due to the magnitude and level of technical detail
used to develop this proposed criteria, the State of New Mexico
Environment Department is seeking, at a minimum, an additional
30 days to complete this review, for a total comment period of 90
days.
Your consideration is greatly appreciated.
Thank you for your comment. The comment period was
extended to be for 90 days.
No edits.
EPA-HQ-OW-
2017-0260-0007
(Anonymous public
comment)
This comment is a request for an extension of the public comment
period for the draft "Updated Aquatic Life Ambient Water Quality
Criteria for Aluminum in Freshwater".
Potable Water Treatment Facilities in Massachusetts may be
subject to Aluminum limits in their NPDESpermits and therefore,
Massachusetts Water Works Association is seeking, at a minimum,
an additional 30 days to complete a review of the proposed
criteria, for a total comment period of 90 days.
237
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 22: Ko^iirdin^ ;i rc(|iios( I'or ;ni
extension on I lie com moil I period
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Criloriii Document
El 'A-IIQ-OW-
2017-0260-0009
(Roger Claff, Senior
Scientific Advisor,
American Petroleum
Institute (API))
API requests a 60-day extension, through November 27, 2017, of
the public comment period for the U.S. Environmental Protection
Agency's (EPA's) Notice of Availability, "Request for Scientific
Views: Draft Updated Aquatic Life Ambient Water Quality Criteria
for Aluminum in Freshwater" (82 Fed. Reg. 35198, July 28, 20
17). These draft criteria, derived through data normalization by
application of a multiple linear regression model, and addressing
the influence of numerous receiving water quality parameters
including pH, DOC, and hardness, are technically complex.
Aluminum chemistry and effects on aquatic life complicate criteria
derivation and implementation. Although API and others have
begun an assessment of the draft criteria, a 60-day extension of the
comment period is warranted to allow the public to provide
meaningful, detailed technical reviews of EPA's criteria derivation
process, and to identify and detail their concerns with criteria
implementation.
EPA-HQ-OW-
2017-0260-0011
(Fredric P. Andes,
Federal Water Quality
Coalition)
On behalf of the Federal Water Quality Coalition, we request a 60-
day extension of the comment period for the draft updated
aluminum water quality criteria. The new criteria, andEPA's draft
guidance, present a series of technically complex scientific issues.
The guidance document, along with other related materials, total
almost 2000 pages. The current comment period does not provide
us with adequate opportunity to perform a careful review of these
materials and provide meaningful, comprehensive comments and
recommendations. Therefore, we believe that an additional 60 days
are needed in order to perform those tasks. We ask that the Agency
grant this request before the comment period expires, so all
stakeholders know that they have more time to complete their
comments. If you have any questions or need any additional
information, just let me know. Thank you.
EPA-HQ-OW-
2017-0260-0015
(Jason D. Bostic, Vice-
President, West
Virginia Coal
Association (WVCA))
The West Virginia Coal Association (WVCA) is writing this letter
to respectfully request a 60-day extension of the comment period
for the U.S. Environmental Protection Agency's (EPA's) Notice of
Availability, "Request for Scientific Views: Draft Updated Aquatic
Life Ambient Water Quality Criteria for Aluminum in Freshwater"
(82 Fed. Reg. 35198, July 28, 2017).
WVCA is requesting a 60-day extension of the public comment
period to allow the coal industry in West Virginia to fully evaluate
the ramifications of the proposed criteria.
238
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I oil lopic 22: Kc^iirdin^ ;i rc(|iicsl I'or ;ni
extension on (lie comment period
I'.I'A Response
Kc\ision l.ociilion in
20IX Aluminum
Crileriii Document
El 'A-11Q-OW-
2017-0260-0017
(John Heggeness,
Surface Water Quality
Standards and
Monitoring Bureau of
Water Quality
Planning, Nevada
Division of
Environmental
Protection)
The Nevada Division of Environmental Protection (XDEPj
requests a 30-day extension to the current review period for the
draft AWQC for exposure of aquatic life to aluminum in freshwater
systems ("Draft Updated Aquatic Life Ambient Water Quality
Criteria for Aluminum in Freshwater.") The NDEP Bureau of
Water Quality Planning needs additional time to adequately
review the technical details of the document and prepare
comments. The revised deadline would be October 26, 2017.
EPA-HQ-OW-
2017-0260-0018
(Steven A. Buffone,
Compliance &
Regulatory Affairs,
Consol Energy Inc.)
CONSOL Energy Inc. (CONSOL) is requesting a 60 day extension
to submit comments on the Environmental Protection Agency
(EPA) proposed Draft Updated Aquatic Life Ambient Water
Quality Criteria for Aluminum in Freshwater, Docket ID: HQ-
OW-2017-0260-0001. CONSOL owns and operated the premier
underground longwall coal mining complex in the United States,
located in southwestern Pennsylvania, and additional legacy
mining properties that generate water discharges that are often
managed under National Pollutant Discharge Elimination System
(NPDES) permits.
CONSOL is requesting this 60 day extension in order to fully
assess the impacts of the draft criteria on our ability to meet our
existing NPDES discharge limits. These 2017 draft criteria are
complex when compared to the 1988 AWQC Criteria, with the
addition of species, the derivation of data through normalization
by application of a multiple linear regression model, and
addressing the influence of numerous receiving water quality
parameters including pH, DOC, and hardness. A 60 day comment
period is not adequate to complete our review and provide
substantial comments to the EPA.
CONSOL appreciates the opportunity to comment on the Draft
Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater and your consideration of our request for
additional time.
239
-------
( oiiimeiil
Nil in hoi'
(Or^iini/iilioii)
Public C oiii moil I on lopic 22: Ko^iinlin^ ;i rc(|iios( I'or ;ni
extension on (ho comment period
I'.I'A Response
Ko\ision l.ociilion in
20IX Aluminum
Criloriii Document
El 'A-11Q-OW-
2017-0260-0019
(Laura Cooper,
Division of Water and
Waste Management
Water Quality
Standards, West
Virginia Department of
Environmental
Protection)
This comment is a request far an extension of the public comment
period far the draft "Updated Aquatic Life Ambient Water Quality
Criteria for Aluminum in Freshwater".
Although this document has been anticipated for some time, the
implications this document will have on water quality standards in
West Virginia warrants a thorough assessment. Due to the
magnitude and level of technical detail used to develop this
proposed criteria, the West Virginia Department of Environmental
Protection is seeking, at a minimum, an additional 30 days to
complete this review, for a total comment period of 90 days.
EPA-HQ-OW-
2017-0260-0024
(Nevada Division of
Environmental
Protection (NDEP))
The Nevada Division of Environmental Protection (NDEP)
requests a 30-day extension to the current review period for the
draft AWQC for exposure of aquatic life to aluminum in freshwater
systems ("Draft Updated Aquatic Life Ambient Water Quality
Criteria for Aluminum in Freshwater.")
The NDEP Bureau of Water Quality Planning needs additional
time to adequately review the technical details of the document and
prepare comments. The revised deadline would be October 26,
2017.
EPA-HQ-OW-
2017-0260-0033
(Lisa D. Daniels,
Acting Deputy
Secretary,
Pennsylvania
Department of
Environmental
Protection (DEP))
The Pennsylvania Department of Environmental Protection (DEP)
is requesting an extension to the public comment period for the
Draft Updated Aquatic Life Ambient Water Quality Criteria for
Aluminum in Freshwater. DEP needs additional time to be able to
more thoroughly review and formulate comments on the draft
recommendations for updating the 304(a) aquatic life criterion for
freshwater aluminum. The draft recommendations are based, in
large part, on data and conclusions from, yet unpublished reports
and studies that are not readily available to DEP staff for this
review. DEP needs time to better understand the basis for using
this new data, and the complexities of the bi-modal nature of
aluminum chemistry and toxicity described in the Multi-variate
Linear Regression Model used in calculating the updated criteria.
240
------- |