United States Nuclear Regulatory Commission - Protecting People and the Environment
Home > NRC Library > Document Collections > ACNW > Meeting Transcripts > 2002 > 131st Meeting - January 8, 2002


                Official Transcript of Proceedings

                  NUCLEAR REGULATORY COMMISSION



Title:                    Advisory Committee on Nuclear Waste



Docket Number:  (not applicable)



Location:                 Rockville, Maryland



Date:                     Tuesday, January 8, 2002







Work Order No.: NRC-166                                Pages 1-37





                   NEAL R. GROSS AND CO., INC.
                 Court Reporters and Transcribers
                  1323 Rhode Island Avenue, N.W.
                     Washington, D.C.  20005
                          (202) 234-4433.                         UNITED STATES OF AMERICA
                       NUCLEAR REGULATORY COMMISSION
                                 + + + + +
                ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW)
                                 + + + + +
                                 TUESDAY,
                              JANUARY 8, 2002
                                 + + + + +
                            ROCKVILLE, MARYLAND
                                 + + + + +
                       The ACNW met at the Nuclear Regulatory
           Commission, Two White Flint North, Room T2B3, 11545
           Rockville Pike, at 11:00 a.m., George M. Hornberger,
           Chairman, presiding.

           COMMITTEE MEMBERS:
                 GEORGE M. HORNBERGER, Chairman
                 RAYMOND G. WYMER, Vice Chairman
                 B. JOHN GARRICK, Member
                 MILTON N. LEVENSON, Member





           ACNW STAFF PRESENT:
                 HOWARD J. LARSON, Special Assistant, ACRS, ACNW
                 RICHARD K. MAJOR
                 LYNN DEERING
                 LATIF HAMDAN
                 SHER BAHADUR
                 AMARJIT SINGH
                 JOHN T. LARKINS
                 RICHARD P. SAVIO
                 CAROL A. HARRIS

           ALSO PRESENT:
                 TIM McCARTIN












                                 I-N-D-E-X
           Opening Statement. . . . . . . . . . . . . . . . . 4
           Discussion of Proposed Amendment to
                 10 CFR Part 63 Adjourn . . . . . . . . . . . 5
           Presentation by Tim McCartin . . . . . . . . . . . 7



















           .                           P-R-O-C-E-E-D-I-N-G-S
                                                   (11:03 a.m.)
                       CHAIRMAN HORNBERGER:  The meeting will
           come to order.  This is the first day of the 131st
           meeting of the Advisory Committee on Nuclear Waste.
           My name is George Hornberger, Chairman of the ACNW.
                       Other Members of the Committee present are
           John Garrick, Milton Levenson and Raymond Wymer.
                       During today's meeting, following the
           planning and procedure session, the Committee will
           discuss proposed revisions to 10 CFR Part 63, discuss
           proposed letters and finalize plans for a Committee
           retreat.
                       John Larkins or Howard Larson is the
           designated federal official for today's initial
           session.
                       This meeting is being conducted in
           accordance with the provisions of the Federal Advisory
           Committee Act.  We have received no written comments
           or requests for time to make oral statements from
           members of the public regarding today's session.
           Should anyone wish to address the Committee, please
           make your wishes known to one of the Committee staff.
                       It is requested that speakers use one of
           the microphones, identify themselves and speak with
           sufficient clarity and volume so that they can be
           readily heard.
                       Before proceeding, I would like to cover
           some brief items of current interest.  It has been
           announced recently that Bill Reamer will become the
           Deputy Director, Division of Waste Management,
           replacing Josie Picone who will become the Deputy
           Director, Office of State and Travel Programs.  Mr.
           Jack Sorenson, ACRS staff Senior Fellow, completed his
           term with the Office on December 31, 2001.  He has now
           joined the Office of Nuclear Material Safety and
           Safeguards to assist in risk-informing activities.
           The trade press recently reported that politicians in
           the Swedish town of Oesthammar have voted to allow
           drilling in bedrock in the community to see if it is
           suitable for a final spent fuel repository.  Mr. Phil
           Justice has been assigned as NMSS liaison for the
           ACNW.
                       We are going to proceed to our discussion
           of the proposed amendment to 10 CFR Part 63 and John
           Garrick is the cognizant member.  I'll turn it over to
           John.
                       MEMBER GARRICK:  Thank you.  We're going
           to hear from the staff on a proposed rule on 10 CFR
           Part 63 which has to do with the specification of the
           probability for unlikely features, events and
           processes or as they are affectionately known as FEPs
           and as I understand it, the Commission has voted this
           proposed rule, but it has not been published yet for
           public comment because the SRM is still being held up
           for reasons that maybe Tim will explain.
                       What is involved here is that the EPA
           Yucca Mountain standard is really -- involves three
           standards, three separate standards:  the so-called
           All Pathway standard, the Human Intrusion standard and
           the Groundwater standard.
                       And in the EPA standard, the concept of
           very unlikely events has been defined as a specific
           number and these are excluded from all three
           standards.  So these numbers, these thresholds become
           very important because they drive the issue of what
           can be considered or what has to be considered and
           what doesn't have to be considered.
                       And for the Human Intrusion and
           Groundwater standard, the category called unlikely
           events may also be excluded, but the EPA left it up to
           the Nuclear Regulatory Commission to define what is
           meant by unlikely.
                       So this is a topic of considerable
           interest to the Committee.  The problem here is that
           it is sort of after the fact advice that something we
           were talking about earlier and it's somewhat difficult
           for the Committee to have had much of an impact on the
           voting of the Commission.  And this is one of the
           flaws that we've spoken about in the past that's
           inherent in this whole advisory process.  But we don't
           want to dwell on that at the moment.  We are going to
           talk about it at our retreat.
                       One of the things that I'm very interested
           in, in this whole discussion is I think maybe an issue
           that's just as important as what these numbers ought
           to be is the issue related to the process of
           assembling or disassembling or aggregating FEPs
           because you can imagine that you might have something
           that's just above an acceptable threshold and through
           redefinition of the FEPs or the scenarios or what have
           you, you end up with a set of conditions that are
           actually below the threshold and I'd be very
           interested in the discussion in being assured that
           that kind of manipulation is protected somehow.
                       So with that, Tim, Tim McCartin, we're
           looking forward to your clarifying where we are in
           this issue and perhaps answering some of our
           questions.
                       MR. McCARTIN:  Hopefully.  Thank you, Dr.
           Garrick.
                       As you noted, we'll be talking about the
           probability of unlikely events.  I would like to make
           a few introductory statements and one would be, as you
           alluded to, although we are anticipating an SRM
           shortly, we do not have one, so this presentation
           should be viewed as the staff opinion.  It isn't the
           Commission's opinion yet until a proposal is
           published.
                       Likewise, I do know that I believe we try
           to and get benefit from participation with the
           Committee and I would say it is unfortunate that we
           did not -- were not able to come to the Committee with
           our proposal prior to going to the Commission.
           However, we were aware of that.  We knew the
           limitations of it.  We were operating with an SRM on
           final Part 63 that directed us to do an expedited rule
           making and gave us a relatively short time period to
           do this.  And so there really wasn't an opportunity to
           interact appropriately or as we typically would like
           to do.
                       However, we think that going out with the
           proposal, rather than getting the Committee's maybe a
           very hurried quick opinion, we're proposing a 75-day
           comment period.  During that public comment period we
           would like to come back to the Committee, discuss the
           public comments we get and get the Committee's
           reaction at that time and certainly we would factor in
           any Committee's opinions, views in the final to this
           amendment.  But it really -- we had approximately 2 to
           3 weeks to do this before getting into the concurrence
           process and so writing this in two to three weeks
           really does not lend itself well to getting a lot of
           input and thereby the Commission saying do an
           expedited rule making which is direction to skip some
           of the things you normally would do.
                       Also, along those lines, I'll repeat it at
           the end, but when we're talking about unlikely events
           in this context, I think it's very important to
           recognized that unlikely is for this application only.
           It is a very good qualitative word.  We aren't
           suggesting that we're defining the word "unlikely" for
           all the uses that might be used in regulation at NRC,
           but merely in this application.
                       Generally, I'd like to give you a little
           of the regulatory background, some of the technical
           background that we thought about and then finally I'll
           give what our proposal for the probability value for
           unlikely events is.
                       I will probably slip into using the
           unlikely events.  The language is actually features
           events and processes and sequences of events and
           processes, but a shorthand way of saying is I just use
           the word "events", but I mean the full spectrum.
                       The background, as actually Dr. Garrick
           alluded to, the EPA standards were finalized on June
           13th of last year.  There were three standards:  the
           individual protection standards, human intrusion and
           groundwater protection.  Very unlikely FEPs are not to
           be considered in any of these calculations.  Unlike
           FEPs are to be considered only for the individual
           protection standard.  They're excluded from
           consideration in groundwater protection and human
           intrusion.
                       The final standards went a little further,
           obviously as proposed in the EPA standard the cutoff
           for very unlikely FEPs was a one chance in 10,000 of
           occurring within the 10,000 years, what we sometimes
           refer to as the 10-8 per year cutoff.  That's for very
           unlikely FEPs.  The probability for unlikely FEPs were
           not defined and they left that to the NRC.
                       In looking at the preamble to the
           standards, it's important to look at how unlikely is
           to be applied.  As I mentioned human intrusion,
           groundwater protection.  They're very specialized
           calculations.  The human intrusion calculation is done
           to look at the robustness of the repository to the
           consequences of intrusion.  Likewise groundwater
           protection is to evaluate the degradation of the
           groundwater resource.  They're very specialized
           calculations.
                       And further, in looking at these
           calculations, the preamble to the standard really
           talks to focusing the calculations on likely or
           expected performance, so the reason for excluding
           unlikely is you're trying to get to what is the likely
           behavior, expected behavior for these two specialized
           calculations.
                       Final part 63 was published in November of
           last year.  In it, the Commission indicated that they
           intended to conduct a separate rulemaking to define
           unlikely and indicated that the numerical value would
           be somewhere between 10-8, 10-4 per year.
                       Our approach in looking at how to define
           unlikely, conceptually we thought it was easier to
           think of three broad categories:  very unlikely,
           unlikely and likely.  Clearly, we already had a
           definition, a quantitative definition very unlikely,
           the 10-8 per year and in looking at those three
           categories we felt it was easier conceptually to think
           of very unlikely and likely.  If you set the values
           for those two, you clearly have the middle defined and
           so looking at a lower bound for what should be
           considered a likely FEP, we felt that somewhere
           between 10-6 and 10-4 range was an appropriate value
           for what could be considered likely as a lower bound
           for likely.
                       Quantitatively, what does that really
           mean?  The 10-6 per year is essentially a 1 percent
           chance of occurring within the 10,000 year period.  We
           felt that a 1 percent chance of occurring is neither
           expected nor likely.  10-5 per year is a 10 percent
           chance of occurring within the 10,000 years.  And then
           if we go to 10-4 per year, it's a very high
           probability of occurring within the 10,000 years.
           Obviously, it's very likely that it would occur.
                       And I know I was accused early on of the
           Goldilocks Syndrome here and one's too low, one's too
           high, the middle one is just about right and the 10-5
           per year would seem to be a 10 percent chance of
           occurring as the lower bound for likely, that if you
           got much lower than that, sort of like 10-6, 1
           percent, that should not be considered likely.  The
           10-4 per year, that would seem to be, you would want
           to be a little lower than that.  It's almost certain
           to occur.  It's very likely to occur.
                       And thus, our proposal, we felt that
           unlikely FEPs would be best characterized by a
           probability range and we're defining the unlikely FEPs
           as that value between 10-5 and 10-8 per year,
           essentially a 10 percent chance of occurring at the
           high end, down to the .01 percent chance which is
           really the 1 in 10,000 over 10,000 years at that for
           the very unlikely.
                       We think that upper bound of 10 percent
           for unlikely seems to be a prudent range for defining
           the unlikely events in the context of these two
           calculations.  Once again, it's not intended as a
           precedent for other applications where the term
           unlikely might be used in NRC regulations.  At the
           beginning of this exercise of writing the amendment,
           we looked at the word unlikely, where is it used?  It
           appears in all kinds of places, other regulations.  It
           really is the context and for this particular
           application it's appropriate.  There are many other
           uses of the word unlikely and we are not setting
           precedent for that.
                       Where are we at?  Right now, we anticipate
           that we will get an SRM shortly and the amendment will
           be published this month, 75-day comment period.  We
           anticipate that we would have a final amendment some
           time in the summer to early fall.  And once again, I
           would like to reiterate that I think during the public
           comment period, as we get some of the comments in,
           there could be an appropriate time, maybe more than
           one time, to come back to the Committee and discuss
           the comments we've gotten and where we think we're
           headed with this amendment.
                       And with that, if there are any questions,
           I'd be happy to try to answer them.
                       MEMBER GARRICK:  Thanks, Tim.  Milt,
           microphone.
                       MEMBER LEVENSON:  You did what is fairly
           easy to do which is to confuse me, Tim.  After you've
           defined that you're going to talk about three things,
           very unlikely, unlikely and likely, you then ring in
           a new one which is very likely.  Is that intended to
           be a different category?
                       MR. McCARTIN:  No, no.  It was just -- the
           10-4 in terms of -- I guess we could have -- well, we
           were looking for the lower bound for unlikely or the
           -- the lower bound for likely and --
                       CHAIRMAN HORNBERGER:  Just take the "very"
           off and you solve the problem.
                       MR. McCARTIN:  The reason "very" is there,
           that in terms of -- it's too likely to be considered
           the lower bound for likely.
                       (Laughter.)
                       MR. BAHADUR:  Is 10-5 likely?
                       MR. McCARTIN:  That's the -- yes, the
           lower bound for likely.  That's what our proposal is.
           The lower bound for likely, the upper bound for
           unlikely.  You've got three categories.  Obviously,
           there's that.
                       MEMBER LEVENSON:  Okay, one other
           question.  From a standpoint of risk informed and I
           know maybe probably it isn't possible for this, but
           I'd like your comment on the idea of defining these
           independent of consequences when they're a basis for
           screening out things.  You're not -- there's nothing
           in here that allows you to throw something out because
           even if it's likely, the consequences are unimportant.
           There's a whole separate set of things, right?
                       MR. McCARTIN:  Right.  It does not look --
                       MEMBER LEVENSON:  You still have the other
           screen, even if something is very likely, but has
           extremely low consequences, still have the ability to
           throw it out?
                       MR. McCARTIN:  Oh, absolutely.  I mean the
           regulation allows for something that does not have a
           significant effect and that would be certainly based
           on low enough consequences or low enough risk.
                       CHAIRMAN HORNBERGER:  That would occur in
           screening the FEPs, is that right?
                       MR. McCARTIN:  That's one area you could
           do it.  You could actually have a FEP come in.  You do
           the analysis you see that the consequences and/or risk
           is low enough that we -- it does not have a
           significant effect on the time or the magnitude of the
           dose.  I can eliminate it.  And that's really the --
           in the rule, it does say that.  Things that don't have
           a significant effect on the timing and magnitude of
           the dose, do not have to be included in the analysis.
                       MEMBER GARRICK:  Ray?
                       VICE CHAIRMAN WYMER:  Once these
           categories are defined, then you get down to the
           really tough question of how do you determine
           probabilities.  Where in the regulations is there
           criteria for doing that?
                       MR. McCARTIN:  Well, the only thing -- the
           one thing we have said is that we, along the lines I
           know Dr. Garrick mentioned that could there be
           mischief by defining things a particular way merely to
           get it a low enough probability to get it out of the
           analysis and we have said that in general for events,
           categories of events and we're expecting a broad
           definition of the event, rather than a very narrow
           definition and maybe a simple example could be that if
           you have seismicity, seismicity as an event occurs.
                       If you wanted to define maybe a very
           narrow range of magnitude of the seismic event, you
           could get down to a very small probability and we're
           not looking at -- we're not expecting a narrow
           definition.  It tends to be more of a broader
           definition, but other than that, there's nothing
           that's part of the NRC review.  It would be looking at
           how a particular event was defined that potentially
           led it to be screened from the analysis.
                       VICE CHAIRMAN WYMER:  So you take somebody
           else's calculation of probability and evaluate it?
                       MR. McCARTIN:  Yes.
                       VICE CHAIRMAN WYMER:  Okay.
                       MEMBER GARRICK:  George?
                       CHAIRMAN HORNBERGER:  Tim, let's see, how
           can I phrase this question?  One of the things that I
           found a little confusing is this notion that you
           presented that should we have a probability range or
           a single number.  At first I thought well, a single
           number can't be what Tim means because that doesn't
           make any sense.  You're not going to pick the square
           root of 2 times 10-6, the single number for an
           unlikely event.  So that isn't what you mean.
                       So what I don't understand is then what
           the distinction is between saying okay, I'm going to
           pick the single number, less than 10-5.  How is that
           different from defining a range?
                       MR. McCARTIN:  I don't believe it is.
                       CHAIRMAN HORNBERGER:  Okay.
                       MR. McCARTIN:  But --
                       CHAIRMAN HORNBERGER:  So in terms of plain
           language, don't you think that you might be doing
           violence to understanding by raising this as a big
           issue?
                       MR. McCARTIN:  Well, what it does is --
           the reason we use the range, we also have that lower
           end cutoff, that unlikely doesn't proceed on to zero,
           if you will, but it stops at the 10-8 and from that
           point on it's very unlikely and it's really just
           creating that here's the range for unlikely and then
           here is likely.  But you're right, it's just --
                       CHAIRMAN HORNBERGER:  In the presentation,
           you can add confusion by saying well, we're not
           picking a single value.  We're picking a range when --
           obviously, that's what you're doing.
                       MR. McCARTIN:  Well, yes.  Maybe a better
           way to say it is it's less than 10-5 and therein is
           less than 10-8.  That might --
                       CHAIRMAN HORNBERGER:  I think you
           mentioned this.  It is important to highlight that
           this decision, as Ray indicated well how do you decide
           on these probabilities.  These would only be the
           unlikely events or event sequences or the whole -- I'm
           using event the same way you said you were going to
           use it.  It's only for the stylized analyses for human
           intrusion and well, basically igneous activity.
                       MR. McCARTIN:  Groundwater.
                       CHAIRMAN HORNBERGER:  Or groundwater
           protection.  So they're not screened out over the all
           pathways dose at all?
                       MR. McCARTIN:  No, no.  The individual
           protection would have the unlikely events, yeah.
           Obviously, very unlikely is excluded from everything.
                       CHAIRMAN HORNBERGER:  Right.
                       MR. McCARTIN:  But the unlikely is
           included in the individual pathway.
                       CHAIRMAN HORNBERGER:  Thanks.  That's all
           I have.
                       MEMBER GARRICK:  Tim, how would you
           correlate FEPs with scenarios?
                       MR. McCARTIN:  I would say that to me it's
           easier to think of the events and an event is like a
           scenario class.  I think they're similar.
                       A scenario class would have some
           collection of FEPs within it.
                       MEMBER GARRICK:  Yeah, yeah.  FEPs also
           contains the word process.
                       MR. McCARTIN:  Yes.
                       MEMBER GARRICK:  So it's not impossible
           for a process to be characterized as a scenario.
                       MR. McCARTIN:  Yes, right.  Generally, to
           date, probabilities, explicit probabilities have been
           assigned primarily to events.  The DOE and ourselves
           have not tried to assign a probability to a feature or
           a process.  It could be done, but to date it generally
           is looking more at the events.
                       MEMBER GARRICK:  Now I guess the other
           question I would want to ask is why are we doing this?
           Why aren't we just focusing on the risk of not meeting
           the standards and just calculate that risk?  Why do we
           want these thresholds?
                       MR. McCARTIN:  Probably as much for
           practicality purposes, I think.  Certainly, the 10-8
           cuttoff for very unlikely, I think you come to a point
           where how low do I want to try to go to estimating
           what might happen?  And I think 10-8 in part, is due
           to practicality of how far I want to go in probability
           space.  10-10, 10-12.  I mean you get to some pretty
           outlandish kinds of things and a lot of speculation,
           depending on how far you go down.  I think 10-8 is
           there, as much for practicality in that it seems to be
           a reasonably low level.  We don't want to go lower.
                       The unlikely events, in my opinion, is
           recognizing that we have two very special calculations
           here, human intrusion and groundwater protection and
           you're trying to get more -- you don't want to go to
           as extreme events as you've considered in the
           individual all pathway.  But you're looking at, in a
           general sense, what's likely to occur.  And it's one
           also of for these calculations will do a simpler kind
           of calculation.  I would say it's one of practicality,
           once again and -- but I don't know, from a risk
           standpoint should you look at a 10-10 event that has a
           consequence --
                       MEMBER GARRICK:  Well, we've run into this
           problem in the large scope risk assessments of nuclear
           power plants and the reason I ask the question about
           scenarios is that you can look upon a risk assessment
           as a structured set of scenarios.  And the question
           was if you impose screening levels, you need to be
           careful to not get in the position that there's 10-4
           events and so what we did was stored all of the events
           that were screened and included them in the aggregate
           calculation of the final risk measure which could be
           something like an off-site dose or a release category
           or a core damage frequency or what have you.
                       But that, at least, protected us against
           any accumulated effect that might come from a class of
           scenarios that met the cutoff criteria, but we didn't
           throw them out.  We kept them in the bottom line
           calculation.  Do you anticipate a similar approach
           here?
                       MR. McCARTIN:  To date, we have not been
           as worried about the completeness argument.
                       MEMBER GARRICK:  Right.
                       MR. McCARTIN:  That -- which is what I
           think you're getting at.  Could we throw away enough
           things that indeed we're really not representing the
           risk of the facility any longer.
                       MEMBER GARRICK:  Right.
                       MR. McCARTIN:  And I believe in the rule
           where we talk about we're expecting the events in
           these types of things to be defined in a broad sense,
           that we have not seen an effect that would, that is
           throwing out big chunks of risk, if you will.  I look
           at say volcanism, seismicity.
                       I mean these things are in and they're in
           a broad sense, they're not things -- we have not seen
           things being tossed out that the cumulative effect
           would be that now we have 10 percent of the risk and
           90 percent of the risk was tossed out.  Certainly, the
           desire was when we said that we were expecting a broad
           definition was trying to get at that fact.  I think
           our review would have to look at it.  We have to look
           at what gets thrown out, etcetera.
                       As you know, the rule also talks to not
           the amendment, but final 63 talks about we want to see
           DOE's basis for what they threw out and I think we
           would have to be sensitive and be able to answer that
           question.  Have we thrown out more risk than we've
           kept in, but right now I think because the categories
           tend to be broad --
                       MEMBER GARRICK:  Of course, another
           opportunity for accountability here would be the
           uncertainties, the uncertainties could be impacted by
           the impact that is judged from the FEPs that have been
           screened out.
                       In principle, I like the idea of a range
           versus a value, although as George says, the threshold
           is a value.  But the range is an interesting concept
           and I think a step in the direction of accounting for
           uncertainty.
                       In the decision to do that did you
           consider going the next step of actually imposing a
           distribution function?  I can imagine a p0 function,
           a probability frequency function that would
           characterize this parameter, this range even more
           fully in a risk sense.
                       Did you consider that at all?
                       MR. McCARTIN:  No.
                       MEMBER GARRICK:  Rather than a limit line
           as in the Reg Farmer limit line idea, you would have
           a limit distribution and that way --
                       CHAIRMAN HORNBERGER:  Does that mean that
           you would partially include the analysis in the human
           intrusion?
                       MEMBER GARRICK:  No, what I'm talking
           about is that if you do a calculation of the frequency
           of occurrence event and you characterize that
           calculation as a probability density function, does
           that function -- is that compatible with the criteria
           and the criteria could be a PDF.  That's all I'm
           saying.
                       CHAIRMAN HORNBERGER:  I guess I don't
           follow that.  It seems to me that what Tim is
           describing is a way to conform to what is stated in
           the EPA rule which is to define when an event sequence
           or whatever, however we want to call it gets included
           in, for example, a human intrusion stylized analysis.
                       MEMBER GARRICK:  Yes, I understand.  But
           all I'm suggesting is that --
                       CHAIRMAN HORNBERGER:  That requires a yes
           or no answer is what I'm saying.
                       MEMBER GARRICK:  Yes.  But I think that if
           you're starting to talk about criteria of what
           constitutes an acceptable value, you could do it in
           the context of distribution functions as well as you
           could do it in the context of ranges, as well as you
           could do it in the context of specific numbers.
           That's all I'm saying.
                       CHAIRMAN HORNBERGER:  Not to define a
           likely event.
                       MEMBER LEVENSON:  Well, in essence, it's
           the same thing.  The EPA cutoff says that you run
           hundreds of analyses --
                       CHAIRMAN HORNBERGER:  No, no.  I
           understand the rule, but it has nothing to do with --
                       MEMBER GARRICK:  George, all I'm saying is
           that you can represent likelihood in a number of
           different ways.  One way you can represent it is with
           the frequency without regard to uncertainty.  Another
           way you can represent it is with the probability and
           inherent in your probability would be an uncertainty.
           But another way you could represent it is in the
           probability of frequency and you could have the notion
           of frequency and you could say there's uncertainty in
           that frequency and I characterize that certainty in
           the form of a probability distribution.
                       And so all I'm suggesting is was there any
           thought --
                       CHAIRMAN HORNBERGER:  But still, it
           strikes me -- suppose you now identify an events
           sequence and you say well, there's a probability
           distribution associated with it and with central
           tendency as 10-6 or 5 times 10-6 per year and some
           dispersion parameter around it.  Is that event likely,
           unlikely or very unlikely?
                       MEMBER GARRICK:  Oh, I would, I guess I
           left out a part.  I would -- this is how I would go
           possibly instead of the concept of likely and unlikely
           and very unlikely.
                       CHAIRMAN HORNBERGER:  Okay.
                       MEMBER GARRICK:  To me, the concept of
           likely,
           very unlikely, etcetera, is an intermediate step 2
           that kind of -- that PDF kind of characterization.
                       CHAIRMAN HORNBERGER:  I understand that
           now.  But what I'm saying is that it strikes me that
           what Tim is faced with doing is conforming with the
           existing EPA regulation which doesn't permit that.
                       MR. McCARTIN:  I guess I would prefer the
           -- in the regulation it be fairly rigid in terms of
           less than 10-5, greater than 10-8, but in terms of
           application and I don't know if this would get to some
           of your desires, would be when I'm looking at
           assigning a probability to a particular FEP, could I
           use a distribution for the probability of that FEP and
           I think the answer is yes.
                       Now the issue would be one of well, is it
           in or is it out?  If I get a little bit over here and
           I think the intent of the rule where you're looking at
           mean kind of behavior, I would say if you use a
           distribution which is characterizing the uncertainty
           in your probability estimate, if the mean of that
           distribution would be -- you would use the mean to
           decide whether you're in or out.  That's something
           that I think we probably will be looking at, possibly,
           in the context of the review plan.
                       VICE CHAIRMAN WYMER:  But the real problem
           is going to occur when you come up with an event which
           is neither very likely nor very unlikely, but likely.
           Then you're on pretty shifting ground.
                       CHAIRMAN HORNBERGER:  No, then it's in.
                       MEMBER GARRICK:  Then it's in, yeah.
                       VICE CHAIRMAN WYMER:  But then you're
           going to apply a judgment.
                       CHAIRMAN HORNBERGER:  No, no, no.  Then
           it's in the analysis.  Then it has to be considered in
           the groundwater protection analysis and in the human
           intrusion, stylized human intrusion.
                       VICE CHAIRMAN WYMER:  Okay, I see what you
           mean.
                       MEMBER GARRICK:  One other comment, Tim,
           here.  You describe these as qualitative evaluation
           guidelines.  In order to establish compliance, you're
           really imposing a very quantitative requirement on the
           part of the analyst, right?
                       MR. McCARTIN:  Absolutely, yeah.
                       MEMBER GARRICK:  Because these thresholds
           are pretty definite.
                       MR. McCARTIN:  Well, you are right and the
           qualitative was more in terms of the preamble in the
           EPA standard.  This is a qualitative word.  Here are
           some of the things they said in a qualitative sense to
           give you an indication of what the intent was and by
           that, if you look at qualitatively the words, they
           were intending to have expected conditions, likely
           conditions and so when I look at -- I've got this
           qualitative word unlikely and the intent as the way we
           read it is to get to likely expected conditions and
           that is what allows -- we then went okay,
           quantitatively, you're absolutely right.
                       We want a very sharp, clear indication of
           how we will implement that and that's why we're --
           there is no doubt there.  I mean we're proposing 10-5
           as the one end and 10-8 at the other and likewise for
           very unlikely, 10-8 is very sharp. There is no
           qualitative nature to that.
                       MEMBER GARRICK:  I think the thing that
           you have to be very much alert to is that the analysts
           don't get so involved in these evaluation guidelines
           that they compromise the real issue here which is
           calculation of the risk of meeting the standard and
           that's -- no, that's the thing that a lot of these
           kind of bottom up considerations tend to do.
                       There is something people can get hold of
           and they run with them and it's one of the reasons why
           we, for example, were against subsystem requirements
           was we want the focus to be the measure of its been
           identified by law, in this case the three standards,
           15 milirem, 15 milirem and 4 milirem.  That's the
           context.  And that's something I think you'd want to
           be very careful with in this.
                       MEMBER LEVENSON:  So is the total thrust
           of this as George just implied a minute ago, to these
           categories, to decide whether or not you do an
           analysis, whether or not it's in the --
                       MR. McCARTIN:  Well, the analysis will be
           done.  The question is what kinds of FEPs do I include
           in the analysis and in reality I mean you can look at
           -- and I'll talk to both cutoffs.  The 10-8 cutoff, as
           you know, I think on the plus side, you can -- do you
           consider all the important things at Yucca Mountain.
           I mean you have volcanism in that calculation.
                       We can't think of many things below 10-8,
           I mean there's things that people --
                       CHAIRMAN HORNBERGER:  A Bolide impact.
                       MR. McCARTIN:  Which?
                       CHAIRMAN HORNBERGER:  A Bolide impact.
                       MR. McCARTIN:  Oh.  Tsunamis.  There's
           things --
                       CHAIRMAN HORNBERGER:  No, no, no.  A
           Tsunami isn't very likely to come over the Sierra, but
           if you look at a probability of a Bolide impact, it's
           not -- it's less than 10-8, but it's certainly not
           zero.
                       There are craters in Arizona.
                       MR. McCARTIN:  Yes, yes.  But I think one
           is sufficient enough to disrupt the repository 300
           meters down, I think has been characterized as less
           likely than --
                       CHAIRMAN HORNBERGER:  I was agreeing with
           you.  It's less than 10-8 if you're screening it out.
           Potentially bit impact.
                       MR. McCARTIN:  Yes.  So I think you've got
           those kinds of things in there.  When you get to what
           should we include in human intrusion and groundwater
           protection, certainly igneous activity at 10-5 falls
           out certainly and then you've got things like rock
           fall seismicity will still be in to some extent.  At
           least qualitatively that's how we see the
           calculations.  What is the impact of this particular
           amendment, that really is the primary effect and we
           think the calculations will be done.  It's just a
           matter of what gets included.
                       CHAIRMAN HORNBERGER:  Tim, it strikes me,
           this is just a follow-up on John's last comment that
           you already have, I think, good experience to draw on
           that goes to two of the things that John mentioned,
           first of all, a distribution of frequencies and second
           of all, what happens, how you make logical decisions
           about whether something is screened and that is you
           look at igneous activity and DOE's analysis as I
           recall right below 10-8, but with a broad distribution
           adn so it doesn't get screened out even though the
           mean or the median of the distribution is below 10-8.
           I think that you don't throw out good judgment when
           you draw these lines and I think that you have good
           experience with that already.
                       MR. McCARTIN:  Right.  I believe DOE's
           latest numbers have it slightly greater than 10-8, but
           it's on that edge, but that's one of the things,
           clearly, that we would be reviewing and looking at,
           things that are screened out that are near that
           boundary, be it the 10-8 or 10-5 and what's the basis.
           So yeah.
                       MEMBER GARRICK:  How important do you
           think these categorizations will become in the
           analysis, the unlikely, very unlikely and likely?
                       MR. McCARTIN:  The understanding we have
           to date would say that it's not very important.  With
           the simple -- with one simple caveat that there are
           some things that the calculation will be somewhat
           easier to do without and the -- not having to include
           igneous activity in the groundwater protection
           calculation and the human intrusion calculation.  It's
           one less thing to have to include there.  The
           calculations we've done to date would say that isn't
           -- wouldn't have made a difference.
                       MEMBER GARRICK:  In the nuclear weapons
           field they have a similar categorization and they have
           a very unlikely, well, they have unlikely, very
           unlikely and extremely unlikely.  Was there any
           discussion or interaction with them on this whole
           process and how it has worked?  They've been doing
           this for a number of years.
                       MR. McCARTIN:  No.  I can tell you when we
           first started on this amendment, we talked to a number
           of people in the Agency both within NMSS and NRR and
           the Office of Research in terms of give us some
           insights on the unlikely, very unlikely, etcetera.
                       MEMBER GARRICK:  Right.
                       MR. McCARTIN:  And the overall response I
           got from everyone was that it really depends on the
           exact nature of the calculation you're doing and the
           consequences and they said -- everyone was somewhat
           unwilling to do much without knowing more about the
           impact and what the application is.  And so I think
           for now, I think we're comfortable defining it in the
           context of Yucca Mountain, in the context of the
           10,000 year compliance period, in the context of human
           intrusion, groundwater -- it's a very specialized
           application and for here, I think it makes sense.
                       There aren't many other people that deal
           with the 10,000 year compliance period in terms of
           looking at consequences and so we didn't get a lot of
           information from the people we contacted, but we did
           not go outside the Agency, I will say that.
                       MEMBER GARRICK:  Any other questions?
                       MEMBER LEVENSON:  It seems to me an
           important factor here, John, and that is that it looks
           like they  no longer have to consider human intrusion
           occurring during a volcano eruption.
                       (Laughter.)
                       MEMBER GARRICK:  We used to call those
           combined loads.
                       (Laughter.)
                       Yes, Latif?
                       MR. HAMDAN:  Yes.  Tim, do you have
           anywhere in the rule some basis for the 10-5, why it's
           not 10-6 or 10-4, for example?
                       MR. McCARTIN:  Well, that's what the
           amendment is doing is proposing 10-5.  In the
           amendment, we certainly talk to the other three
           values, 10-4, 10-6 and give our rationale for it, but
           this is the proposal.  And once again, when we -- I'll
           say we go up before the Commission and even though
           OGC, the legal arm of NRC, for a proposal for
           rulemaking it's just that.  Is this a reasonable
           proposal?  That's not to say it's going to end up here
           at 10-5, but it seems and I think people agreed that
           it seemed a reasonable value to go out for public
           comment.
                       Will it end up there?  We've given our
           basis why.  We'll see what the public comment says and
           likewise, I think for the Committee and once again,
           this will be the third time, but I still want to say
           we apologize for not coming early on, but I think this
           will give you a longer time to think through this.
           It's our first cut.  That's why we go out for public
           comment and maybe 10-5, maybe there's a more
           appropriate number, but we'll see what happens during
           the public comment period.
                       MEMBER GARRICK:  Any other comments from
           audience or staff or anybody?
                       Okay, I think that satisfies us.  Thank
           you, Tim.
                       MR. McCARTIN:  Sure.
                       CHAIRMAN HORNBERGER:  Thanks very much,
           Tim.  Anything that anybody wants to bring up that we
           could accomplish in 10 minutes or should we break for
           lunch adn reconvene?
                       So because John and I have a meeting with
           Commissioner Merrifield at 1 o'clock, what I would
           suggest is that Ray and Milt could get us started --
           I don't know if either of you have read this yellow
           letter, I haven't yet.  But there are questions.  The
           big thing is to look through this research report and
           try to come up with the key things that we need to say
           in our research report.  So my suggestion is you
           should go ahead and have a read through that adn the
           two of you could discuss it.  That one and this white
           one that Dick Savio handed around, research.
                       MR. SINGH:  I haven't passed out the
           research report yet.
                       CHAIRMAN HORNBERGER:  Oh okay, so it's not
           around.
                       MR. SINGH:  But I have copies.
                       CHAIRMAN HORNBERGER:  Okay, so what I
           suggest is Dick, why don't you give copies to Ray and
           Milt and --
                       MEMBER LEVENSON:  When will you be back?
                       CHAIRMAN HORNBERGER:  Probably around
           1:30.  Okay?  So we'll start our letter writing
           session around 1:30 and we don't need to be on the
           record at all this afternoon?  Is that right?
                       So this will end the record keeping.
           Adjourned.
                       (Whereupon, at 11:54 a.m., the proceedings
           were concluded.)



Page Last Reviewed/Updated Thursday, March 29, 2012