United States
Environmental Protection
Agency
Office of Radiation Programs
(ANR-459)
• EPA/520/1-89-031
January 1990
&EPA
Comments and
Response to Comments
NESHAPS for Radionuclldes
Printed on Recycled Paper
-------
-------
40 CFR Part 61
National Emission Standards
for Hazardous Air Pollutants
EPA 520/1-89-031
Comments and Response to Comments
Environmental Impact Statement
for NESHAPS Radionuclides
BACKGROUND INFORMATION DOCUMENT
January 1990
U.S. Environmental Protection Agency
Office of Radiation Programs
Washington, D.C. 20460
-------
-------
Preface
The Environmental Protection Agency is promulgating National
Emission Standards for Hazardous Air Pollutants (NESHAPs) for
Radionuclides. An Environmental Impact Statement (EIS) has been
prepared in support of the rulemaking. The EIS consists of the
following three volumes:
•*
VOIiUME I - Risk Assessment Methodology
This document contains chapters on hazard
identification, movement of radionuclides through
environmental pathways, radiation dosimetry,
estimating the risk of health effects resulting from
expose to low levels of ionizing radiation, and a
summary of the uncertainties in calculations of dose
and risks.
VOLUME II - Risk Assessments
This document contains a chapter on each radionuclide
source category studied. The chapters include an
introduction, category description, process
description, control technology, health impact
assessment, supplemental control technology, and cost.
Tt has an appendix which contains the inputs to all
the computer runs used to generate the risk
assessment.
VOLUME III - Economic Assessment
This document has chapters on each radionuclide source
category studied. Each chapter includes an
introduction, industry profile, summary of emissions,
risk levels, the benefits and costs of emission
controls, and economic impact evaluations.
Copies of the EIS in whole or in part are available to all
interested persons; an announcement of the availability appears in
the Federal Register.
-------
For additional information, contact James Hardin at
(202) 475-9610 or write to:
Director, Criteria and Standards Division
Office of Radiation Programs (ANR-460)
Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
IV
-------
LIST OF PREPARERS
Various staff members from EPA's Office of Radiation Programs
contributed in the development and preparation of the EIS.
Terrence McLaughlin
James Hardin
Byron Hunger
Fran Cohen
Albert Colli
Larry Gray
W. Daniel Heridricks
Paul Magno
Christopher B. Nelson
Dr. Neal S. Nelson
Barry Parks
Dr. Jerome PusJcin
Jack L. Russell
Dr. James T. Walker
Larry Weinstock
Chief, Environmental
Standards Branch
Health Physicist
Economist
Attorney Advisor
Environmental
Scientist
Environmental
Scientist
Environmental
Scientist
Environmental
Scientist
Environmental
Scientist
Radiobiologist
Health Physicist
Chief Bioeffects
Analysis Branch
Engineer
Radiation
Biophysicist
Attorney Advisor
Project Officer
Author/Reviewer
Reviewer
Author/Reviewer
Author/Reviewer
Reviewer
Author/Reviewer
Author
Author
Reviewer
Author/Reviewer
Author/Reviewer
Author
Reviewer
An EPA contractor, S. Cohen and Associates, Inc., McLean, VA,
provided significant technical support in the preparation of the
EIS.
-------
-------
TABLE OF CONTENTS
1. INTRODUCTION 1
2. COMPLIANCE ISSUES ". 2
2.1 POINT OF COMPLIANCE , 2
2.2 ACCIDENTS . 4
2.3 REPORTING REQUIREMENTS 4
2.4 MODIFICATIONS AND NEW CONSTRUCTION 5
3. USE OF COMPUTER MODELS 7
3.1 GENERAL COMMENTS 7
3.2 USE OF DOSE RESPONSE MODELS IN RISK ASSESSMENT ... 9
3.3 THE CAP-88 RISK ASSESSMENT CODE 14
4. SOURCE CATEGORY SPECIFIC COMMENTS . 24
4.1 DOE FACILITIES 24
4.2 NRC-LICENSED AND NON-DOE FEDERAL FACILITIES .... 31
4.3 URANIUM FUEL CYCLE FACILITIES 41
4.4 HIGH LEVEL WASTE FACILITIES 48
4.5 ELEMENTAL PHOSPHORUS FACILITIES 49
4.6 COAL-FIRED UTILITY AND INDUSTRIAL BOILERS 55
4.7 INACTIVE AND LICENSED URANIUM MILL TAILINGS .... 59
4.8 DOE RADON SITES . 67
4.9 UNDERGROUND URANIUM MINES . 69
4.10 SURFACE URANIUM MINES ..... 74
4.11 PHOSPHOGYPSUM STACKS 76
Vll
-------
-------
1. INTRODUCTION
On March 7, 1989, the EPA published in the FEDERAL REGISTER (54
Fed. Reg. 9612) proposed National Emission Standards for
Hazardous Air Pollutants (NESHAPs) for radionuclides emitted to
air from 12 source categories. The proposal requested public
comments on the proposed NESHAPs and the specific risk management
approaches that were used to develop them. Informal public
hearings were held in Washington, DC and Las Vegas, Nevada to
give interested parties an opportunity to present their views,
and written comments were solicited. Comments were received from
almost 300 individuals and organizations representing state and
federal agencies, public interest groups, industry and private
citizens. This document addresses the comments relating to
compliance procedures, modeling, risk assessment, and those
specific to each source category. The EPA's response to comments
addressing significant legal and policy issues are presented in
Section VII of the preamble to the final rule on radionuclides,
published at 54 Fed. Reg. 51654.
-------
2. COMPLIANCE ISSUES
2.1 Point of Compliance
Concerns regarding this topic generally fell into two groups;
those agreeing with determination of compliance based on maximum
radionuclide ambient air concentration at the nearest residence
but often disagreeing with EPA's exposure calculations as
unrealistically high or low, and those who stated that some other
point should be used.
2.1.1 Regulations based on the nearest exposed individual at
maximum concentration appear to conflict with standards in Title
10 of the Code of Federal Regulations ("CFR"); these require
controls which restrict radiation levels in any non-restricted
area.
Response: The limits established by the NESHAPs apply to any
member of the public exposed to ambient air emissions of
radionuclides. As a practical matter, for most facilities this
will be the nearest exposed individual at maximum concentration.
Similarly, implementation of standards under 10 CFR to any non-
restricted area assure that no member of the public receives a
dose greater than the standards allow. Although the
implementation is different and the facility will sometimes have
to meet both standards, the NESHAP or the 10 CFR regulations that.
requires the lowest emissions will be controlling for that
facility. In the context of NRC-licensed facilities, the EPA
anticipates that any inefficiency that arises from the additional
NESHAP implementation scheme will be mitigated through
cooperation agreements entered into by the EPA and the NRC.
2.1.2 The NESHAP should be based on effective dose equivalent at
the off-site point of maximum concentration regardless of whether
a person currently lives there. This is necessary due to the
limited ability of dispersion models to accurately predict
concentrations at any given location and also because a residence
may in the future come to be located at the point of predicted
maximum ground-level concentration. Moreover, as a matter of
policy, NESHAP decisions should be based on the conservative
assumption that the nearest resident lives at the facility
boundary or at the point of highest ground level concentration.
Alternately, it has been suggested that a "critical group"
approach be utilized as preferable to the moving target
represented by the maximally exposed individual.
Response: The EPA's implementation procedures specify that the
limit'applies to the most exposed member of the population. This
comports with the EPA's belief that its compliance and
implementation modeling and procedures are capable of reasonable
accuracy and are practical to use. This also implements the
-------
EPA's policy that NESHAPs protect real persons. Moreover, since
compliance is on an annual basis, should in the future some
individual come to leave his or her residence located in the area
of maximum concentration, that change in the MIR will be
accounted for. The implementation procedures not only allow but
require, the regulated community to specify the locations of
nearby individuals. The compliance models then assess dose to
the individual located at the point of highest environmental
risk. The EPA believes this is consistent with its authority
under the CAA.
2.1.3 Adjustments to the models or calculations so as to more
accurately reflect the exposure of the maximum exposed individual
to the activity being regulated are appropriate. These could
include consideration of occupancy factors and building
sheltering, temporal variations, particle size, air
concentrations, etc. Monte Carlo modeling techniques of
occupancy factors and other variables would give best estimate of
risk and also the distribution of probable values about the risk
estimate.
Response; The EPA's assessment and compliance codes do account
for particle sizes and:differing air concentrations due to
dispersion as well as occupancy factors and building sheltering,
but not temporal variations. The EPA has also performed
uncertainty analyses using Monte Carlo techniques. The results
indicate that while residency factors have a large influence on
the estimated risk (factors of 10 to 15), the risk estimates used
by the EPA are within 95 percent confidence interval.
2.1.4 The proposed regulation appears to limit airborne
radionuclide emissions from a given facility. The wording on
this is not very clear or consistently stated throughout the
supplementary documents.
Response; The rules were developed to apply to individual sites
where the release is from stacks or vents, and individual sources
where the releases are from area sources such as mill tailings.
2.1.5 The hypothetical, maximally exposed individual is such an
arbitrarily conceived concept that it serves no useful analytical
purpose, and is assumed to be at the point of maximum exposure.
Response; The EPA disagrees. The maximum exposed individual
identified in the risk assessment for 12 source categories, is
not hypothetical as the EPA used all resources available, within
the time constraints, in an attempt to ascertain where an
individual actually lives and at what level that individual is
exposed. This included demographic surveys, company reports, and
-------
U.S. Geological Survey maps. In so doing, the location of the
maximum exposed individual is not the point where maximum
exposure occurs, but the location where an actual individual
lives who receives the maximum annual exposure from the source.
2.2 ACCIDENTS
2.2,1 The final rule should specify the basis for coverage of
accidents, any special provisions concerning treatment of
accidents for compliance purposes, and the types of accidents
included in EPA's supporting assessments for the proposed rule.
Accidents are a special case and should be evaluated in the ample
margin of safety checklist. Estimated radionuclide releases due
to accidental release should be specifically included in the
demonstration of compliance to the NESHAP and in the dose/risk
assessment for new and modified sources.
Response; The issue as to what constitutes an accident and
whether accidents must be differently treated from other
hazardous pollutant emissions to ambient air, is not addressed by
the CAA and has not been considered in the development of these
standards. Nor are accidents anticipated to be considered when
the EPA evaluates applications for modification or new
construction. Therefore, accidents that result in the release of
radionuclides in excess of the standard would constitute a
violation that is enforceable by the Agency.
2.3 REPORTING REQUIREMENTS
2.3.1 EPA should specify more frequent record-keeping and
compliance reporting requirements.
Response; The EPA believes that the specified record-keeping and
annual reporting requirements are adequate. EPA is not aware of
any additional environmental or safety benefits that would be
derived from requiring additional records or more frequent
reporting.
2.3,2 NE~SHAP standards for radionuclide emissions should apply
for the calendar year rather than "... any period of 12
consecutive months...".
Response; The EPA disagrees as this measure is taken to more
accurately reflect whether a facility is emitting excessive
radionuclides into ambient air and thereby jeopardizing public
health.
2,3.3 • In seeking EPA approval for proposed construction and/or
modification activity, the applicant should only be required to
describe to EPA the proposal and its calculated increased
emissions to ambient air. Some of the reporting requirements are
-------
not clear and some may be needlessly complex. These should be
clarified and, where possible, simplified through use of a 1
percent threshold compliance reporting limit and one-time
reporting where appropriate.
Response; Applications for construction or modification are
required to supply the.information necessary to identify the
activity and to justify the estimated impact upon public health
due to the potential increase of radionuclide emissions.
Reporting requirements have been simplified wherever possible.
Appropriate threshold compliance reporting limits have been
established.
2.4 MODIFICATIONS AND NEW CONSTRUCTION
2.4.1 The proposed threshold value of 1 percent of the standard
for potential emission increases from modifications or new
construction is unreasonably low. EPA approval should not be
required unless the increase will exceed 10 mrem/yr. In the
alternative, a 1 percent threshold based on a standard of 25
mrem/yr would provide considerable administrative relief.
However, it should be noted that, although prior approval to
construct or modify is not required for small dose contributors,
start-up notification is still required and the annual report to
the EPA must contain all the information in the application so
there is no savings in paperwork.
Response: The EPA believes the threshold value of a potential
emissions increase of 1 percent of the 10 mrem/y ede standard,
above which EPA approval to construct or modify is required, is
reasonable and proper. This has been added to the final rule as
a result of comments to the proposed rules and should provide
considerable administrative relief; trivial sources need not
apply, but sources releasing significant amounts of radionuclides
would be required to do so.
2.4.2 Permit use of the COMPLY code to determine whether an
application need be made to the EPA for approval to construct a
new source or modify an existing source.
Response; COMPLY may be used for this purpose.
2.4.3, The interpretation that any increase in radionuclide
emissions triggers requirements of 40 CFR Part 52, PSD is
especially troublesome; a threshold should be established,
consistent with the approach for other priority air pollutants.
Response; The Agency is aware of this issue and is considering a
threshold of that kind.
-------
2.4.4 For the section "Distinction between Construction and
Modification," the construction of a new building is now defined
as new construction at the facility. Additional clarification
should be added to this section. For example, is the expansion
of an existing building to allow replacement of an existing
process with new state-of-the-art equipment, which will not
increase emissions, defined as new construction? One could say
the expansion is a new building and that this constitutes new
construction. Assuming that this is not new construction, if a
new building to house the process is built adjacent to the
existing facility, is this new construction?
Response; EPA is preparing detailed guidance for its regional
offices and the states in dealing with situations such as these.
If the facility is in doubt, it should contact the EPA regional
office for clarification.
-------
3. USE OF COMPUTER MODELS
3.1 General Comments
3.1.1 Use of AIRDOS-EPA, RADRISK, and other EPA-approved codes
to demonstrate compliance should be permitted. The RADSGREEN
Model is equivalent to COMPLY and should be accepted by the EPA
as an alternative to COMPLY; documentation is available from the
Ohio EPA on request. Any computer program used to demonstrate
compliance should be peer reviewed, validated if appropriate, and
meet nuclear quality assurance (NQA-1) requirements. The EPA
should include a provision for approval of alternate programs.
Response; The EPA has devoted considerable resources and time to
the development of the CAP-88 and COMPLY models for determining
compliance. The intent of the COMPLY code is to assure the
protection of public health while minimizing the data collection
and input burden on affected licensees. While the EPA believes
that these codes are sufficient for all licensees, it has
provided for alternative compliance procedures to be used once
they are submitted to the agency for review and approval. The
EPA codes are available for peer review and have been widely
distributed to the regulated community. There is no benefit in
requiring compliance codes to meet nuclear quality assurance
requirements.
3.1.2 Facilities with multiple sources should be modeled in a
multiple source mode with radionuclide emission parameters
representative of each individual source; AIRDOS-EPA model cannot
accommodate sources at more than a single physical location.
Response; The AIRDOS-EPA model does co-locate multiple stacks.
For compliance purposes, where only the exposure of the maximum
individual is required, multiple code runs can be made to more
exactly calculate the contribution to dose from widely separated
sources to a single location. Guidance on performing such
multiple calculations is available by contacting the EPA's Office
of Radiation Programs.
3.1.3 In general, area sources are not monitored by measuring
the radon flux; most of the time boundary measurements are made.
A more uniform approach would be to base the limit on the total
curies released per second per site; another method would be to
base the limit on the concentration of radon in the air at the
site boundary.
Response; The EPA is fully aware that radon releases from area
sources are only infrequently monitored by direct measurements on
1 the source. The standards that the EPA has promulgated for such
sources (i.e., inactive phosphogypsum stacks, uranium mill
tailings disposal sites, and DOE Radon sites) specify design
-------
basis radon flux rates per unit area, to be confirmed by post-
remediation monitoring of the flux on the source.
3,1.4 Approach A should be modified to include the^NAS
recommendations of measurements of actual radionuclides rather
than following the current practice of placing undue reliance on
general models.
Response; At the levels established by the NESHAPs,
concentrations of many radionuclides may be difficult to
distinguish from natural background levels or below the limits of
detection of even state-of-the-art detection systems. Therefore,
the Agency will continue to rely primarily upon modeling
techniques to assess the risk of radionuclides released into the
ambient air.
3.1,5 Existing air emission monitoring equipment and methodology
should be allowed to remain in service for DOE's and NRC-licensed
facilities provided emissions from the source do not account for
more than 1 percent of the permitted value for the facility.
Response; The EPA has reviewed the existing monitoring and
analytical procedures for both DOE and NRC-Licensed facilities.
Based on these reviews, the EPA has determined that existing
procedures, as specified or referenced in each licensee's
technical specifications, are satisfactory and may be used.
3,1.6 There are no instruments or proven techniques available
for monitoring some of the isotopes at the detection levels
necessary to comply with the proposed standards.
Response; Absent specific examples, the Agency is unable to
determine the accuracy or significance of this comment. However,
the NESHAPS are promulgated with the provision that alternative
methods of determining emissions may be submitted to the Agency
for review and approval.
3.1.7 The EPA should allow use of certified test results as an
alternative to the adjustment factors in Table 1 of Appendix D.
Response; The Agency believes that most licensees will be able
to demonstrate compliance simply by using the approval adjustment
factors. However, the EPA will allow the use of factors based on
"certified" test results when the pertinent data relating to the
certification and details of the certification tests are
submitted to the agency for review and approval.
3,1.8 Regulated facilities should be required to acquire and
utilize in emissions models site specific meteorological data.
Response; Sensitivity analyses of meteorological data sets have
shown that the variation between data sets is relatively small.
8
-------
The EPA sees little benefit to the public health and safety to be
gained by requiring site-specific meteorological data versus data
from the nearest station.
3.1.9 Population data for performing dose/risk assessments are
based upon United States census tract data. What base year
census data is used to perform these analyses? Are the data
adjusted to account for population growth in the period after the
census was taken?
Response: The assessments use 1980 census data for the 0-80 km
population estimates, augmented where noted by site-specific data
for the nearby individuals. No adjustment to the census data has
been made.
3.1.10 The EPA numbers are not real deaths, nor should they be
considered as such. The numbers are nothing more than a
statistical calculation and have little or no relevance in
biological or public health terms.
Response; The EPA is fully aware of the statistical nature of
the risk numbers that it calculates. However, it strongly
disagrees that the estimates have no biological or public health
significance. There is a broad consensus in the radiation
protection field that the approach used by EPA is prudent and
appropriate for risk calculations in the radiation field. In
addition, this methodology has been reviewed and accepted by the
SAB.
3.2 Use of Dose Response Models in Risk Assessment
3.2.1 The linear non-threshold Dose Response Model is overly
conservative and does not represent scientific consensus when
used to extrapolate to the extremely small doses allowed in the
proposed rulemaking. BEIR III specifically cautions against
extrapolation of its risk estimates to such low doses (i.e., less
than 100_mrem/year) due to the high degree of uncertainty. No
scientific studies to date have been able to prove or disprove
that health effects occur at doses below a few hundred millirem
per year. Consequently, risk coefficients applied to such low
doses become so uncertain as to be meaningless
Response; Although the EPA acknowledges the comment that linear
non-threshold dose response model is often considered overly
conservative and not representative of a scientific consensus,
some dose-response relationship must be assumed or dose and risk
cannot be effectively determined. (ICRP #26, par 27,28). The
Agency believes this assumption is reasonable and realistic and
-------
has more adequate scientific support than for any other
hypothesis. The Agency has also tried to quantify the sources ot
uncertainty in the risk factors produced and has indicated the
range of uncertainty in the risk factor. Model uncertainty is
one of the uncertainties subsumed in this range.
3.2.2 The EPA has ignored recommendations made by its own
Science Advisory Board. The EPA is in error in treating risk
estimates as scientific fact. Conservative best guess estimates
must be clearly identified as such when presented to the public
and decision-makers. In general, the EPA's risk methodology
fails to represent the scientific community inasmuch as it
ignores other models which more closely represent scientific
observations. The preferred approach is to present a range of
values corresponding to various accepted dose response models so
that decision-makers can be informed as to the lack of certainty
in risk estimation.
Response; The Agency believes it has followed the scientific
recommendations of the SAB to the greatest extent possible. The
central value risk coefficients, response projection model, and
range of risk used by the Agency have been endorsed by the SAB.
Moreover, the Agency disagrees with the suggestion that it has
ignored superior models or that it has engaged in improper
decision-making.
3.2.3 The EPA uses a lifetime fatal cancer risk of 400 x 10 6
per rem. The justification of this value is insufficiently
supported and does not agree with risk coefficients proposed by
BEIR III, ICRP, NCRP.
Response; The risk factor of 400 x 10"6 fatal cancers per rem is
the estimate developed using the BEIR III linear, relative risk
model. It would also be about the number developed using the
linear relative risk model, risk coefficients from the UNSCEAR
1988 Report and a dose rate effectiveness factor of 3.
The ICRP risk estimate is over 18 old and does not represent the.
best of the current scientific information on the subject. The
NCRP has not made a numerical estimate of cancer for total body
gamma exposure.
3.2.4 The EPA's lifetime risks (Table 3) are inconsistent with
previously published lifetime occupational risks.
Response; Agency estimates of lifetime risk of cancer should not
be expected to be consistent with occupational risk estimates.
The Agency estimates population risk, birth to age 110 in a
lifetable; occupational risk estimates are for ages., 18-65.
10
-------
3.2.5 Although Volume I of the BID addresses teratogenic and
genetic risks, it is not clear whether these risks were
considered in combination with the carcinogenic risks in the
proposed emission standards.
Response; The Agency calculates genetic and teratogenic risk
factors. In the NESHAPs assessments the genetic and teratologic
risks were inconsequential compared to both the risk of
carcinogenesis and the overall uncertainty in the estimates.
They were not specifically quantified in the risk assessment.
3.2.6 EPA's risk estimates do not include updated information
for Japanese A-bomb survivors with regard to mortality data and
changes in dosimetry, nor do they reference more current risk
estimates such as BEIR V, UNSCEAR 86 and UNSCEAR 88.
Response; The Agency is familiar with the new data from Japan
addressed in Shumizu et al (Rad. Research 118: 502-524 (1989)
and RERF Technical Report 5-88 (1988), Stather, et al (NRPB
publication, NRPB-R226 (1988)) and the UNSCEAR 1988 Report. The
revised BID will show how this new data may affect Agency
radiation risk estimates. However, the Agency is awaiting a
review and analysis of the BEIR V Report with risk coefficients
calculated for North Americans, before attempting to revise
Agency risk coefficients. The final BID (or FEIS) discusses the
genetic risk estimates given in the UNSCEAR 1986 and 1988
Reports.
3.2.7 The EPA's reason(s) for electing an RBE of 2.7 for gammas
for genetic effects is not sufficiently substantiated.
Response; Although the support for an RBE of 2.7 is not
overwhelming, additional reference that the genetic effects of
alpha emitters are no greater than those due to acute x-rays can
be found in the UNSCEAR 1986 Report (par 402 page 87 and par 428
page 91). The risk from acute gamma or x-ray exposure was
reduced by a factor of 3 to give the estimate due to chronic
exposure, as noted in the BID. This is equivalent to using an
RBE of 3.
3.2.8 For teratogenic effects, the new DS 86 dosimetry suggests
a threshold for the developmental period of 8 - 15 weeks in the
range of 10 to 20 rads.
Response; The Agency agrees. However, the conclusion is too
uncertain to require a revision of risk coefficients at this time
(Otake, et al, RERF Technical Report TR 16-87). For instance,
only nineteen cases of mental retardation, distributed across six
dose categories in the 8 to 15 week exposure group, comprise that
data base. Perhaps the new cases of mental retardation
identified (Otake, et al, RERF Technical Report TR 16-87), or the
new data on IQ decrement (Otake, et al, RERF Technical Report TR
11
-------
2-88; Schull, et al, RERF Technical Report TR 3-88) will help
resolve this problem. Current scientific consensus on
appropriate models for radiation risk projection has selected the
relative risk model. The NCRP model is an absolute risk model.
In addition, the NCRP risk factor is based on 1975 vital
statistics. Generally speaking it is of historic interest but
has been superseded by newer risk models and theories.
3.2.9 The EPA fails to explain why it believes the NCRP 78 radon
risk estimate is outside the lower bound. Therefore, the EPA's
radon risk coefficient of 360 x 10"6 per WLM may be on the high
side since EPA did not include the NCRP value (1-91, 1-143).
Response; The NCRP 78 model was rejected because (1) the assumed
temporal dependence on risk—exponential decay of absolute risk—
does not appear to be consistent with the epidemiological data
(see BEIR IV); (2) the risk model is not properly normalized,
further reducing the risk estimate by about one-third; and (3)
the model fails to take into account the evidence for a synergism
between smoking and radon.- The assumed time dependence with its
incorrect normalization leads to a substantial underestimate of
the risk for the general population. Because the model does not
consider smoking, it assigns the same risk to nonsmokers as
smokers; as a result, its risk estimate for nonsmokers actually
exceeds that obtained with the BEIR IV or ICRP 50 model.
3.2.10 The EPA's radon risk coefficient of 360 x 10"6 per WLM is
based on BEIR IV and ICRP 50. Both BEIR IV and ICRP 50 estimates
are relative risk models and assume a multiplicative interaction
with smoking. However, the models differ with regard to time
since exposure. This results in BEIR IV (and the EPA) adopting a
methodology which is biased toward underestimation of exposure
and hence overestimation of risk.
Response: This comment has been considered in light of both
comments that say that the exposure of the US miner cohort was
underestimated, and those that say the exposure is overestimated.
However, the U.S. cohort is only one of four analyzed in BEIR IV
or one of three in ICRP 50. The BEIR IV and ICRP 50 models are
the authors best fit to their preferred model. While neither may
be proven correct in the future, they are the current best
estimate models.
3.2.11 The EPA, by using BEIR IV risk coefficients, does not
adequately account for smoking as a confounding variable in
estimating the lung cancer risks posed by exposure to radon.
Since smoking was more likely prevalent among miners and since
smoking is a potent lung carcinogen, the validity of the
measurements of the much weaker radon dose response signal in the
range of 0 - 300 WLM is in doubt.
12
-------
Response; The EPA is aware that it is possible that a weakness
to BEIR IV rests in the fact that smoking may be a confounding
variable. When full smoking data is collected for the various
cohorts, which has not been done as of this date, reanalysis may
allow more detailed account to be taken of the radon-smoking
interaction.
3.2.12 The miner exposure data used by BEIR IV are not measured
values, only a small fraction of miner exposure data represents
measured exposure. It is likely that these calculated exposure
values overestimated true exposures.
Response; While not all exposure estimates are based entirely on
measured data, there is no evidence that the calculated estimates
either overestimate or underestimate exposure.
3.2.13 Of the four miner cohort data sets used to estimate the
radon risk coefficient, there are two data sets for which there
is an excess risk at zero excess exposure; incorrect dose
assignment has been-postulated as one possible cause. For the
Eldorado miner cohort, the dose assignments are much lower than
was actually experienced which leads to inflated risk
coefficient. Thus, a 54% excess risk at zero exposure is
included in the dose-response plot. The Malmberget data from
Sweden also show as excess risk at zero dose intercept. These
two data sets should, therefore, not be used until these problems
are corrected by the renormalization of data.
Response; The Agency disagrees. The commenter provides neither
adequate justification nor detailed explanation capable of
supporting its position.
3.2.14 For radionuclides other than radon, the EPA distinguishes
between fatal and non-fatal effects. The EPA has not explained
why it assumes all radon induced lung cancers are fatal.
Response; Lung cancer is fatal with 95% or more of cases dying
within a year or so of diagnosis. For the draft BID, only fatal
lung cancers were calculated and listed in the case of radon
exposure. In the case of low-LET exposure or deposited
radionuclides, for many organs, 50% or less of the cases are
fatal, so both fatal and non-fatal cancers exist. This is
reflected in the BID in Tables 6-8 and 6-9 where mortality and
incidence are tabulated by organ and age.
3.2.15 The uncertainties associated with the calculated risk
estimates are not reflected once the risk estimate is carried
forward for use in making regulatory decisions.
13
-------
Response: The. EPA disagrees. The uncertainties and other limits
to risk assessment are fully articulated and considered in the
decisionmaking of the final rule.
3.2.16 Dr. Radford provided observations regarding total risks
for the Ontario uranium miners and the Czechoslovokian uranium
miners. There populations are showing very similar risks.
Instead of the risks coming down as purported in BEIR IV, the
risk estimates continue to go up or, at least, stay the same.
This data is backed up by bigger studies, more complete and
extensive, but of shorter duration.
Response: We agree that the tempered dependence of risk
following radon exposure is still uncertain. Further followup of
all miner cohorts are desireable and should continue in order to
settle this question.
3.2.17 The value of 730 deaths per 106person-WLM attributed to
BEIR III cannot be found in BEIR III. The actual figure, as
noted by the EPA, is 440 deaths. Tables 1-1 and 2-13 of BEIR IV,
and Table 6-12 of the BID are misleading.
Response; The 730 number is based on the effective working level
months, a concept EPA no longer uses. - _
3.2.18 The reference in the draft BID at 6-36 to the risk factor
of 830 fatalities per million person-WIM as being an Atomic
Energy Control Board of Canada (AECB) estimate for lifetime
exposure to Canadian males is incorrect; the value cannot be
found in the cited source, the estimate was not accepted by the
AECB, and the so-called best estimate is really the upper end of
the range of risk values.
Response: EPA modifies the AEC BC value (see section 6.4.3, vol..
1 of the BID.
3.2.19 The EPA risk analysis should utilize the 40+ years of
exposure data based on real people and real situations is
available in the records of DOE and NRC (and predecessor
agencies) (1-163)
Response; EPA's risk factors are based on human epidemiological
studies of many kinds, including the atomic bomb survivors. The
Agency knows of no definitive studies of NRC and DOE personnel
that would significantly improve this data base.
3.3 The CAP-88 RISK ASSESSMENT CODE
3.3.1 The AIRDOS code is limited in its treatment of multiple
release points and area sources. These limitations include the
facts that all area sources are treated as circular, multiple
14
-------
release points (whether stack or area sources) are co-located at
the origin of coordinates, and it is questionable if the code can
be used for receptor distances which are less than 2.5 the radius
of the effective circular source.
Response; The EPA's CAP-88 assessment codes are limited with
respect to multiple release points and do treat all area sources
as circular, but have the advantage of practicability.
Sensitivity studies have shown that these modeling
simplifications have little impact on estimated population
exposures. Nevertheless, with respect to maximum individual
exposures, the co-location of all sources at a given site may
sometimes either over- or under-estimate doses and risks
depending upon the release location selected. The CAP-88
assessment codes may be used to assess exposures within 2.5 times
the effective radius of an area source.
3.3.2 Releases are treated as point sources in calculating
exposures which could result in an incorrect estimate of
potential exposure for sites with many sources or covering large
areas.
Response; As noted in II.2.1, the accuracy that is obtained from
the CAP-88 codes is more a function of carefully selecting the
release location for sites with multiple release points. The
treatment of dispersion for area sources is not affected by the
size of the area source.
3.3.3 Dose is calculated at the point of highest air
concentration at ground level which for large sites results in
dose estimates to the public at locations within restricted areas
of the site.
Response; The commenter is mistaken. The maximum individual
exposures and risks reported in the FEIS are either at the point
of maximum off-site dose where detailed demographic data are not
available, or at the occupied location of highest risk where such
data are available.
3.3.4 Use of a Gaussian Plume Dispersion Model limits the
application of the model in simulating transport and dispersion
over complex terrains where terrain-induced flows can result in
significant horizontal and vertical variations in wind speed and
direction.
Response; The EPA is aware of the limitations of the Gaussian
Plume model for complex terrain, and is working to develop
assessment models that provide alternative dispersion models.
However, until such time as these models are working and
documented, the Agency will continue to rely upon the estimates
obtained from the AIRDOS/DARTAB/RADRISK models. Validation
studies of the dispersion portion of the code, which have
15
-------
included sites located in complex terrains, have shown that the
predicted air concentration are sometimes greater and sometimes
lower than field measured values.
3.3.5 There are no provisions for surface perturbations, such as
strong relief or surface roughness, nor particle resuspension.
Response; Although the CAP-88 codes do not directly account for
these factors in estimating environmental concentrations, in
general, these are second order corrections that do not
significantly affect the estimated concentrations and risks.
3.3.6 The set of radionuclides available for use (RADTAB data
set) does not include all radionuclides produced by all
facilities to which the model must be applied, specifically
accelerator/research reactor products and those produced in
medical uses are not included.
Response: The CAP-88 code includes all radionuclides likely to
be produced in airborne emissions from accelerators, research
reactors and from medical uses of radionuclides.
3.3.7 The assumptions regarding emission stack height,
temperature, flow rate, etc. make no provisions for site-
specific.
Response; The CAP-88 assessment codes do allow for the input of
site-specific data regarding release height, stack temperature,
and/or flow rates. While the codes do not allow for both
momentum and buoyant plume rise, this is generally not a
significant limitation. Where default values are used for these
parameters it is due to the fact that the EPA has not been able
to develop and substantiate site-specific data.
3,3.8 The food pathway has no provision for considering pigs,
goats, goat's milk, or poultry.
Response; The Agency believes that consideration of cow's milk
and beef for the ingestion pathway is sufficient to estimate the
likely exposures resulting form the consumption of all
foodstuffs.
3.3.9 The EPA's computer model should be upgraded to reflect
state-of-the-art as recommended by the Science Advisory Board.
Response; The EPA is working on the development of its
assessment codes. However, given the deadlines established by
the Court's, this rulemaking cannot be delayed until the new
codes become available.
3.3.10 CAP-88 should include a screening procedure to be used to
determine whether a detailed analysis should be performed.
16
-------
Response; The EPA disagrees. The CAP-88 code is approved for
use by DOE Facilities. The quantities of radionuclides handled
at these sites and the existence of site-specific demographic and
meteorological data make it reasonable to use the CAP-88 code to
determine the levels of exposure caused by airborne emissions.
3.3.11 The COMPLY program may indicate that the same facility is
both in compliance and not in compliance, depending on how the
data is described. For example, treating an entire stock of an
isotope as one form will enormously change the compliance levels.
Response; Since the facility is required to put the data into
the correct form of isotope, EPA does not consider there to be a
problem with determining compliance.
3.3.12 Very conservative assumptions have been made in an effort
to keep the model simple; the result is that the calculated
population dose is significantly higher and less accurate than
that calculated by other methods currently in use.
Response; COMPLY does not calculate doses that are significantly
less accurate than other methods that are currently in use.
Major differences in dose estimates stem from the fact that ICRP
2 methodology is required by the NRC for certain licensees.
3.3.13 Monte Carlo methods can be used to take into account
uncertainty in the calculation of potential risk to the MIR posed
by radon emissions (from mill tailings).
Response; The Monte Carlo technique has been employed to
estimate uncertainty to the MEI posed by radon emissions. The
results have been presented in Volume I of the final BID (FEIS).
3.3.14 For U-238, Th-230, Ra-226, and Pb-210, the EPA's risk
factors are greater than those given by the ICRP by an average
factor of 2.5.
Response; EPA's risk factors for the above radionuclides are
based on organ specific risk models (chapter 6 of Vol. 1 of the
final BID) and are believed to be realistic for long term chronic
exposure. They are higher than those used by the ICRP.
3.3.15 EPA has accepted ICRP models for alpha-emitting bone
seeking nuclides which yield incorrect dose/risk values.
Response; EPA believes its dose/risk models for alpha-emitting
bone seeking radionuclides are reasonable. (Chapter 6 of Vol. 1
of the BID).
3.3.16 In its discussion of organ weighing factors, the EPA does
not specify which organs correspond to "remainder of organs".
17
-------
Response; The remainder category includes all other ICRP target
organs that were not listed.
3.3.17 The EPA's dosimetry model for immersion dose to direct
plume exposure ignores high energy gamma dose from overhead plume
for locations close to an elevated release.
Response: This is true, but this is seldom an important
consideration. When it could be, we do not use an elevated
release for compliance calculations.
3.3.18 The EPA does not apply a dose rate effectiveness factor
(DREF) which may reduce risk for low-LET doses by a factor of 2
to 10.
Response; The Agency is not persuaded that it is prudent to use
a low dose rate effectiveness factor for low LET radxation with
respect to all kinds of human cancers. This issue is discussed
in detail in Section 6.2.3 of Vol. l of the final BID.
3 3.19 The fx values for short-lived Pu-238, 240, 242 are
different than those published by the ICRP in Publication 48.
Why has the EPA adopted these values?
Response; These values were used in our 1975 guidance for _
protecting the public from transuranics. However, the EPA is
currently revising its values to conform_with the ICRP 48
recommendations, which are numerically similar.
3.3.20 The EPA's assumption of a 570 gram lung target region
conflicts with the ICRP's recommendation of averaging over the
entire 1000 grams of lung tissue.
Response; The EPA averages the dose over the entire lung using
ICRP contemporary models. However, for estimating the risks, the
EPA considers only the dose to the pulmonary region.
3.3.21 There is an inherent contradiction in the EPA's
assessment methodology. It assumes an individual spends 24
hours/day outside for 70 years in calculation of risk from<
particulate radionuclide emission, however in calculating indoor
radon exposure risk, it assumes 75% of the individual's time is
spent indoors.
Response; As a practical matter, there is no contradiction. EPA
assumes very small particulars enter residences without
hinderences, behaving like air.
3.3.22 The EPA's assumption that the radon decay product
concentrations in the air transported from the radon source to
the receptors is in significant state of equilibrium with the
parent radon is very conservative.
18
-------
Response; The EPA calculates the equilibrium fraction as a
function of distance. This simplification, which uses a uniform
wind speed to relate distance to time, greatly reduces the
calculational resources needed to compute precise equilibrium
fractions. The Agency does not believe that this simplification
greatly over- or under-estimates the equilibrium fraction at a
given location.
3.3.23 The definition for "effective dose equivalent" is
inaccurate - "effective dose equivalent is not a "risk-weighted
average" but the sum of the risk-weighted organ dose equivalent
commitments.
Response: The definition has been corrected in the final
documentation.
3.3.24 The EPA emission factors and the NCRP screening model
assumes that the nearest resident is at 10 meters and food
production at 100 meters. This negates the assumption on page
9617 that stack height and area/facility size have negligible
effects because they will have effects at these distances. In
real life, we have not seen residents living at 10 meters or food
production at 100 meters from facilities releasing significant
quantities of radionuclides.
Response: The commenter is mistaken, neither the EPA emission
factors nor the NCRP's screening models are based on the stated
assumptions. The EPA did base the Table of Allowable Quantities
on the assumption that no individual resides within 10 meters and
no food is grown within 100 meters. Facilities, if they exist,
which do not meet this condition, are prohibited from using the
Table to determine compliance. Levels II through IV of the
COMPLY code to consider building wake effects.
3.3.25 Does EPA assume that the regulation is geared towards
protecting people in residences only based on the 10 meter
release point? What about offices, factories, etc. - are they
considered occupied residences?
Response; The regulations cover any member of the general
public. The Table of Allowable Quantities was developed to
minimize the burden of demonstrating compliance on small
licensees while assuring the every member of the public received
the full protection provided by the standards.
3.3.26 Section IV.D.2. apparently does not address precipitation
which may have a significant affect on the dispersion of
pollutants in some parts of the country.
19
-------
Response; Level 4 of the COMPLY computer program has been
revised to take precipitation into consideration.
3.3.27 While the emission factor specified may be appropriate
for known volatile compounds of H-3 or iodine isotopes, we do not
believe that it could be defended for elements which have no
known volatile or gaseous compounds. This factor should be re-
evaluated and reduced where appropriate.
Response; The release fractions were developed as a means of
minimizing the burden of demonstrating compliance while assuring
the public health is protected. Thus, the emission factors
represent upper-bound limits. If a licensee cannot demonstrate
compliance on the basis of the release factors and the effluent
control adjustment factors, then monitoring is both appropriate
and required.
3.3.28 The measurement methods in the proposed rule should be
used as guidance only, i.e. acceptable but not required methods.
Radiation measurement technology is constantly changing and
facilities should have the explicit permission to adopt new
better techniques as they become available.
Response; The Agency has amended the measurements methods to
include those already approved by the NRC. However, it has not
dropped the requirement that use of other methods be approved by
the Administrator.
3.3.29 The SAB recommends against "deliberately producing biased
measurements" when demonstrating compliance, that is the result
that would occur using COMPLY.
Response; An uncertainty analysis, using a Monte Carlo
simulation technique, in support of this rulemaking suggest that
the CAP-88 codes produce median risk (or dose) values on an
annual basis. While the lower levels of COMPLY are conservative,
Level 4 produces results that are not biased and are similar to
those calculated by the CAP-88 codes when the receptor is outside
the building wake zone. At close distances, COMPLY calculates
different air concentrations due to inclusion of building wake
effects.
3.3.30 considering that accidental releases are included in
annual emissions, the EPA should include some additional
provisions in its compliance procedures for dealing with
accidents.
Response; Accidental releases that are not catastrophic can be
adequately handled by the present codes. If the accident is
catastrophic, then compliance with the NESHAPS would be mute and
the EPA's Protective Action Guidelines should be applied.
20
-------
3.3.31 The COMPLY code does not employ a finite modeling
approach but calculates only doses from air immersion/inhalation
which will tend to underestimate doses from an elevated release
and overestimate doses from ground releases.
Response; To the extent that the EPA understands this comment,
it disagrees. The COMPLY code can be applied to both ground
level and elevated releases, and calculates doses based on air
immersion, ground-surface contamination, inhalation, and
ingestion.
3.3.32 The COMPLY computer code has no provisions for site-
specific meteorology and topographical information.
Response; The commenter is misinformed. The COMPLY code has not
provision for site-specific meteorological data at the lowest
levels. Level 4 of COMPLY allows for using site-specific
meteorological data in the form of a modified wind rose showing
how frequently the wind blows from a given direction at a given
wind speed. Although the code is unable to consider
topographical details such as surface roughness, it does take
into consideration building wake effects. Building wake effects
will be more important than other terrain characteristics for
most facilities in the NRC-Licensed category.
3.3.33 The COMPLY code does not allow for stability classes for
calculating dispersion.
Response; Consistent with the recommendations of the NCRP, the
COMPLY code assumes stability class D. This assumption makes a
small difference in most cases and was adopted because it greatly
simplifies the use of the code.
3.3.34 The COMPLY code only provides for one farm at one
distance and that distance is applied in all 16 sectors.
Response; This assumption has been adopted to simplify the code.
3.3.35 Tne COMPLY code has no provision for time dependence to
accommodate for cows on pasture versus cows on stored feed.
Response; Simplifying assumptions have been intentionally
incorporated into the COMPLY code. Some of these assumptions
tend to overestimate risk (residence at one location for 70
years), while other assumptions may .underestimate risk to
selected populations (infants and children). The dose calculated
by COMPLY is not an actual dose but a dose for determining
compliance with the NESHAP.
21
-------
3.3.36 The COMPLY code accepts only information as to the
nearest receptor in each of 16 sectors. The closest receptor may
not always be worst case.
Response: The nearest receptor is at the maximum air
concentration in COMPLY because the Gaussian dispersion factors
use the concentration at the point where the plume touches down
for distances located closer to the point of the release.
3.3.37 The EPA must clarify instructions as to the meaning of
"maximum annual air concentration". This could be interpreted to
mean location with highest annual average air concentration,
location with highest 95th percentile Chi/Q or even location with
highest single concentration reported during the year.
Response: No clarification is necessary. The "maximum annual
air concentration" can be interpreted as the point closest to the
point of release. Although the plume may come down beyond this
point, the diffusion factors used provide for a constant
concentration from the point of release to the point where the
plume touches down.
3.3.38 Requiring the use of the COMPLY code by the nuclear
industry disregards the existence of a carefully developed and
very rigorous methodology already in existence for calculating
risk to offsite individuals, and will result in different values
being calculated and reported to EPA than are reported to NRC
under guidelines of NUREG 0133, Regulatory Guide 1.109.
Response; Much of the NRC's existing framework is outdated,
incomplete, and inconsistent. For material licensees and power
reactors, the NRC uses the outdated ICRP 2 methodology to
calculate doses. Further, building wake effects, accidental
releases, and the contributions from radon decay products
(notably Po-210 and Pb-210) are not included. For uranium
milling facilities, the NRC used ICRP 26/30 methodology, while
the EPA's AIRDOS methodology is used for conversion plants. It
would be difficult for the EPA to be consistent with this system.
3.3.39 The COMPLY code and thorough documentation should be made
available for evaluation and to allow demonstration of compliance
with the regulations.
Response; The COMPLY code is extremely well documented. The
methodology and equations used through Level 3 are from the
NCRP's Commentary No. 3. The COMPLY code and Level 4 are
22
-------
described in the EPA's Background Information Document
"Procedures Approved for Demonstrating Compliance with 40 CFR
Part 61, Subpart I. The Agency has not releases the source code
for COMPLY to preclude the distribution of unauthorized versions.
3.3.40 The COMPLY code's requirement to calculate the building
wake factor at distances beyond which there is a wake is
unnecessary and should be deleted.
Response: The requirement is based on the recommendations of the
NCRP in_its Commentary No. 3. The Agency has no information
indicating the correction is unnecessary.
3.3.41 The COMPLY computer code should be modified to allow use
of the radioisotope and curie file for each release point so that
if the program needs to be rerun these data could then be
available without re-entry.
Response; The COMPLY computer program is intended for
compliance. The EPA does not encourage many reruns of the code
to achieve compliance.
3.3.42 In running the COMPLY computer program, it is not clear
what is meant by "multiple emission vents". Can one emission
point that represents the average flow conditions and accounts
for all radionuclides emitted be used rather than considering 10
release points separately? Dow believes an operator should be
allowed to aggregate emission points. This should not effect
final results, but will make the running of the program much
easier.
Response; Separate release points can be used if they are not
far apart in comparison to the distance to the receptor.
3.3.43 Additional means for determining compliance should be
considered by the EPA.
Response; The EPA is allowing provisions for approving
alternative models that are equivalent to COMPLY.
3-3.44 The EPA assumes that radionuclides build up in the
surrounding area for 100 years, rather than a more realistic 25-
year period, before the exposed population comes into existence.
This increases estimated ground concentrations, surface doses,
and vegetation concentrations by a factor of three.
Response; EPA uses a 35 year half-life value for long half-life
radionuclides on the ground, which modifies the ground build-up
in surrounding areas.
23
-------
4. SOURCE CATEGORY SPECIFIC COMMENTS
4.1. DOE FACILITIES
4.1.1 Basis for the Standards (legal/procedural issues)
4 i 1.1 DOE proposes under Subpart H a 25 mrem/yr effective dose
equivalent standard for emissions to the air from DOE facilities.
The selection of an "acceptable level of risk" for air emissions
at 25 mrem/yr effective dose equivalent is more consistent with
the recommendations of national and international radiation
protection standards-setting organizations than the levels
considered by EPA in the proposed rule.
Response: The Administrator has considered the recommendations
of both the national and international advisory committees in _
reaching decisions on what constitutes safe with an ample margin
of safety. The 10 mrem/y EDE level established by the NESHAP
provides the protection of public health with an ample margin of
safety required by the Act.
4.1.1.2. in Section VII.A.3 of the BID, it is difficult to
determine the basis for the levels of emission control proposed
for DOE facilities. The proposal notes that DOE is well within
the present NESHAP limitation of 25 mrem/yr, produces essentially
no risk, either to individuals or the public at large, and is
reducing emissions further. Given these facts, it is difficult
to identify any basis for a change in the current NESHAP limit to
10 mrem/yr under Approach A. Risks are certainly not known
within a factor of 5, and there is essentially no risk difference
between 10 and 25 mrem/yr.
Response; The assessment of the risks posed by DOE_facilities
indicates that doses are generally well below the limits
established by the existing NESHAP. The change from 25 mrem/y
whole body and 75 mrem/y critical organ to. 10 mrem/y EDE_reflects
the changes that have occurred in health physics and radiation
protection. EPA has been urged to change to the new unit of
effective dose equivalent for some time. For most facilities,
the change in the NESHAP is not a significant increase or
decrease in the stringency of the older limits, but is
essentially equivalent.
4.1.2 Dose and Risk Calculations and Analysis
4121 The BID used emissions based on DOE's 1986 report of
emissions and meteorological data taken from nearby weather
stations. It is DOE's position that site specific methodology
should be used when available. The EPA's use of site specific
24
-------
data in conjunction with the less representative meteorological
data from adjacent weather stations is a contributing factor in
the technical errors in EPA's analysis of impacts from DOE's
sites.
Response: The EPA uses site-specific meteorological data for DOE
sites when such data can be obtained in the STAR format required
by the assessment codes. Changes to the meteorological data
between the draft arid final EIS are documented in Appendix A of
Volume II.
4.1.2.2 The principal release point for the Hanford facility is
stated to be 61 m above the ground; however, a stack height of 10
m was assumed in the risk assessment. It is not possible to
determine whether the assumed 10-m stack height for Reactive
Metals, Inc. is appropriate It is difficult to believe that a
good approximation of the Y-12 stack height could not be
determined. How was the other information, such as which
building is the major effluent source, the effluent filtration
systems, etc., obtained? It is clear that the assumed flow rate
of 200 cf. is an unrealistically low value. There are two values
for U-234 in Table 2.5-1. Table 2.5-4 shows 6000 people with
lifetime risks exceeding 1E-4, but Table 2.5-3 indicates that the
maximum lifetime risk is 8E-5. The effectiveness of the proposed
cleanup depends directly on the fraction that is tritiated water
vapor; the basis for the statement that "much" of it is in that
form is not given. The sum of the U-234 releases listed for
individual Y-12 buildings disagrees with the value for the
facility that is given in Table 2.5-1.
Response; The EPA is limited in its assessments to the data in
the DOE reports. Discrepancies in the DEIS have been corrected
based on these and other technical comments and information
received during the comment period. The discussion of applicable
control technologies and effectiveness is based upon analyses
performed by PNL for the Agency during 1983-1984.
4-1.2.3 It is stated in Section 2.7 that releases are expected
to be double the 1981 values. Examination of the previous BID
shows that releases of U-234 and U-238 have both declined from
0.113 Ci/y in 1981 to 0.02 Ci/y in 1986. For some reason the
doses did not decrease proportionately (88 mrem to the lung in
1981 to 19 mrem to the same tissue in 1986). The heading for the
second column of .Table 2.7-5 is inconsistent with the table
title; the same dollar values are called "HEPA Filter
Installation Cost" in Table 2.7-6. The first paragraph under
Table 2.7-6 doesn't make sense. If the total costs equal AE
costs plus all other costs and the AE costs are 25% of all other
costs, then the total costs will be 5 times the AE costs.
25
-------
Response; The discrepancies noted have been corrected in the
FEIS. The fact that the risks estimated for this facility have
not decreased exactly with the reduction in emissions is due to
use of a slightly different location of the maximum individual
and minor changes in the assessment codes.
4.1.2.4 It is not clear from the discussion on page 2-72 that
the Ar-41 releases at Brookhaven were assessed using the BMRR
stack height of 45 m (plus plume rise), or whether they were
included with some other sources released at much lower levels
(10-18 m) .
Response;
heights.
In the FEIS, the site was modeled using actual release
4.1.2.5 Sixty-one percent of the dose from operation of the
Battelle-Columbus facility is attributed to K-40. Many of the
nuclides in Table 2.19-1 are naturally occurring. The fact that
these nuclides were reported in effluent air samples doesn't mean
that they are effluents due to facility operation.
Response; The Agency assumes radionuclides reported in the DOE's
Effluent Information System are due to process releases.
4.1.2.6 The population around the Oak Ridge National Laboratory
is given as 850,000; a better estimate of the population is
600,000 (p.2-8).
Response; The EPA's estimate is based on results obtained from
the SECPOP computer code. Based upon the latitude and longitude
provided for each facility, the code estimates the 0-80 km
population using 1980 census data.
4.1.2.7 The new regulation is in terms of the effective dose
equivalent, but only organ doses are calculated in the BID. The
BID should estimate the effective dose equivalent so that these
can be directly compared with the new regulation.
Response; As explained in the FEIS, the Agency bases its
decision-making not on doses but on risks. The EDE is chosen for
the NESHAPS limits to allow facilities to implement the standards
using methodologies with which they are familiar.
4.1,2.8. A suggested rewrite of section 2.3.4 follows.
2.3.4 Supplementary Controls
2.3.4.1 LAMPF Main Stack
26
-------
Air activation products from the LAMPF target cells and beam stop
are exhausted through the main stack, which is located near the
center of the experimental areas of the east end of the half-mile
long linear accelerator. Total emissions as measured at the stack
in 1988 were 121,000 Ci. Over 99% of this activity was from
short-lived radionuclides; the main contributors were the 2-min
half-life oxygen-15 (58%), 20.4-min carbon-11 (25%), and 10-min
nitrogen-13 (13%). The only radionuclide that is longer-lived
than carbon-11 was argon-41 (110 min), which contributed 0.4%.
The maximum individual annual dose for a member of the public for
1988 from LAMPF emissions, as calculated by AIRDOS-EPA/RADRISK
and using a 30% reduction for shielding by buildings, was
approximately 9 mrem. However, long-range plans for LAMPF call
for up to 40% increase in annual integrated beam current over
1988, so future maximum doses in the 10 to 15 mrem/year range are
likely if no corrective action is taken.
Several potential methods for reducing the airborne radioactivity
attributable to LAMPF operations have been considered. The large
air flow to the LAMPF main stack (currently about 16,000
cu.ft./min) makes it very difficult to use any existing
technology to remove the gaseous activation products from the air
stream. The most realistic approach would be to provide
significant holdup time to allow decay of the short-lived
components before release. One favorable scheme would be to
utilize a long duct of slowly moving air to provide an extended
decay time. For example, an 8-ft diameter corrugated metal pipe
that is 8000 ft long with a flow of 10,OOO cu.ft./min would yield
a transit time of 40 min. The (roughly) estimated construction
cost of this air holdup system is $3,000,000 (FY-92 dollars)
including connection of the existing exhaust system, suitable
trenching and dirt shielding for the pipe, and a new stack at the
end of the delay line. Table A.I presents the reduction in
radionuclides emissions as a function of holdup time for the
major constituents.
Implementation of the 40-min air holdup will result in a
reduction of the stack emissions by a factor of about 13, with
carbon-11 as the main component remaining at about 0.26 of its
original concentration (25%). An additional dose reduction for
both employees and the public should be realized with this
installation because the most sensible construction plan would be
to extend the decay pipe eastward on the long narrow LAMPF mesa,
resulting in a stack location that produces lower dose to the
surrounding population.
27
-------
Table 1. Effect of hold time of the release of principal air
products from LAMPF.
Half- Percent of Total Decay Factor
Life Activity
Radionuclide fmin) before holdup 20-min Holdup 40-min Holdup
Oxygen-15 2.0 58% 0.001 1E-6
Carbon-ll 20.4 25% 0.51
Nitrogen-13 10.0 13% 0.25
0.26
0.06
Response; The FEIS reflects the emissions from DOE facilities
for 1986, the last year for which complete data were available at
the time that the analysis was made. If future activities at
LASL will result in doses exceeding the limits established by the
NESHAP, then additional controls and/or programmatic curtailments
will have to be implemented to assure that the public receives
the protection provided by the standards. The discussion of
control technology provided by the commentor indicates feasible
emission controls are available to the facility operator
4.1.2.9 The example dose/risk assessment for the Rocky Flats
Plant contains numerous errors, beginning with the radionuclides
and their activities that were used as the emissions source
terms. These are not the nuclides nor the activities that were
reported by the Rocky Flats Plant in its 1986 air emissions
report. The distance to the nearest nearby individuals also is
incorrect; the distance used is within the Plant boundary. If
this assessment is to be used as the technical basis for the
rule, significant revision is necessary.
Response; The assessment of the RFP was based on the emissions
reported in the DOE's Effluent Information System. Attempts were
made to resolve discrepancies between that data base and the
individual facility's annual operating reports. However, time
considerations did not allow full resolution of all such
discrepancies. The estimated exposures and risks from the RFP
were not critical to the decision making for this source
category.
4.1.2.10 The proposed rule does not specifically state whether
the 10 mrem per year effective dose equivalent limit is based on
committed dose or annual dose. ' Although probably implicit in the
AIRDOS/RADRISK calculation methodology, which is specified, we
suggest that the basis needs to be explicitly stated in the rule,
because it is a significant element with regard to radiation
protection.
Response; The standard is for the committed dose that results
from one year's emissions.
28
-------
4.1.2.11 What are the errors associated with the measuring of
radionuclides and flow rates at the control devices? How are
these errors compensated for in the compliance demonstration?
How many unmonitored points of emission exist at these
facilities? What are the estimated emissions from these
unmonitored points of emission? How do the measured radionuclide
emissions compare to radionuclide losses based upon facility-wide
mass balance calculations? The estimation of radionuclide
emissions include the estimate of a control device efficiency
factor. How reliable are these factors? What is the procedure
used to verify these factors? Since control device efficiency
tends to deteriorate over the lifetime of the equipment, how
often are these factors re-assessed? Are the emission rate
calculations based upon the most favorable, the most
conservative, or an average efficiency factor?
Response; The EPA bases its risk assessment of DOE Facilities on
the emissions reported by the individual sites to the DOE's
Effluent Information System. The vast majority of the reported
emissions are based on measured release rates. The
implementation procedures for the NESHAP require the facilities
to determine emissions based on approved monitoring, sampling,
and analytical methods. The error associated with these
measurements are relatively small. EPA belief, based on
engineering studies and general knowledge concerning the release
potential of emission points, is that unmonitored points of
emission do not release significant amounts of radionuclides.
4.1.3 Control Technology
No Significant Comments.
4.1.4 Level of Proposed Standards
4.1.3.1 It is inappropriate and inconsistent with the guidance
of radiation-protection organizations to state that the dose
limits apply to any member of the public. Rather, the dose
limits should apply to reference individuals in critical
population groups.
Response; The standards apply to all members of the general
public. However, the implementation procedures calculate doses
based on "reference man".
4.1.5 Compliance and Implementation Procedures
4.1.5.1 EPA should delete the one-time reporting requirement in
revising the standard, or, at least, exempt DOE sites that
routinely operate at 20% of the standard or below.
29
-------
Response; DOE facilities have already submitted their initial
report. The rule only requires the submission of annual reports
on operations.
4,1.5.2 The point of release in the annual reporting
requirements should be the process areas at DOE facilities rather
than the actual points of release since some larger facilities
have multiple process areas with many separate release points.
Response: EPA does not object if all the release points of a
single building are combined as one weighted average release
point, provided that this is noted in the annual report.
4.1.5.3 The EPA should establish a threshold approach for annual
reporting consistent with other source categories; we recommend a
threshold for compliance reporting at 1 percent of the standard.
Response; EPA is requiring report from all DOE facilities that
release radionuclides. The Agency believes that the limited
number of DOE facilities make these reports practical and useful.
4.1.6 Costs of Compliance
4.1,6.1 Many of the estimates of cost are incomplete and are
therefore underestimates. The level of detail provided varies
greatly by source category. The principal reference regarding
costs for Section 2 "(Mo86)" is not included in the reference
list (there are, however, several references listed which are
never called out in the text). The basis for deciding when to
consider dose reduction alternatives varies from category to
category. In Section 2, it is clearly related to effective dose
equivalent levels; however, in other sections, there is no
estimate of effective dose equivalent (or even of lung dose) and
the criteria appear to be risk level or number of predicted
deaths.
Response; EPA calculated the cost of controls for those DOE
facilities with the highest estimated individual risks. Since
doses and risks are correlated, using doses does not result in a
different analysis than if risks levels are the selection
criterion. Cost estimates were based on reliable information and
only considered in making the ample margin of safety decision.
4.1.7 Other Comments
4.1.7.1 Environmental assessments, prepared by DOE for
construction and operation of new radionuclide sources, should be
reviewed by the EPA prior to issuance of NESHAP approval.
30
-------
Response: DOE facilities will have to apply to EPA for approval
to construct a new facility or modify an existing one in such a
way that will result in increased emissions, as specified in the
regulations of 40 CFR Part 61.
4.2. NRC-LICENSED AND NON-DOE FEDERAL FACILITIES '
4.2.1 Basis for the Standards (legal/procedural issues)
4.2.1.1 Facilities which use limited amounts of sealed source
devices for industrial purposes should be exempted from the
proposed rule. Properly constructed and maintained sealed
sources do not leak and, hence, do not emit radionuclides into
the atmosphere.
Response; The EPA agrees, and facilities that only use or
possess sealed sources are exempt.
4.2.1.2 The regulatory and record keeping burden imposed on
facilities utilizing very small amounts of nuclear material would
be excessive and is out of proportion when viewed in a holistic
sense.
Response: The EPA has spent considerable time and resources to
develop the tiered set of compliance procedures to minimize the
burden on small users while still assuring that the public health
and safety are protected. The Agency does not consider it to be
an unreasonable or excessive burden to require facilities to
account for the amount of material that they possess in the
course of a year, especially since they already have procedures
in place to assure that all such material is properly logged into
and out of their facilities.
4.2.1.3 Clarification is needed as to whether or not the rule
will apply to nuclear facilities operating under the Atomic
Energy Act (AEA) of 1954. These are regulated by the NRG; the
proposed regulations would be duplicative, unnecessary, and
unlikely to produce any significant reduction in the already low
levels of risk.
Response; The rule will apply to nuclear facilities operating
under the AEA of 1954, as specified in the applicability
provisions of the regulations. The CAA provides a different
regulatory framework resulting in standards different than those
promulgated under the AEA. These NESHAPS will provide a greater
level of safety for the public for routine emissions of
radionuclides.
4.2.1.4 Accidental releases should not be subject to the
standard and may conflict with ongoing interagency efforts by
EPA, FEMA, NRC, DOE, and other federal agencies to develop
31
-------
guidance for accidental releases of radionuclides from federally
owned or federally licensed nuclear facilities.
Response; Accidental releases are subject to the NESHAP, and can
result in a violation of the standard. However, the procedures
for implementing these NESHAPS do not involve specific measures
directed at accident prevention. The risks and releases from
accidents are not to be considered in granting approvals to
construct or modify facilities. However, to the extent that non-
routine emissions are to be expected from the operations at the
facility they will be considered. EPA is not planning to develop
any guidance for accidental releases which will conflict with
other interagency efforts to develop guidance for accidental
releases.
4,2.1.5 The EPA should limit the rulemaking on Subpart I to
establishing an emission standard and rely on the NRC to enforce
the standard under existing NRC regulations and requirements,
since this proposed rule could render the U.S. uranium mining,
milling, and fuel-fabrication industries less competitive
nationally and internationally.
Response; EPA hopes to establish an Memorandum of Understanding
(MOU) with NRC to reduce any duplicative requirements in the
implementation of these standards, consistent with CAA Section
112.
4,2.1.6 NRC licensees should be allowed to take exception to the
EPA guidance until such time that comparable NRC guidance is
available.
Response; All regulated licensees will be required to follow EPA
compliance procedures. EPA has received input from NRC in the
development of the compliance guidance.
4.2,1.7 The NRC does not believe that it will have a sufficient
basis to enforce the resulting standard for any of the approaches
in this proposed rule under the current NRC-EPA Memorandum of
Understanding since compliance with even the 10 mrem/yr dose
limit cannot be validated by currently available measurement
techniques.
Response; The implementation procedures do not require
demonstration of compliance to be based' on measurements. One of
the reasons for the development of the COMPLY code is that
environmental monitoring at the levels mandated by the NESHAPS is
not always feasible.
4.2.1:8 There is a definite need for a State to have the ability
to assume responsibility to locally administer the NESHAPS
program; especially Agreement States.
32
-------
Response; EPA will encourage the States to apply for delegation
of authority to administer the radionuclide NESHAPS.
4.2.1.9 It is unclear whether the proposed rule is intended to
cover releases of NARM from facilities licensed by NRG.
Response; The rule applies to all air emissions of radioactive
materials released from facilities covered by the standard.
4.2.1.10 It should be stated in the subject Rule that if there
is an inconsistency between the requirements or standards of the
yet to be promulgated 40 CFR Part 191, Subpart B, regarding
disposal, and those of the subject rule, emissions of
radionuclides to the air resulting from spent fuel or high-level
radioactive waste disposal shall not exceed those established for
NRC-licensed facilities in Subpart I of the subject rule.
Response; There can be no inconsistency between 40 CFR Part 191,
Subpart B and these NESHAPS since high level nuclear waste
disposal operations are not covered by a NESHAP.
4.2.1.11 1-131 used for cancer therapy should be exempt from all
regulation and compliance requirements within the scope- of the
proposed standard. Such an exemption could be justified on
humanitarian grounds and substantiated with a benefit/risk
analysis which clearly demonstrates the value of this activity in
the preservation and extension of human life.
Response; The EPA disagrees, with the suggestion of an
exemption. These standards will not prevent anyone from getting
the 1-131 therapy they need.
4.2.2 Dose and Risk Calculations and Analysis
4.2.2.1 Hospital stacks and vents are normally on top of the
building and hospitals are typically multi-story buildings. The
assumption of a 6- or 15-m release height has no basis in
reality. Nearby individuals within 100 or 150 m will clearly be
in the building's wake.
Response; The FEIS reflects changes in the assumed stack height
for the generic assessment of hospitals.
4.2.2.2 The source term for the large hospital doesn't include
1-131 or other nuclides that might be used for tests and
research.
Response; The assessment in the FEIS has been changed to include
1-131. The assumed source term is based on reported releases. -
33
-------
4.2.2.3 Section 3.3.2.2 states that "actual site data were used
for the risk assessments" but the "stack heights used were all 15
m." Perhaps it is a remarkable coincidence. Tables 3-7 and 3-8
are not consistent; 2 E-4 is given as the maximum lifetime risk.
Most of the dose from Facility D is due to noble gases; reducing
the radioiodine dose by a factor of 100 would still leave an
effective dose equivalent of about 7 mrem/y (for the assumed
conditions).
Response: It is a coincidence that the stack heights were all 15
meters for the radiopharmaceutical manufacturers.
4.2.2.4 Table 3-19 is not consistent with Table 3-18; Facility C
has calculated lifetime risks greater than 1E-4.
Response: The distributions of risk are based on extrapolations
from typical facilities to the entire segment. Thus, there is
some inconsistency in the levels of risks between these
distributions and the estimates made for specific facilities
evaluated due to their greater potential for causing significant
risks. The FEIS has been amended to clarify the level of risk
associated with each of the segments.
4.2,2.5 There are no tables of numbers of people exposed at
various lifetime risk levels for fuel fabrication, source
material licensees, incinerators, or shipyards/DOD reactors.
Response; The distributions are included in the FEIS.
4.2.2.6 Table 3-32 shows no risks above 1E-4 which is
inconsistent with estimates presented previously and with the
text on the same page (3-29).
Response; An asterisks (*) has been inserted in the risk
distribution tables in the FEIS where there is insufficient
information to precisely quantify the risk. In this case it is
known that only a very few people are exposed to those risks.
4.2.2.7 The EPA should allow use of the COMPLY computer program
as an alternative to determining compliance.
Response; The COMPLY code is approved^for use by NRC-Licensed
and Non-DOE Federal facilities.
4.2.2.8 Risk estimates are inappropriate due to failure to
include relevant model parameters such as occupancy time,
shelter, plume buoyancy, site-specific factors, especially those
which' apply to urban facilities.
34
-------
Response; The EPA performs site-specific analyses when it is
appropriate. However, for the NRC-Licensed source category the
number of facilities makes this impractical. As for appropriate
model parameters, the EPA's generic assessments are either based
on facilities defined on the basis of typical sites or are actual
(reference) facilities. The EPA assessments assume 100 percent
occupancy time, and assume a buoyant or momentum driven plume
rise, as appropriate.
4.2.2.9 It is unclear how or why the EPA would propose such
concentration limits in light of the ongoing NRC effort to revise
10 CFR 20. We recommend that the EPA discontinue its efforts to
promulgate this rule, but, failing that, that its implementation
be synchronized with the NRC's efforts regarding 10 CFR 20.
Response; The EPA concentration limits are explicitly developed
for the limits imposed by the NESHAP and the limiting conditions
imposed on their use. They are in no way connected with, nor
should they be confused with, the MFC's that the NRC has used and
is revising in 10 CFR 20.
4.2.3 Control Technology
No Significant Comments.
4.2.4 Level of Proposed Standards
4.2.4.1 EPA has argued that the 10 millirem limit is actually
less stringent than the 25/75/25 provisions of 40 CFR Part 190,
but this is not necessarily true.
Response; EPA believes that for most situations the NESHAP is
somewhat less stringent than the Uranium Fuel Cycle standard, but
recognizes there are situations where this would not be the case.
4.2.4.2 It is not clear what the licensee must do if the
radionuclides in use are not listed in Tables 1 and 2. The limit
for MO-99 in Column 4 of Table 1 needs to be corrected.
Response; The tables now include more than 400 radioisotopes.
If a facility has an isotope that is not listed, it should
contact the EPA for assistance. The Mo-99 limit has been
corrected.
4.2.5 Compliance and Implementation Procedures
4.2.5.1 Meeting this limit and its rules for implementation will
require the Army to undertake extensive development of standing
operating procedures (SOPs) to provide guidance to Army
commanders. The proposed rule appears to lack sufficient,
scientifically sound basis for such an effort.
35
-------
Response; The EPA has prepared implementation guidance for the
regulated community. There is no reason for the Army to
undertake an extensive program to develop SOPs, but should simply
use that guidance.
4.2.5.2 DOE recommends that EPA designate to the NRG the
responsibility for implementing and enforcing compliance with
radionuclide emission standards for NRC-licensed facilities. The
NRC currently regulates radionuclide emissions under existing 10
CFR Parts 20 and 50 and implements the EPA's regulations for
radionuclide emissions under 40 CFR Part 190 and Part 191,
Subpart A.
Response; The EPA and the NRC are engaged in discussions to
decide whether or not the NRC will enter into an MOU to implement
the NESHAP.
4.2.5.3 NRC-licensed facilities should be afforded the same
flexibility to use AIRDOS as is afforded to DOE facilities.
Response; AIRDOS is a much more difficult model for facilities to
use and EPA is concerned that its widespread use would result in
a large number of inaccurate reports. DOE facilities are allowed
to use AIRDOS because they have been trained how to use AIRDOS
correctly.
4.2.5.4 NRC-licensees should be allowed a suitable
implementation period that will allow development of site
specific models and procedures to demonstrate compliance.
Response; The development of the implementation models has
already been completed by the Agency. Since the models were
developed to require the minimum of site-specific information,
there is no reason why additional time beyond that provided for
by the Act is required.
4.2.5.5 Since the data required for demonstration of compliance
with current NRC requirements will not be substantially changed,
the current reporting methodology should be allowed to satisfy
all regulatory reporting criteria (i.e., both NRC and EPA) until
such time as a unified reporting and calculational format is
developed.
Response; The NRC reporting requirements for licensees are not
uniform. The EPA's reporting requirements have been developed to
minimize the burden on licensees. The COMPLY code automatically
generates the required report. In addition, NRC rules do not
require compliance with a specific dose limit of 10 mrem/y to any
member of the public and are, therefore, not applicable.
36
-------
4.2.5.6 For sources with very small inventories, the owner or
operator should be permitted to use process or source knowledge
to demonstrate that potential releases are below the 0.1 percent
continuous sampling level.
Response; EPA has developed a framework to allow facilities
using small quantities of radionuclides to estimate emissions.
These estimated emissions are then used to demonstrate
compliance, either through the use of tables or the computer
model COMPLY.
4.2.5.7 The proposed sampling and analysis plan is so different
from the existing NRC radionuclide emissions reporting procedures
that industry would be required to keep two sets of books. The
monitoring plan proposed by the EPA conflicts with the NRC plan
and would require major re-engineering of existing monitoring
apparatus and substantial capital investment.
Response; Upon review of the NRC's requirements for monitoring
and analysis at major fuel cycle facilities, the EPA has
determined that these licensees can continue to use the
monitoring plans specified in their license. For non-fuel cycle
facilities, the compliance procedures specify calculational
techniques that are believed to allow licensees to demonstrate
compliance without changes to their facility or capital
expenditures.
4.2.5.8 It is difficult, if not impossible, to predict the type
and amounts of radionuclides that might be used in new.medical
facilities and universities.
Response; The standard does not require precise prediction but
the reporting of past actions. However, some estimate of future
uses can be made from prior history. These estimates should
allow any facility to establish its own procedures to ensure
compliance.
4.2.5.9 Problems were encountered with nuclide inventories,
waste form, site configuration, and emissions in trying to
evaluate the prescribed compliance procedures for low level waste
sites. Many of the emissions are not point source releases, some
possible nuclides are not listed even in compliance level 4, and
the COMPLY code and instructions do not address area or volume
sources or packaged wastes.
Response; While the COMPLY code does not explicitly account for
area sources, such area sources can be assessed by assuming the
release occurs at the center of the area. The radionuclides that
may be assessed using COMPLY have been increased to over four
Jiundred.
37
-------
4.2.5,10 The principle shortcoming in regulating the doses
listed in 54 FR 9612 comes in difficulty of measuring
concentration levels as listed in Table 2, Appendix E. For some
particular isotopes, the average natural background radiation in
various parts of the country exceeds this limit. Demonstrating
compliance would be difficult under the EPA's proposed limit.
Response; The EPA's implementation procedures do not call for
environmental monitoring, in part because of the difficulties
noted by the commentor.
4.2.5.11 The EPA must provide more flexible and more realistic
methods for demonstrating compliance than are currently provided
by the COMPLY computer code. Otherwise, it is likely that no
mill can operate without major process or facility modification.
Response; The COMPLY code provides a realistic estimate of the
annual dose to the maximally exposed individual. At level 4, the
COMPLY program is complex enough to account for the conditions at
a mill. No process changes should be required.
4.2,5.12 The EPA should develop guidance documents and training
for the designated contacts to deal with computer problems and
other problems such as emission estimates and area and volume
sources.
Response; The EPA intends to provide training and implementation
guidance for all agencies with compliance responsibilities.
4.2.6 Costs of Compliance
4.2.6.1 NRC licensees will be faced with increased paperwork and
costs from dual NRC/EPA inspections and requirements for analyses
and reporting, applications for facility modifications, and
unfeasible or expensive sampling requirements.
Response; The EPA does not agrees that there will be a
significant increased burden for licensees. EPA procedures have
been designed to limit the burden to licensees. In addition, EPA
hopes to establish a Memorandum of Understanding with NRC to
further reduce any increase in burden to licensees.
4.2.6.2 Compliance with Approach D would be impractical and, if
such a limit were imposed, the production of vital radioisotopes
such as Mo-99, 1-131, Xe-133, 1-125, P-32 and others would have
to be discontinued. Since 1 in 4 hospital patients in the United
States have some nuclear diagnostic test performed with one or
more of these isotopes, Approach D would have detrimental effects
on the general health and welfare of the nation.
Response; The EPA has not promulgated the limits proposed under
Approach D.
38
-------
4.2.6.3 The EPA should balance the minimal risks associated with
exposure to low levels of radiation with the risk of inhibiting
technological advances in biomedical research.
Response;^EPA does not believe that the standard will inhibit
advances in biomedical research, since current operations comply
with the standard.
4.2.7 Other Comments
4.2.7.1 The final rulemaking should clarify the status of low-
level radioactive waste disposal facilities; the intent of the
CAA and Subpart I would appear to focus more on operations and
stored wastes than on properly closed disposal sites.
Response: Closed disposal sites are covered under the NESHAP.
4.2.7.2 The applicability of Subpart I to temporary work sites
and outdoor work sites must be further evaluated; the standard
applies only to licensed facilities.
Response; The standard applies to all NRC-licensees and includes
all activities under the control of a licensee whether indoor or
outdoor, permanent or temporary.
4'2'7!3 . The annual possession worksheet may overestimate
quantities onsite at any particular time since it does not
account for quantities being shipped out; license possession
limits^may be a more realistic quantity for calculation of
potential release. The concentration worksheet does not
accumulate concentrations of the same nuclide from different
stacks or vent; this will underestimate quantities released.
Response; Since the NESHAP imposed an annual dose standard, the
relevant quantity is the amount of material handled in one year.
The EPA_recognizes that this is different than the NRC's
possession limits. The possession limit is not a useful tool
because it only limits the total amount of radioactive material
the licensee can possess at any one time. Some facilities only
use a small fraction of their possession limit, while others can
use many times their possession limit during the year.
The implementation system does not account for material that is
shipped out, in those cases where the package has been opened.
The_system of release fractions is based on the amount of
radionuclides that would escape when a larger quantity of
material is used. If EPA was going to allow deductions for
material shipped out then it would have to use a different, and
higher, set of release fractions. The compliance procedures do
account for the additive exposures from multiple release points.
39
-------
4.2.7.4 EPA needs to consider the potential impact of the three
proposed dose limits on the development of new low-level waste
disposal sites. Having to demonstrate compliance with release
limits of 3 millirem/year or 0.03 millirem/year could limit the
viability of engineered alternative disposal methods now being
considered by States.
Response; EPA has considered the effect of the standard on all
licensees at the second, ample margin of safety step. The final
result of the two step process is a standard of 10 mrem/y ede.
4.2.7.5 It is not clear whether the effluent discharge
concentrations in Table 2 of Appendix E are applicable to
releases from only one stack or whether they represent averaged
or summed concentrations in those cases where the licensee
maintains several such release points.
Response; The worksheets in the guidance document indicate that:
(1) each nuclide released from the same stack or from different
stacks must be summed; and (2) the same nuclide released from
more than one stack should not be summed although the highest
concentration from any stack should be used. This level of
detail is not provided in Appendix E.
4.2.7.6 Will the NESHAP limits override the current MPC or
proposed BAG limits of 10 CFR Part 20?.
Response; Only when they are more restrictive.
4.2.7.7 The proposed regulations inappropriately group
NRG/Agreement State-regulated facilities in the same category.
Modeling assumptions are based on the largest emitters, meaning
that nuclear power plants, hospitals, and universities are all
considered and treated equally. Hospitals, universities and most
radiopharmaceutical firms emit far smaller amounts of a very
limited inventory of radionuclides than do nuclear reactors. The
EEA's claim in VII-B-1 that these facilities emit large numbers
of radionuclides is not necessarily true.
Response; The EPA's assessment of the facilities that are
licensed by the NRG show that the doses and risks received by the
individuals at greatest risk are roughly comparable across many
of the segments including power reactors, radiopharmaceutical
manufacturers, and hospitals. Even though individual licensees
may emit only a handful of radionuclides, as a group NRC-
licensees do emit a large number of different radionuclides.
40
-------
4.3 URANIUM FUEL CYCLE FACILITIES
4.3.1 Basis for the Standards (legal/procedural issues)
4.3.1.1 If any such new regulations are deemed necessary, before
any new radiation standards are promulgated the EPA should
implement the Memorandum of Understanding with the NRC to develop
such a regulation on a joint basis, taking account of both the
adequacy of the existing limits wherever possible and the need
for consistency in methodology even where new standards are felt
to be appropriate.
Response; EPA and NRC are currently discussing the possibility
of creating a MOU which would help to reduce the regulatory
burden on all NRC-licensees, including fuel cycle facilities.
EPA hopes that these discussions can be brought to a successful
conclusion. Unfortunately the court-ordered deadline for the
promulgation of these rules did not allow enough time for these
discussions to be completed. NRC has been actively involved in
the development of the rulemaking. EPA has reviewed and approved
current NRC sampling and analysis procedures for fuel cycle
facilities.
4.3.2 Dose and Risk Calculations and Analysis
4.3.2.1 It isn't clear why the fuel fabrication facilities are
not analyzed on a site-by-site basis. There are very few
facilities compared to other source categories which are all
analyzed on a plant-by-plant basis.
Response; Given the level of risk estimated for the fuel
fabrication facility with the greatest potential for emissions
and the limited time available to the Agency under the Court
order, it was determined that site-by-site analysis of this
segment of the fuel cycle was not needed.
4.3.2.2 It is not reasonable to analyze only one
"representative" of 100 reactors. In addition, one has only to
look at the GSDs for the release rates (Tables 4-23 and -24) to
realize that there are problems with this assessment. Apparently
no consideration was given to whether the reactor operated for a
long or short period during the year, whether it was a new plant
or an old one, etc. One of the features of AIRDOS is that many
different radionuclides can be analyzed, so why were surrogates
for radionuclides used? It isn't clear that the surrogate
release values were chosen to be comparable on a dose equivalent
basis. A reasonable sample of facilities should be analyzed on a
site-by-site basis using a credible set of source terms. Onsite
meteorological data are available at all the plants. In fact,
analyses for the plants have already been performed and can very
likely be found in the same reports used to construct Table 4-31.
41
-------
Response; -Given the time and resources available to the Agency,
the decision was made to evaluate power reactors based on model
reactors typifying facilities with emissions near the mean. Data
reported by individual plants were used to estimate the upper
bound of individual risks. In the way the model plants were
defined, the operating history of the individual plants for that
year were not relevant. Surrogates, chosen on the basis of the
contribution to dose, were only used to define the emission rates
for other radionuclides. Appendix A to Volume II shows the
actual source terms that were run. Within the time and resource
limitations, on-site meteorological data in a suitable format
were not available to the Agency. Finally, while assessments of
all plants have been made, and are indeed made annually by PNL,
the assessment methodology and dose/risk factors are not
consistent with those used by the EPA.
4.3.2.3 The maximum individual risk, which is slightly higher in
the calculation than the EPA preferred value of 1 x 10 lifetime
risk, is to a hypothetical maximum exposed individual, with many
conservatisms included in the calculation, exposed to the
effluent from a mill. Use of such a hypothetical worst case
calculation to conclude that the real risk is high enough to
warrant additional regulation for the whole UFC is not warranted.
Response; The level of the risks calculated for the facilities
in the UFC source category are but one of the factors that
influenced the Administrator's decision to promulgate a NESHAP.
As noted previously, the Act provides important mechanisms that
provide additional assurance that the public health will be
protected from routine operational releases.
4.3.2.4 As currently written, the standard indicates that the
EPA believes, and assumes, that a plant being regulated will
operate right at the limit every moment of its operating life.
This is an erroneous assumption.
Response; The EPA makes no such assumption. In fact, the Agency
believes that most plants will operate as far below the limits as
is practical.
4.3.2.5 Risk handling should consider risk/benefit trade-offs
and the relative risks of alternatives to the risk-producing
activity, including the costs and risks inherent in the
termination of that activity.
Response; The EPA agrees these consideration are appropriate for
the second step, determining ample margin of safety.
4.3.3 Control Technology
No Significant Comments.
42
-------
4.3.4 Level of Proposed Standards
4.3.4.1 The NCRP concluded in September, 1987 that about one-
tenth of one percent of the radiation dose to the average person
in the United States results from the uranium fuel cycle. This
is hardly an amount worthy of additional regulation especially
given the fact that average natural background radiation levels
vary by 100 mrem in a year across the U.S..
Response; EPA regulations under the Clean Air Act are designed
to protect the maximally exposed person as well as the average
person.
4.3.4.2 Use of the EPA methodology, because of its many
conservatisms, effectively results in significantly lowering the
standard limit.
Response; The EPA believes that its compliance procedures result
in reasonable estimates of actual exposures.
4.3.5 Compliance and Implementation Procedures
4.3.5.1 Estimated doses to the maximum exposed individual off-
site are routinely calculated by our facilities and reported to
the NRC using a carefully developed system of sampling, analysis,
and reporting. The NRC's system of inspecting and enforcing
compliance is well developed. The data from our facilities
produced in compliance with this rigorous methodology demonstrate
that the emissions have represented and continue to represent an
acceptable level of risk.
Response; The EPA has approved existing sampling and analytical
methods for UFC facilities. However, since the NRC's
calculational techniques are not uniform, and some are based on
ICRP 2 methodology, the EPA has determined that doses must be
calculating using its compliance procedures. EPA hopes that it
and NRC will conduct negotiations to create an MOU to define the
roles each will be responsible for in implementing the NESHAP for
NRC-licensed facilities.
4.3.5.2 The proposed additions to 40 CFR Part 61 do not contain
any provision to accommodate variances for unusual operations.
It should recognize that temporary or unusual operating
conditions may exist where standards could be exceeded for a
short period of time and continuing operation may be in the
public interest. Existing regulation allows such variances.
Response; The EPA limit is from exposure to total emissions over
the course of a year. This should provide adequate operation
flexibility.
43
-------
4.3.5.3 The EPA's proposed reporting system would result in
redundant and conflicting recordkeeping and reporting
requirements.
Response; The EPA believes that its recordkeeping and reporting
requirements are reasonable and do not impose an unreasonable
burden on licensees.
4,3.5.4 Use of the prescribed simplistic COMPLY methodology for
all plants is frequently inappropriate because of the extensive
site-specific data employed in the current NEC mandated
methodology. Demonstration of compliance should be allowed for
nuclear power plants by employing existing NRC-approved methods.
This existing methodology would more accurately represent the
doses and risks to "real" people in the environment.
Response; As noted above, the EPA has prescribed the COMPLY
methodology because the manner of calculating doses is not
uniform under the existing NRC regulatory framework and in some
cases relies on outdated ICRP 2 methodology.
4.3.5.5 EPA's proposed compliance procedures in this rulemaking
reflect a total disregard of Section 122(c) of the Act and the
EPA-NRC memorandum of understanding implementing that provision.
Response; The EPA disagrees, the compliance procedures have been
developed with the input of the NRC.
4.3.5.6 Should a radionuclide NESHAPS for nuclear power plants
be promulgated, the EPA should include provisions for a long
implementation period.
Response; The EPA knows of no reason why power plants require a
long implementation period. The compliance procedures have been
developed and are extremely straightforward, and are ready for
implementation 90 days after the effective date of the standard
as required by Section 112.
4.3.5.7 Satisfying both NRC and EPA prescribed methods may be
difficult, and perhaps impossible. It would be very difficult to
monitor flow or sample for iodine being released from a
pressurized water reactor vent, but continuous flow measurement
and sample collection would be required by the current EPA
proposal. Apparently any plan, other than continuous off-line
sampling, would not be allowed by the proposed rules; if one
wished to employ a different method, prior approval would be
required. There is no basis given for this added requirement.
44
-------
Response; The final rule permits the determination of releases
using the monitoring provisions, calculational techniques, and
analytical procedures incorporated in each fuel cycle facility's
technical specifications.
4.3.5.8 Section 61.104 requires an annual report to include
certain details regarding the facility which for our plants would
be voluminous and unchanging. Such requirements are
inappropriate for a large facility with unchanging process
parameters, such as nuclear power plants.
Response: The EPA agrees that the report could be voluminous.
However, the information requested is essential to enable the
Agency to make a valid determination of whether or not the
facilities are indeed complying with the standard. Given modern
reproduction methods, facilities should be able to reproduce and
submit this information on an annual basis.
4.3.5.9 Paragraph 61.107 should be changed such that an
application under Paragraph 61.107 would not be required for
construction of a new facility or modification of an existing
facility if it has already been evaluated and accepted by the
NRC.
Response; The EPA disagrees, EPA has a legal responsibility to
ensure that the new facility will meet the requirements of its
regulations; NRC approval is not designed to provide that
assurance.
4.3.5.10 The 40 CFR Part 61 Proposed Rulemaking does not include
any discussion of accidental release or associated
limits/guidelines. If emissions during accident conditions are
subject to this UFC NESHAP we recommend concise guidance be
provided by the EPA as an integral part of any rulemaking.
Associated EPA Protective Action Guidelines should be revised as
necessary.
Response; Accidental releases are governed by the standard and
can result in a violation of the standard. EPA intends to
develop procedures to assist the regulated community in
determining whether they are in compliance in accident
conditions. EPA's Protective Action Guidelines will be reviewed
to determine whether they should be revised to include discussion
of radioactive NESHAP compliance.
4.3.6 Costs of Compliance
4.3.6.1 The EPA's proposed standards would waste industry and
government resources demonstrating and evaluating compliance at
facilities that are already subject to stringent standards and
compliance requirements.
45
-------
Response; These standards will provide additional protection to
the public from the risks of radionuclides from routine releases
without requiring the wasteful expenditure of time or money from
the regulated community.
4.3.6.2 The estimated cost is greater than $1,000,000 per plant.
This substantial sum of money will not reduce risk to the public
and will not significantly improve effluent quantification.
Response; The estimate of $1,000,000 per plant is wholly
unsubstantiated and the EPA is not aware of any costs associated
with this NESHAP beyond some procedure modifications and the
minimal time needed to determine compliance and prepare the
required reports.
4.3.6.3 The EPA has not provided a valid cost-benefit analysis
in the record.
Response; This level of these standards was not determined using
cost/benefit analysis and no such analysis is required.
4.3.6.4 Stringent effluent limits could increase significantly
the price of nuclear fuel as a result of increased costs in the
UFC facilities which would also be required to install expensive
equipment having no commensurate benefits. Thus, stringent
effluent limits could lead to substantial increases in electric
rates for equipment which does not substantially increase public
health and safety.
Response; The EPA does not have any information to indicate that
fuel fabricators will have to install additional effluent
controls to comply with the NESHAP.
4.3.6.5 The dose limit of 3 mrem/yr may force other nuclear
facilities throughout the nation to operate under load
restrictions. A reduction in nuclear power generation could
impact the nation's economy, and impede the nation's efforts to
reduce chemical emission pollutants.
Response; The promulgated limit is 10 mrem/y EDE.
4.3.7 Other Comments
4.3.7.1 Making the citizen suit provisions of the Clean Air Act
available will not provide a significant benefit. The industry
has pervasive regulatory oversight. It is very difficult to
imagine a situation where a plant would have a continued
violation that could adversely affect public health and safety
and then for the violation to continue after the NRC has been
notified.
46
-------
Response; NRC rules do not require that routine emissions be
kept low enough to ensure that no member of the public receives a
dose greater than 10 mrem/y ede. In addition NRC does not
require its licensee to make such calculations as would be
required to find out whether or not a facility provides this
level of protection to the public. Therefore, it cannot be
assumed that the NRC would prevent members of the public from
receiving doses in excess of the acceptable level, of 10 mrem/y
ede. In addition, EPA recognizes the Congress believes that
provisions establishing the right of citizen suits do provide
real additional protection to the public.
4.3.7.2 The EPA's analysis should consider the impacts of the
alternative fuels which would substitute for the UFC; there is no
discussion of increased CO2 emission rates, the greenhouse effect
or global warming.
Response; The EPA's analysis does not consider these effects as
there is no information available to indicate that the
promulgated rule will have an adverse impact on the generation of
electricity by nuclear reactors, causing any increase in such
effects.
47
-------
4.4 HIGH LEVEL WASTE FACILITIES
4.4.1 Basis for the Standards (legal/procedural issues)
4.4.1.1 The EPA should not regulate air emissions from WIPP and
MRS under the Clean Air Act. In the DEIS the EPA determined that
the facilities that are covered under the High-Level Waste
Disposal Facilities source category (i.e., the Waste Isolation
Pilot Plant, High-Level Monitored Retrievable Storage (MRS)
facility, and the Waste (HLW) repository) are designed with
state-of-the-art effluent control systems enhanced by performance
requirements of the waste forms and packages. The DEIS indicates
that each of these facilities projects a lifetime cancer risk to
individuals in the vicinity of less than 10~6. However, the EPA
proposed no NESHAPs only for HLW disposal. We believe it was the
EPA's intent not to propose NESHAPs for the WIPP or an MRS
facility, as well; however, the EPA did not specifically exclude
WIPP and MRS from the proposed regulation. We request that the
EPA clarify its statement not to propose NESHAPs for WIPP and MRS
consistent with its position not to propose NESHAPs for HLW
disposal.
Response; EPA is not setting a standard regulating air emissions
from high-level waste disposal facilities.
4.4.2 Dose and Risk Calculations and Analysis
No Significant Comments.
4.4.3 Control Technology
No Significant Comments.
4.4.4 Level of Proposed Standards
4.4.4.1 In consideration of the factors and rationale that DOE
has used in determining and recommending an acceptable or safe
level, no further reduction below the safe level is required to
provide an ample margin of safety.
Response; The Administrator has determined that the releases
from the HLW Disposal source category do not require a NESHAP to
assure the protection of the public health.
4.4.5 Compliance and Implementation Procedures
No Significant Comments.
4.4.6 Costs of Compliance
No Significant Comments.
48
-------
4.5 ELEMENTAL PHOSPHOROUS FACILITIES
4.5.1 Basis for the Standards (legal/procedural issues)
4.5.1.1 We feel it is inappropriate to reduce levels of Po-210
emissions below the level of regulation finalized in the February
6, 1985, rulemaking.
Response; The Court decision under which this rulemaking has
been conducted imposed a two-step methodology for determining the
level of emissions that protect the public health with an ample
margin of safety. The risk level that the Administrator has
determined is acceptable requires limiting Po-210 emissions to
less than the amount allowed under the previous NESHAP.
4.5.2 Dose and Risk Calculations and Analysis
4.5.2.1 The meteorological data used by the EPA in the
assessment for Soda Springs is in error. The predominant wind
direction is not out of the north blowing directly toward the
Soda Springs population center, but rather out of the
south/southeast and south/southwest blowing away from Soda
Springs. This information, obtained from our onsite EPA-approved
meteorological monitoring station, was provided to the EPA.
Response; The draft assessment used a data set from the on-site
meteorological tower. Subsequent review indicated that the data
set was invalid. The additional data provided by the plant was
not in a form suitable for input to the assessment codes.
Therefore, meteorological data from the nearest airport was used
in the final assessment. Review of this data set indicates that
the predominate wind directions closely approximate those
reported by the plant.
4.5.2.2 The population in the 80-km radius of the Monsanto Soda
Springs plant should be 75,000 rather than 100,000.
Response; The population cited is based on the results obtained
from the SECPOP computer code which uses the 1980 census data.
4.5.2.3 EPA's assumptions of class Y solubility for the
radionuclides and a sensitive lung mass of only 570 g have
resulted in gross overestimates of dose. Also, EPA used the f1
factors of class D for Pb and class W for Po, thereby treating
the Pb-210 and Po-210 as insoluble when applied to the lung dose
and then treating them as soluble compounds when applied to the
systemic dose. EPA cannot have it both ways. Also, ICRP
recommends 1000 g for the lung mass rather than 570 g.
Response; EPA's use of Y solubility for radionuclides is based
on solubility tests conducted on samples collected from calciner
stacks at elemental phosphorus plants. The use of a 570 gram
49
-------
lung mass is done because the estimate of risks is based only on
the dose to the pulmonary region of the lung. Since almost all
the risks from radionuclide emissions for elemental phosphorous
plants is due to the radiation dose to the lung, the fl factors
used do not significantly affect the risk assessment.
4,5.2.4 The use of unreasonable default values by the EPA for
environmental weathering, soil density, dairy and beef cattle
populations and vegetable gardens is indefensible. Environmental
weathering is usually represented with a 14-day half-life (the
EPA used 35 years). The EPA chose a soil density of 220 kg/m
where 1,000 to 1,200 kg/m3 would be more appropriate.
Response; EPA uses a 14 day half-life for the removal or
weathering of radionuclides deposited on vegetation. The 35-year
half life referred to in the comment applies only to the removal
of radionuclides from soil. In addition, the EPA value for soil
density is used for only top 15 cm of soil, for which such value
is appropriate. Finally, none of the factors mentioned could
significantly affect the risk assessment for elemental
phosphorous plants since almost all the risk is from inhalation
and exposure to the lung.
4.5.3 Control Technology
4,5,3.1 Calciner technology used at the various elemental
phosphorus plants are different and must be considered
individually. For example, the grate calciner at FMC typically
has only 0.5% of the feed in the off gas while others can have as
much as 25%. Also, the grate calciner generates up to 2.5 times
as much off gas as a nodulizing kiln. Such process differences
require much different design criteria when selecting appropriate
control technology.
Response; The estimates of performance for venturi scrubbers and
wet ESP systems are based on actual particle size distributions
collected at the Monsanto and FMC facilities and do not assume
that the distributions are normal or the same for both types of
calciners. The SD/FF and HEPA filter systems are relatively
insensitive to particle size. The performance estimates for
these systems are conservative for the particle size
distributions found for both types of calciners.
4.5.3.2 The information presented in the BID on control
technology is based on assumptions which are not supportable.
The stack data collected on FMC's calciners do not support, the
pressure drop vs. emission assumptions used in the BID.
Response; The relationship between pressure drop and venturi
scrubber performance is well established in the industry and is
used by most control device manufacturers in system design.
Because the Po-210 is concentrated in the fine particle fraction
50
-------
at FMC, the removal efficiency is quite low for pressure drops of
6.5 and 10 in. w.c. Consequently the difference in emissions
from these systems is estimated to be quite small (less than 10
percent). The differences cannot be measured reliably because of
inherent process variation and uncertainty in the sampling and
analysis methods. Comparable emissions rates such as those found
at FMC are expected given these uncertainties.
4.5.3.3 Using venturi scrubbers, the reduction in Po-210
emissions stated in Table 6-17 of the BID are not achievable at
the stated pressure drop. At a cost of $7 million, FMC installed
new venturi scrubbers on both caleiners that were capable of a
pressure drop of 10 inches w.c. Based on this data, the capital
costs presented in Table 6-19 of the BID for venturi scrubbers
appears to be low.
Response; The achievable reductions depend in part on the
fraction of the radionuclides associated with the small particle
sizes (see response to comment IV.5.3.4). The costing procedures
used to develop the estimates presented in the BID are those
outlined in the EAB cost manual. These procedures have been
extensively used by the EPA to develop air control device costs
for regulatory assessments and generally are adequate to provide
reasonable costs with the ±30 percent accuracy. The wet
scrubber costs are based on standard manual estimates adjusted
for high alloy construction. The cost estimates for venturi
scrubbers compare reasonably well (10 to 40 percent difference)
to costs of systems installed by the industry within the last 2
years. The EPA believes that the capital cost estimates used for
the regulatory analysis are reasonable.
4.5.3.4 Of the four systems, HEPA filters would seem to be the
least desirable. There is no data on the performance of HEPA
filters on combustion systems or high-temperature furnaces.
Installation downstream of a wet scrubber would appear to greatly
shorten the life of disposable filters. This technology would
have to be carefully examined before a system could be designed
for consideration.
Response; HEPA filter systems were selected as one of the high-
efficiency particulates control systems, although they have not
been used on elemental phosphorus plants. However, they have
been successfully used to control radionuclide emissions from
uranium process plants and fine particulate emissions from high-
volume air recycle streams at lead battery plants. Also, HEPA
filters were selected because they provide a much greater level
of control than is provided by the other control alternatives
that were evaluated.
A major operating cost for HEPA filters is filter replacement,
and cost estimates for HEPA filters include the costs of filter
replacement and waste disposal. As a result, although the
51
-------
capital costs for HEPA filters is the lowest of the four control
technologies evaluated, the total annualized cost is higher than
the other technologies.
4.5.3,5 Controls for Po-210 at elemental phosphorous plants may
not provide adequate protection from the abnormally high Rn-222
emanation rate.
Response: Radon-222 releases were included in the source terms
evaluated for these plants. The evaluation indicates that the
radon-222 is not a major contributor to the total risk.
4.5.4 Level of Proposed Standards
4,5.4.1 The regulation of radionuclides from elemental
phosphorus plants is not warranted.
Response; The EPA disagrees, its risk assessment demonstrates
that current emissions do not result in an acceptable level of
risk to public health and do not protect public health with an
ample margin of safety.
4.5.4.2 In the February 6, 1985 rule-making, the EPA
Administrator correctly established that the risks from
radionuclide emission from elemental phosphorus plants are "very
small" and the limit of 21 Ci/y of Po-210 from calciner stacks
effectively limited emissions.
Response; During this reconsideration, the Administrator
reexamined the risks from elemental phosphorous plants and has
determined the conclusions in the February 6, 1985 rule-making
were incorrect and that 21 Ci/y does not provide the level of
safety with an ample margin required by the law and the Court's
guidance on determining the appropriate levels.
4.5.4.3 The Fort Hall Reservation is highly exposed to, and
affected by, radionuclide emissions from elemental phosphorus
plants. Thus, we urge EPA to adopt the emission standard of 0.6
Ci/y of Po-210 proposed in Approach C which would provide ample
health risk protection and also represents the application of the
best available control technology of a high energy scrubber wtr.ch
has proven to greatly reduce Po-210 emissions. We feel that
Approaches A and B do not provide the surrounding community T,ith
an ample margin of safety required to protect the public he?.lth
and therefore would not be in keeping with the intent of the'_
district court remand decision or Section 112 of the Clear Air
Act.
Response; The Administrator has determined that a NESFAP limit
of 2 Ci/y will assure the protection of public health with an
ample margin of safety.
52
-------
4.5.5 Compliance and Implementation Procedures
No Significant Comments.
4.5.6 Compliance Costs
4.5.6.1 Information is not presented on the ability of a wet ESP
or a spray dryer/fabric filter to remove less than 0.5 urn Po-210
particulates. Therefore, pilot work would be required. FMC
estimates installed costs of $40 - 80 million for the wet ESP's,
much higher than presented in Table 6-19 of the BID, and $25 - 50
million for the spray dryer/fabric filter system.
Response; The fractional efficiency for the wet ESPs are based
on 1) data from measurements of a wet ESP system applied to a
calciner at an elemental phosphorus plant that included particle
sizes below 0.5 micrometers, and 2) accepted design calculations.
These fractional efficiency estimates did account for removal of
particles less than 0.5 micrometers in size. In general, fabric
filtration technologies have been found to be insensitive to
inlet particulate matter loading and particle size distribution
as long as properties of the inlet particulate matter allow good
cake formation on the bags. Past experience with the SD/FF
system with the lime slurry on similar process streams have shown
this system to be particularly insensitive to the particle size
distribution in the exhaust gas streams. These systems have
demonstrated the ability to remove volatile metals such as
arsenic from combustion source and metallurgical process exhaust
gasses at greater than 99 percent efficiency. The EPA believes
that the estimates of 99+ percent removal used for the analysis
are reasonable.
4.5.7 Other Comments
4.5.7.1 There is no indication that EPA used any of the ambient
monitoring data from the Southeast Idaho Radiation Exposure Study
to verify the AIRDOS dispersion model predictions. The EPA's
Proposed Guidelines for Exposure Assessment (49FR46304) states,
'"When the estimates of environmental concentrations are based on
mathematical models, the model results should be compared^to
available monitoring data, and any significant discrepancies
should be discussed. Reliable, analytically determined values
should be given precedence over estimated values whenever
significant discrepancies are found. The monitoring data from
the study indicate that the Soda Springs area currently has
essentially background levels of Po-210 in the ambient air.
Response; EPA did review and evaluate data from the Southeast
Idaho Radiation Exposure Study in developing standards for
elemental phosphorous plants.
53
-------
4.5.7,2 The EPA did not adhere to the stack test methods for
point source particulates as described in Method 5 (40CFR60) when
obtaining a sample for radionuclide analysis. Dilute nitric acid
was used as a rinse of the probe after the acetone, and nitric
acid was also used in the impinger solutions instead of a
specified grade of water. The approved method states that no
modifications shall be made without the Administrator's approval.
Response; The slight modifications to Method 111 were made to
assure complete recovery of the polonium-210 and to provide a
more accurate measure of polonium-210 emissions.
4.5.7.3 The particle size characterization of emissions is
inadequate for exposure assessment purposes. The basic problem
is that the measurement technique used could not accommodate the
particle size distribution encountered. The 0.3 um diameter
assumed in the EPA assessment was obtained by an extrapolation
beyond measured data. Thus, the 0.3 um diameter is suspect and
we feel it could easily be 0.2 um or less.
Response; Differences in particle size between 0.3 and 0.2
micrometers would have only a small impact on the estimated
risks.
4.5.7.4 We ask you to waive the emission testing requirements
for both Occidental Chemical Company's elemental phosphorus units
at Columbia, TN based upon favorable emission test results.
Furthermore, if the waiver of emission testing is granted, we
assume that the requirement to install a device for measuring the
phosphate rock feed to the kiln during service testing is also
waived.
Response;
testing.
The new rule no longer allows for waivers from
54
-------
4.6 COAL-FIRED UTILITY AND INDUSTRIAL BOILERS
4.6.1 Basis for the Standards (legal/procedural issues)
4.6.1.1 Regulation of coal-fired boilers under the Approach D
alternative is unwarranted. Regulation of any radionuclide
source category under Approach D is unrealistic and unjustified
because of the conservative assumptions in this approach and
because the releases fall well below those considered to be an
acceptable risk to the public. Approaches A, B, and C were
determined not to warrant regulations of radionuclide emissions
from coal-fired boilers, even at the ample margin of safety
decision.
Response;
boilers.
The EPA has not promulgated a NESHAP for coal-fired
4.6.1.2 Emissions from coal-fired boilers are presently
regulated under the National Ambient Air Quality Standards
(NAAQS) for particulate matter. In addition, the new larger
coal-fired boilers have to meet NSPS requirements for other
pollutants such as particulate, sulfur dioxide, and other
hazardous pollutants.
Response; The EPA is aware of the existing regulations that
apply to coal-fired boilers, their existence was considered in
determining that standards would not be issued.
4.6.1.3 If the Agency intends to follow a two-step decision-
making approach, and therefore make a "best estimate" of risk as
part of the initial step, it must also reduce conservatism in the
following areas: (1) average U-238 concentration in fly ash, (2)
stack height and temperature profiles, (3) food transfer factors,
and (4) the linear, non-threshold dose response model.
Response; The value for uranium in fly ash is based on the
geometric mean value of uranium in coals and enrichment factors
based on monitoring data. The stack heights and temperature
profiles used are also based on the actual plant parameters. The
basis of the linear no-threshold dose response assumption and the
food transfer factors are discussed extensively in Volume I of
the FEIS.
4.6.2 Dose and Risk Calculations and Analysis
4.6.2.1 What is the meaning of the enrichment factor for the two
radon nuclides in fly ash? Does it mean that the radon wasn't
released during combustion? The post-combustion concentration of
radon in fly ash is 20 times greater than the original
concentration in the coal.
55
-------
Response; The enrichment of radon in the fly ash would be more
accurately stated as the enrichment of radon in the effluent off-
gas. The factor of 20 is based on the EPA's monitoring results.
4.6.2.2 While it is clearly not reasonable to expect the Agency
to perform site by site evaluations of 1200 coal-fired power
plants, the extrapolation from four facilities to 1200 seems to
be inadequate. This may be due to the fact that the extrapolation
procedure is not described explicitly. Similarly, the
extrapolation for the industrial boilers is even greater and
involves more uncertainty. It would be more appropriate to take a
larger representative sample of both categories of boilers and to
use those results to estimate the country wide impact of those
facilities.
Response: The methodology employed to extrapolate risks from
"representative" utility plants and a single large industrial
boiler was chosen based on the time and resource restraints
imposed on the Agency by the Court order. The EPA believes .that
these estimates are accurate enough for decision-making purposes,
and doubts that evaluation of even a larger fraction of
facilities would result in a significant change in the estimate.
4.6.2.3 Other assessments of this question have been performed.
The results of the EPA assessment should be compared to those and
the differences between the results should be discussed and
rationalized.
Response; The assessment has been compared with other
evaluations, notably those performed by OAQPS. Reasons for the
differences in the estimated risks have been identified. These
include differing methodologies for determining locations of
nearby individuals, differences in accounting for total uranium
releases, and differences in the versions of the assessment codes
used.
4.6.2.4 Apportioning radionuclide emissions based on the number
of boilers fails to account for, among other things, units with
different capacities and units with different levels of
particulate control technology; radionuclide emissions can be
apportioned more accurately based on annual particulate
emissions.
Response; Use of plants with typical particulate emissions in
the assessment obviates the need for explicitly accounting for
differences in capacities and particulate control technology.
The extrapolation method allows for the accounting in demographic
variations, and these can be at least as important as the other
factors mentioned.
56
-------
4.6.2.5 The EPA assumes a population of 17.2 million residing
within a radium of 80 kilometers; yet only two coal-fired power
plants in the U.S. have a surrounding population in excess of 10
million.
Response; The reference plants evaluated were chosen to reflect
typical and extreme emissions characteristics for plants located
in urban, suburban, rural, and remote locations. Since the risks
in terms of deaths/y were calculated for each location segment
separately and normalized to the total population of the United
States, the choice of a plant with a very large surrounding
population does not greatly effect the estimate.
4.6.2.6 The very fact that existing electric utility generating
facilities are being and will continue to be replaced by new
electric utility generating facilities which must meet 40 CFR
Part 60 Subpart D(a) anyway, means that the risk of death due to
radionuclide emissions from coal-fired boilers will drop
irrespective of 40 CFR Part 61 Subpart U. By forcing retrofits
on older boilers, thus requiring them to operate longer in order
to retrieve the retrofitting costs, it will stymie their
replacement.
Response; The EPA acknowledges that emissions are unlikely to
increase given the NSPS. This is one factor considered in the
decision not to promulgate a NESHAP for this source category.
4.6.2.7 The EPA's estimates are based on numerous conservative
assumptions, for example, 1) individuals reside in the
predominant wind direction at a distance of 750 meters from the
plant, 2) a large fraction of the foodstuffs consumed by the
individual are grown at that location, 3) individuals reside at a
single location for 70 years at the point of maximum exposure.
Given these many conservatisms, EPA's risk estimates can easily
be an order of magnitude or more higher than more realistic
assessments.
Response; The EPA does not consider these assumptions
necessarily "conservative" in using reference plants to assess
the risks for source categories where the number of facilities
make site-by-site assessment impractical. With upwards of 55,000
coal-fired boilers and their location in both rural and urban
areas it is likely that there are individuals residing in the
predominate wind direction, that some individuals do obtain a
considerable fraction of their food locally, and that there are
individuals who reside in the locations of maximum environmental
concentration for their entire lives.
57
-------
4.6.3 Control Technology
No Significant Comments.
4.6.4 Level of Proposed Standards
No significant comments
4.6.5 Compliance and Implementation Procedures
4.6.5.1 The EPA is proposing to mandate a legal impossibility by
requiring the installation the needed retrofit equipment required
for compliance in a two year time span.
Response; No NESHAP is promulgated for this source category.
4.6.6 Costs of Compliance
4.6.6.1 Most coal-fired boiler owners would have to spend a
considerable amount of money to replace control equipment in
order to achieve the proposed emission standard. The
radionuclide reduction achieved in lower particulate emission
rates would not justify the considerable expense that would be
incurred. Public utilities in particular indicated that they
would have to retrofit fossil fuel plants to a baghouse
technology for particulate/radio-nuclide control or install
electrostatic precipitators.
Response; No NESHAP is promulgated for coal-fired boilers.
58
-------
4.7 INACTIVE AND LICENSED URANIUM MILL TAILINGS
It is expeditious and seems reasonable to combine the response to
comments for these two source categories because: 1) most
commentors provided comments on both active and inactive mill
tailings piles; 2) comments often pertained to both categories of
tailings piles; 3) the piles are similar in many respects; and
4) both categories of piles are included within one industry.
4.7.1 Basis for the Standards (legal/procedural issues)
4.7.1.1 The change from a design standard to a performance
standard is inconsistent with the codified rules and was rejected
by both the EPA and the NRC when the UMTRCA rules were
promulgated.
Response; The final NESHAP for disposal is an emission standard,
as required by the Clean Air Act.
4.7.1.2 There is no justification as to why the Clean Air Act
standard requires a synthetic liner under the tailings ponds.
Response; The requirement for a synthetic liner comes from the
UMTRCA rulemaking. Since the requirements of UMTRCA must be met
for future impoundments, this approach is reasonable. In
addition, EPA has a responsibility to make sure that the rules it
promulgates under one statute do not result in unnecessary
transfer of pollution into another medium, in this case
groundwater.
4.7.2 Dose and Risk Calculations and Analysis
4.7.2.1 Radon concentration and risk should be directly
proportional to the source term. However, while the source term
at the Bluewater mill increased by 1.3 (28.3 to 37.4 KCi/y) the
radon concentration at the nearest residence increased by a
factor of 1.8 (1.8 to 3.3 pCi/L). These discrepancies should be
resolved before the final rule-making.
Response; The assessment in the FEIS has been amended to reflect
the new or revised information obtained and developed since the
DEIS was published. The EPA appreciates the efforts of the
industry and DOE to provide additional information. Where the
validity of the information could be confirmed, it has been
incorporated into the FEIS. For example, the FEIS has been
revised to reflect the anticipated relocation of piles.
4.7.2.2 The relationship, 1 pCi/m2-s of radon per pCi/g of Ra-
226, is inaccurate, varies by an order of magnitude, and results
in an overestimate of the flux. Flux from mill tailings piles
should be based on flux measurements.
59
-------
Response; The EPA recognizes that the 1 to 1 radium to radon
correlation is an approximation. However, there is no scientific
consensus on what the value should be. EPA believes that this is
a reasonable estimate for situations where mill-specific data are
unavailable.
4.7.2,3 The assessment was performed at the Panna Maria tailings
pile assuming that all tailings were dry and exposed, while much
of the tailings pond is covered by one foot of a water deposited
clay.
Response; The assessment in the FEIS reflects the information
provided by the commentor. EPA has determined that the
evaluation of all dry conditions contained in the draft EIS does
not provide, given the disposal activities under UMTRCA, a
particularly realistic estimate of future exposures and risks.
4.7,2.4 The height of the Panna Maria tailings dam is 15 m, not
the 1 m assumed in the assessment.
Response; The 1 meter release height is appropriate for
estimating emissions from a volume source.
4.7.2,5 The EPA has grossly overestimated radon emissions by
assuming that the radium is evenly distributed throughout the
tailings pile.
Response; Lacking mill-specific data, the EPA assumed the even
distribution of radium throughout the tailings. This assumption
only affects that estimate of the source term during operating
and standby periods. For the disposal period, the assumption of
uniformity is roughly correct since the entire pile will be
drying and exposed. The assumption is irrelevant to the
estimation of the post-disposal source terms.
4.7.2.6 The heading on the table indicating required cover depth
in meters appears to be incorrect. It should indicate depth in
feet. The DOE requests further information regarding the
calculational methodology used to construct this table. For
example, was RAECOM used to project cover thicknesses?
Response; The column heading is correct. The depths of cover
were calculated using the methodology set forth in Appendix B of
the EIS. The RAECOM model was not used to project needed depth
of cover. Depending on local conditions, the absolute cover
depths for any individual pile may be over- or under-estimated.
However, the EPA feels that the aggregate costs are sufficiently
accurate for decision making.
60
-------
4.7.3 Control Technology
4.7.3.1 The EPA has not derived the tailings pond size limit
based on a thorough study of the emissions, but instead the 40-
acre size seems quite arbitrary.
Response; The 40-acre limit was chosen to reflect best current
practice, as exemplified by the Sweetwater impoundment.
4.7.3.2 Studies at the Shirley Basin site indicate much greater
reduction in radon emissions vs. soil cover thickness than
estimated by EPA in Table 9.12 of the BID. For example, 1.5 m of
random fill reduced the radon emanation to levels
indistinguishable from background (about 1 pCi/m2-s) . Thus, it
appears that EPA's cover thickness calculations are grossly in
error.
Response: In assessing the costs of various alternative disposal
fluxes the absolute depths of cover and costs are not as
important as the incremental changes. As stated in the FEIS, the
depths of cover for all sites are based on the assumption of
sandy soil and other parameters such as moisture content set
forth in Appendix B. Facilities with access to clay soils will
be able to meet the disposal flux requirements using thinner
covers than estimated. Also as required under the UMTRCA rules,
the cover must be designed to meet the flux limit for 1,000
years.
4.7.3.3 Continuous disposal of mill tailings has not been used
by the uranium industry; it would present serious operational
problems and excessive costs. Immediate disposal of tailings
under the continuous disposal method may not always be practical
given the limited construction seasons and the size of the
impoundments.
Response; The Agency is aware that continuous disposal has not
been used at existing mills. However, it believes that the
technology is applicable, as demonstrated by its proposed use at
the San Miguel mill that was licensed but not constructed. This
lack of practical experience is part of the reason why mill
owners can choose between continuous and phased disposal.
4.7.4 Level of Proposed Standard
4.7.4.1 The proposed EPA radionuclide NESHAPS will cause the
closure of the domestic uranium milling industry. The shut down
of the mills would cause strict reliance on foreign-source
uranium which could be devastating to our energy and defense
independence. In addition, the dismantling of the industry will
violate Congressional findings in the Atomic Energy Act that a
viable uranium industry is essential to our national security,
61
-------
result in a further reliance on fossil fuels which create
significant health hazards by their use, result in further loss
of employment to industry workers and workers in service
industries related to the producing industry, and create more
social and economic upheaval in the mining communities that have
already suffered greatly from the industries' demise.
Response; The NESHAPS will not result in the closure of the
domestic uranium milling industry. The uranium industry is not
economically viable due to large uranium inventories, a non-
increasing demand for uranium, and the ability of foreign
suppliers to produce uranium at lower cost. Recognizing these
facts, the Administrator is required by the Act to establish
NESHAPS that protect the public health without considering the
possible closure of the plants in the affected industry. Closure
considerations may only be addressed in the second step of the
process when the ample margin of safety is determined.
4.7.4.2 The EPA has failed to demonstrate that there is a
significant risk due to radionuclide emissions from uranium mill
tailings piles. Many industries posing greater risks are
ignored, e.g. soil tillage, water supplies, building materials,
natural gas and background.
Response; EPA risk assessment demonstrates the significance of
the risks presented by uranium mill tailings piles. The EPA's
on-going programs will determine whether or not regulation of
other activities, including those mentioned by the commentor,
require regulation.
4.7.4.3 The public health is protected from radon emissions from
uranium mill tailings piles with an ample margin of safety under
the existing regulatory requirements of UMTRCA. New EPA
regulatory requirements under the Clean Air Act are duplicative,
unnecessary, and would serve no useful purpose. The UMTRCA
standard has not been shown to be unsatisfactory.
Response; The Administrator has determined not to defer to other
regulatory authorities. The NESHAP for tailings disposal
establishes both a time frame for disposal, citizen suit
provisions and monitoring requirements that are not provided by
UMTRCA.
4.7.4.4 The most stringent of the proposed standards is
unrealistic because the soil used to cover the pile would exceed
the limit of 0.02 pCi/m2-s. The background flux in the area is
about 0.5 pCi/m2-s In addition the proposed standards of 6
pCi/m2-s to 0.02 pCi/m2-s for disposal of uranium mill tailings
are not enforceable because the natural variability is greater
than the proposed standard, thus rendering the model for
predicting compliance unreliable.
62
-------
Response; The final NESHAPS establishes a design flux through
the cover of 20 pCi/m2-s.
4.7.4.5 There is no need for a new standard since EPA has not
demonstrated that 0.12 fatal cancers per year that would result
from the current radon emission standard is not a acceptable
risk.
Response: The fatal cancers per year is just one of the risk
indicators that the Administrator considers. Individual lifetime
fatal cancer risks must also be found to be acceptable with an
ample margin of safety.
4.7.4.6 A standard of 2 pCi/m2 is both feasible and reasonably
achieved. The EOF believes that the standard must be set at 2
pCi/m2-s even under Approach A because of the requirements to
establish an ample margin of safety based on "feasible"
technologies capable of diminishing the risk.
Response; EPA selected an emission limit of 20 pCi/m2-s as the
NESHAP since it protects public health with an ample margin of
safety. The lower limit of 6 pCi/m2-s was considered for the
standard but was not selected because it would result in little
reduction of either individual risk or population effects when
-compared to the relatively large costs , of implementation.
4.7.5 Compliance and Implementation Procedures
4.7.5.1 The proposed standard is not written as an average, but
as an instantaneous rate. This is patently unfair and punitive.
The standard may be exceeded for short periods of time during
severe climatic conditions, while the annual average may be an
order of magnitude less than the standard allows. This forces
the cover to be designed for the most severe conditions.
Response; The final NESHAP establishes a long-term average flux.
Monitoring is required to provide assurance that the installed
cover meets the emission limit. The EPA's experience with radon
flux measurements establishes the fact that the average obtained
from a single monitoring period is sufficient to account for
variations over time when the number of samples approaches 100.
4.7.5.2 If promulgated, the requirement for covering mill
tailings to reduce releases of radon would in fact create a
greater risk for the workers from accidents than the risk radon
emissions are believed to pose for the general public.
Response; While covering the piles might result in risks to the
workers, the DOE's experience with disposal under the UMTRCA
standard has been that accidents are experienced far less
63
-------
frequently than estimated. Further, while there will be some
construction related risks, the benefits of covering the piles
will be accrued over time periods of 1,000 years or longer.
4.7.5.3 Several troubling aspects of this regulation are (1) how
is the term "cease to be operational" defined, and (2) what
action will the EPA take if an operator reports noncompliance.
Response; The definition of operational has been clarified in
the final rule. The EPA will, as required by the Act, take
appropriate enforcement action in the case of non-compliance.
4.7.5.4 The EPA should recognize and accept alternative work
practices which are capable of meeting the standards.
Response; The work practices for operating and standby
impoundments are based on .what the Agency believes are least-cost
alternatives.
4.7.5.5 The limitations of Method 115 are unduly restrictive and
skew the results to periods of high emanation rates rather than
indicating average rates.
Response; Monitoring radon flux during.or immediately after
periods of rainfall or during periods of freezing temperatures
are known to produce unreliable estimates of the average emission
rate. Furthermore, an operator may select a multi-period
measurement approach to demonstrate compliance.
4.7.5.6 Consideration must be given to the fact that some
tailings piles have already been stabilized; there will be
substantial health, safety and environmental costs if these have
to be modified.
Response; The Administrator did consider the fact that some
piles have been stabilized in determining the ample margin of
safety in the final rule. .
4.7.5.7 Flexibility must be built into any of the proposed
regulations; 2 years is a very short time for the design, review
and approval by agencies, and construction of new impoundments
required under Approaches A and B; mills not already meeting the
new technology under C would have to shut down; under Approach D,
time would be required to develop new technology for tailings
disposal.
Response; The final NESHAPS for operating and standby periods
are believed to give operators sufficient time to plan and gain
regulatory approval of future impoundments. The two year limit
for completing disposal of tailings at inactive and inoperable
mills is compelled by the Act and is not within the
Administrator's discretion to change or negate.
64
-------
4.7.6 Compliance Costs
4.7.6.1 The EPA estimates that the proposed regulations would
cost up to $37 million additional capital costs for operating
mill tailings piles and $200 to $300 million additional capital
costs for the disposal of uranium mill tailings. The EPA fails
to include the cost of reclamation of the 12 piles it classifies
as active. Imposing these capital costs on this industry could •
serve to close down an industry that the Congress has stated is
vital to the national security.
Response: The FEIS shows the appropriate capital costs for
disposal of all piles, work practices at all operable
impoundments, and single cell, phased, and continuous disposal
technologies at future impoundments. These costs including
reclamation were included in the DEIS as well.
4.7.6.2 Compliance costs would have the effect of shutting down
the uranium milling industry which would be an adverse effect
requiring the EPA to deem the proposed regulations to be a "major
rule".
Response; The EPA disagrees that the final rules will serve to
shut down the industry. The compliance costs for the NESHAP will
only slightly increase costs over the already existing
regulations.
4.7.6.3 Given the small incremental reduction in health risks
associated with the EPA's regulatory requirements for existing
piles and the critical state of the uranium production industry,
the costs are too high.
Response; The EPA estimates that the NESHAP for operating and
standby mills will have a total annual cost of approximately
$1.25 million. The Administrator found these costs to be
reasonable in the second step of the rulemaking process.
4.7.6.4 The DOE, which has experience in stabilizing tailings
piles, has testified that the costs to stabilize existing
tailings piles may be on the order of $1-2 billion.
Response; The DOE's cost estimates include monies for research
activities, technology development, and community participation.
These costs will be incurred by DOE regardless of the standard
promulgated by the Agency. The EPA's estimates, as noted above,
are for disposal only, and are based on a single soil type and
assumed moisture content. Thus, while the explicit costs for a
given mill might be over- or under-estimated, the EPA believes
that the aggregate costs are reasonably accurate and provide
sufficient basis for decision making.
65
-------
4,7.6.5 EPA does not address the question of long term funding
for testing, reporting, and possible remediation of the piles
after closure.
Response; The final rule requires one-time monitoring to
establish compliance with the design flux, thus there are no
long-term monitoring and reporting costs. If additional
remediation is required, based on test results, the EPA expects
that the costs will be borne by the responsible firms, in the
case of Title II sites, and by DOE and the states in the case of
Title I sites.
4.7.7 Other Comments
4.7.7.1 The EPA's basic research into and comprehension of
fundamental issues in uranium milling appear inadequate. The EPA
gives a confusing and incomplete account of the milling process
for uranium ore.
Response; The EPA disagrees. The process description referred
to is merely a summary of the extensive descriptions that have
been published by the Agency and the NRC.
4.7.7.2 If the proposed regulations are promulgated without
modification, the uranium industry will require the exemption
provided in Section 112 (c)(2) of the CAA.
Response; The EPA does not believe that the effect of the
standard would require such action. However, application for an
exemption is available to such facilities who determine that they
cannot meet the requirements of the NESHAP.
4.7.7.3 The survey of the U.S. uranium industry is incomplete;
many major operators are omitted.
Response; The Agency believes that it has identified and listed
all licensed and inactive tailing impoundments that have resulted
from the recovery of uranium by conventional acid- and alkaline-
leach methods.
4.7.7.4 If the uranium mills are unable to operate, this will
effectively end domestic uranium production. Both underground
and surface uranium mining are dependent on mills to process
their ore.
Response; The EPA agrees that if all mills close that mines will
not be able to find domestic sources of milling. However the
minimal costs associated with the NESHAP for operating and
standby mills should not force closure of the industry.
66
-------
4.8 DOE RADON SITES
4.8.1 Basis for the Standards (legal/procedural issues)
4.8.1.1 A DOE memorandum indicates that the EPA may have chosen
not to regulate DOE radon emissions in its previous rulemaking
due to its belief that the silos at Fernald (FMPC) were not
vented. However, the constant seepage of radon from the silo
cracks and the purposeful 1986 venting of radon demonstrate the
necessity of such regulation.
Response: The final rules establish a NESHAP for DOE Radon sites
including FMPC.
4.8.2 Dose and Risk Calculations and Analysis
4.8.2.1 The EPA should look at the combined health impact caused
by radon and uranium from the FMPC. The level of control should
result in an estimated risk no greater than the risk tolerated in
the 1983 proposed rulemaking.
Response; The final NESHAPS for DOE Facilities and DOE Radon
Sites establish limits to assure protection of public health with
an ample margin of safety. The two standards cannot be properly
compared or combined. The DOE Radon standard covers the disposal
of waste material, the DOE facilities standard covers emission
from operational emissions. While the disposal standard is
designed as a limit to be used by DOE in its CERCLA cleanup of a
waste problem, the other standard is designed around monitoring
of operation and annual reporting. Any attempt to combine the
two would cause confusion in implementing standards applicable to
disposal for operations and visa versa.
4.8.2.2 The radon emissions from the K-65 silos at the FMPC in
Fernald have not been routinely measured. How were these
emissions estimated given the deteriorated state of the
containment and the variation of emissions due to the heating of
the silos? How were accidental releases included in the emission
estimates? What other radon emissions occur at the FMPC site?
Have radon emissions been estimated from radioactive materials
deposited in the soil within the confines of the DOE facilities?
Response; The emissions from the K-65 silos were estimated, as
discussed in Chapter 10 of Volume 2 of the FEIS, on the basis of
the Ra-226 content, the area of the silos, and theoretical
equations for diffusion through concrete. The Agency is aware
that its estimate is only a rough approximation. The EPA is not
aware of any other major radon sources at the FMPC site.
4.8.3 Control Technology
67
-------
No Significant Comments
4.8.4 Level of Proposed Standard
4.8.4.1 The radon emissions rules are expressed in curies
released per unit area per unit time. These units are adequate
for mill tailings but not for manufacturing or experimental work
that involves uranium, thorium or radium. Dose limits such as
those employed in the DOE operations rules should be used, if EPA
intends to regulate emissions of radon from such operations.
Response; The NESHAP for DOE Radon sites applies only to radon
emissions from storage and disposal facilities. A flux standard
per unit area per unit time is appropriate for such sites which
are area sources.
4.8.5 Compliance and Implementation Procedures
No Significant Comments.
4.8.6 Costs of Compliance
No Significant Comments.
68
-------
4.9 UNDERGROUND URANIUM MINES
4.9.1 Basis for the Standards (Legal/procedural Issues)
No Significant Comments
4.9.2 Dose and Risk Calculations and Analysis
4.9.2.1 Although EPA's modeling incorporates either the buoyancy
effect or momentum of the air discharged from a mine vent, it
does not include both effects when the air being discharged is
hot.
Response; The AIRDOS code will not assimilate both buoyancy and
momentum factors in the same calculation. Since buoyancy will
predominate over momentum, it was used in the underground mine
dose and risk calculations.
4.9.2.2 The use of 1980 census data taken during the boom period
will overestimate the exposure to the nearby individual at Sheep
Mountain No. 1 mine.
Response; Demography data that were obtained during a site visit
to Sheep Mountain No. 1 show that the nearest individual is at a
distance of 5,200 m, similar to the 6 km distance reported by
U.S. Energy Corp.
4.9.2.3 AIRDOS predicts Ra-222 exposures and risks assuming the
receptors in each sector were always at the center line at the
plume. Proper modeling would use sector averaging which
disperses the plume uniformly in each downwind sector.
Response; AIRDOS has the capability of using either the plume
center line concentration or sector averaging. The latter was
used in these analyses.
4.9.2.4 No effective dose equivalents are presented in the draft
BID. The EPA needs to clearly state the methodology used to
determine radiation dose and risk from the radon concentrations
at the location of the maximum exposed individual, and allow
review of the regulation to verify the calculated dose and
associated risk.
Response; The methodology that the EPA used to develop the
exposure and risk estimates is exhaustively described in Volume I
of the EIS. The Court imposed deadline under which this
rulemaking was conducted does not allow for extending the period
for review and public comment. However, the Agency made every
effort to provide interested parties the background information
as quickly as possible.
69
-------
4,9.2.5 The environmental transport of radionuclides under the
CAP-88 program assumes a flat plain with no variations. Thus,
the assessment is inadequate when the mine is located in
mountainous areas or when the mine exhaust vent is at a different
elevation than the 80 km population.
Response: The AIRDOS (CAP-88) computer code used in this
assessment does assume a flat terrain and has no provisions for
accommodating the conditions stated in the comment. However,
this situation does not consistently over- or under-estimate the
risks.
4.9.2,6 Radon emissions from uranium mine sites do not present a
significant risk because radon daughters disappear into the
background within short distances from the sources and few, if
any, people live within the critical area of the exposure.
Response; The data provided purporting to substantiate the
assertion that the daughters are indistinguishable from
background within a few feet of the vent ignore the plume rise;
i.e., the measurements are meaningless since they were taken at
the wrong environmental location. The Agency does not ignore
risks to persons simply because they live in remote locations
with few other individuals subject to the risks.
4.9.2.7 Errors or omissions in characterizations of specific
facilities which could affect the dose and risk calculations and
assessments were pointed out; additional site-specific
information was provided for use in the analyses.
Response; The EPA appreciates the additional information and has
incorporated as much of it as possible into its FEIS.
4,9.2.8 The ICRP 1987 values used in the risk analysis in the
EPA's proposal have not been proven to be scientifically correct
and therefore are inappropriate as a basis for regulations.
Response; The Act recognizes the range of scientific opinion and
uncertainty, and explicitly requires the Administrator to
establish NESHAPS even when scientific certainty is not possible.
The ICRP values are within the range of scientifically credible
estimates, and the Administrator considers that they, along with
the other estimates of radon risk provide an adequate basis for
regulation.
4.9.3 Control Technology
4.9.3.1 The underground uranium mine standard requiring 30 m
stacks is unreasonable. Stack height is not a significant factor
in release when the air is directed vertically. The air is
usually discharged at a substantial velocity and often possesses
70
-------
buoyancy due to its temperature and humidity which gives an
effective release height substantially above stack height.
Response; The final NESHAP establishes a dose standard.
Operators are provided full discretion in the method or
combination of methods chosen to assure compliance with the
standard.
4.9.3.2 EPA should identify the design practices or technologies
(alluded to in Sections 2.3.3.2 and 2.4.2) by which new mines can
meet the proposed standards.
Response; Techniques such as retreat mining and careful
placement and orientation of mine vents are just two of the
design practices that could be employed. The final NESHAP
permits the greatest possible operational flexibility in meeting
the exposure limits.
4.9.3.3 Stacks attached to fans would have to be designed to
withstand the environment, particularly the wind. These stacks
will create additional air resistance which will increase the
fans operating pressure and decrease the fans operating quantity.
This decrease in air quantity will increase the miners' exposure
to airborne radiation.
Response; XEf stacks are used to meet the NESHAP limits, careful
design and sizing will be necessary to assure that they are
suited to the environmental conditions and do not result in
unacceptable exposures of miners. If mines find stacks to be
unacceptable then they can use other methods to meet the
standard.
4.9.4 Level of Proposed Standards
4.9.4.1 The proposed regulations will either severely limit the
production of existing underground uranium mines or will shut
them down. The elimination of domestic uranium production will
force total reliance by the U.S. on uranium from other countries
that could effect national security (nuclear powered naval fleet)
and the commercial production of electricity by nuclear power
plants.
Response; The final NESHAP will not cause a shutdown of the
uranium industry. While the closure of some mines is possible,
the final standard provides full flexibility in the method or
methods selected to achieve compliance. In addition several
mines are already in compliance with the standard.
4.9.4.3 The 5,000 Ci limit is not justified or proper. Most
mines will not be able to comply with such a limit and most which
comply will do so by dividing an ore body into separate mines
instead of a single mine.
71
-------
Response; The final NESHAP establishes a dose standard, with
choice of compliance strategy left to the discretion of the
operator.
4.9.4.4 The EPA should conduct a very careful and thorough study
to determine the total impact the proposed laws will have on the
uranium mining industry, and especially the adverse effects it
will have on the regional economy when a mine is forced to shut
down.
Response; Under the requirements of the Act and the instructions
of the Court, the Administrator can only consider the impacts of
the limits required to assure protection of public health in the
second step of the decision making process. The NESHAP for
underground mines is established at the level required to protect
public health, and more stringent standards were rejected during
the ample margin of safety step, in part, due to consideration of
the adverse impacts of closure on local communities.
4.9.5 Compliance and Implementation Procedures
4.9.5,1 The EPA should consider continuous monitoring using a
"track-Etch" type detector with quarterly exchange to determine
annual emission rates. As an alternate method, grab samples
(Lucas cells) on a quarterly basis should be allowed with a
record of mine operating conditions at time of sampling. Another
method of sampling would be to take 24-hour measurements every
3rd or 4th day. This would give about 90 days of radon values
per year and enough data to average the exhausted radionuclides,.
Response; Section 1.2 of Method 115 provides for the use of
alternative methods upon prior approval of the Administrator.
Information should be submitted demonstrating that the
alternative method will provide emission data equivalent to the
approved method.
4.9.5.2 When radon gas is emitted from a mine or tailings pile,
it dissipates into an unmeasurable quantity above background
within a few feet. Thus, the building of 30 m stacks or exhaust
vent holes will not have any measurable effect on public exposure
after a few feet.
Response; Radon emissions at locations near mine vents
significantly exceed background levels. While concentrations at
the levels required to protect public health may be below
measurement levels given the ambient concentrations of radon,
this in no way negates the fact the standards that are lower than
background levels are required to protect public health with an
ample margin of safety.
72
-------
4.9.5.3 The EPA should recognize and accept alternative work
practices which are capable of meeting the standards.
Response; The final rule provides the maximum degree of
operating flexibility in meeting the standard by establishing a
dose standard for radon emissions from underground uranium mines.
4.9.5.4 The monitoring specified in method 115 is inapplicable
to situations found at some underground uranium mines, e.g.,
sporadic or intermittent vent fan operation, high humidity, etc.
Response; Method 115 provides flexibility for mine operators by
establishing methods for both periodic and continuous operations.
4.9.5.5 EPA's definition of an active mine does not take into
account the temporary ventilation of idle mines when workers
enter them for inspections or maintenance purposes.
Response; EPA rule regulates all emissions from the ventilation
systems of a underground uranium mines. Ventilation for
inspections and maintenance is included in the standard, because
the radon emitted by these operations add to the health risk to
the public.
4.9.5.6 With respect to mines, the Rule is not ..feasible. To
protect miners, MSHA requires extensive ventilation to exhaust
radon from active mineworking areas in order to provide fresh,
clean air. The rule would reduce the fresh air-exhaust flow
through bulkhead areas resulting in higher exposure levels for
underground mine workers.
Response; The rule provides sufficient flexibility to the mine
operators to allow for the protection of both workers and the
public.
4.9.6 Compliance Costs:
4.9.6.1 The EPA should consider the large costs of adding 30 m
stacks to exhaust vents at mines that are competing in a very
depressed market. This will force some mines to close.
Response; The final NESHAP allows the operator to select the
most cost-effective compliance strategy.
4.9.6.2 The assessment of costs ignores higher prices in long-
term contracts, maintenance costs during shutdowns, and higher
unit costs of lowered production rates.
Response; The FEIS attempts to evaluate these additional costs.
However, such costs vary widely since there are significant
differences in the emissions and operations of the different
mines.
73
-------
4.9.7 Other Comments:
4.9.7.1 The EPA failed to consider mines under development,
particularly in the Arizona strip.
Response; The EPA's assessment is of currently operating mines.
The Agency is not aware of any changes in mining methods that
will cause the emissions from mines developed in the future to be
significantly different than from currently operating mines. In
any event these mines will be covered by this standard.
4.9.7.2 No individuals live at the locations designated by EPA
for the maximum exposed individuals at Pigeon and Kanab North
mines.
Response: EPA has changed the distances to the maximum exposed
individuals at Kanab North and Pigeon mines to 30,000 m and
24,000 m, respectively, reflecting the distances to Fredonia, AZ.
4.9.7,3 The attached is a report on radon measurements of actual
vent holes. The point of maximum exposure here is the center of
the vent hole. After a few feet, the radon daughter
concentration falls off to negligible levels.
Response; The measurements were taken between 4 feet and 10 feet
from the ground within 8 feet of the vent, which extended 4 to 5
feet above the ground. It is obvious that the radon from the
vent was passing well over the point of measurement due to plume
rise. .
4.10 SURFACE URANIUM MINES
4.10.1 Basis for the Standards (legal/procedural issues)
4.10.1.1 The mining of uranium ore is conducted under general
license from the Texas Department of Health and regulated under
Texas rules for the control of radiation. The statement to the
contrary in 12.1.2.2.3 of the BID is incorrect. There are
regulations for mine reclamation by the Texas Railroad Commission
which specify the radiation levels on the surface of the mine at
the time of bond release.
Response; According to cognizant state government personnel the
statement in the FEIS is correct.
4.10.2 Dose and Risk Calculations and Analysis
No Significant Comments.
4.10.3 Control Technology
74
-------
No Significant Comments.
4.10.4 Level of Proposed Standards
4.10.4.1 The Nuclear Regulatory Commission and the Agency have
funded several studies to monitor radon concentrations near open
pit uranium mines over the last decade. These reports show that
anything lower than the existing standard will be extremely
difficult to control without increasing the health risks to the
very people who are required to control the problem.
Response; No NESHAP is promulgated for surface uranium mines.
4.10.5 Compliance and Implementation Procedures
4.10.5.1 The monitoring specified in method 115 is inapplicable
to situations found at some surface uranium mines.
Response; No NESHAPS is promulgated for surface uranium mines.
4.10.6 Control Technology and Compliance Costs
No Significant Comments.
4.10.7 Other Comments
4.10.7.1 There are only two open-pit uranium mines in production
in the United States. These mines are required to help produce
the nation's uranium requirements to fulfill nuclear power needs
for industry and the country's defense establishment.
Response; The Agency is aware that currently only two
conventional open-pit uranium mines are in operation.
75
-------
4.11 PHOSPHOGYPSUM STACKS
4.11.1 Basis For the Standards (Legal/procedural issues)
4.11.1.1 The ORP's plan to propose a standard for phosphogypsum
stacks was not based on a determination that such a regulation is
necessary and appropriate for protection of public health. We
understand that the ORP agreed to propose a standard solely in
order to obtain an unopposed extension of time to the court-
imposed schedule in Environmental Defense Fund, Inc. vs. Lee M.
Thomas.
Response; Although the commentor is correct that EPA agreed to
propose a standard for phosphogypsum, it is not true that it was
done "solely" to get an extension. The Agency was already
working on this source category and was considering it for
regulation. EPA is under no court order to promulgate a final
standard for this category. The decision to do so is based
solely on the risks presented by phosphogypsum stacks to public
health.
4.11.2 Dose and Risk Calculations and Analysis
4.11.2.1 The risk estimate source term should include a
recognition of the radon not emitted from the stack area due to
the natural radon the stacks and associated ponds seal off.
Response; Correcting the source term for background radon flux
(0.2 to 0.3 pCi/mz-s) from the area beneath the stack would
result in an insignificantly small reduction in risk estimates.
4.11.2.2 The BID states, "For one section of Florida, it is
estimated that the number of persons exposed is overestimated by
a factor of seven, while the risks are understated by a factor of
three." A similar but less severe problem exists in the
assessment for the multiple stacks in southeastern Idaho. Better
models that would resolve this overlap problem should be used.
Response; New modelling approaches are needed to estimate the
risk from multiple sources. However, until they become
available, the existing models provide the best estimates of
risks. The case study of over-lap in the central Florida area
was conducted to place an upper-limit on the degree of
underestimation of individual risks. As pointed out, multiple
sources do not affect the estimate of committed fatal cancers, as
the effects are simply additive. The only effect of this
treatment of multiple sources is on the estimate of maximum
individual risk. Since the phosphogypsum stack with the maximum
individual risk is in Louisiana, any error this might have caused
is insignificant.
76
-------
4.11.2.3 The EPA has performed the assessment analysis assuming
a release height of one meter while phosphogypsum stack heights
range from 10-60 m. We believe an appropriate and easy-to-use
adjustment would be to assume the height of release is one-half
of the physical stack height. This would result in a more
realistic assessment.
Response; The correct modeling of a non-uniform volume source is
uncertain. The 1 meter release height was chosen to assure that
the risks to nearby individuals are not significantly
underestimated. Moreover, as explained in the EIS, the
sensitivity study that was made shows that even if the commentor
is right, the estimated risk to nearby individuals would only be
significantly overstated for one stack.
4.11.2.4 The EPA must use site-specific data in its analysis of
radionuclide emissions from phosphogypsum stacks.
Response: The EPA used site-specific data at all phosphogypsum
stacks where such data was available. This included stack size
and shapes, flux measurements, radionuclide content, stack
conditions, etc., made available by on-site visits, company
replies, and The Fertilizer Institute (TFI). In some case
information provided during the comment period resulted in
changes to the inputs to the risk assessment.
4.11.2.5 The EPA's assumptions concerning stack geometry are all
no more than approximations. It is doubtful whether any one
actual stack exhibits the EPA's assumed geometry.
Response; The EPA exerted considerable effort to obtain site-
specific information to use in computing the source terms for the
phosphogypsum stacks and information obtained from discussions
with company employees. From our observations and input from the
company officials, we developed a generic stack with length (not
height as TFI states) twice the width and a 1:3 slope to the
sides. Most stacks we have seen do approximate a rectangular
configuration. We made many inquiries to companies, as well as
to TFI, for specific stack data, including stack dimensions and
slope of the sides. We received an excellent response to our
inquiries from some companies, and either poor or-no response
from others. It is not easy to obtain this information from most
companies, as indicated by TFI's efforts to get the slope of the
sides of stacks as reported in their comments. Apparently they
were able to obtain the slopes for only 27 of the 63 existing
stacks. When we had specific stack data it was used in the
assessment. For those stacks that specific information had not
been made available, we had no choice but to use generic
parameters. .
77
-------
4.11.3 Control Technology
4.11.3.1 Allow maximum flexibility in determining how much
reduction is required at any given facility and how it is
achieved. For example, it may not be necessary to cover the
entire stack to meet the standard.
Response: The standard does allow such flexibility.
4 11 3.2 The radon flux from soil cover will affect the depth of
cover necessary to attain a radon flux goal, and would increase
the costs over that estimated by the EPA. This will require
site-specific consideration.
Response: Site specific considerations will be necessary in all
cases to assure compliance with the standard. Soil cover should
contain approximately background levels of radium and add
insignificantly to the radon flux.
4 11.3.3 The soil moisture assumed in the cover by the EPA is
about 3 times greater than that seen in Florida soils, which
would double the thickness to achieve^the same emanation rate
increasing the soil cover cost by 100%.
Response; The estimated soil moisture used by the EPA is based
on an empirical correlation that uses annual average rainfall and
evaporation rates.
4.11.3.4 The haulage distance for dirt to cover the stacks
cannot be based on a generic evaluation (10-mile round trip), but
must be determined on a case-by-case basis.
Response; EPA did not attempt to estimate costs for each pile.
Rather a generic pile was developed based on average conditions.
This approach is adequate for rulemaking since it is the
incremental costs that are important in considering progressively
more stringent controls. These incremental costs are much less
sensitive to errors in the generic pile than total costs.
4.11.3.5 The rule should allow consideration of alternative risk
control methods other than covering the stack with dirt.
Response; The operator is free to use any method that will
permanently meet the radon flux standard from the stacks.
4.11.3.9 The EPA should recognize that 1) phosphogypsum in north
Florida has 50 - 75% less radium than assumed by EPA based on the
radium content of rock from other Florida areas; 2) a soil
density of 1000 to 1200 kg/m3 would be more appropriate that the
220 used; and 3) environmental weathering is commonly represented
by a 14-day half-life rather than the 35 years used.
78
-------
Response; The FEIS has been amended to reflect the lower radium
content of north Florida rock. The soil density used is
appropriate for phosphogypsum stacks. The 1000 to 1200 kg/m3
cited reflects the specific gravity of typical soils while the
200 kg/m3 used by the EPA reflects the bulk density of typical
soils.
4.11.4 Level of Proposed Standards
4.11.4.1 The EPA should allow an alternative standard of 0.5
pCi/L radon at the stack's edge.
Response; This standard would be very difficult to measure on an
average annual radon concentration basis considering that
background concentrations in central Florida range from 0.14-5
pCi/L and the uncertainties associated with measuring ambient
airborne radon concentrations. For example, consider the case of
Seminole Fertilizer Corporation's south stack which is surrounded
by reclaimed land that resulted in a measured base perimeter
radon concentration of 0.7 pCi/L (see EPA 520/5-88-021).
4.11.5 Compliance and Implementation Procedures
4.11.5.1 The EPA failed to consider the radiation exposure to an
individual caused by the soil which would be placed on top of the
stack. This is important in evaluating Approach D (1-60).
Response; Because there is very little radium in soil that
should be used for cover it would not have a significant effect
on health and it will have little effect on compliance with the
standard since Approach D was not selected.
4.11.5.2 We do not believe more than 20 samples for each
category (sides, beaches, roadways, etc.) are warranted. It
would be better to require more than one set of measurements
since the flux is affected by meteorological conditions.
Alternatively, the number of measurements should be based on the
stack area. For example, on a 9 ha stack this would amount to
one measurement per 300 sq m, while the EPA's long-term study
(EPA 520/5-88-021) made one measurement per 78,000 sq m. We
recommend making one measurement per 10,000 sq m.
Response; The number of samples required for each category is
based on EPA's report 520/5-88-021.
4.11.5.3 Section 3.1.2 requires a minimum of 300 measurements to
characterize the radon flux from the phosphogypsum stack. This
is contrary to the conclusion in the EPA's background document, A
Long-Term Study of Radon and Airborne Particulates at
Phosphogypsum Stacks in Central Florida.
79
-------
Response; The cited EPA document considered only the loose, dry
area of the top surface. The rule considers the total stack
consisting of 5 regions. The conclusion was made that 100
measurements adequately defines a mean flux over a large area
with large flux variability, while fewer measurements will
adequately define a smaller area with less variability.
4.11.5.4 The EPA should make it clear that emissions limits are
applied to closed stacks and not to operating stacks.
Response; The final rule clarifies this point.
4,11.5.5 Retesting should be required on a non-regular basis to
better provide for testing over all seasons of the year. Also,
if initial testing indicates non-compliance, the operator should
be allowed to conduct three additional tests at three-month
intervals and report the results of each test as well as the
average radon flux for all four tests. Compliance should be
based on the four-test average.
Response; The final rule provides flexibility by allowing for
either periodic or continuous measurement.
4,11.5.6 The proposed rule does not specifically identify what
would constitute a stack no longer in use and therefore subject
to the standard. However, the language of proposed Section
61.203(d) indicates that EPA may view the discontinuation of the
actual placement of phosphogypsum on the stack with the cessation
of the use of that stack. This view is not correct. The rule
should not restrict a company's ability to use the surface area
of a phosphogypsum stack for water management purposes even
though the stack or a portion of the stack is no longer regularly
used for the placement of phosphogypsum.
Response; The definition of when a stack is subject to the
NESHAP has been clarified in the final rule. The rule specifies
that stacks used for water management are still in operation and
do not need to be tested.
4.11.6 Compliance Costs:
No significant Comments
4.11.7 Other Comments
4.11.7.1 No 1987 flux measurement results were used in the
assessment. The EPA used only measurements made in 1988.
Response; The 1987 data were not used because EPA was uncertain
where the measurements were made, what procedure was used, and
they were few in number.
80
-------
4.11.7.2 The EPA assumed a stack side slope of 1 to 3, whereas
the side slopes of the Swift Creek and Swannee River stacks are
1.2 to 1 and 1.5 to 1, respectively.
Response: EPA regrets that this information was not received in
time for the final risk assessment.
81
-------
------- |