United States
              Environmental Protection
              Office of
              Radiation Programs
              Washington, DC 20460
EPA 520/3-80-009
October 1980
Population Risks
Uranium Ore Bodies

                     EPA REVIEW NOTICE

    The Office of Radiation Programs, U.S. Environmental
Protection Agency, has reviewed this report and approved
it for publication.  Mention of trade names or commercial
products does not constitute an endorsement.

High-level Waste Environmental                        EPA 520/3-80-009
Standards Program
Technical Support Document
      Population  Risks from  Uranium Ore Bodies
                     W. Alexander Williams

                          October 1980
                     Office of Radiation Programs
                 U.S. Environmental Protection Agency
                       Washington, D.C. 20460

    The Office of Radiation Programs, U.S. Environmental Protection
Agency, carries out a national program to evaluate individual and
population exposure to ionizing and nonionizing radiation and to
promote development of controls for the protection of public health
and the environment.

    This report is technical support for EPA's high-level
radioactive waste standards program; it estimates the radiological
releases and impact of unmined uranium ore.

    The Office of Radiation Programs invites readers to report
omissions or errors, submit comments, or request further information.
                                  OFFICE OF RADIATION PROGRAMS
                              U.S. ENVIRONMENTAL PROTECTION AGENCY


    I am Indebted to Dr. Cheng-Yeng Hung who programmed the solution
to the diffusion equation and provided the enrichment factors used
in this work.  I thank the distinguished scientists outside EPA who
kindly reviewed this report prior to publication.  Within EPA
Dr. Abraham Goldin, Dr. James Neiheisel, Mr. Nels Nelson,
Mr. Dan Hendricks, Dr. Tin Mo, and Mr. Jay Sikhanek made many
helpful suggestions, which I gratefully acknowledge.  Mrs. Linda
Martin, Mr. Tom Kelly, and Miss Bertha Clow provided editorial and
organization assistance, for which I thank them.


    This report estimates the minimum radiological releases  and
potential impact of deep-lying uranium ore, so that they may be
compared with projected releases and impacts from radioactive
waste.  Uranium concentrations and groundwater flow rates are used
as input data for three models developed by EPA for analyzing the
impact of high-level radioactive waste.  One set of data is  obtained
from some ore bodies which are being mined by the in situ solution
process.  Another, the minimum impact case, is obtained by using
conservative data on uranium concentrations in uraniferous
groundwaters in conjunction with a model aquifer developed by EPA.

    The estimated releases from actual ore bodies prior to mining
range from .19 to 1.1 curies per year for 226Ra<  por a generic
model ore body with parameters chosen to minimize the impact, a
release of 1.2xlO~4 curies 226Ra per year is estimated.  To take
into account the different size of these ore deposits, the impacts
can be adjusted or normalized to the amount of ore from which a
100,000-metric-ton heavy metal nuclear waste repository would be
derived.  After normalization, the estimated yearly releases range
from 7.4 millicuries for the minimum case to 30 or 300 curies for
the actual ore bodies.  This corresponds to .023 fatal cancers per
year for the minimum case and 100 to 1000 fatal cancers per year for
the actual ore bodies using EPA environmental  pathway and dosimetry

    The approach used in this work can serve as a starting point for
making more detailed assessments using other assumptions.

Preface                                                    iii
Abstract                                                     v
Introduction                                                 1
Method                                                       3
Environmental Impact Analysis—Basic Assumptions             8
Generic Model of Natural Ore Body                           10
Site-Specific Case Studies                                  14
Discussion                                                  19
Conclusion                                                  22
References                                                  24
I.    Uranium, Daughters, and their Retardation  and  Dose
       Conversion Factors                                    5
II.   Data Used for Individual Case Studies                  12
III.  Estimated and Normalized Health Effects  for  Site
      1 Specific and Generic Uranium Ore Bodies              15


    Many authors have compared the health hazard from unmined
uranium ore with that from nuclear waste.  Several (Cohen,  1977;
Hamstra, 1975; and the review by Voss, 1979) believe that the hazard
of high-level radioactive waste after decay of several hundred years
is less than that of unmined uranium ore.  Their comparisons used
hazard indices which do not indicate how these materials behave in
the natural environment.

    This assessment estimates the releases and impact of actual,
natural uranium ore bodies, rather than just a measurement of
potential impact.  For this purpose, I have used an assessment
methodology incorporating models that EPA is using to assess the
impacts of high-level waste repositories.  In addition to examining
a few actual ore bodies, this report presents an approach for
estimating the minimum radiological impact from unmined uranium
ore.  The report deals only with the population risk from unmined
ore bodies and not with the risks and benefits from the nuclear
generation of electricity or from other uses of nuclear energy.

    Most of the uranium ore bodies in the contiguous portion of the
United States are secondary deposits of waterborne uranium in

permeable geologic formations.  The source of the uranium was
igneous, uranium-bearing rocks, which were oxidized or weathered and
later leached by water.  The dissolved uranium moved with the
water.  In some cases the waterborne uranium reached the ocean; in
others the solution infiltrated the ground and either the uranium
was deposited in permeable rock that was strongly reducing or the
solution was mixed with agents  (such as phosphates or vanadates)
forming insoluble uranium salts.  The factors that control the
deposition of uranium are diverse but interrelated:  the
oxidation-reduction potential  (Eh), acidity (pH), presence of
complexing and/or precipitation agents, and the  presence of
adsorbing solids, such  as marine muds,  iron oxyhydroxides, clays,
and organics.  Despite  this  diversity of mechanisms of uranium
deposition,  almost  all  U.S.  uranium reserves currently commercial  at
production costs  not exceeding $50 per  pound are found in coarse
clastic rocks  (DOE,  1979), where uranium was deposited from  ground
water.  Qidwai and  Jensen  (1979) and Deffeyes and MacGregor  (1980)
discuss  the  formation  of  these "roll front" sandstone deposits  in
further  detail.

    After  deposition  of uranium, the daughter products will
accumulate  according  to the  laws of radioactive  decay.   If the  ore
deposit  is  in  a  permeable geological medium,  it  is  possible  for both

the uranium and its daughters to dissolve into water moving through
the deposit (See, for example, Fix, 1955; Phoenix, 1959; Germanov
and Panteleyev, 1968; Kaufmann et al., 1975; Langmuir, 1978;
Capuano, 1978; Lueck, 1978; or Runnel Is et al., 1980).  This
dissolution can also be selective; activity ratios of 234u/238u
in groundwater as high as 12 have been reported (Fleisher and Raabe,

    The remainder of this report presents a methodology for
estimating the radionuclide releases from an ore body, the transport
of the radionuclides, and the resulting effect on people.  The
methodology is applied to a model ore body designed to estimate a
lower bound of its effects and also to actual ore bodies to get a
more realistic, but probably high, effect estimate.


    The method used in this paper consists of applying three
models.  One considers the transport of radionuclides from an ore
body to a stream via groundwater.  Another considers the movement of
radionuclides in the biosphere once released to a stream.  The third
estimates the effect of released radionuclides on people.  Uranium
concentrations and groundwater flow rates can be used as input data
to these three models.

The geosphere transport, biosphere transport, and dosimetric models
(Smith et al., 1980a and Smith et al.,  1980b) were developed
initally to analyze the consequences of high-level nuclear waste
burial.  The biosphere transport and dosimetric models consider
various means of contamination reaching people.  These include
direct exposure, drinking water, eating fish, eating  irrigated food
(crops, beef, milk), and breathing air.  Table I summarizes the
results of  Smith et al.  (.1980b), who estimated the number of fatal
cancers that would result  in  their model population for  each curie
of  various  radionuclides entering  a stream.  They gave 3.1
cancers/curie of 226Ra  and  1.3 cancers/curie of 234U.  in this
analysis  I  consider only the  impact of 226Ra.  Since  226Ra  is
reconcentrated  in  groundwater (as  discussed  later)  its effect will
be  larger than  that of  any other daughter  of 238U.  This simpli-
fication  will not  affect a minimum impact  estimate, since the
contribution  of other  uranium daughters can  only  increase the
radiological  impact.   Radon-222  (the most  significant air pollutant
 in  uranium ore)  has  a  half-life  of 3.8 days.   Radon  in groundwater
would probably  decay  before reaching  the surface.   Since all
 nuclides  except 226Ra  are  ignored, my estimate of  health effect
 will be low.

     One set of data used in this modeling is taken from  several  ore
 bodies being exploited by the in situ (solution mining)  process.

                                Table I
               Uranium, Daughters, and their Retardation
                      and Dose Conversion Factors
                                                   Fatal  Cancers/Ci
4.5 x 109y
2.5 x 105y

7.8 x 104y
1.6 x 103y
3.8 days


   aSource:  Arthur  D. Little  (1979);  originally  from Denham  et  al.
   bSource:  Smith et  aU  (1980a,  1980b).
   cRetardation  factor:
       Time  in years a given radionuclide  takes to travel  10  feet
       Time  in years groundwater takes to  travel  10 feet
 The  amount  of  uranium  leached from  an ore body per year before mining
 began  is  the product of the annual  volume of water moving through  it
-times  the measured  premining concentration of uranium  in the
 groundwater.   Another  set of inputs  is obtained by using lower bound
 data for  uranium concentrations  in  uraniferous groundwaters (Fix,
 1955;  Germanov and  Panteleyev, 1968; Langmuir, 1978; Capuano, 1978)
 in conjunction with EPA's model  aquifer, which is one mile long
 (Smith et al.,  1980a).

    The total annual amount of water moving through an ore body is
given by the velocity of the water times the porosity of the ore body
times the cross-sectional area.  The quantity of uranium leached by
the groundwater will be the product of the amount of water and the
concentration of uranium in the water.  All of this data is routinely
measured for economic deposits.

    The dissolved uranium and daughters will move down the hydrologic
gradient slower than the groundwater because they react chemically
and physically with the rock.  Retardation factors  in a desert soil
are given  in Table  I for 238u and  some of  its daughters.  If one
assumes a  groundwater velocity of  10 meters per year, 230jh win
travel  10  meters  in 50,000 years;  22*>Ra win move IQ meters in 500
years;  and 222Rn, because of  its short half-life, will decay before
it moves 10 meters.  Since it takes a significant fraction of their
half-lives for the  230Th and  the 226Ra to  move this short dis-
tance,  they will  decay  before they are transmitted  long distances  in
the  groundwater.

     Uranium-238  and its daughters  are present  in  all  the  aquifer
since uranium's  long  half-life  permits  it  to move  long distances  from
the  source ore  body.   Radium-226  is found  at higher activity concen-
trations  in the  groundwater  than 238u because  of  differential

adsorptions of elements by the rocks (Burkholder and Cloninger,  1977,
1978; Lester et al., 1975).  Approximating the decay chain by
238y + 234y + 230jh -»• 226Ra, and using the approximation
method outlined by Rogers (1978), the ratio between the 22*>Ra
activity and the initial 238U activity released will be less than
or equal to 39 when the retardation factors are those in Table I, the
travel time of the groundwater is 290 years (the shortest time to
travel one mile of any reported for the specific cases later
examined), and the preferential enrichment ratio of 234u to 238y
is 12 (the largest reported by Fleisher and Raabe (1978)).  The exact
solution for the four-member decay chain shown above is very
difficult.  However, the solution of a three-member chain has been
solved by Lester et al. (1975) and has been applied to this system:

    Uranium-238->-  Thorium-230'•*• Radium-226

    The enrichment factor of 226Ra relative to 238U is highly
dependent on the retardation factors, aquifer velocity,  and travel
distance.  The solution of Lester et al. (ignoring dispersion)  shows
an activity for 226Ra that is 33 times higher than the initially
released 238U activity using the data shown above.  The activity
enrichment factor  will  be used for some cases but may overestimate
the enrichment.  Since this enrichment factor changes when the


retardation factors, distance, or travel time change, an enrichment
factor of 10 will be used for the minimum case, to make certain this
case is conservatively  low.

Environmental Impact Analysis—Basic  Assumptions

    To analyze the  impact, one must know the boundary conditions  at
the ore body and the geochemical characteristics of  the aquifer.   It
is difficult to determine these characteristics precisely.   I  assume
that (1)  solubility considerations  limit the uranium concentration in
groundwater at one  part per  billion,  (2) uranium leaves the  ore body
at a rate of solubility times water flow,  (3)  all of the uranium
released  from the ore  body reaches  the  stream  at the same  rate it
enters groundwater, (4) 226Ra reaches the  stream at  activity levels
either 10 or 33  times  higher than  the 238|j  activity, and (5)  one
curie  of  226Ra released to the  stream causes a commitment  of 3.1
fatal  cancers.

     This  model  assumes that  solubility  controls the  concentration of
 aqueous  uranium  in  the ore  at  one  part  per billion.  The concen-
 tration  of uranium in the groundwater of ore bodies  has  been shown to
 decrease  down  dip,  after water  has moved through  the ore body
 (Germanov and  Panteleyev, 1968  and Phoenix, 1959).   This  argues

strongly that thermodynamics are more important in limiting the
uranium concentration rather than the kinetics discussed by
Grandstaff (1976) at higher concentrations.  For any given daughter
of 238u it is reasonable to assume that after several million
years, the daughter will be in equilibrium with the parent as they
both migrate in an aquifer, except that the daughter might migrate
faster or slower as discussed above.  Groundwater chemistry can
change radically along an aquifer and this can change the amount of
dissolved uranium.  For example, a rise in the solution CO^
pressure from 3x10-4 atmospheres to 10-2 atmospheres increases
the solubility of uraninite (for Eh values above -.05 Volts) by over
1000 times (Langmuir, 1978).  Despite the uncertainty of the data,
there is evidence that the groundwater in some ore bodies is
saturated or even supersaturated with some uranium mineral  species
(Boberg and Runnells, 1971; Lueck, 1978; Runnel Is et al., 1980).

    If one assumes that uranium is dissolved by groundwater at a
known rate per year and travels 1  mile to a surface stream, the
annual release rate to the stream will eventually be the same as the
release rate from the ore body.  The release rate of 238y win
control the release rate of the daughters to the stream, because the
daughters dissolved directly from the ore body do not migrate fast
enough to reach the stream prior to decay.  Although a travel

distance of 1 mile has been used throughout this work, the long
half-life of 238u obviously permits it to be transmitted for much
longer distances in groundwater.

    In order to compare the impact of ore bodies with that of a
high-level waste repository, the two must be put on a comparable
basis.  The amount of fuel used by a 1000 MWe reactor in 1 year is
about 35 metric tons enriched uranium; therefore, a 100,000-metric-
ton heavy metal repository will contain waste equivalent to 2900
reactor years of operation.  Because 1140 metric tons unenriched
U30g  are the  annual requirements for 5.3 1000 MWe reactors (EPA,
1973), the  annual requirement per reactor is 215 MT unenriched
ILO .  Therefore, a 100,000-metric-ton heavy metal repository
inventory was derived from 620,000 metric tons U30g (equal to 215
metric  tons per reactor year times 2900 reactor years).

Generic Model of Natural  Ore Body

     I propose a generic  model  to  estimate the minimum impact  of an
 unmined  uranium ore body.  The  data  used for this case  is  summarized
 in Table  II.

    The model  ore  body  is located  in an  aquifer  that  has  a slower
 velocity  than those of  the  actual  ore  bodies considered later.  The


model aquifer is the upper, or overlying, aquifer  in  the model
high-level waste repository developed by EPA  (Smith et  al.,  1980a  and
Arthur D. Little, Inc., 1979).  The aquifer has the following

    horizontal permeability = 10~4 cm/sec,
    gradient = .01,
    porosity = 15%, and
    thickness of strata = 30 meters

    The cross-sectional width of the model ore body is 3.7
kilometers, perpendicular to the groundwater flow.  This width has
been selected because 1) it falls within the range of actual ore
bodies listed in Table II, and 2) it is the approximate width of a
high-level waste repository (as described by Smith et al., 1980a and
Egan, 1978).  One would expect a roll front deposit to form perpen-
dicular to the flow of groundwater.  The interstitial velocity in
this generic ore body will approximate 2.1 meters/year.  Since the
width of the ore body is 3.7 kilometers, the annual flow of water
through the hypothetical ore body will be 35,000 cubic meters.  If
this entire volume of water leaves the ore body with  a concentration
of one part per billion of uranium, the amount of uranium leached per
year will be 35 grams/year (12 microcuries/year of 238y).  A

                                                  Table II

                                Data  Used for  Individual  Case Studies
Model Case
Nine Mile Lake
Natrona Co., Wyo.
Johnson Co., Wyo.
Highland (Case 1)
Converse Co., Wyo.
Highland (Case 2)
Converse Co., Wyo.
of Ore Strata
Cross Sectional
Width of Ore Body
Uranium Reserves
(metric tons U30g)
5,900 - 6,800
450 - 1,400
  ^Measured values  range from .640 to levels  not  detectable.
  bAssumes all  groundwater flow is to the southeast  and  ignores the effects from dewatering for mining operations in
the area.   This includes ore reserves only from the  solution mining operations.
  cThis includes ore and reserves from the entire Exxon  Highland operation and ore and  reserves from the adjacent
Morton Ranch project and assumes that all water flow is  to the southeast.  This  may not be true because mine dewatering
has changed the site hydrology.  Reserves are as  reported in the environmental reports  of the projects.

uranium concentration of 1 part per billion is lower than that
measured in uranium mines and in uraniferous aquifers  (Phoenix,  1959;
Langmuir, 1978; Germanov and Panteleyev 1968; Kaufmann et al.,  1975;
Runnel Is et al., 1980) and is the lowest value reported for
uraniferous aquifers by Fix (1955).  It compares favorably with  the
thermodynamic values of Capuano (1978), and Langmuir (1978).

    It might be argued that the location of the model ore body  in an
aquifer is inappropriate since the presence of the aquifer would tend
to maximize the impact by providing a ready source of geological
transport.  However, over 9(Tpercent of known U.S. uranium reserves
(at production costs not exceeding $50 per pound) are found in coarse
clastic sediments (DOE, 1979), terrains in which aquifers would be

    Additionally, it is possible that 226Ra can be reconcentrated
into aquifers while it and its parents migrate, as discussed
previously.  To make certain that the model produces a lower bound,
an activity enrichment factor of 10 will be used for this case,
rather than the 33 obtained from the decay and diffusion equations
discussed earlier.  Using this factor, the projected yearly release
of 226Ra to the stream is 120 microcuries, which produces a
commitment of 3.7xlQ-4 fatal  cancers per year (1.2xlO-4Ci  x 3.1

cancers/Ci) (Smith et al., 1980a and Smith et al., 1980b).  If other
radionuclides in the decay chain were considered, the impact would be
larger, but this fact does not affect this minimum estimate.  The
release and impact of the model ore body are summarized  in Table III.

    For the sake of comparison with the actual ore bodies discussed
subsequently, the model ore body is assumed to have reserves of
10,000 metric tons U308.  This corresponds to an  ore body 50 meters
thick, 30  meters high,  and 3.7 kilometers wide, with an  ore grade of
.09%  (assuming  the density of the  host  is 2g/cc).  This  ore grade
compares favorably to the average  of  .07% for ore reserves  (at costs
not exceeding $50 per pound) reported by DOE  (1979).  If the ore
reserves were smaller,  the  impact  per ton of U^Og would  be  larger
after normalization.  This normalization process  puts the effects of
different  size  ore bodies on a comparable basis.  This adjustment is
done  by  dividing  the  effects of  an ore  body by  its reserves and
multiplying the quotient  by 620,000 metric tons,  the amount of
uranium  from which a  100,000-metric-ton heavy metal repository was

Site+Specific Case Studies

     The  Nine Mile Lake, Irigary,  and  Highland sites have been
 selected for site-specific  analysis,  since field  data  are available

                               Table III
       Estimated and Normalized Health Effects for Site Specific
                     and  Generic  Uranium Ore Bodies
Generic Model
Ore Body
(Minimum case)
Nine Mile Lake
Natrona Co., Wyo.
Johnson Co., Wyo.
Highland (Case 1)
Converse Co., Wyo.
Highland (Case 2)
Converse Co., Wyo.
Estimated Fatal
Cancers from
Radium Released
(per year)
  aEstimated yearly fatal health effects for ore from which a 100,000 MTHM
radioactive waste repository was derived.
from them.  The adsorption of the radionuclide by the geological
medium plays an important role in retarding the migration of
radionuclides.  The results of laboratory tests on the retardation
factors indicate that this parameter may vary greatly because of the


chemical characteristics of the solution and the geological and
geochemical characteristics of the geological medium.  The
retardation factors used are shown in Table I.  These are for a
typical western desert soil with low-to-moderate cation exchange
capacity.  This sandy-to-sandy-loam soil contained about one
milligram of free CaC03 per gram (Oenham et al., 1973).

    The site characteristics for the specific cases studied are
summarized  in Table II.  I used premining hydrologic and chemical
data obtained from some ore bodies now being mined by the  in situ
solution method.  These data are probably superior to that from
other  kinds of  uranium mines.  In Table III the cancer commitment  is
summarized  in terms of the ores from which  a  100,000 metric-ton-
heavy  metal repository was derived.  The cancers were obtained by
multiplying the 226Ra releases in curies/year by 3.1 cancers/curie
(Smith et  al.,  1980a, and  Smith et al.,  1980b).  The contribution
from  all  other  isotopes  is relatively  insignificant because of the
large enrichment  of 226Ra  relative to  238y  in the  groundwater.
Since only premining  data  has  been used  in  this work,  it does not
assess the environmental  impact  of current  mining  operations.  These
actual ore bodies are in  arid  regions  where the population density
 is lower than  the population  density  in  the EPA model.


Nine Mile Lake:

    Using premining data from the Rocky Mountain Energy Company's
Nine Mile Lake in situ solution-mining operation, the ore body
released about 190 millicuries of 226Ra per year before mining
began.  This release causes .59 fatal  cancers per year using EPA's
model pathways and populations.  The reserves shown in Table II will
permit operation of the proposed 227 metric tons per year plant for
9 years (Rocky Mountain Energy Co., 1979).

Irigaray Project:

    Using data from Wyoming Mineral Company's Irigaray in situ
solution-mining project, an annual release of 430 millicuries of
226Ra per year occurred in the ore body prior to mining.   An
average of 1.3 fatal cancers per year  might be expected from this
release using EPA's model pathways and populations (Wyoming Minerals
Corp., 1978).

Highland Project:

    The Highland Project of Exxon Minerals is divided into two
cases.  In one of these only the release from the ore bodies
currently being mined by the in situ method is considered (220


millicuries of 226Ra per year); in the other the release of all
the ore bodies in both the Highland and adjacent Morton Ranch
projects are assessed (1100 millicuries of 226Ra per year).  These
releases would cause a yearly commitment of  .68 and 3.4 fatal
cancers.  The site specific chemical  and hydrologic data developed
for the original Highland Mill have been supplemented with more
recent data from the solution-mining  proposal.  Mine dewatering
operations have changed the local hydrology  of the area.  Because of
this  uncertainty, the first available report on hydrology has  been
used, although later references show  very different measurements
(see, for example, Final Environmental Statement for Highlands
Uranium Solution Mining Project,  pages 2-14  and 2-16, Exxon Minerals
Co.,  1978).

    The  stated uranium reserves for the second case  are probably  too
 low because  the  price of  uranium  has  increased substantially  since
 the Highland mill  was  licensed.   Of the ore  bodies  at the Highland
 area, 1  to 3 million  pounds U^Og  are  reported  as  the Highland
 solution-mining  reserves,  32  million  pounds  U30g  are from the
 Highland  surface operations,  and  21 million  pounds  are  attributed to
 the  adjacent Morton Ranch mines.   These reserves  total  about  55
 million  pounds  or 25  thousand metric  tons.



    Measured concentrations in uraniferous areas are generally
higher than the assumed concentration (one part per billion) of
uranium in the model.  This is the same concentration used in some
risk analysis cases for a model high-level radioactive waste
repository (Smith et al., 1980a); increases or decreases in this
solubility-limited concentration would affect uranium in both models
in a parallel manner.  Although the concentration of uranium in a
saturated solution can be lower under some conditions (Langmuir,
1978), field measurements and some thermodynamic studies show this
one-part-per-billion limit to be a reasonable uranium concentration
in the reducing down-dip groundwater (Kaufmann et al., 1975; Fix,
1955; Phoenix, 1959; Germanov and Panteleyev, 1968; Capuano, 1978).
Additionally, factors which might vary,  such as changes in cancers
per rem, retardation factors, or target  populations, would affect
both the repository and ore body impacts in a parallel, although not
necessarily identical, manner.

    The estimated effects from the actual ore bodies considered have
great uncertainty.  This is because (1)  trace concentrations of
uranium are difficult to measure and vary within the ore, (2) small
changes in Eh can produce large changes  in dissolved uranium


concentrations, (3) the groundwater transport and dosimetric models
are, in part, hypothetical (Smith et al.,  1980a) and  (4) the
retardation factors used are from one medium and might not reflect
the range of values that exist  in nature.  The  area of largest
potential uncertainty  is in  interpreting the available data for
uranium solubility, which can vary by several orders  of magnitude.

    The release and impact assessments  for the  actual ore bodies
should be looked  upon  as high estimates for the groundwater pathway
because the uranium concentrations used are higher than one might
expect down dip.   Although the  actual  impact of these ore bodies
could be higher,  due to  limitations of  the transport  and dosimetric
models,  it  is  more likely  that  the  impact  would be  lower, because
uranium  solubility decreases as the groundwater migrates down  dip
into  a more reducing geological environment  (Phoenix, 1959  and
Germanov and Panteleyev,  1968).  Since  the reported  uranium
concentrations are averages, they may  include  samples taken from
both  the oxidizing and reducing side  of the  ore body.  It  is  my
judgment that  the averages reported  for the  actual  ore  bodies are
overestimates  for this reason.   Additionally,  it  is possible  that
drilling in the ore body,  to obtain samples,  introduced  oxygen,
causing  an  increase in the uranium concentration.   Therefore,  these
 estimates  can  be  looked upon only as  a high  impact estimate,  since


the high uranium concentration tends to maximize the estimated
releases and impacts.

    The releases from unmined uranium ore bodies to groundwater are
"chronic" or continuous.  If one postulated or assumed various
accidents (such as drilling into the ore), the impact might be much
larger, if no environmental or institutional controls are assumed.

    The impact estimate for the actual ore bodies is over a thousand
times higher than that from the model ore body.  The actual effects
from deep ore bodies would probably fall between these estimates
because the uranium concentrations reported from the actual ore
bodies are not typical of the down-dip concentrations.  If the
uranium concentration is raised by a factor of 120 (the highest
reported by Fix (1955)) and the reconcentration factor is raised by
a factor of 3.3, the estimated effects from the model ore body fall
near the range found for the actual ore bodies.

    The differences in the impacts can be explained: a) there are
factors which make the effects from the model  ore body smaller:
solubility of uraninite (2 orders of magnitude),  interstitial
aquifer velocity (factor of 2 more), enrichment of radium-226 in
aquifer relative to parents (factor of 3), and cross-sectional width


of the ore body (factor of 2); b) there are factors which tend to
make the effects from the model ore body  larger than that of the
actual ore bodies examined:  estimated uranium reserves  (factor of
3, but possibly more or less); and c) there are factors  which do not
affect the actual and model ore body estimates significantly:
thickness and porosity of ore  strata (each  less than a factor of 2).

    One obvious source of bias is the fact  that the actual  ore
bodies discussed are of the relatively large  size  that make uranium
milling economically feasible. A recent  study by  the Department of
Energy (1979)  indicates that  small ore bodies are  far more  numerous
than  larger  ones.   Because  information concerning  the reserves,
mineralization,  location, etc.,  of ore bodies is  generally
considered proprietary,  it  is  possible that consideration of only
large ore  bodies for this  analysis has  introducted other, unknown
biases.  Despite this,  it  is  likely  that  the  groundwater impact of
the  uranium  ores,  from which  a 100,000-metric-ton heavy  metal waste
repository was  derived,  lies  somewhere  between  the two estimates.


     A simplified hydrologic model  along  with  environmental  transport
 and  dosimetric models  has  been used  to  assess the environmental


impact of several actual uranium ore bodies.  Although  the  impacts
are only approximate, it is likely they represent the upper  impact
of ore bodies through groundwater pathways.  The relatively  large
impact can be explained by the existence of ore bodies  in aquifers.
If ore bodies were not formed in aquifers, the effects  of unmined
ore would be both more difficult to assess and much lower.  Using
this model the estimated fatal cancers vary for three actual ore
bodies from a low of .59 to 3.4 fatal cancers/year; these fatal
cancers result from estimated releases ranging from 190 millicuries
to 1100 millicuries of 226Ra per year.

    A model ore body model  would release 120 microcuries of 22f>Ra
per year.  This would be the expected minimum impact from an ore

    When normalized to the expected health effects of the ores from
which a 100,000 metric-ton-heavy metal repository was derived, the
number of fatal cancers ranges from 2.3x10-2 per year for the
model case to 102 to 103 per year for actual ore bodies.  It
appears likely that the actual impact of individual ore bodies lies
within this range, due to the assumptions made in the analysis.



W. W. Boberg and D. D. Runnel Is (1971) Recconnaissance Study of Uranium
in the South Platte River, Colorado.  Economic Geology, 66, 435-450.

Harry C. Burkholder and Michael 0. Cloninger (1977) The Reconcentration
Phenomenon of Radionuclide Chavin Migration.  Battelle Pacific Northwest
Laboratories Report BNWL-SA-5786, April 1977, 30pp.

Harry C. Burkholder and Michael 0. Cloninger (1978) The Reconcentration
Phenomenon of Radionuclide Migration.  American Institute of Chemical
Engineers Symposium Series, 74, 83-90.

Regine M. Capuano  (1978)  Preliminary Analysis of the Formation of
Uranium Roll Deposits  as  a Result of Reactions between Circulating
Fluids  and an Arkose.  Economic Geology, 73, 308.

Bernard L. Cohen  (1977) The  Disposal  of Radioactive Wastes from Fission
Reactors.  Scientific American, 236,  21-32.  Also  see: High-Level
Radioactive  Waste  from Light Water  Reactors.  Rev. Mod. Physics, 49, 1.

Kenneth S. Deffeyes  and  Ian  D. MacGregor  (1980) World  Uranium
Resources.   Scientific American,  242,  66-76.


D. H. Denham, D. A. Baker, J. K. Soldat, and J. P. Conley (1973)
Radiological Evaluations for Advanced Waste Management Studies.
Battelle Pacific Northwest Laboratories Report BNWL-1764, 42pp.

Daniel J. Egan, Or. (1978) Risk Assessment in Support of Environmental
Standards:  EPA's High-level Radioactive Waste Standards.  Presented at
the 71st Annual Meeting, American Institute of Chemical  Engineers,
Miami, Florida, November 16, 1978.

Department of Energy (DOE) (1979) Statistical Data of the Uranium
Industry.  Grand Junction Office Report GJO-100 (79), January 1, 1979.

Environmental Protection Agency (EPA)  (1973)  Environmental  Analysis  of
the Uranium Fuel Cycle.   Part I - Fuel  Supply,  U.S.  Environmental
Protection Agency Report EPA-520/9-73-003-B.

Exxon Minerals Company (1977) Supplemental  Environmental  Report for
Highland Uranium Solution Mining Project and  U.S.  Nuclear Regulatory
Commission, 1978, Final  Environmental  Statement, for Highland Uranium
Solution Mining Project  (NUREG-0489),  NRC Docket 40-8102.  Also see
United Nuclear Corp.,  Environmental  Report  -  Morton  Ranch Mill, 1976.

Philip F. Fix (1955) Hydrogeochemical  Exploration  for Uranium.   U.S.
Geologica^ Survey Professional  Paper,  300,  667-671.


Robert L. Fleisher and 0. G. Raabe (1978) Recoiling Alpha-emitting
Nuclei:  Mechanisms for Uranium-series Disequilibrium.  Geochim.
Cosmochim. Acta, 42_, 973-978.

A. I. Germanov and V. M. Panteleyev (1968) Behavior of Organic Matter in
Groundwater During Infiltrational Epigenesis.  Internat. Geology Rev.,
JO, 826-832.

D. E. Grandstaff  (1976) A Kinetic Study  of the Dissolution of Uraninite.
Economic Geology, 71_, 1493-1506.

J. Hamstra  (1975) Radiotoxic Hazard Measure for Buried Solid Radioactive
Waste.   Nuclear Safety,  16,  180-189.

Robert F. Kaufmann,  Gregory G.  Eadie,  and Charles R. Russell (1975)
Summary of  Groundwater  Quality  Impacts of Uranium Mining and Milling  in
the  Grants  Mineral  Belt,  New Mexico.   U.S. Environmental Protection
Agency Technical  Note ORP-LV-75-4,  70pp.

 Donald Langmuir (1978)  Uranium  Solution-Mineral Equilibria  at Low
Temperatures with Applications  to Sedimentary Ore Deposits.  Geochim.
Cosmochim.  Acta,  42, 547-569.


D. H. Lester, G. Jansen, and H. C. Burkholder (1975) Migration of
Radionuclide Chains through an Absorbing Medium.  American Institute of
Chemical Engineeering Symposium Series No. 152, Absorption and Ion
Exchange, 7J[, 202.

Arthur D. Little, Inc. (1979) Technical Support of Standards for
High-Level Radioactive Waste Management.  Volume C, Final Report for EPA
Contract No. 68-01-4470, EPA Report EPA 520/4-79-007C.

Steven Lynn Lueck (1978) Computer Modelling of Uranium Species in
Natural Waters.  Thesis presented to University of Colorado (abstract).
Also see Lueck, Runnells, and Markis, Computer Modelling of Uranium
Species in Natural Waters:   Applications to Explanation.  1978 Joint
Annual Meeting of the Geological Society of America (with 10 other earth
sciences professional groups), October 23-26, 1978, Toronoto,  Canada

D. A. Phoenix (1959) Occurrence and Chemical Character of Ground Water
in the Morrison Formation.   U.S. Geological Survey Profesional Paper
320, 55-64.

H. A. Qidwai and M. L. Jensen (1979) Methodology and Exploration for
Sandstone-Type Uranium Deposits.  Mineralium Deposita, 14, 137-152.


Rocky Mountain Energy Company (1979) Nine Mile Lake Project Environ-
mental Report, January 1979, U.S. Nuclear Regulatory Commission Docket

V. C. Rogers  (1978) Migration of Radionuclide Chains in Groundwater
Nuclear Technology, 40, 315-320.

Donald D. Runnel Is, Ralph Lindberg, Steven L. Lueck, and Gergely Markos
(1980) Applications of Computer Modelling to the Genesis, Exploration,
and  In-Situ Mining of Uranium and Vanadium Deposits.  New Mexico Bureau
of Mines Memoir or± the Grants Mineral Belt, in press.

C. Bruce Smith, Daniel J. Egan, W. Alexander Williams, James M. Gruhlke,
Cheng-Yeng Hung,  and Barry  Serini  (1980a) Population Risks from Disposal
of High-level Radioactive Wastes  in Geologic Repositories.  U.S.
Environmental Protection Agency Report 520/3-80-006.

J. Michael  Smith, Ted  W. Fowler,  and  Abraham S. Goldin (1980b)
Environmental Pathway  Models for  Estimating Population Risks from
Disposal  of  High-level  Radioactive Waste  in Geologic Repositories.  U.S.
Environmental Protection Agency Report EPA 520/5-80-002.


J. W. Voss (1979) Safety Indices and their Application to Nuclear Waste
Management Safety Assessment.  Battelle Pacific Northwest Laboratory
Report PNL-2727,  prepared for the U.S.  Department of Energy under
contract EY-76-C-06-1830, 67pp. and three appendices.

Wyoming Minerals  Corporation (1977) Environmental Report for Irigaray
Project and U.S.  Nuclear Regulatory Commission, Final Environmental
Statment, for Irigaray Solution Mining  Project  (NUREG-0481), NRC  Docket

1. Report No.
 EPA 520/3-80-009
3. Recipient's Accession No.
 Title and Subtitle
  Population Risks  from Uranium  Ore Bodies
                                                 5. Report Date
                                                   October 1980
  W. Alexander Williams
                                                 8. Performing Organization Rept.
 Performing Organization Name and Address
  Office  of Radiation Programs (ANR-461)
  U.S. Environmental  Protection Agency
  Washington, D.C.   20460
                                                 10. Project/Task/Work Unit No.
                                                 11. Contract/Grant No.
2. Sponsoring Organization Name and Address
  Office of Radiation Programs (ANR-461)
  U.S.  Environmental Protection Agency
  Washington, D.C.  20460
                                                 13. Type of Report & Period
15. Supplementary Notes
16. Abstracts
   This report  estimates  the minimum  radiological  releases and potential  impact of
   deep-lying uranium ore,  so that  they may be  compared with  projected  releases and
   impacts from radioactive waste.  Uranium concentration and groundwater flow rates
   are used as  input data for three models developed by EPA for analyzing the impact
   °f high-level  radioactive waste.   One set of data is obtained from some ore bodies
   which are being mined  by the in  situ solution process.  Another, the minimum impact
   case, is obtained by using conservative data on uranium concentrations in uranifer-
   ous groundwaters in conjunction  with a model  aquifer developed by EPA.
17. Key Words and Document Analysis.  17a. Descriptors

   unmined  uranium ore
   radiological releases
   uranium  in groundwater
   population effects
17b. Identifiers/Open-Ended Terms
17c. COSATI Field/Group
18. Availability Statement
   Release unlimited
                                      19.. Security Class (This
                                                           20. Security Class (This

           21. No. of Pages

                                                            22. Price

   » V.S. GOVERNMENT miNTDJO OFFICE : 1980- W1-08Z/M'
                                                    THIS FORM MAY BE REPRODUCED
                                                                                 USCOMM-DC B26S-P74