UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                           WASHINGTON, D,C, 204SO
           10,  1988

The Honorable Lee 11. Thomas        EFA-SAB-RAC-89-003
Administrator
U. S. Environmental Protection Agency                                o".eg OF
401 M Street S.W.                                               ™e *OM,N)STR*TOR
Washington, D.C. 20460

Dear Mr. -Thomas:

     The Radiation Advisory Committee of the Science Advisory Board has
reviewed the Office of Kadiation Program's plans for revising the technical
basis for the Badionuclides MESHAP.   The Science Advisory Board's Dose
and Risk Subcommittee sent you separate reports on Low-LET radiation risks
and on risks associated with radon.  This letter transmits the report of
the Sources and Transport Subcommittee,

     The Director of the Office of Badiation Programs presented its
approach to the revisions in the May 23, 1988 memorandum,  "Badiation Bisk
Assessment Methodology*' and in the June 21, 1988 memorandum, "Review of
Clean Air Act Risk Assessments by Radiation Advisory Cksnnittee." Staff
from the Office of Radiation Programs supplemented these memoranda with
presentations at the open meeting of July 13-15, 1988 .  Members of the
public provided extensive written and oral public connent  on technical
issues.

     In considering whether the Office of Badiation Programs approach was
state-of-the-art and scientifically defensible, the Subcommittee addressed
many issues including;  the accuracy and completeness of the technical
data, the validity of the modeling approach, the relevance of the data and
nodeling to the objectives, the presentation of results, and uncertainty.

     Of the numerous findings by the Subcommittee, we wish to highlight
three which we believe to warrant the most serious attention by the Agency:

     1.   Portions of the AIRDOS-EPA methodology are no longer state-of-
          the-art, and must be updated to incorporate important recent
          advances in  modeling radiohuclide transport through environmental
          pathways.

     2,   To date, EPA's treatment of modeling uncertainties has been
          qualitative rather than quantitative although state-of-the-art
          methods for estimating uncertainty are available,

     3.   Best estimates (defined on page 9 of the report) with appropriate
          uncertainty statements should be used in all risk assessments.
          The "best" estimate should be statistically defined, according
          to the target population or  individual and the shape of the
          uncertainty distribution.

-------
                                  - 2 -
     To correct these deficiencies, the Subcommittee urges the
to make use of qualified groups and individuals to help implement immediate
and long-term improvements in model structures, uncertainty and sensitivity
analyses, and model validation.  Results from evaluations of similar
radiological assessments are available which the Agency could use now
to guide its immediate activities.  Innger-term efforts should involve a
substantive upgrading of radionuclide transport codes and ensure that the
methodology gains and maintains a state-of-the-art status.

     Detailed recaimendations which deal with these and other topics are
found in the report.

     These concerns aside, the Subcenndttee ccmnends the Agency for its
intentions to present radiation consequences as a function of risk
level, as in the benzene example cited in the presentation; for the initial
steps taken to validate the atmospheric dispersion code within AIRDOS-EPA;
and for the use of simplified models for initial screening in the case of
compliance procedures,

     The Subconinittee hopes the Office of Radiation Programs will incorporate
this advice into the Background Information Document and reminds the
Agency that the Radiation Advisory Committee has asked to review Volunes
I and II of the new Background Information Document when they are available.

     In considering the results of this review it is important to recall
that very similiar findings and recotroendatlons were offered to the Agency
by the Science Advisory Board in January 1984.  The apparent lack of
responsiveness on this matter by the Office of Radiation Programs during
this four year period is of grave concern to the Science Advisory Board.
It is the opinion o£ the Board that action is required now to assure that
future reviews will yield evidence of a more defensible scientific basis
for regulatory decisions on radionuclide emissions.

     The Subconinittee appreciates the opportunity to conduct this scientific
review,  v& request that the Agency formally respond to the scientific
advice transmitted in the attached report.

                                Sincerely
                                                .„
                                Norton Nelson, Chairman
                                Executive Coroittee
                                                 Board
                                                 \
                                William J. schull,j
                                Radiation Advisoi
           Chairman
           Ccranittee
                                       ion
 Enclosure
   Chairman
and  Transport Subccoinittee
  Advisory Committee
ccx  j. Moore
     EX Clay
     R. Guimond

-------
"NA11QNAL QUSSIGN STANDARDS FOR HAZA1DOUS AlE POLUUmMTS (NESteP)
                    SUNMEDS FOR MDIQNUCLIDES"
                 REVIEW OF ASSESSMQW tffilHODOLOGIES
                 Sources  and Transport Subcommittee
                              of the
                    Badlation Advisory Gocmittee
                U.S. Environmental Protection Agency
                       Science Advisory Board
                           November  1988

-------
                  u. s. ENviRONMEwm PROTECTION

                                  NOTICE
     this report has been written as a part of the  activities of  the
Science Advisory Board, a public advisory group providing extramural
scientific information and advice to the Administrator and other  officials
of the Environmental Protection Agency.  Hie Board  is  structured  to
provide a balanced expert assessment of scientific  matters related to
problems facing the Agency.  This report has not been  reviewed  for approval
by the Agency and, hence, the contents of this report  do not necessarily
represent the views and policies of the Environmental  Protection  Agency,
nor of other agencies to the Executive Branch of the Federal government,
nor does mention of trade names or commercial products constitute
a reconmendation for use*

-------
                                 ABSTRACT
The Environmental Protection Agency's Office of Radiation Programs described
Its plans to update the technical basis supporting the National Bnission
Standard for Hazardous Air Pollutants (NESHAP) for radionuelides.  Plans
relating to sources of radtonuclides in the environment, transport
modeling, exposure, sensitivity analysis, and uncertainty analysis were
described in a series of briefings at public meetings and documents
including Radionuclides, BackgroundInformation Document for Final Rules
(1984) and two memoranda ftorn tfe tiirectot of the Office of iadiation
programs "Radiation Risk Assessment Methodology" May 23, 1988 and "Review
of Clean Air Act Msk Assessments by Radiation Advisory Committee,"
June 21, 1988.

The Sources and Transport Subeonntittee of the Science Advisory Board's
Radiation Advisory Conmittee reviewed these plans.  Major findings and
reconroendations were made regarding the state-of-the-art of the transport
model (AIRDOS-EPA)» uncertainty and sensitivity analysis, model validation,
and the use of best estimates in risk assessment.  The Subcommittee found
that portions of the AIRDQS-EPA methodology are no longer state-of-the-art,
nor are they completely defensible from a scientific viewpoint because
important advances in modeling radionuclide transport have not been
incorporated.  Because treatment of modeling uncertainties in radiation
risk assessment by the Office of Radiation Programs has not been quantitative
or rigorous, the assessments cannot be scientifically evaluated.  The
Subcommittee reconnended that best estimates with appropriate uncertainty
statements should be used in all risk assessments.  The "best" estimate
should be statistically defined, according to the target population or
individual and the shape of the uncertainty distribution.

-------
                   u.s.  Wfmmmn&L PROTECTION
                          SCIENCE AOTSQIff BOARD
                       RADIATION AWISOKf CtMHTTEE
                    SOURCES AND TRANSPORT SUBCOMMITTEE
                                  ROSTER

Chairman

Dr, John Till
        Radiological Assessments Corporation
        Route 2, Box 122
        Nesses, South Carolina  29107

Manbers/Constiltants

Dr. James E. Martin
        University of Michigan
        School of Public Health
        Ann Arbor, Michigan  48109

Dr* H. Robert Meyer
        Chem-Nuclear Systems
        Uranium Mill Tailings Remedial
          Action Project
        Post Office Box 9136 c/o M£-F
        Albuquerque, NM   87119

Dr. Steven L. Simon
        Department of Environmental Sciences
        CB-7400
        Rosenau Hall
        University of North Carolina
        Chapel Hill, North Carolina  27599

Mr, William L. Terapletoti
        Battelle Pacific Northwest
        Post Office Box 999
        Riehland, Washington  99352

Mr. Paul Voilleque
        Science Applications International Corporation
        101 South lark Avenue - Suite 5
        Idaho Falls, Idaho  82402

Dr. F. Ward Whicker
        Department of Radiology
          and Radiation Biology
        Colorado State University
        Fort Collins, Colorado  80523

-------
Executive Secretary

Mrs. Kathleen White Conway
        Science Advisory Board (A-101F)
        U.S. Environmental Protection Agency
        401 H Street S.W.
        Washington, DC 20460  '

Staff Secretary

Mrs. Dorothy Clark
        Science AiJvisory Board  (A-101F)
        U.S. Environmental Protection Agency
        401 M Street S.W.
        Washington, DC 20460

Director

Dr. Donald Barnes
        Science Advisory Board  (A-101)
        U,S. Environmental Protection" Agency
        401 M Street S.W.
        Washington, DC 20460

-------
                            TMLE OF
1.0     INTRODUCTION                                                    1
2.0     OVERALL APPBOACH TO 1HE USE OF DMA AND                         3
        MODELS IN THE RISK ASSESSMENT
        2.1    Use Dose/Risk Assessment Models for                       3
              Deriving the Radionuclide NESHAP
        2.2   Objectives of Assessment Calculations                      3
        2*3   Input/Output Parameters                                   4
        2,4   Perspective on and Understanding of                       4
              Calculated Health Risks
        2.3   Limitations of Dose Assessment Codes                       4
              on Mainframe Computers
        2,6   Recannendationa on the Use of                             5
              Models for Radionuclide NESHAP
3.0     SCIENTIFIC BASIS OF RISK ASSESS^! MEIHDDOLOSY                  7
        3,1    Model Structure                                           7
        3,2   Model Validation                                          8
        3.3   Model Uncertainty                                         9
              (with definition of "best estimate")
               m
        3.4   Parameter Sensitivity                                     9
        3.5   Ifodel Documentation and Accessibility                    10
        3*6   Recoamendations on Models Used in the                     10
              Eadionuclidee Risk Asessment
4,0     USE OF SITE-SP1CIFIC EM3A                                      11
        4.1   Data from Other Federal Programs                         11
              Should be Incorporated in the BID
        4,2  Site-Specific Parameters and Measurements                 12
             Should be Used Whenever Possible
        4.3  Recomendations for the Aquisition and                    12
             Use of Site-Specific Data

-------
5.0     UNCEKQUM                                                   13
        5,1  The Role of Uncertainty in Risk Analysis                  13
        5.2  Improvements in the Estimates of Uncertainty              14
             5,2,1   Sensitivity Analysis                              14
             5.2.2  Parameter Variability                             14
             5.2.3  Propagation of Uncertainty                         15
        5.3  Recconendation Regarding the Estimation                   15
             of Uncertainty
6.0     MDDEI5 FCR CCMPLIANCE APPLICATIONS                             16
        6.1  Application of Simple Models                             16
        6.2  Recomendation on Alternative Compliance                   16
             Screening Model Development
7.0     KEEERBKXS                   "                                 17
APPENDIX A   May 23, 1988 Memorandum from R. J. Guimond
             to Donald Barnes: "Radiation Risk Assessment
             Methodology", May 23, 1988*
APPENDIX B   June 21, 1988 Memorandum from R. J. Guimond
             to Donald G. Barnes; "Review of Clean Mr Act
             Risk Assessments by Radiation Advisory Coraaittee,"
             June 21, 1988.

-------
1.0  INTRODUCTION

     The Science Advisory Board's Radiation Advisory Committee initiated
this review because revision of the "National Bnission Standard for
Hazardous Mr Pollutants;  Standards for Radionuclides" (NESHAP) is an
important activity which could benefit from the use of new data and
scientific techniques developed in the last five to ten years,  this report
will generally refer to that standard as the "Kadionuclides NESHAP".

     the Radiation Advisory Coumittee formed the Sources and Transport
Subeofflflittee to conduct the review.  The roster for this Subcommittee
appears at the front of this report.  The Subcommittee based its review
on two memoranda (see appendices) from the Director of the Office of
Radiation Programs (GBP) with their attachnaits (1,2), oral presentations
by OBP staff at the July 13-15, 1988 meeting, and public comments.

     The objective of this review was to examine the scientific basis for
the evaluation of source terms and radiological assessment models that
will be used in the revisions to the Radionuclides NESHAP Background Information
Documents scheduled for completion late this winter*  The Subcommittee
review of the risk assessment methods was scheduled at this time to
assist the Agency in meeting its court-mandated deadlines for issuing a
proposed rulemaking of February 28, 1989.

     The following members of the public provided ccuments on July 13, 1988;

          Dr. Donald Scroggin of Beveridge and Diamond PC
              on behalf of the Idaho Mining Association

          Dr. Leonard Hamilton of Brookhaven National Laboratory

          Mr. Joe Baretincic of DC Corporation
              on behalf of The Fertilizer Institute

          Dr. Edwin Still of Kerr-Md3ee
              on behalf of the American Mining Congress

          Mr. Louis Cook of Chevron Resources Corporation
              on behalf of the American Mining Congress

          Mr. Tony Thompson of Perkins Cote
              on behalf of the American Mining Congress

          Dr. Douglas Chambers of SUES Consultants
              on behalf of the Americang Mining Congress
              and The Fertilizer Institute

-------
                                  - 2 -

The Subcommittee appreciates these public comments,  which were well prepared
and technically enlightening, and believes the information provided should
be considered by the Agency in its ongoing revision to the N1SH6P Background
Information Docunents.

     Staff from the Office of Radiation Programs briefed the Subcomnittee
on planned changes to the methodology and data bases that will ultimately
be incorporated into the Background Information Document for the radionuclide
NESHAP.  However, since the Office of Radiation Programs is under severe
time constraints, the Subcommittee was not able to review the results of
calculations or revisions to methodologies that will be used.  Such
results may not become available until late winter.   Therefore, key issues
and recommendations of the Subcommittee are based on its review of previously
documented methods, the appended memoranda and oral presentations by the
Office of Radiation Programs staff.

     Since no formal issues were raised by the Agency in preparation for
this review, the Subcommittee, after studying the supporting docunents
and listening to briefings, identified five major topics for discussion.
These topics along with specific recommendations follow*

-------
                                  - 3 -


2.0   OVERALL APPROACH TO 1HE USE OF W& AND MODELS IN TH£ RISK

2.1   Use Dose/Risk Assessment Models for ^ Deriving the Badicnucllde
      NfeSHAP                  -   '

      The Subcommittee focused on the extent to which models should be
used in lieu of efforts to obtain measured data; whether the model should
be usable~~far both, deriving a standard and determining compliance; and
the manner in which input/output data are presented, especially the
output data regarding risk distribution and uncertainty.

     The Subcommittee concurs with statements of the the Envirooiental
Engineering Committee in its June 1, 1988 draft resolution on modeling (3),

          The use of mathematical models for environmental decision-
     making has increased significantly in recent years.  The reasons
     for this are many, Including scientific advances in the under-
     standing of certain environmental processes, the wide availability
     of computational resources, the increased number of scientists
     and engineers trained in mathematical formulation and solution
     techniques, and a general recognition of the power and potential
     benefits of quantitative assessment methods,  Ihe increased
     reliance on mathematical models is evident within the U.S. Environtnental
     Protection Agency (EPA) , where integrated environmental release,
     transport, exposure, and effects models are being developed and used
     for rulanaking decisions and  regulatory Impact assessments.

     Despite its appreciation of modeling, the Subcommittee believes that
measured data -best represent source strengths and environmental concentrations
and also near -source atmospheric and environmental concentrations from
sources subject to complex diffusion (such as near a building complex or
large gypsum or uranium tailings pile) .  the use of measured source data
for elemental phosphorous plants is a good example of a case in which EPA has
successfully benefited from this approach,  fclhere such data are not
available or cannot be obtained on the schedule required, it is appropriate
to use assessment models.

2,2
     Although the 1984 Background Information Doament (4) describes in
Volume 1 Chapter 6 methods to model the movement of radionuclides tteaugh
environmental pathways, it fails to identify clearly the specific objectives
of the calculations.  Examples of assessment objectives are: the calculation
of the maximum effective connitted dose equivalent Do the average individual
in an exposed population,  the effective dose equivalent per individual
in the most exposed population group and the probability that the average
dose in a critical group does not exceed a predetermined value.  Although
the methodology for various objectives may be similar, input data will
differ substantially depending on whether average or highly conservative
estimates are desired.

-------
                                  - 4 -

     The Subcommittee believes that ultimately it Is necessary to estimate
the expected number of health effects in the population as a consequence
of routine emissions (including predictable seasonal and episodic releases)
and to be able to relate this to an expected exposure level.  The uncertainty
of the estimated health risks inherently incorporates the uncertainty in the
exposure level.  Uierefore, full disclosure of the source and transport
ttncertainty way help quantify the total risk uncertainty and provide
additional input that can be used in setting, emission standards.

2.3  Input/Output Parameters

     The Subconinittee is concerned about both how input/output parameters for
dose/risk models are chosen and about the actual parameters selected.  This
concern stems from the knowledge that data and it's interpretations which are
clearly and thoroughly presented are more easily understood, more accurately
interpreted, and more readily related to other ccraaon data or studies.

2.4   Perspective on and Understanding of Calculated ifealch Msks

     It is essential to provide scientific data and analyses to the
scientific comnunity, to the risk management decision-maker, and to the
public in ways which show that often the calculated health effects may be
derived for a population at very low individual risk.  One effective tool
for this purpose is presenting the population distribution of the calculated
risk by individual risk level as is being considered in the draft revised
benzene standard doctraents (5).  A decision to ignore very low individual
risk levels is clearly risk management rather than risk assessment;
however, the data should be available to decision makers in a way that
provides the perspective necessary for informed judgments.  Similarly,
comparisons of these estimated risk levels with other commonly encountered
and accepted risks is necessary for perspective*

2.5  Limitations of Dose Assessment Codes on Mainframe Copguters

     The Subcommittee understands that the Agency is proceeding to develop
a replacement code for AIRDOS-EPA.  These new models will be embodied in
a Conputerized Radiological Risk Investigation System (CBRIS) on mainframe
machines.  Models implemented on mainframe computers are generally
inaccessible to all but a few specialists, are difficult to modify, and
are expensive.  Ihe restriction on accessibility limits interaction with
peer and interested user groups with the result that state-of-the-art
methodologies rarely get widely implemented in a timely manner.  Current
generation microcomputers are approximately equivalent in power to late
1970's mainframe machines on which current EPA dose assessment codes were
written.  It has been demonstrated that many transport ana dose models
can now be  implemented on current generation personal computers.

     The advantage of dose assessment models implemented on microcomputer
systems is  that they can routinely be made available for peer-rewiew.
Such  interactions would likely result to significant state-of-the-art
improvements being made to the Agency's methodology at no cost  to the .
Agency.

-------
                                  - 5 -

2.6  Reconnendationson the Use of'fiodels' for Radionuclides 'MESHAg

     Clearly the use of measured data as the basis for the Radionuclldes
NESHAP is preferable to calculations! models whaiever it can be reasonably
obtained because there is no need to estimate exposures if real and
representative data are used.  For example, measured ambient air concentration
are more defensible than an estimate of air concentration based on an
approximate source term and an atmospheric dispersion model,  Mien used
with care, models can be and are a necessary tool for deriving and complying
with the Badionuclides NESHAP;  however, attention should be given to
uncertainties and the  presentation of model inputs and outputs in understandable
and useful formats.  The Subcommittee makes the following recommendations
concerning the use of models for the radtonuclidea NiSHAP:

     1.  The EPA should use site-specific measured data on source terms
     and environmental concentrations especially for sources that represent
     complex assessment situations where current models fail.  EPA should
     also use site-specific measured data, or at least generic study
     results, where available, for other input parameters to the models*

     2.  Where sufficient data are not- available, the EPA must apply
     updated state-of-the-art ealculatianal models to its derivation of
     the radionuclides NESHAP,  To do so, EPA must intensify its efforts
     to employ current and state-of-the-art nodels for each major model
     conponent used to determine the risk to public health from various
     radionuclide emissions sources.  EPA muse also incorporate both
     recent advances in modeling methods and the results of validation
     studies In environmental transport and plume dispersion models*

     3.  The EPA must clearly state the objectives of the risk assessment
     calculations.  The Subcommittee recommends clarifying both the
     objectives of the assessments and the steps necessary in the ecological
     and dosimetric modeling to meet those objectives.  Statements of
     objectives are necessary to provide information regarding the intended
     conservatism or realism of the assessment calculations.  The
     clarification of objectives will also serve as a guide in making
     decisions to use conservative or realistic model assumptions and
     in the choice parameter values.  Specifying the objectives will be
     invaluable in justifying the choice of parameter values, thus making
     the results more defensible*

     4.   The Subcommittee strongly suggests that dose estimates be
     realistic, relevant to defined populations, and accompanied by
     a quantitative statement of uncertainty which can be propagated
     into the dose-risk framework. Scenarios can be used as part of
     this approach.  For example, if continuous exposure at a certain
     location is part of the scenario, the occupancy factor  is  fixed (at
     1UO%) and only the variations in the other parameters contribute  to
     the uncertainty estimate.

-------
                             - 6 -

5.  Hie EPA must clearly display input:/output parameters used in
the calculational models for the iadionuelides NESHAP risk assessment.
Particular attention should be given to the population distribution of
calculated health effect estimates among the population at risk.
These estimates must be displayed as a table showing the distribution
of risks over the population, broken down by such categories as:

     a. the individual risk level,

     b. the size of the population subgroup at that risk
        level, and

     c, the estimated incidence of particular effects that
        occur at the given individual risk level in the particular
        population subgroup.

The Subcommittee strongly supports the presentation of calculated
risk data for the ladionuclides NESHAP standards in a format similar to
that in the memorandum on benzene (5).

When preparing supporting documents for the Eadionuclides NESHAP, EPA
should display all assumptions, Input parameters and research and
studies upon tthich they were based*  The presentation of uncertainties
(See Section 5.0) will also contribute to greater credibility and
understanding of the risk assessment process,

6.  All dose assessment computer codes for radionuclides should be
developed for use on microcomputers unless code size and complexity
requirsients justifies the use of mainframe machines.  Such codes
must be made readily available for review by outside peer, expert,
and user groups.

-------
                                  - 7 -

3.0  SCIENTIFIC BASIS OF RISK ASSESSMENT METHODQljQSY

     Portions of the AIIDQS-EPA methodology are no longer state-of-the-art,
and must be updated to incorporate important recent advances in modeling
radionuclide transport through environmental pathways.   The current
transport methodology was state-of-the-art and scientifically defensible
some 5-10 years ago.  However, EPA's general methodology, and the radionuclide
transport sections of AIRDQ8-EPA in particular, have not changed substantiveiy
since that time.  Many advances have been made in the field of modeling
radionuclide transport within the last five years but 1PA has not incorporated
such advances into its own methodology*  Examples of such advances that
are not currently reflected in AIEDOS-EPA are discussed below under the
categories of model structure, model validation, model  uncertainty, parameter
sensitivity, and model documentation and accessability,

3.1   Model Structure

     The food chain portion of AIRDOS-EPA is a steady-state model adapted
from earlier codes such as HEKiES(6)and formulations in the U.S, Nuclear
Regulatory Commission Guide 1,109(7).  The main differences between
AIRDOS-EPA and these earlier methodologies involve the  choice of certain
parameter values.  The depositlon-ingestion sections of AIBDQS-EPA are
baaed on straightforward formulae that are well-documented and generally
accepted.  The choices of parameter values are generally based on relevant
scientific literature.

     The Subcommittee favors the use of dynamic models  because there are
distinct disadvantages of steady-state models such as AIRDOS-EPA.  For
example, predictions of steady-state models only apply  accurately to chronic,
constant release scenarios.  In practice, enissions from many types of
facilities are not constant, but rather episodic or seasonal.  Furthermore,
steady-state models are not well-suited to handle the very marked seasonal
changes in climatological conditions, agricultural practices, and food
distribution patterns.  Finally, steady-state models are not fully testable
because many data sets are in the form of time-series measurements, which
cannot be directly compared to steady-state model predictions.  In short,
steady-state foodchain models are limited in application, not realistic,
and not readily subject to direct validation.  Several  dynamic foodchain
models have been developed outside EPA, including RAGTIME (8), 1CCSYS (9)
RADFOOD(10)and PATHWffif(H).  These codes incorporate the dynamic processes
necessary for more realistic simulation of radionuclide transport ttawigh
the environment.  Dynamic models of course, do handle chronic, steady-state
release easily, and they are not difficult to structure or program.

     Numerous parameters which are known to vary considerably in time
and space are treated as constants to AIRDOS-EPA*  These include, for
example, the pasture intake of dairy cows, the foliar interception
fraction, and the source fraction of various foods to people in a particular
locale.  Recent, more updated models, have successfully dealt with these
variations  to produce more realistic estimates  (11).

     Several basic pathways which are frequently  important  in the  natural
environment are not included  in AIRDOS-EPA.  For  exanple, resuspension of

-------
                                  . 8 -

recent deposits cm the soil by wind or ocher disturbance is a. very important
process for arid and semi-arid environments, especially sfaan longer-lived
radionuclides are involved (11).  Sensitivity analyses for many simultaneously
varying parameters reveal that resuspension can be,  in some cases, a
dominant process affecting dose to man (12),  Another example of omitted
pathways in AIRJDOS-EPA is soil ingestion, both by cattle and people,
especially children. Itese phenomena, and numerical  estimates, are
well-doaiaented in the literature (11,13)*  The relative importance of soil
ingestion is usually small, but under some circumstances, this pathway
can be considerably more important than others,  "these pathways, by their
omission, may in sorae cases offset the gaierally conservative choices of
parameter values in AIWBS-EPA.

     As another example of shortcooings in model structure, to the atmospheric
diffusion portion of AIRDQS-EPA, the code does not deal adequately with
cotiplex terrain and building wake effects,  furthermore, the use of the
harmonic mean of morning and afternoon lid heights throughout the day was
questioned by the Subcommittee.

3.2  tfedel Validation

     Efforts by 13PA to validate or test the accuracy of AIRDQS-iPA appear
to, have been minimal, especially for'that portion of the model subequent
to dispersion which treats deposition, environmental transport and ingestion.
The Gaussian plume model portion, however, has been  compared to real data
sets with encouraging results* for which EPA should  be conmended.  Without
a good deal of effort to validate as many steps in the risk assessment
calculations as is possible, the results will always be subject to criticism
by the public, as well as the scientific community*   For exanqsle, there
has already been considerable criticism by representatives of industry
who claim that due to over-conservatism in the assessment models, the
regulatory standards are unreasonably restrictive.  Others are likely to
look for the other extreme, arguing that standard are too permissive.
Without convisctag model validation data, !PA will be unsure of their
degree of conservatism or accuracy and therefore have continual difficulty
in defending some of its regulatory positions.

     Hie field of radlonuclide transport model validation is relatively
new, but rapid advances are currently being made in  the U.S. and in
mmerous other countries*  We are rapidly progressing froa peer-review
and model comparison exercises to real-world coraparsions between model
predictions and independent field data sets.  Because of its scrutiny
by peers and the courts, the PAlHWff model received  fairly exhaustive
validation testing some five years ago with data sets made possible by
extensive foodchain sampling programs in the western U*S, during the
latter part of the weapons fallout era (14).  More recently, the BICHDVS
(Biospheric Model Validation Study) effort was initiated by the Swedish
Government and has matured into a truly international effort, involving
some 15-20 nations.  The BIQMQVS program gained exceptional momentum
from the Chernobyl accident, which resulted in the accumulation of
extensive data sets from at least a dozen sites world-wide.  Active U.S.
participants have not included people from EPA.  A similar model validation
effort has been initiated as a Coordinated Research Program by the  International
Atomic liiergy Agency, but again, without EPA participation.

-------
                               m 9 -.

     Clearly, model validation is crucial to the achievement of public
and scientific credibility of risk assessments.  A reasonable, workable
methodology exists, as do numerous data sets*  There is ample opportunity
for EPA to bolster its effort in the area of model validation.

3.3  Model Uncertainty

     The general field of risk assessment is rapidly moving away from the
practice, of giving single, perhaps wrst-caie estimates, to that of,'
jarwHing best estimates along vidh" YVtatement of the uncertainty of
tnat best* estimate.Uiis new, evolving practice reflects the attempt
by scientific modelers to exercise complete honesty and full disclosure
in arriving at dose or risk estimates.  All model structures and parameter
values have inherent and unavoidable uncertainties which owe to real-world
complexity and variability, as well as to a lack of knowledge, data, or
both.  Therefore all model predictions contain corresponding uncertainties,
Without rigorously derived  uncertainty estimates, the credibility of
dose or risk values cannot be judged.  My enlightened reviewer will
likely assign a very low credibility to an estimate not accompanied by a
statement of uncertainty*

     In the case of AIRDQS-EPA, it is clear that little or no formal
propagation of uncertainties through the methodology has been carried
out.  While data with which to construct uncertainty distributions on
many parameters is lacking, it is still reasonable to construct such
distributions, reflecting the actual degree of ignorance on the part
of the modeler.  Methods for propagating uncertainties through radicwuclide
transport models are available (15,16), as are published estimates of
uncertainty for many critical transport parameters (12,17,18,19)*

3.4   Parameter Sensitivity

     An important aspect of model evaluation is that of xmderstanding the
relative degree- to which individual processes or parameters affect the
model prediction, and the degree to which uncertainty in a parameter
affects model output uncertainty*  A sensitivity analysis can be carried
out simultaneously with an uncertainty analysis (12), or it may be done
independently.  Modeling is seldom perfect, so as long as needs justify
and resources permit, modelers should strive Co continually evaluate and
improve their models*  Conducting a series of sensitivity analyses is the
most efficient way to reveal the most influential pathways and parameters»
and thus to guide the expenditure of resources and effort for the sake of
model improvement.  Sensitivity analysis techniques are readily available
and have been successfully applied to dose assessment models  (11,12,20),
It is not evident to the Subcommittee that EPA has made any substantive
effort  in the area of sensitivity analysis related to the Eadionuclides
N1SHAP  or,  in particular, AIRDQS-EPA.
   The Subcommittee defines  "best  estimate"  as the  arithmetic mean in the
   case of normal distributions and the geometric mean (median)  in the
   case of log-normal distributions.   The best estimate for other distribution
   shapes requreg specific statistical definition to avoid confusion.

-------
                                  - 10 -

3.5   Model Documentation and Accessability

     The current EPA codes for radiation dose assessment are not clearly
and concisely documented, nor are Chey readily available for outside use
or peer review,  this hinders progressive evolution of the codes because
independent critique and input is made difficult.  It is useful to concisely
document models in the open literature so that they can be openly examined*
Such documentation should include a clear statement of the objectives of
the models, including a definition of the target individuals or population
groups to *hich the output applies aa well as a careful exposition of and
justification for the model structure and parameter values.  The advances
in the power and speed of personal computers have been shown to make
possible their use for many complex models.  The ability to distribute
models enhances the process of positive model evolution.

3.6  Seconipetujlaticos on Models Used in EPA's Radionuclide Risk Assessment

     1.  The Office of Radiation Programs must become state-of-the-art in
its risk assessment methodologies.  Ihe transport portions of AIRDOS-EPA
need extensive revisions.  Methods already developed by other groups for
model validation, uncertainty and sensitivity analysis need to be incorporated.
This task nay be accomplished taost efficiently by establishing a close,
continuing working relationship with a group or individuals acknowledged
to be current in these fields.  Immediate use of uncertainty estimates and
validation exercises from other transport models is essential if EPAs short-term
goals for NESHAP development are to be achieved.  For the longer-term,
EPA should develop its own capabilities with the help of others and
participate more actively to national and international meetings devoted
to these topics*

     2.  The Office of Radiation Programs should carefully define the
generic individuals and/or populations to which its risk assessments are
targeted and carefully articulate these definitions in the Background
Information Document and other relevant documents.

     3.  The dose/risk assessments conducted by EPA must provide best
estimates (as defined on page 9) along with statistically appropriate
measures of uncertainty.  The probabilitites of individuals receiving
doses or risks at various fractions or multiples of the best estimates
should be clearly revealed in all numerical presentations.

    4.  As the Office of Radiation Programs develops new software to
accomplish dose/risk assessments, codes compatible with personal computers
should be encouraged.  This strategy is not only coat-effective, but  it
facilitiates future improvements, communication capabilities, and credibility
within the public and scientific community.

     These recoraaendations are consistent with, and in some cases almost
identical to, those developed during the Science Advisory  Board's 1984
review (21).

-------
                                  _ 11 _

4.0  THE USE OF SITE-SPECIFIC DATA

4.1 Data from Other Federal Programs Should be Incorporated in the BID

     The EPA is noc always using the most appropriate data available in the
performance ot radionuclide NESHAP development.  Since preparation of the
Background Information Document in 1984, a great deal of new data, of
significant potential value to this 
-------
                                  - 12 -

4*3  Recommendations for the Aqulsition and Use of Additional Data

     1.  The EPA should initiate a thorough survey of all current remedial
action programs sponsored by the government.  !he survey should identify
key personnel within each project capable of quickly providing the relevant
data,  Tne Oak Ridge National Laboratory's report, Remedial Action Contacts
Directory. M»uld be a good starting point, (24)

     2.  The EPA should request immediate access to other federal data
relevent to the Badfonuclides NESHAP work.  Uiese data include the following.

          a.   Radioactive particulate concentrations.

          b.   Nonradioactive dust concentrations,  (Supplementary, for
               comparison purposes),

          c.   Radon and other radioactive gas concentrations.

          d.   Meterological data.

          e.   Radionuclide concentrations in the specific source material
               (e.g., tailings or gypsum stacks)*

          f.   Particle size information (pile and airborne).

          g.   Solubility information (standardized lung fluid tests).

          h.   Quality control information defining conditions -under which
               the data were collected and analyzed.

     3,  The EPA should use existing data sets to correct the results of
AIRDQS-EPA for specific sites.  For example, environmental monitoring
data provided by the Mount Taylor representatives and the New Mexico
Environmental Improvement Division study of radon concentrations in the
Grants New Mexico area, would provide a basis for evaluating the results
of AIRDQS-EPA predictions for that specific mine's emissions.  EPA should
perform similar corrections for all other facilities for liiich measured
concentration data are available.

     4.  Data sets acquired from outside sources must be inspected
carefully for systematic quality control errors, to allow evaluation of
the accuracy of results employing that information.

-------
                                   - 13 -

5.0
5,1  The Role of Oncertainty ^' in ,'Msfc' Analysis

     Because quantitative estimates of the uncertainty provide very important
information tQ~the decision maker and others concerned with Che risk
decision, the Subcommittee recommends in strongest terns that EPA make
quantitative estimates of the uncertainty associated with the Agency's
dose and risk estimates.  Calculations of uncertainty clarify the reliability
of the central estimate and provide information essential to understanding
the reliability of the estimate,

     In 1984, the Science Advisory Board recommended (21) that the Office
of Radiation Programs explicitly present uncertainties as part of the
radionuclides risk assessment.  The Office of Managanent and Budget (25) »
the Office of Science and Technology Policy (26) » the National Science
Foundation (27) » and the EPA Administrator (5) , have further emphasized
the need for defining uncertainties in risk assessments.

     the October 1984 Background Information Document (4) summarizes
(Vol. I, p. 7-29) some sources of uncertainty and the "reasonable" accuracy
which was stated to be a factor of three to four*  Ihe problem with this
qualitative approach is that there is no way to substantiate the stated
range even though a "factor of three to four" may correctly describe the
accuracy of dose calculations to represent typical members of the population.
This assumed range of error in the source term and environmental transport
is close to that estimated  for the dose response models (e.g., QKP's
risk estimate of 120-750 lung cancers per million person WW» with a
central estimate of 300, implies an uncertainty factor of 2.5). (28)
Because the uncertainty in the source term and environmental transport
models is believed to be of the same order of magnitude as that for
dose-response models , uncertainty estimates for source terms and transport
play an important role in establishing the total uncertainty of the
calculations of health effects*

     The uncertainty statement , however, must have an interpretation that is
understood and preferably is of use in decision making.  The uncertainty
estimate is more than simply a statement about lack of knowledge.  Given
the proper conceptual framework, e.g. establishing probability distributions
of parameters based upon expert judgement or data, the uncertainty estimate
can be used to express the probability that the true dose does not exceed
a specific value.  This framework enables the uncertainty estimate to be
used in a meaningful way for decision making*

     To avoid misleading the decision-maker, uncertainty statements should
also be accompanied by a discussion of what the model does and does not
include.  To the extent that a model omits certain pathway or processes ,
it is incomplete, however, uncertainty analysis cannot assess the
completeness of a model.  Because uncertainty analysis, can only reflect
the pathways a«i/or processess accounted for in the model, it cannot

-------
                                  . 14 -

defensibly compensate for the emissions of pathways and/or processes.
For example, to AIRDQS-EPA the absence of relevant pathways (as identified
in Section 3.0, page 7) cannot be adequately accounted for by uncertainty
analysis without arbitrarily inflating the total uncertainty in an indefensible
manner,

5*2  Improvements in the Estimates of uncertainty

     It is essential that the EPA progress from qualitative, estimates
of uncertainty to soundly based numerical estimates that cover all
portions of the calculations.  The need for estimates of the uncertainty
in the risk assessment results was identified in the initial review of
the SAB in 1984 (21).  Although some qualitative and numerical estimates
were given for portions of the source, transport, and dose calculations
in the 1984 BID, the overall uncertainty in the estimates of dose was not
evaluated in an integrated and focused manner (4).

     The QKP has stated its intention to again provide qualitative estimates
of uncertainty and believes them to -be adequate (2),  The Subcommittee
strongly disagrees because the proposed QBP approach is not state-of-the-art
and the argument that it is too difficult to perform a quantitative
evaluation is not valid.  Currently the capabiliity to perform Msnte
Carlo calculations yielding probability distributions for the dose estimates
is widely available on personal computers.  Techniques for these stochastic
calculations have been described and used by several other groups in
similar evaluations of dose from particular  sources of radionuclides
released to the environment.  (References 12,15,16,17,20, and 29, for
example)*  The available techniques and desktop calculations! capabilities
permit the improvements recommended below to be accomplished in a timely
manner.

5.2,1  Sensitivity Analysis; The Agency must  perform sensitivity analyses
to identify ther most critical parameters for the important exposure pathways
for the various source categories.  The EPA has already identified some
critical exposure pathways as the result of dose calculations presented
in the 1984 Background Information Docuaent.  (See Table 7.6-1, Volume 1
page 7-28 and the assessments for specific source categories in Volime 2.)
For most of the categories» inhalation is the critical pathway.  Food
chain transport was found to be important for the "DOE facility" category
and may also be important for the  t®C licensee and other federal facility"
and the "uranium fuel cycle" facility categories.

5.2.2  Parameter Variability: The EPA must define the distributions of
the most important parameters identified in the sensitivity analyses.
The problems in establishing reasonable probability distributions are
often less difficult than expected for several reasons.  This procedure
can be facilitated by establishing the maximum conceivable range of
values and the estimate of central tendency,  Multiplicative models  have
been shown not  to be extremely sensitive to distribution shape,  a finding
 that can be confirmed by modifying distribution types and comparing
 results produced by the various assumed distributions.  The EPA must

-------
                                  - 15 -

also define the distributions of the measurements or estimates of the
source terns for the various categories.

5.2.3  Propagation t^of _ttacertainty: For each source category,  the EPA
should perform Sonte "Carlo calculations to determine the dose distribution
that results from variations of the critical parameters.  Note that these
calculations can often be performed separately on a personal  computer
without running AIRDOS-EPA repetitively.  This is accomplished using a
reduced model which explicitly considers only the critical variables.

     The reduced model should yield nearly the same final result as the
coi$>lex model.  Hottte Carlo calculations are then performed for those
variables to generate frequency distributions for the estimated dose.
Calculations such as those performed by Dr. Chambers and submitted as
part of his testimony on July 13, 1988 exemplify what can be done (30),

     It is also possible that analytical error propagation methods may
work sufficiently well for simple exposure pathways.  Tnose pathways,
e.g. inhalation, that are not modeled by a large mmber of parameters or
processes may be especially amenable to this treatment.  It appears that
for 8 of the 11 source categories in dP's June 21 aemorandtin (2), inhalation
may be the main exposure pathway.  The principal differences between the
pathways would be the variability of the source term and local meteorology.

     When assumptions must be  made regarding the shape of input parameter
distributions, the uncertainties of parameters will also reflect the lack of
knowledge of environmental processes.  Uncertainty statements should also
be made for systematic errors which result in model bias.  Model bias was
seen for individual sites in the comparison of AIRDOS-EPA with measured
values (31).  For any one site, the predictions were consistently above
or below the taeasured values.  Such comprehensive evaluations significantly
contribute to the ability to make quantitative uncertainty statements.
For example, the spread of predictions after adjustment for the observed
bias can be used as an estimate of the uncertainty in downwind air concentrations,
at least for the sites considered in the comparison.

5,3  EBomeatime^d^_he Estimation of tftrartauity
     The Subcommittee strongly recommends that the EPA make quantitative
estimates of uncertainty for the risk assessment for each source category.
These uncertainty estimates and their bases need to be presented as part
of  the Radionuclides NESHAP.

     The EPA does not have time to conduct a comprehensive quantitative
uncertainty analysis of AIBEQS before publication of the revised proposal
in  February 19S9.   It is, however, both possible and desireable for the
Agency to make some interim quantitative estimate of uncertainty based on
studies of similar  models.  Therefore, the EPA should acquaint itself with
ongoing and completed studies of uncertainty in environmental transport
models, report the  nature of the uncertainties studied and their magnitudes,
and discuss those findings and models in relation to AIRDOS.  The sensitivity
analysis, studies of parameter variability, and propagation of uncertainty
identified above will take longer to complete and should therefore be
started promptly so that they may be used in the final regulations.  Experienced
people could be  realistically expected to complete such work within  two  years.

-------
                                  - 16 -

6.0  ropEis. TOR.
6,1  Application of Simple Models

     The commercial and non-conmercial use of radtonuclides Is licensed
by U.S. Nuclear Regulatory Commission (NiC).  A large proportion of these
licensees involve the use of small quantities of radionuclides which
likely represent a very small risk to the public.  The Subcommittee
believes that the series of computer codes presently employed by the
Agency for the Radionuclides NESHAP are complex and virtually unavailable
to most scientists and other users because they are on main frame computers.
These limitations for demonstrating compliance must be recognized.  Ihe
Subcommittee believes that an approach originally reconmeuded by the
National Council on Radiation Protection and Measurements (NQRP) , of
applying the most simple models first, followed by a more coop lex model,
if necessary, is appropriate (32).  It is, therefore encouraging to
note that the Agency recognizes that other simple, user-friendly and less
costly model programs are available P can meet the same objectives, and
would be more appropriate to demonstrate compliance.  However, a formal
process must be established for comparing the results of any alternative
methodologies with that of the IPA's to facilitate their approval and use,

     A tiered approach which meets this criteria is being proposed for determining
compliance using Annual Possession and Mr Concentration Tables, application
of Level II and III of the NCRP Screening Model (33) and/or EPA's microcomputer
Code (COMPLY).  This methodology appears to be based on sound environmental
transport and radiation protection principles,  however, the Subcommittee
has not specifically reviewed these methods in any detail for such compliance
applications.

6,2  Recommendations on Alternative Cknpjj.anee Screening Model DevelOEjaent

     The Subcommittee recommends that EPA develop criteria for the evaluation
of alternative compliance models and publish a process for gaining their
approval.  The Subcommittee strongly supports the IPA's proposed tiered
approach for NBC License compliance and recommends its application for
the Eadionuclides NESHAP.  The Subcommittee also strongly encourages EPA
to subject these compliance procedures to peer-review.  High priority
must be given to making the proposed methodologies available to users in
a timely manner.

     The Subcommittee also encourages EPA to apply the same philosophy
and approach, i.e. simple models first, followed by more complex methods,
where appropriate to assess compliance for categories of sources other
than radionuclides,

-------
                                  - 17 -

 7.0  REFERENCES

  1,  R.J. Guimond, Memorandum to Donald Barnes;  Subject:  "Radiation Msk
      Assessment Methodology", May 23, 1988,

  2.  R. J. Gutmond, Memorandum to Donald G. Barnes; Subject: "Review of
      Clean Air Act Risk Assessments by Radiation Advisory Committee,"
      June 21, 1988.

  3.  Science Advisory Board, Environmental Engineering Gonmittee,
      SAB-*EEC Resolution' on' the ' Ose' of ' Ifej^emtical'ttedels ' for' Integrated
      lOTsi .Assessment in Rule* and ^ision-Ilgd^ at_EgAf Draft by
      Mitchell Snail dated June 1988.

  4.  Envitomental Protection Agency* 1 984    Kadlonuclideg . Background Information
      Document for. Ji-^faiLeB t Volumes I and I'f.                    '"'* M™L- .....
  5.   Lee M. Thomas, Memorandum on "Proposed Benzene NESHAP Decisions and
      Limitation of Issue  to Section 112 of the Clean Air Act," April 5, 1988,

  6,   Fletcher,  J. F, and  W. L, Dotson* 1979,  HUMES; A Digital Computer
      Code  for Estimating  Regional Radiological Effects from the Nuclear
      Power Industry.   Battelle Northwest Laboratory.  HEEL-IKE-? 1-1 §8

  7.   U.S*  Nuclear Regulatory Commission. 1977. Calculation of Annual
      Doses to Man from Routine Releases of Reactor Effluents for the
      Purpose  of Evaluating Compliance with 10 CFR Part 50, Appendix I,
      Regulatory Guide  1*109, Revision 1.

  8.   G.G.  Killough and F.O. tfoffman* 1988,  Validation of the M5TIME87
      Dynamic  Food-Chain Model Against Fallout Datafrom the Chernobyl
      Accident.  J. Tem^afeia£1Sci.' €3J*SS^''"'?a^^*r' .  '      *
                                         ,,,:             .
  9.   H.G.  Paretzke, P. Jacob,  fcijpBller. and S. Prohl. 1988.  Concept
      and Validation Studies of the Real-Time Seactor-Accidesnt Consequences
      Assessment Model "ECCJSfS". Abstract in Radiation Protection Practice.
      Volume I.   Proceedings of the 7th IIPA. Congress.  Author's address;
      GSF-Institote for Radiation Protection, P-8042, Ifeuherberg,
      Federal Republic of Germany.

10,   Koch, J. and J,  Tadmor.   1986.  RABFOQD-A. Dynamic Model for
      Radioactivity Transfer Through  the ftnan  Foodchain.  Realtii
      Phys. 50:721-737,

11.   Whicker, F, W. ,  and T. S3. Kirchner.  1987.  PATHWHi  A Dynamic
      Food-Chain Model to Predict Radionuclide  Ingestton After Fallout
      Deposition.  Health Phys. 52 (6): 71 7-737,

-------
                                  - 18 -

12.   Otis,  M.  D.   1983*  Sensitivity and Uncertainty Analysis of Che PATHiS
      Radixmuclide Transport Model.  Ph.D.  Dissertation, Colorado State
      University,  Fort Collins,  CO,
18.
13.   La^oy,  P.  K.  19B7,  Estimated soil ingestion  rates for use  in risk
      assessment.   Risk Anal.  7:355-359.

14.   Kirehner,  T.  B.  and F. W.  Whicker.   1984.  Validation of PATHWAY:
      A Sinulation  Model  of the  Transport  of ladionuclldes Ifarough
      Agroeeosystens.  Icol. Modelling,  22:21-44.
15.   R.  0.  Gilbert.  1984.  Assess ing 'Uncertainty' in Pollutant, transport Models.
      1RANS-STAT,  Statistics for Environmental Studies, Novaaber 27,
      PNL-SA-12521,

16.   F.  0.  ftoffiaan and 1,  H. Gardner.  1983.  Evaluation of Uncertainties to
      Radiological Assessment Models in Radiological Assessment, J. £» Till
      and H, E. Meyer (Eds.),   Nuclear Regulatory Commission report,
      NURBS/CR-3332.

17.   Hoffinan, F,  0.  and C. F*  Baes, III.   1979.   A Statistical Analysis
      of Selected  Parameters for Predicting Food  Chain Transport and Internal
      Dose of Radionuclides. Oak  iidge National  Laboratory, Oak Ridge, TN»
      NURffi/CR-1004/OBNL/NUiiEHM-282.
Iman, R. L. and J* C, Helton.  198S.   An investigation of uncertainty
and sensitivity analysis techniques  for conpiter models.   Risk Anal.
8:71-90.
19.   Breshears,  D. D. 1987.   Uncertainty and sensitivity analyses of simulated
      concentrations of radionuclides  in milk. M.S.  Thesis. Colorado State
      University, Ft. Collins.  95p.

20.   R. L. Man, J. C. Helton,  and J* E* Campbell.   1981.  An Approach to
      Sensitivity Analysis of Computer Models, Part  II  -  Banking of Input
      Variables,  Response Surface Validation, Distribution Effect and
      Technique Synopsis, Journal of Quality Technology,  Jl(4): 232-239.

21.   Science Advisory Board Subcommittee on Risk Assessment  for Radionuclides
      Report on the Scientific Basis  for IPA's Proposed National Qaission
      Standards for Hazardous Air Pollutants for Mdionuclides, August 1984.

22.   Mayer, H. R., C. Begley and C,  Daily,  "Held  Instruments Developed for
      Use on the UMJffiA. Project," Proceedings of  the  Waste Management  1987
      Annual Meeting, University of Arizona, Tucson, March 1987.

23.   Ifeyer, H. R. and C. Daily, "QA Verification Procedures  in Uranian Mill
      Tailings Processing Site Remedial Action", Proceedings  of the American
      Society for Quality Control, Second Topical Conference  on Nuclear
      Waste Management Quality Assurance, Las Vegas, Nevada,
      February 9-11, 1987

-------
                               - 19 -
24,  Knox, !. M. (publisher).  tenediai^ Aetiotx' Cbptacts ' Directory
     QMNL/TH- 10807. May 1988. Available" from tfie Remedial Action  Program
     Center at Oak Ridge National Laboratory*

25.  Office of Management and Budget, Letter from Wendy Gram (CMS) to
     Lee Thomas (EPA), August 12, 1986.
26.  Office of Science _ and Technology Policy,  Oftjg. Comments.' on gA Guidelines
     for Carcinogenic Risk Assessment, July 15, Y986.     ' "            "

27.  National Science Foundation,  Letter from David T* Kingabury (NSF) to
     James Kamihachi (EPA), July 25,  1986.

28,  Science Advisory Board, Sibcamittee on Risk Assessment,  Letter to
     Lee M. Thomas (EPA), July 11, 1988,

29,  G, Schwarz and F. 0. Hoffinan. 1980. "Inprecisions of Dose Predictions  for
     Badionuclides Released to the Envirortoent:   An Application of a
     Monte Carlo Simulation Technicque, fegiroiC' jnteri^iqnal , 14:  289-297.
30*  Douglas B» Chambers, Presentation ' to Sources ' ^d ^ Transport 'Subcociaittee
     of the Badiatioti Advisory' Ooami'ttee"»'' Icicnee Advisory Board", ""
     U.S« Environiaental Protection Agency, Jtily 13,  1988.

31.  S, K, Beal, et al. » Cqroari_8gn_ of^Ai&I^_-El>A Predict iona of Ground-
     Level Airborne JMionuctrdecehtra^^
     Manuscript dated October 1987.

32.  National Council on Radiation Protection and Measuraients (NCEP) .
     1 984.  Radigl^lcal Assessment;  Predicting die' Transport,
     Moaccumiiation,, and^ Uptake by Man of MdiotBicIia^" Ifelias_ed_. tp__the.
     'ESwironiient. ...... tJORF Report. N6V 76'.              '

33.  National Council on Eadiation Protection and Measurements (NCRP),
     1986.  .Scr^iiittTedttiiqugs for Detepnininj^ Compliance with
     Er^rcrapgital Staidards , Releas^ of RadianucTi^es to the Atmosphere,
     NCRP CkUBentary kfo.Tt                      '                '

-------
                                                      APPENDIX  A
                                                                       >
                UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                              WASH1XGTON, D.C 20460          "

                                 MAY 2 3 1988
                                              ;  i
SUBJECT;  Radiation  Risk  Assessment Methodology  \
                                        1 *  i y ^""""^  " ti
FROM;     Richatd J, Guimond, Director  L. »J-*'* .Jv-'
          Office of  Radiation Programs  (ANR-458), "~

TO:       Donald Barnes,  Director
          Science Advisory Board (A-101)


     At our April 12,"  1988, meeting on radionuclide HESHAPS,
we agreed to supply  past  background documents used to support
NESHAPS rulemaking.  Attached .for transmission to the Radiation
Advisory Committee of  the science Advisory Board (SAB) ace copies
of the background information documents produced in support of
the various Clean Air  Act radionuciide rulemafcings.  A copy of
the latest document  describing our risk assessment methodology,
to be used in support  of  a low-level radioactive waste management
standard, is also attached,

     The risk assessment  methodology that will be usei to
develop new background information documents will be virtually
identical to that used in the past with respect to source,
dispersion, and pathway modeling.  However, we propose to incor-
porate a dose-risk factor range of 120 to 1200 fatal cancers per
million person-rem to  account for the uncertainty in that factor.
The central estimate of risk for whole-body, low-let radiation to
the general public will be determined by using a risk factor of
400 fatal cancers per  million person-rem,  corresponding to the
linear, relative-risk  model in BEIE III*  The whole-body risk
will be allocated among the Various target organs,  consistent
with an organ specific relative risk model for all cancers other
than leukemia and bone cancer,

     Also, we propose  to  base the radon risk estimates on the
preferred model contained in BSIR IV.  We will send you another
memorandum which expands  on our proposed treatment of radon and
requests your comments.
            .  *
     other modifications  to the methodology will compute the
effective dose equivalent, as defined by XCRP, and the radon
equilibrium ratio as a function of distance from a radon source.

-------
     Of  lesser importance, we propose to make adjustments in our
thyroid  risk estimates in  light of current information as
summarized in NCR? Report NO, 80,  First, the estimate of 20%
mortality for radiation-induced thyroid cancer would be changed
to 10%.  The 20% figure relates to mortality from _all thyroid
cancers; however, there is ample evidence that the types of
thyroid  cancer induced by  ionizinf radiation have a mortality of
only about 101.  Second, 1-131 would be considered to be one-
third as effective as x-rays for induction of thyroid cancer,
rather than one-tenth, as assumed previously.  The data regarding
this question are incomplete and somewhat conflicting—one animal
study has shown 1-131 to- be considerably more effective than
previously thought *

     11  is extremely important that we obtain your review of our
current  risk assessment methods and our proposed changes to these
methods by August 1, 1988,  This date is made necessary by our
plans to finish the recalculation of risk assessments by early
September in, order to have decision documents ready for Agency
and Administrator reviews this fall.  We will make every attempt
to incorporate youc comments as we proceed*  lowever, our
schedule is inflexible due to a court-mandated proposal date of
February 2S, 1389.  If we receive your comments after Aufust 1,
198S, we may not be able to utilize them in performing the risk
assessment which will support*'the development of the proposed rule
although it may be possible to take note of your comments in the .
preamble to the proposed rule and consider them for the final rule,
which has a court-mandated promulgation date of August 31, 193f.

     If  the Radiation Advisory Committee has any questions about the
attached material or our approach to risk assessment, please let me
know*

5 Attachments

cc:  Gordon Burley (ANK-458J
     j. William Gunter (ANR-4SQ)
     TerrenceA, McLaughlin (ANfi-460)

-------
            Uriied S;SKS
            Environmental Protection
            Agtncv
Office of
Ridiaiion Programs
Washington. D-C. 20460
CA .
October
  0. "-;--v^2"
&EPA
            Background information
            Document For Final  Rules
            Volume I

-------
                                 COHfENtS
Figures
Tables

1.   INTRODUCTION
                                                                    Page

                                                                    viii


                                                                    1-1

               History of Standards Development                     1-1
               Purpose of the Final Background Information Document 1-2
               Scope of  the Final Background Information Document
               E?A's Computer Codes
2.   CURRENT REGULATORY PROGRAMS AMD STRATEGIES

          2. 1  Introduction
          2.2  The International Commission on Radiological Pro-
                 tection and the National Council on Radiation
                 Protection and Measurement*
          2.3  Federal Guidance
          2.4  The Environmental Protection Agency
          2.5  Nuclear Regulatory Commission
                    2.5.1  Fuel Cycle Licenses
                    2.5.2  Byproduct Material licenses
          2.6  Department of Energy
          2.7  Other Federal Agencies
                    2.7.1  Bureau of Radiological Health
                    2.7.2  Mine Safety and Health Administration
                    2.7.3  Occupational Safety and Health
                             Administration
                           Department of Transportation
                     2.7.4
           2.. 8   State Agencies
3.   HAZABD

           3.
           3.
           3.
           3,
           3.
             I  Bv-iteBcJ^hae fcdiation
             2  "Evidence that Radiation I* Mtrragenie
             3  Evidence that Radiation Is Teratogenic
             4  Uncertainties
             5  Summary of  Evidence That Radiation is a Carcinogen,
                 Mutagen,  and Teratogen  «^,
4.   EMISSION  OF RADIOHUCtlDSS IHTO HE All

           4.1   Introduction
           4.2   Sources of Radionuclide Releases into the Air
 1-2
 1-4

 2-1

'2-1
2-2
2-8
2-10
2-12
2-12
2-13
2-13
2-14
2-14
2-14

2-14
2-14
2-15

3-1

3-1
3-6 '
3-9
3-10

3-11

4-1

4-1
4-2
                                    iii

-------
                          CONTENTS (continued)

                                                                    Page

                    4.2.1  Department of Energy (DOE)  Facilities    4-2
                    4,2.2  Nuclear Regulatory Comnissiou (KRC)
                             Licensed Facilities and non-DOE
                             Federal Facilities                     4-13
                    4.2.3  Coal-Fired Utility and Industrial
                             Boilers                                4-19
                    4.2.4  Underground Uranium Mines                4-19
                    4.2.5  Phosphate lock Processing and Wet-
                            Process Fertilizer Plants                4-20
                    4.2.6  Elemental Phosphorus Plants              4-22
                    4.2.7  Mineral Extraction Industry Facilities   4-23
                    4.2.8  Uranium Fuel Cycle Facilities* Uranium
                             Mill failings. High-level Hast*
                             Management                             4-2?
                    4.2.9  Low-Energy Accelerators                  4-31
          4.3  Radionticlide Releases                                4-31
                    4.3.1  Department of Energy Facilities          4-32
                    4.3.2  NIlC-Licensed Facilities and Non-DOE
                             Federal Facilities                     4-32
                    4.3.3  Coal-Fired Utility aad Industrial
                             loilers                                4-35
                    4.3.4  Underground Uranium Mines                4-35
                    4.3.5  Phosphate Rock Processing aad Wet-
                            Process Fertilizer Plants                4-36
                    4,3*6  Elemental Phosphorus Plants              4-36
                    4,3.7  Mineral Extraction Industry              4-37
                    4.3.8  Uranium Fuel Cycle Facilities, Uranium
                             failings, ligh-Level Waste Management  4-38
          4.4  Uncertainties                                        4-41

5.   RETUCTIOM OF DOSE AND RISE                                     5-1

          5.1  Introduction                                         5-1
                    5.1.1  Emission Control technology              5«1
                    5.1.2  Work Practices                           5-1
                    5,1,3  Impact of Existing Regulations on
                             Strategies for Reducing Emissions      5-2
          5*2  Emission  Control Technology                          5-2
                    5.2.1  Scrubbers                                5-3
                    5.2.2  Filters                                  5-4
                    5,2.3  Mechanical Collectors and Electrostatic
                             Precipltators                          5-10
                    5-2.4  Charcoal Adsorbers                       5-10
                    5.2.5  Miscellaneous Emission Control Equipment 5-13
          5.3  Work Practices                                       5-15
          5.4  Summary of Emission Reduction Strategies             5-18
          5.5  Uncertainties in Evaluation of Control  Technology
                  Efficiencies                                        5-18
                                    tv

-------
                          CONTENTS (continued)

                                                                    Fage

6.   MOVEMENT OF MBICMJCLIDES THROUGH tSVXXORKEHXAL PATHWAYS       6-1

          6.1  Introduction                                         6-1
          6.2  Dispersion of Radiotmclides through  the Mr          6-3
                    6.2.1  Introduction                             6-3
                    6.2.2  Air Dispersion Models                    6-6
                    6*2,3  Uncertainties in Atmospheric Dispersion
                             Modeling                               6-8
          6.3  Deposition on Atmospheric Badionuclides               6-9
                    6.3.1  Introduction                             6-9
                    6,3.2  Dry Deposition Model                     6-9
                    6.3.3  Vet Deposition Model                     6-9
                    6.3,4  Soil Concentration Model                 6-10
                    6.3.5  Uncertainties                            6-11
          6.4  Transport through the Food dials                     6-11
                    6.4.1  Introduction                             6-11
                    6.4.2  Concentration in Vegetation               6-12
                    6.4.3  Concentration in Meat and Milk           6-13
                    6,4.4  Summary •                                 6-14
          6.5  Calculating the Environmental Concentration of
                 Radlonuclides:  The AIRDOS-EPA Code                 6-14
                    6.5.1  Introduction                             6-14
                    6.5.2  AIlDOS-EtA   .                            6-15

7.   RADIATION DQSIM1TII                                  .          7-1

          7*1  Introduction                                         7-1
          7.2  Definitions                                          7-1
                    7.2.1  Activity                                 7-1
                    7.2.2  Exposure and Dose                        7-2
                    7,2.3  'External and Internal Exposures          7-2
                    7.2.4  Dose Equivalent                          7-3
          7.3  Dosimetry Models                                     7-3
                    7.3.1  Internal Doses                           7-3
                    7*3.2  External Doses                           7-8
                    7.3*3  Effects of Decay froduct*                 7-8
                    7.3.4  Dose Kate Estimates                      7-9
          7.4  EPA Dos* Calculation                                 7-9
                    7,4.1  Dose Rates                               7-9
                    7,4.2  Exposure and Usage                       7-10
          7.5  Uncertainty Analysis                                 7-11
                    7,5.1  Dose Uncertainty Resulting from Indi-
                             vidual Variation                       7-12
                    7,5.2  Dose Uncertainty Resulting from Age      7-14
                    7.5.3  Dose Uncertainty Caused by Measurement
                             Errors                                 7-23
          7.6  Distribution of Doses in the General Population      7-23
          7.7  Suwsary                                             7-29

-------
                          CONTENTS (continued)

                                                                    Page

8.   ESTIMATING THE RISK OF HEALTH EFFECTS RESULTING FROM RABIO-
       NUC1IDE AIR EMISSIONS                                        8-1

          8.1  Introduction                                         8-1
          8,2  Cancer ftiak Estimates for Low-lET Radiation          8-3
                    8.2.1  Assumptions Heeded to Make ii.sk
                             Estimates                              8-4
                    8.2,2  Dose Response Functions                  8-4
                    8.2.3  The Possible Effects of Dose Rate on
                             Radiocarcinogenesis                    8*6
                    8.2.4  Risk Projection Models                   8-7
                    8.2.5  Effect of Various Assumptions on the
                             Numerical Risk Estimates               8-9
                    8,2,6  Comparison of Cancer Risk Estimates for
                             Lo*r-LET Radiation           .           8-10
                    8,2.7  EPA Assumptions about Cancer Risks
                             Resulting from Low-LET Radiations      8-12
                    8.2.8  Methodology for Assessing th* Risk of
                             Radiogenic Cancer                      8-13
                    8,2,9  Organ Risk*                              8-14
                    8,2,10 Methodology for Calculating the Pro-
                             portion of Mortality Resulting
                             frea Leukemia                          8-19
                    8.2,11 Canter Risks Hue to Age-Dependent Doses  8-20
          8.3  Fatal Cancer Risk Resulting from ligh-LET Radiations 8-21
                    8,3.1  Quality Factors for Alpha Fartlcles      8-22
                    8.3.2  Dose Response Function                   8-22
                    8.3.3  Assumptions Made fey EPA for Evaluating
                             the Dose from Alpha Particle Emitters  8-23
          8,4  Estimating the Risk Resulting from Lifetime
                  Population ExposttfWfe-fcQB R«4on-222 Progeny        8-25
                -  •  t»frfl - €h*r*c*uit lag Exposures to the General
                   3r '       Population vis-a-vis^Undergro^ftid
                             HiMrs                   ~             8-26
                    8.4.2  The 1FA Modal                            8-28
                    8.4,3  Comparison of iisfc Estimates             8-29
          8,5  Uncertainties in Risk Estimates for Radiogenic
                  Cancer                          *                 8-33
                    8.5.1  Uncertainty in dfe Dose Response Models
                             Resulting from Bias in the A-bomb
                             Dositjetty                              8-33
                    8.5.2  Sampling fariation                       8-3?
                    8,5.3  Uncertainties Arising  from Model
                             Selection                              8-3?
                    8.5.4  Suamary                                  8-39
          8.6  Other Radiation-Induced Health Effects               8-41
                    8,6.1  Types of Genetic Harm  and Duration  of
                             Expression                              8-41
                                     vi

-------
                          CONTENTS (continued)
                    8.6.2  Estimates of Genetic Harm Resulting from
                             Low-LEf Radiations                     8-44
                    8.1.3  Estimates of Genetic Harm for ligh-LET
                             Radiations                             8-50
                    8.6*4  Uncertainty in Estimates of lUdiogenetic
                             Ham                                   8-50
                    8.6.5  The EPA Genetic Risk Estimate            8-54
                    8.6.6  Teratogenic Effects                      8-56
                    8.6.7  Nocstochastic Effects                    8-63
          8.7  Radiation Risk - A Perspective                       8-63

9.   SUMMARY Of DOSE AND RISK ESTIMATES                             9-1

          9,1  Introduction                                         9-1
          9,2  Doses and Risks for Specific Facilities       ,   "•   9-2
          9.3  Overall Uncertainties                                f-2
    .   .             9*3.1  Emission and Pathway Uncertainties       9-2
                    9.3.2  Cose Uncertainties                       9-5
    .-,.  : -• •      9.3.3 .llsk teetrttintles    ,  . ,r         • .     9-5
                    f,3.4  Overall Uncertainty                      9-5

Addendum A - Computer Codes Used by EPA to Assess Doses from . .
             Radiation Exposure                                     A-l

Addendum I - Mechanics of the Life Table Implementation of the
         , ', c Risk Estimates      •           .•      • •         " i ,   B-l
                                    vli

-------
Hadon-j22 ffgom PQE Facilities

o    There ate about 5 facilities with radon-222 emissions
     due to uranium ore residues remaining from the former
     Manhattan Engineering District Sites,

o    Major sites are the Fernald cite and the Monticello
     tailings pile.

0    All sites will be assessed using site specific data.

o    Source term data will be somewhat uncertain due to
     fugitive emissions; control technology and cost data
     good,

o    This is new work.

Coal-fired Boilers

o    There are about 1200 utility boilers, and tens of
     thousands industrial boilers,

o    Boilers will be characterized and grouped and model
     boilers developed.

o    Number of people at risk will be uncertain;
     considerable exposure overlap is expected due to the
     large number of boilers.

o    Latest QAQPS data on risk and emissions to be used.

0 r a n i urn Fuel
o    There are approximalty 100 major nuclear power
     stations that require approximately 30 to 40 support
     facilities o£ various kinds.

o    Previous analysis in 1975 is obsolete.

o    sites will be assessed based on models*

o    Uncertainty is moderate due to model approach,

o    This is new work.

-------
High-level Radioactive Haste
o    HO facilities are in existence.
o    Previous EPA work under Atomic Energy Act authority
     judged sufficient.
o    This category is given low level of effort.
Phosphogypsum files
o    All sites (80) will be assessed using site specific
     data.
o    Data will be good; uncertainty moderate largely due to
     uncertainty in emissions data.
o    This is all nev work.

-------
For Each Source Category

T the extent possible, we will provide for each category
the following information:

* Individual fatal cancer risks based on site specific
  meteorology, demographics, and
- The number of people at risk of fatal cancer by
  range of risk and incidence.

- Feasibility and cost of controls and resulting risk
  reduction.

- Health effects in addition to fatal cancer,  to
  the extent known.

- uncertainties.

-------
                                          Agr««»tnt
 Environmental fiottction  Ajtncy
 Office of ladlation Frogt**«
 401 M St., SV
 ViBhlntton ' DC  20660
                          U.S. D*part«*nt  of  £o«r§y
                          Oak Xtdf* Optt*tlon» Office
                          P.O. Box £  Oak  tidf*. TN  37831
                                         .
 AtictiBcnt of li*k  to  Population of the U.S. from ladoa £vi«ti0ac froa
               Stack*.
 Christopher Welion  (ANR-460)
 Office of Radiation  Progt»»*
 Environ»«ot*l Prottcc ionAjency
 Wiihinston, DC  20460  FTS 47S-I640
                           Frank 1. 0*Donne1
                           Oak tidg* National Laboratory,  (450QS,
                           MS-102. P.O.  Sox I
                           Oak iidta, TN  37131 FTS 626-2132
IQ ,
                         Jam 30,
                           l.
 S«e  Attached Scop*  of  Work
<9 ttatwiat* A«(M»itt i«v axft Tfwtttvr M
 leeftOftv Arf of  1912&U.S. C. AftfJiK a
                                       sendtd. Clean Air  Act 19
                                                     u
                                                     7  Federal..
                                  AMOUNT
                                       Tm|
                                                                  AMI *
-------
   k Assignment 1-41, Cnange 1

   Scope of Work Amendment

     Section V. Scope o£ Work, under item 1 add the following:

     The contractor shall prepare a detailed examination of cost
     effectiveness for all control options for phosphogypsum
     stacks.  This shall include cost estimates using actual
     data from representative stacks in the industry.
     The contractor shall also prepare a-R*gulitosy Impact
     Analysis (fctA) in support qf the culemaking activities for
     phosphogypsum stacks.  The^'llA shall include the industrial
     profile prepared under WAl-41.  In addition, the contractor
     shall prepare an economic inpact analysis and a
     cost-benefit analysis for all control options.
        i
     The -RIA shall also include summary discussions of emission
     .levels, risk levels, feasible control options, and an
     examination of the possible misuse of phosphogypsum and the
     benefits of preventing this misuse.  The" RIA shall also
     include a Regulatory Flexibility Analysis.  ThrJtIA shall
     be of adequate scope and depth of analysis to support a
     major rulemaking.

     Section V. Scope of ttork, under item 4 add the following:

     the contractor shall prepare an evaluation of the work
     performed by other program offices within EPA as described
     in WA 1-41.

*  Economic Analysis Report Change

   Sectio'.> VI.  Reports is changed as follows:

   Draft outline for EA chapters:  5 copies,  7 days after W.A.
                                   Assignment Change 1 is issued,

   Schedule foe EA chapters;       same as above.

   Draft EA:  20 copies, October 1, 198S

   Final EA:  20 copies, December 1, 1988

-------
r
Chapter 3.  Phosphogypsum Stacks

Scape of Work

Task 1:  Risk'Estimates
     The Contractor snail prepare a report,  foe use in tne 3:2,
     on the frequency distribution of nsfc levels from
     Pnosphogypsurn scacks in the U.S. for use in the Background
     Information Document (BID).  The Contractor shall use
     £PA-approved assessment models,  such as ISC/LT, in
     consultation witn tne Tas* Manager,  to  compute the
     frequency distribution, as well  as existing risk estimates
     (using AIROOS/DARTA3) whicn are  available from E£R?.

     In preparing the report, the contractor shall address the
     following technical issues;

1.1  Make adjustments foe the variation of radon decay product
     equilibrium fraction as a function of distance.  Current
     estimates assume a constant 70%  equilibrium fraction.  This
     adjustment will lower the% risk to populations within 20
     kilometers of each stack."

1.2  Compute the correct number of people at each risk level for
     Phosphogypsum stacks that are co-located.  This will
     involve the resolution of two problems;  1)  Sumraing the
     radon exposures to individuals from multiple piles,  and 2}
     correctly counting the populations exposed to each pile
     without douole-counting those populations exposed to
     multiple piles.  The contractor  shall examine the problem
     of considering the varying equilibrium  fractions to
     individuals exposed to multiple sources at varying
     distances and determine if a practible  solution exists,
     The contractor shall incorporate this solution in
     consultation with the Task Manager.

1.3  Evaluate the effect of using an elevated height for the
     radon release from 3 model phosphogypsum stack.  Current
     A1RD0S/DARTAB estimates used a ground level release on the
     assumption that this would correctly account for the plume
     downwash caused by the wake effect.

1.4  Make recommendations regarding the calculation of the
     source term for phospnogypsun sticks.  The current
     estimates are based on half the flux for the top layer of
     phosphofypsum, which accounts for the ponded area.
     However, this does not account for the  reduced flux on the
     sides of the stack, which have crusted  over and generally
     have a flux about 20% of the top layer.  Also* the
     Contractor shall examine the effect of  calculating the
     source term based on flux characteristics averaged over the
     lifetime of the stack.

-------
Task 2:  Evaluation  of  Misuse of Phosphogyosum

     Tne Contractor  snail  prepare a report for use in ti-,e BID
     and RIA tftat  examines* -.*.* potential for misuse of raater.al
     in pftosphofypsw^  staco.  Tr.is T=V ae a significant procle
     foe stacks  in Califorr,; $t  wr-.lcr. reporcedr/ nave disappe
     from oeir.c  used for  soil ;or.cxv»oner.  Pnosphogypaur, ears
     also DS misused foe  zryw^Ii procu-ction.  -Tnis report shall
     -oe of sufficent scope and quality to support a Regulatory
     Irapact Anilysis f or" r.ilaaa^ir.j activities involving an
     industry cost of  $100 r.illi-sn,

E s ti mat ed L_e ve.lr o f i: SJt f o,: r.

     Task it  1000 iaooc  -ours
     TasK 2;  1000 laoor  nc-r=

Schedu le_ and  6  Jt s
     Draft Report;   Jaly  .,  ;9S3,  5 copies
     Final Report:   August  1,  19SS, 5 copies

-------

-------
                          SCOPE OF  WORK
            Computer Science Corporation  -  Las Vegas
                    Contract I6S-01-7365-Q39

Title:    DESIGN AND CODING OF AIRDOS-EPA, VERSION 2

Background;

     The Criteria1and Standards Branch,  DIP, needs to establish
     an upgraded version of the codes used for dose and risk
     estimation for standard-setting activities.  These upgraded
     code will incorporate the latest refinements in calculation
     methodologies,, be more flexible and easy to use, and
     generate output more immediately useful to Agency decision
     makers.  In order to make these modifications, the current
     version of mainframe codes AIRDOS and OARTA3 require three
     general modifications:  incorporation of new assessment
     methodologies, enhancement of input and output procedures,
     and generation of presentation-quality graphics.

Description of Work:

A)  'Incorporation of New Assessment Methodologies:

Under the direction of EPA experts, the contractor shall modify
the existing mainframe AIRDQS/DARTAB codes to:

1)   vary equilibrium fraction of radon decay products as a
     function of distance.

2)   Calculate "Effective Dose Equivalent" ac-ording to ICRP26
     and 30 methodology.

3)   Incorporate new dose and risk factors,

4}   Recompile 'selected sections of code for more efficent
     operation.

5}   (tentative*)  calculate national impacts of radon

6)   {tentative*}  Allow for multiple sources,  not co-located.

7)   (tentative*)  Calculate building wake effects.

          * Methodologies for items marked "tentative*1 are
          presently being developed by the Bioeffects Analysis
          Branch (BABJ,  Coding of these items will require that
          a satisfactory methodology be developed by BAB.

-------
B)  Enhancement of Input and Output;

1)   Set up the codes to run in full screen/interactive
     fashion.  The code should prompt the user to input data in
     a straight-forward and logical manner* preferably using a
     menu format similar to that used by the AIRDQS-PC cade.
     The code shall be v'Eiy user-friendly and forgiving of
     errors,  if practicable* the code should run in real time
     and not batch node.  Assessments for individual facilities
     should be easily accessible from a menu or directory, so
     that a modified run can be easily made.

2)   Allow the user to select meterologieal data from a menu.
     Set up a data base of data sets that can be accessed
     easily.  Code an identification in each meterological data
     set that will allow the program to identify the source of
     the data and the proper format.

33   Allow the user to select from existing population grids or
     generate new ones easily.

4}   Refine output for each assessment such that it succinctly
     summarizes inputf file names and important dose and risk
     data.  Be able to print out any and all location tables and
     other output from a menu.

5)   Have the code wake sure that distances -sleeted match the
     population grid, if applicable.  Set up to run population
     and individual assessments at the same -ime (the codes must
     now be ran separately to alter imported food fractions and
     distances,}

6)   Store ihe output from each run in a master data base that
     will a.^iow for recalculation of doses and risks if the
     factors* change, and to do graphical output summaries for
     ALL assessments on demand (described later).

7)   Generate output in three ways:  for each facility that is
     assessed, for all facilities in a source category, and
     across all source categories.  Categories will be further
     broken down into Radon and Non-radon groups.  There are now
     a total of 11 source categories:

     Non-Radon:

          NEC licensees, DOE facilities, High Level Waste
          Facilities, Uranium Fuel Cycle facilities, Elemental
          Phosphorous Plants, Coal Fired Boilers.
     Radon;
          Underground and Open-Pit Uranium Mines, Active Mill
          Tailings, Disposed Mill Tailings, Radon from DOE
          facilities, Phosphogypsum piles.

-------
8)   Include in the output for tach facility the following items:

     a)   Do a synopsis on just a few pages that summarizes
          facility name, user input, date, run number,  and file
          names used and selected output.  The selected output
          should be the five highest organ doses and effective
          dose equivalent for individual (mrem/vr)  and
          collective assessment {PR/YS5, the maximum individuals
          lifetime risk, total fatal canctrs/yr, and a  table
          showing number of people at various risk  levels.  The
          table of people/risk should be modified to include
          total number of deaths/yr due to this risk or higher,
          take the risk level down to 1E-10, and print  risks in
          x.x EXX format.

     b)   Plot isopleths of individual dose on a map of the
          facility and surroundings (Scale to be determined).
          Include population grid information on this plot, so
          that approximate numbers of people at various doses
          can be easily seen.  Put- a legend at the  bottom of the
          graphic showing facility-name, scale, etc.

9)   Include in the output for each source category the
     following items:

     a)   show the number of facilities, total r^mber of people
          at various risk levels, total population  within 80 km
          of all facilities (assuming no overlap),  total fatal
          cancers/yr, total effects/yr, the highest maximum
          individual risk {and identify the facility with the
          highest risk).
               *
10}  include in the output for a summary across all categories
     the following items;

     a)   A summary of risks, showing the categories, number of
          sources in each, highest individual risk, fatal
          cancers/yr, and total population with 30  km.

     b)   The total number of people at various risk levels for
          each category, arranged graphically so that all
          categories can be easily compared in a visually
          appealing manner.  A grouping of 3-D colored
          histograms may be appropriate here.

     c)   A graphical ranking of highest maximum individual risk
          for all of the categories, with EPA-supplied
          uncertainty bars around the risk points.   The total
          number of deaths/yr shall be incorporated in  a
          notation for each category.

-------
Item 10 (continued)


10)  d)   A repeat of the above with  various  dose  standards
          superimposed, to show the categories that  would be
          affected by various dose standards.

     e)   A repeat of the above that  shows  deaths/yr instead of
          naximum individual dose.
     f)   (tentative*) National  impacts  of  radjajo  from fehe
          assorted radon categories.          •-*ps-4"

C)   Graphics iackage:                                ""/' '
                                             , „ .«rtllt,%.l • , j,__. • -Ittaii'.; \
                                             ' *,j]2J*J  * ' ' *I  ^2™. ._

     Add capability to produce presentation-quality  griphics
     su»raarizing the dose and, risk  assessments  for facilities,
     categories and across all categories to .be used by Agency
     decision makers.

-------
  Document 3436G
                  OUTPUT FOR AIRDOS-PLOS/MAINFRAME
  ID No.:
        Date/Time:_
          Run No.:"
••- Facility:_

   Address:.

      city:
 State?
ZIP
  Source Category:_

  COMMENTS:
           Year:
  INDIVIDUAL ASSESSMENT:

  For RADON ONLY .'
  Exposure in
 .pCi/1 at that location
  Lifetime Fatal Cancer Risk

  POPULATION ASSESSMENT;

  For RADON ONLY:

  Exposure in Person-WLM/Yr
  Total Fatal Cancers/Year
Location;  _ SO0 meters North

ForNon-Radon;

Organ Doses in mrem/y
ICRP effective dose equivalent
Lifetime Fatal Cancer Risk
For Non-Radon:

Per son-Rent/Year
Total Fatal Cancers/Year
  FREQUENCY DISTRIBUTION OF RISK- LEVELS:
                          Total in
Risk
interval
le-l to le-2
le-2 to le-3
le-3 to le-4
le-4 to le-5
le-5 to le-6
< le-6 tc le 7
-tfr—7 tti l6JS"8"""
la— A. t~n. lr 9
-3r Q (u Tp»in
Total in
Interval
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
— 5f1f¥
Interval
or Higher
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
1f**X
   Fatal Cancers
   per Year fro»
   this Interval
     xxxx
     xxxx
     xxxx
     xxxx
     xxxx
     xxxx
     xxxx
     xxxx
     xxxx
  FC/Y from
  this Interval
  or Higher
   xxxxx
   xxxxx
   xxxxx
   xxxxx
   xxxxx
   xxxxx
   xxxxx
   xxxxx
   xxxxx

-------
SITE INFORMATION:

MET data from:

  Temperature:
     Rainfall;
Mixing Height:
SOURCE TERM

Nuclicle
                             Page 2
        Pocatello, Idaho  1965-1969
        mm:         HDR:	CODE:
        19  *C
        11   cm/yr
        1100 meters
SET:
             Stack No:
Class  AMAD  1      2
  Area/Stack Height;
Area/Stack Diameter:
 Plume Rise (units)j

POPULATION ARRAY:   Latitude:
                           	  Longitude-:	 .	
500
N XXXX
NNW XXXX
NW XXXX
W XXXX
WSW XXXX
SW XXXX
SSW XXXX
S XXXX
4,
DISTANCES
1000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

USED
2000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

«*
3000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

4000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

FOR MAKIMUM INDIVIDUAL
1000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

10000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

20000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

30000
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX
XXXX

ASSESSMENT !
FOOD SUPPLY FRACTIONS:

       POPULATION ASSESSMENT:

       Local  Regional  Imported
Veg.:   xxxx     xxxx      xxxx
Meat:   xxxx     xxxx      xxxx
Milk;   xxxx     xxxx      xxxx
REFERENCE FILE NAMES FOR ASSESSMENT:
                           INDIVIDQAL ASSESSMENT:

                           Local  Regional  Imported
                           xxxx     xxxx      xxxx
                           xxxx     xxxx      xxxx
                           xxxx     xxxx      xxxx
Prepar File: MGUCAAR._CAA88.ELSMPHOS(FMCCONCJ
STAR Array:  MGUCAAR7CA¥IIVSTARL1B(XY28945)
Population:  MGUCAAR.CAA88.POPLIB(POCATELL)
Radrisk File:CBNRAC5.CAA84.RADRISK.V'8'4Q1RBP

-------
                             page 3

               Optional fables (selected by user)

Environmental Transport Variables:

(do as now printed in AIRDOS)

     fraction of radioactivity retained after washing:  .5
     Infestion of Produce
     Ingestion of ...
     m 9 * *
     Buildup time in soil

(consult  with Barry on what variables to include?  will decide
case-by-case!)

Radiqnuclide-specific environmental transport variables;

For each  NUCLIBfi;  ANLAM, Scavenging Coefficent,  Deposition
velocity. Gravitational Settling Velocity»LAMSUR

Heterological Pata:

(Using a  convention of wind PROM the direction and CLOCKWISE
ordering  of directions):  Arithmetic Average Wind  Speeds,  Wind
Rose, Harmonic Average Wind Speeds, Stability Array,  Surface
Roughness length, Height of Wind Measurements (meters), Average
wind Speed

AGRICULTURAL ARRAYS:

as they currently appear in A1RDQS  do NOT put in  water arrays^

CHI/Q tables:

as they appear in AIROOS, but put direction and distances  in
ENGLISH,  not numbers!

CONCENTRATION TABLES;

Wind                         pCi/cu*meters  Dry Dep*   Gnd  Dep.
Toward  Distance   Nuclide    (not cm3M     RATE      RATE
North   5000       Po-210    2342            234        234

(deposition rates in terms of cubic meters, not centimeters)

Input values for Radionucljde-Independent Variables*

(as they now are printed in AIRDQS)

INPUT DATA FOR NUCLIDE XXXXX

as now printed in AIRDOS but DON'T print AIRDOS dose conversion
factors;   include Buildup Factors and parents.

-------
4 '  *
                                 page 4
                      Optional  Tables  (continued)
    DOSS /RISK Conversion factors from SARTfti
    (at present, units are not printed,* can w* put units in?)
    Organ dose weighting factors used
    POSE/RISK Location Tables (offer a logical menu here)
    RN-222 Working i^evel Tables;
                           Equil.
    Direct ion  Distance   Friction  WLH
    North      1000        .36       xxx     xxxxx
    (equilibrium fraction and WLM will have to be computed)

-------
     Frequency  Distribution  of
           Individual Risks
   (total number
   at this risk
   or higher
        otal number
       at thts risk
       or higher
                          10
    Itf4   1Q'3
 Elemental
Phosphorous
   Plants
 Active
  Mill
Tailings

-------
 Proposed
Standards:
  500
mrem/y'
  100
mrern/y

  25
mrern/y
   1
mrem/y



Elem.   Umn.
Phos.    Fuel
       Cycle
                            NRC    001 .    Coil
                                           Fired
                                          Boilers
                                        10
                                                      rt
                                        10
                                                    10
                                                    10

-------
•i * t
      pCi/1
  — M«ri:;;;:;-;»i:i

pCi/1
        1
      pCi/1

                             Tsiall
                                                          ::•:
                                                         .-:;i*;*;»:»
                                                         •:-:«:-*'!-:
                                                                          10
                                                                            -2
            Activt    wnim ~    Open
              Hill     9nwnd     pn
           Tailings  Ur*nfllm   Mines
                                              Disposed  Phospho-    DOE
                                                Mill     Gypsum   Radon
                                              Tail inns   Stacks

-------