&EPA
Office of Radiation & Indoor Air
Radiation Protection Division (6608J)
Washington, DC 20460
EPA 402-R-05-009
   August 2006
               Uranium  Location
               Database Compilation
          TECHMQLQGCailVI UI.'.Nfl H Hr.HIN.M Pf DCCUHBHG HHaDftCTIVE fcMTEHIH-5
           TENOR

-------
                                 Table of Contents

Abstract	1
Background	1
Introduction	2
Previous EPA Reports	3
Uranium	4
Data Discussion	5
Duplicate Record Identification and Reliability Value Assignment	11
Data Sources	16
References	23


                                        Tables
Table 1.   Mineral Commodities with Uranium Associations	5
Table 2.   Component Databases	6
Table 3.   Number of Locations Quality Assured Against USGS Mapped Mines	9
Table 4.   Data Source Ranking By State	15
Table 5.   Distribution of Reliability Values	15


                                       Figures
Figure 1.   Western Uranium Locations From the EPA Uranium Location Database	20
Figure 2.   Western Uranium Locations From the USGS MRDS Database	21
Figure 3.   Density of Western Uranium Mines Using the MAS/MILS Database Portion
          of the Uranium Location Database	22

-------
       URANIUM LOCATION DATABASE COMPILATION
Abstract

The Environmental Protection Agency's (EPA) Radiation Protection Division works to address
hazards posed by technologically enhanced naturally occurring radioactive materials (TENORM). As
one part of EPA's efforts to characterize risk from TENORM sources and to identify where
TENORM problems may exist, we have been investigating the potential environmental hazards of
wastes from abandoned uranium mines in the western United States. Between the 1940s and 1990s,
thousands of uranium mines operated in the United States, mostly in the western continental U.S.,
leaving a legacy of potential radiological and chemical hazards.

In order to help us identify where potential problems may occur, we have compiled mine location
information from federal, state, and Tribal agency partners to develop a database that can be used
with geographic information system (GIS) software. Most mines producing uranium as a primary
commodity are, or were located, in Colorado, Utah, Wyoming, New Mexico and Arizona, and are
typically on federal and Tribal lands. The current number of locations associated with uranium, as
identified in the EPA database, is around 15,000. Of these uranium locations, over 4,000 are mines
having documented production.

Uranium mines, particularly conventional type operations, have the potential to become health
hazards if they are not appropriately closed.  Three uranium mines presently are on the National
Priorities List (Superfund), while others are in the EPA Comprehensive Environmental Response,
Compensation and Liability Information System (CERCLIS) hazardous waste database. The
database does not reflect the current reclamation status of the approximately 15,000 uranium
locations.
Background

The focus of this Uranium Location Database (ULD) compilation is the western United States.
Because most uranium mining occurred in the western United States, and this Agency effort
coincided with a Colorado Plateau initiative in the Environmental Protection Agency's (EPA's)
Region 8 office in Denver, Colorado, the initial database compilation efforts were focused there.
Working cooperatively with the Bureau of Land Management (BLM), Forest Service (FS), EPA
regional offices, Navajo Nation, and state agency offices, multiple western state databases have been
incorporated into the master database.  The U.S. Geological Survey's (USGS) Minerals Availability
System/Minerals Industry Location System (MAS/MILS) and Mineral Resources Data System
(MRDS) (McFaul  et al. 2000) databases are also included (uranium locations identified in the eastern
U.S. and Alaska are solely based on these two databases).  Even though these national data sets are
presented, it is the different state and more local federal databases that make the data in this effort
more comprehensive than the previously released national data sets.

Efforts were made to eliminate redundant records, and steps were taken to determine the accuracy of
the data and their reliability. A master database is included that represents the result of these efforts.
However, the  individual databases are included separately as well. The database was included as part
of the peer review of two other technical reports by EPA (U.S. EPA 2006a,  2006b) discussed briefly
below.

-------
In this document, we provide some background on uranium mining, list the contents of the master
database and the component databases, present some maps (see figures at the end of the document)
for comparing the extent of the databases, and include a limited discussion on the data.  Separate
documents provide additional information on the database, including the metadata for the database.
Some of this information, e.g., the metadata, is taken from documentation of the compact disc (CD)
and has been edited and provided in this document for ease of use and readability.
Introduction

This technical report was developed as part of a larger effort to examine the potential hazards of
wastes generated during the mining and processing of uranium, and particularly those wastes known
as Technologically Enhanced Naturally Occurring Radioactive Materials (TENORM). Information
collected and analyzed in that effort will be presented in two additional reports. The first of those
volumes, entitled Technologically Enhanced Naturally Occurring Radioactive Materials (TENORM)
from Uranium Mining: Volume I: Mining and Reclamation Background (U.S. EPA 2006a) will
provide background information on the occurrence, mining, and reclamation of uranium mines and
mills. The second volume entitled Technologically Enhanced Naturally Occurring Radioactive
Materials (TENORM) from Uranium Mining: Volume II: Investigation of Potential Health,
Geographic, and Environmental Issues of Abandoned Uranium Mines (2006b), will evaluate, in a
general way, potential radiogenic cancer and environmental risks posed by abandoned uranium
mines.

In these technical reports, Naturally  Occurring Radioactive Materials (NORM) is defined as:
Materials which may contain any of the primordial radionuclides or radioactive elements as
they occur in nature, such as radium, uranium, thorium, potassium,  and their radioactive decay
products, that are undisturbed as a result of human activities. Radiation levels presented by
NORM are generally referred to  as a component of "natural background radiation".

The term Technologically Enhanced Naturally Occurring Radioactive Materials (TENORM)
is defined as:  Naturally occurring radioactive materials that have been concentrated or exposed
to the accessible environment as a result of human activities such as manufacturing, mineral
extraction, or water processing. "Technologically enhanced" means that the radiological, physical,
and chemical properties of the radioactive material have been altered through having been processed
(or beneficiated) or disturbed in a way that increases the potential for human and/or environmental
exposures. The definition of TENORM used in EPA's reports does not include Atomic Energy Act
materials.

Under the Atomic Energy Act, the U.S. Nuclear Regulatory Commission (NRC) regulates operations
which produce and concentrate uranium and thorium. In accordance with terminology of the Act, the
NRC has defined in 10 CFR 40.4 "source materials" as 1) uranium or thorium, or any
combination thereof, in any physical or chemical form or (2) ores which contain by weight one-
twentieth of one percent (0.05%) or more of: (i) uranium, (ii) thorium or (iii) any combination
thereof. Source material does not  include special nuclear material. It also defines the "by-
product materials" (wastes) of those operations as tailings or wastes produced by the extraction or
concentration of uranium or thorium from any ore processed primarily for its source material
content, including discrete surface wastes resulting from uranium solution extraction processes.
Underground ore bodies depleted by such solution extraction operations do not constitute
"byproduct material" within this definition. Byproduct materials are also regulated by the NRC.

-------
However, certain types of waste from conventional mining of uranium (surface and underground
mining) are not subject to NRC regulation, and are considered to be TENORM. Thus, while these
reports include information about uranium extraction, processing methods and wastes, only the
wastes from conventional mining are considered to be TENORM, and subject to EPA and state
agency oversight.
Previous EPA Reports

The U.S. Environmental Protection Agency has previously issued reports on the uranium mining
industry in response to congressional mandates and programmatic needs. In 1983, EPA published its
Report to Congress on the Potential Health and Environmental Hazards of Uranium Mine Wastes
(U.S. EPA 1983 a, b, c), as required by the Uranium Mill Tailings Radiation Control Act of 1978.
This study provided an important overview of the characteristics and generation of uranium mining
TENORM wastes during a period when the uranium mining industry was still near its production
peak. A subsequent 1985 Report to Congress on Wastes from the Extraction and Beneficiation of
Metallic Ores, Phosphate Rock, Asbestos, Overburden from  Uranium Mining, and Oil Shale (U.S.
EPA 1985), carried out pursuant to requirements of the Resource Conservation and Recovery Act of
1976 (RCRA), as amended, provided additional risk information and characterization of uranium
mining waste. In 1995, EPA issued the Technical Resource Document: Extraction and Beneficiation
of Ores and Minerals:  Uranium as a technical update to provide a means of evaluating wastes that
were exempt from or subject to regulation under RCRA (U.S. EPA 1995).

During the period 1989 to 1993, EPA worked on a draft scoping report (SC&A 1993), which
compiled information on TENORM in several industries, including uranium mining. A preliminary
risk assessment was also developed for certain public and occupational exposure scenarios to the
known radiation levels in those industries. Comments received on the draft from industry, as well as
EPA's Science Advisory Board (U.S. EPA 1994), resulted in further revisions of the scoping draft,
though it was ultimately decided that a final report would not be issued.

Following a review of EPA's guidance for TENORM by the National Academy of Sciences (NAS
1999a), EPA's response to the NAS study (U.S. EPA 2000), and discussions with EPA's Science
Advisory Board (SAB), EPA's Radiation Protection Division decided that a further review of the
current hazards associated with uranium mining TENORM was warranted. The SAB (U.S. EPA
2001) agreed  with EPA's intent to make TENORM documents useful to a broad audience, but also
recommended that the  whole life cycle of a TENORM source, in this case uranium extraction, be
considered beyond regulatory or inter-agency considerations, and that the impacts of non-radiological
contaminants  also be examined in the Agency's technical reports. In addition to most sources of
TENORM, EPA has authorities for environmental standard setting under the Uranium Mill Tailings
Radiation Control Act, cleanup of hazardous waste sites which currently include some former
uranium mines, and assistance to Native Americans that has included assistance in environmental
reviews of proposed in situ leach (ISL) facilities.

ISL operations, as well as uranium mills, and mill tailings impoundments are regulated by the NRC or
its Agreement States. Many of the physical and chemical processes used at uranium mills are the
same as those which extract uranium at ISL operations. While wastes from the ISL operations and
mill tailings are not legally considered TENORM in the United States, this phase of the uranium fuel
cycle is described in the reports, and their locations included in the database in part because radiation
protection standards for the tailings impoundments may have applicability to waste disposal for
uranium mine TENORM wastes. Additionally, the NRC has decided to  allow mill operators to

-------
dispose of wastes other than tailings in the impoundments, which is a possible disposal route for some
currently unreclaimed conventional uranium mine TENORM.
Uranium

Uranium is a common element in nature that has for centuries been used as a coloring agent in
decorative glass and ceramics. Uranium and its radioactive decay products are ubiquitous in nature,
and contribute to natural background radiation found everywhere. In fact, it is important to note that
many of the natural occurrences of uranium present radiation hazards without any disturbance from
miners. By far, the greatest uses of uranium have been defense and electric power generation. The
advent of nuclear weapons and nuclear power in the United States resulted in a full-blown exploration
and mining boom starting immediately after World War II, making uranium the most important
commodity in the mining industry. The uranium production peak spanned from approximately 1948
to the early 1980s (U.S. DOE/EIA 1992). Some uranium mining continues in the United States, and
relatively high-grade resources in other parts of the world are being mined to meet continued demand.
Through the first half of 2005 the industry had generated over 358,000 metric tons (MTs) of uranium
(U3O8) to foster U.S. dominance in nuclear weapons technology, and later to feed the growing
number of commercial power plants that utilized the enormous energy  contained in the uranium
nucleus.1

Another legacy of uranium exploration, mining, and ore processing were many unreclaimed land
workings left behind where the uranium concentration in rock was either found or thought to be
economically recoverable. Thousands of miners and prospectors, as well as large mining companies,
searched the United  States for mineral deposits concentrating the valuable metal, echoing the
California gold rush 100 years earlier. In many instances before the 1970's, they left behind
unreclaimed and exposed wastes elevated in radioactivity from uranium and its radioactive decay
progeny, potentially exposing people and the environment to its hazards.

Most uranium mining in the United States took place in the expansive Colorado Plateau region
straddling the Four Corners where Utah, Colorado, New Mexico, and Arizona meet and in Wyoming.
However, uranium mining occurred in other areas throughout the western U.S., and in some eastern
states as well.
Uranium Associations with Other Metal Mining

Quite typically, beginning in the 1940s, uranium mines would open based on the detection of
radioactivity at the site and identification of uraniferous mineralization. While some deposits were
mined solely for their uranium content, others produced a variety of other minerals, which co-exist
with the uranium minerals (Table 1). In some cases, exploitation of uranium minerals was secondary
to producing another mineral found in greater abundance, commanding a better market price, or less
expensive to produce; nevertheless, their combined economic value contributed to the success of the
mining venture.
1  Data compiled from U.S. DOE/EIA 2003, 2003b, 2005b.

-------
The presence of radioactive minerals was sometimes unexpected, unknown, or ignored in producing
one or more minerals at a mine. Many mine sites operated prior to the 1940s, and even after, have not
been recognized for the inherent hazards potentially posed by radioactivity in the discarded waste
rock or subeconomic ore piles. The geological emplacement or geothermal phenomena that formed
other valuable minerals may have concentrated radioactive minerals as well, or the process of mining,
beneficiation, and milling may have resulted in a concentration of the radioactive minerals in the
waste. In some instances, the mineral(s) being mined may have radioactive elements included in their
molecular structure that impart radioactivity to the ore or even the finished product.
                  Table 1. Mineral Commodities with Uranium Associations
                   Several mineral ores often have TENORM-associated wastes
                     resulting from the co-occurrence of uranium and radium.
                  Aluminum (bauxite)
                  Coal (and coal ash)
                  Copper
                  Fluorospar (fluorite)
                  Gypsum
                  Molybdenum
                  Niobium
                  Phosphate (phosphorus)
                  Potassium (potash)
                  Precious metals (gold, silver)
                  Rare earths: yttrium, lanthanum, monazite, bastanite, etc.
                  Tin
                  Titanium (leucoxene, ilmenite, rutile)
                  Tungsten
                  Vanadium
                  Zircon
                                      Source: U.S. EPA 2003.
Data Discussion
A number of organizations allowed EPA access to their databases and the original information is
provided as part of the publication. The component databases used to compile the uranium location
master database are listed in Table 2, with notes regarding the verification and accuracy of the data.
A complete listing of all data sources evaluated, including details regarding documentation and
processing, is included at the end of this report. Because there were numerous sources of data used in
this compilation, however, efforts were made to reduce mine duplication and compare with existing
data sources for accuracy.  This was done, in part, by comparisons with U.S. Geological Survey
(USGS) topographic maps at different scales.  We also attempted to develop some indication of the
reliability of the data, such as the availability of documentation. The result of this effort produced the
master database and composite shapefile that can be used as a layer with geographic information
systems (GIS).

-------
                                   Table 2.  Component Databases
 Database
                                  Source
 Number of
  Records
DB2
Colorado Bureau of Land Management Abandoned Mine Land Inventory. Point
locations matched quadrangles very well. A high percentage of mines from the EPA
data set matched mine names from the quadrangle maps. Point locations generally
matched within 100 to 200 meters.
     535
DB4
U.S. Geological Survey Mineral Resource Data System (MRDS) from U.S.
Geological Survey Mineral Databases-MRDS and MAS/MILS by McFaul et al,
U.S. Geological Survey Digital Data Series DDS-52 Two Discs, 2000.  Location
points rarely coincided with USGS mine positions although where they did, points
were within 100 - 200 meters.
    1729
DBS
U.S. Geological Survey MAS/MILS. U.S. Geological Survey Mineral Databases-
MRDS and MAS/MILS by McFaul et al, U.S. Geological Survey Digital Data Series
DDS—52 Two Discs, 2000.  The point locations showed good positional accuracy
when compared to the USGS mines.  They were generally within 100 - 200 meters of
mine locations on the USGS quadrangles.  The mine names did match when they
were ascertainable. In one portion of the study area reviewed, there was a noticeable
trend whereby MAS/MILS mine locations were directly on top of section numbers
placed within PLSS sections. This might be an indication that the mine was located
to the nearest section centroid.
8478 (4078 of
  which are
known mines)
DB6
Utah Bureau of Land Management Abandoned/Inactive Mine Land Inventory. The
database lacks mine names. Point locations match many unnamed mine features
found on quadrangle maps.  Many points appear to be within 150 meters or less.
     193
DB7
Utah Abandoned Mine Reclamation (AMR) Database, Utah Department of Natural
Resources.  The database lacks specific mine names.  The points were also in
question due to the references to mine structures e.g. shafts, tunnels, drill holes,
waste dumps, etc. Accuracy was difficult to determine due to the high number of
closely grouped locations. Many points appear to be within 50 meters or less.
     549
DB11
Navajo Lands Project through U.S. EPA Region 9 and the U.S. Army Corps of
Engineers. Location points did not coincide with any quadrangle mine points, but
there was some field verification. This database has been updated since the
information was compiled for this effort and the update was not included in the
composite database.
     887
DB12
State of Arizona. Location points rarely coincided with USGS mine positions.
When there were nearby points they were within 100 - 200 meters of the USGS
designated mines.
     41
DB13
U.S. Forest Service. Location points were generally very accurate, generally within
50 - 200 meters, however 15% of the records had spatial coordinates that placed the
locations outside of intended scope of the data set (Arizona and New Mexico). No
names were assigned to any locational points.
DB14
U.S. Bureau of Land Management. Location points were generally very accurate,
generally within 150 - 200 meters. Names were rarely assigned to locational points,
but were accurate when applicable.
    276
  In addition to these databases, one new mine in Nebraska, the Crow Butte Mine, has been included as a one record
  database separate from the master database.

-------
Database
DB15
DB16
DB18
DB19
DB20
DB21
DB22
DB23
DB24
DB25
Source
South Dakota Abandoned Mine Lands Inventory. No mines on USGS maps to
compare with, and no apparent match with MAS/MILS mines in same location.
However, information provided with this database indicated that seventy of the sites
on U.S. Forest Service property were field verified with a GPS unit.
Califomia-U.S. Forest Service Mines. No matches apparent.
New Mexico Mines Database produced by New Mexico Bureau of Geology and
Mineral Resources. Names match, perfect alignment with USGS mines. Locations
greatly outnumber mines portrayed on USGS map.
Wyoming Abandoned Mine Land. No documentation was received with this data
set. No apparent match with mine names on USGS maps, or with mine names.
Nevada Bureau of Mines and Geology. Where USGS mines available, name
matches were noted. Locations seemed to be off by at least 200 meters.
Texas Mines from Adams & Smith Report. No apparent match for either names or
locations, but data include mine names.
Dakotas Mines from U.S. Atomic Energy Commission Map, 1967. No names on
USGS mines to compare with. Better source for this area than MAS/MILS or
MRDS.
Montana State Library. Some name matches where available. Some locations 0-1
km off from name matches. Locations may have been drawn from MAS/MILS as
their spatial locations are highly correlated.
Inactive Mineral Production Sites - University of Texas. Perfect match with
MAS/MILS; no USGS mines to compare with.
Railroad Commission of Texas Uranium Mines. Local area specialists combined
aerial photography with some field-truthing in order to verify mine locations.
Number of
Records
36
17
1531
119
73
26
181
8
7
101
Verification Using USGS Maps

The basic approach to assess locational accuracy was to compare ULD locations to U.S. Geological
Survey mines (which includes mines of all types), located on maps by means independent of those
used to locate mines from the source databases, on 1:100,000 and 1:24,000 scale maps. In keeping
with the EPA goals for creating quality information, the methodologies, applied here to quality assure
the location data, were a simple effort, in the absence of field verification information, to determine
the relative accuracy inherent in the locational attributes (i.e., latitude and longitude) of each data
source.

Due to the large study area, ICF Consulting, Inc. conducted quality assurance tests of samples of the
total location data set. The sample selection was partially random, however, there was a bias in the
sample, in that the areas chosen for quality assurance were characterized by being areas of higher
location densities than areas not chosen for the sample.  The exact locations within these areas were
unknown prior to being assessed and were therefore random in nature.

The sample set was chosen in this manner in order that the greatest number of locations could be
analyzed. If location selection had been totally random, then the number of locations assessed would
have been lower and possibly less representative in terms of locational accuracy.  Excel spreadsheets
entitled By24K.xls and BylOOK.xls are included in the GIS data directory. These spreadsheets show

-------
the pivot tables that were used to determine which USGS quadrangles held the most locations, and
which set of USGS quadrangles would cover all databases most evenly. These tables were created by
spatially overlaying the locations collected in fiscal year 2002 with the USGS 24K and 100K
quadrangle indices. The results were brought into Excel so that we could develop a pivot table
showing how many locations, from each data source, were located in which quadrangles. It should be
noted that although DB23 and DB24 are not included in these spreadsheets, they were quality assured
since most locations in these data sets were in pre-selected quadrangles. The number of locations
quality assured from each database and the methods and results are detailed below.

A point coverage of USGS mines appearing on USGS 1:100,000 scale 30 by 60 minute series was
created by digitizing on a CalComp digitizing tablet. Of the 126 maps that cover the States of New
Mexico and Arizona, 62 maps were digitized. Sixty-two additional 1:100,000 scale maps were
digitized for the remaining part of the study area. This is in addition to the 18 quadrangles that were
digitized by URS-Techlaw to QA a subset of the study area in Colorado and Utah under an earlier
part of this effort.

By taking a total of the locations in each quadrangle then selecting the quadrangles that contained a
high density of mines ICF derived a new group of mines for comparisons. This new group
represented roughly half of all the ULD mine records. A point was generated for each mine
symbolized on the maps in an Arclnfo coverage and mine names were entered where available. This
resulted in a geographic coverage with 8,343 USGS mine locations. Of these locations 5,280
locations had names. These mine locations from the USGS maps were buffered to an eighth, quarter,
and half of a mile in accordance with the QA/QC methodology established by URS-Techlaw. This
QA method proved less useful in areas of few and scattered uranium locations, as USGS maps do not
show all mine locations, and only report names for less than one-quarter of the locations that they do
show.

The locations shape file was converted to a coverage and intersected with the 1/8, 1/4, and 1/2 mile
buffers. These were inspected, initially manually, and later by using a simple automation tool
developed in MapBasic precisely for this task, for name matches with the USGS mines. For New
Mexico and Arizona, this test revealed that 97 of 501 named locations examined were accurate within
an eighth of a mile, 131 were accurate within a quarter mile and 143 were accurate within a half mile.
For Utah and Colorado, this testing revealed that 50 locations were accurate within an eighth of a
mile, 56 were accurate within a quarter mile, and 64 were accurate within a half mile. This testing
resulted in only a limited number of points being checked; therefore a second method of QA/QC was
performed. For the remainder of the states, a shape file (converted to a Maplnfo tab file for
processing) was created and intersected with the  same buffers.

When this process was automated by a MapBasic program, name matches were found for 28
locations within an eighth of a mile of a USGS mine, 31 locations within a quarter-mile of a USGS
mine, and  36 within a half-mile of a USGS mine. Mine location matches,  regardless of name matches
were found to number 90 within an eighth of a mile of a USGS mine, 163 within one-quarter of a
mile, and 428 within a half-mile. Of the total 6,603 locations  checked, 5,036 had names.

The second method of QA/QC involved a visual proximity test between the final locations coverage
and USGS 1:24,000 quadrangle maps. Due to the large number of quadrangles involved, a sample of
locations was selected for QA/QC.  The sample was determined by selecting quadrangles where the
majority of all mines were located,  as limited by the number of quadrangles which  could reasonably
be reviewed in the contract period. A third criterion is that locations from each of the data sets is
represented in the QA subset. From the sixty USGS  1:24,000 quadrangles that were included in the

-------
QA/QC process, 3,343 locations were sampled for New Mexico and Arizona. Of the 6,603 locations
in the remainder of the study area, 2,168 locations were analyzed.

The point locations for the mines were drawn on-screen with the Digital Raster Graphic (DRG) 1:
24,000 scale quadrangles displayed in the background. These were compared with the locations of
mining activities on these USGS maps. This allowed for an assessment of positional accuracy
between the two data sets. This process was also partially automated, this time using Arc Macro
Language on UNIX workstation Arclnfo version 7.x.  The automation tool allowed the operator to
quickly view one DRG at a time for all 60 DRGs reviewed, with a thematic mapping of all locations
color-coded by data source. The information in Table 3 is a summary  of the results according to data
source.

         Table 3. Number of Locations Quality Assured Against USGS Mapped Mines
                                  (See text for discussion.)
Database
DB1
DB2
DBS
DB4
DBS
DB6
DB7
DB11
DB12
DB13
DB14
DB15
DB16
DB17
DB18
DB19
DB20
DB21
DB22
DB23
DB24
Locations Reviewed
at 1:100,000 scale
56
459
4
1919
6391
127
903
751
329
2475
1885
36
5
21
1155
120
62
43
145
8
43
Locations Reviewed
at 1:24,000 scale
none
none
none
324
1684
none
none
173
108
1034
1387
28
8
5
507
91
16
19
100
8
19

-------
Completeness

Completeness, in the sense of whether or not all uranium deposits are listed in this composite
database, can never be fully determined.  We have attempted to better understand how complete the
database may be by comparing the deposits in the composite to the authoritative source of producing
mines for Texas, available from the Texas Railroad Commission (RRC) in database 25.  All of the
mines in this source are represented in some fashion in the composite database, therefore the
composite appears to be complete in this regard.  In fact, the reverse finding, that more records of
locations in Texas are present in the composite database than in the RRC database, highlights the fact
that the composite contains records that may represent individual mine features, "dogholes," and
other features which may not meet the definition of a producing mine, but which could be the site of
uranium mining wastes.

In terms of specific results from the RRC-ULD uranium location comparison, regardless of the
existence of redundant and compound mine names in the RRC data set, all 71 locations were found in
the ULD.  If Texas is reflective of the success of building a complete data set for the other states, the
ULD should be an extensive repository of potential uranium mine locations. As mentioned above, the
ULD contained many more records of locations in Texas than the RRC data set: 363 to 71.  (The
other databases with Texas data include MAS/MILS, University of Texas, and Texas Department of
Health.) Many of these represent duplicates related to the 71 RRC mines (about one-third).  However,
many more (two-thirds), represent spatially unique deposits that are possibly indications of areas
where exploration was performed but production may never have begun.
Accuracy

As for a preliminary assessment of the locational accuracy of the ULD, of the 130 locations in the
ULD that appeared to match the 71 deposits in the RRC database, the average distance between the
RRC mine and the potential matches or duplicates in the ULD was approximately 1200 meters. The
duplicate identification process discussed below was applied to the entire ULD. The RRC mines
were simply added as though it was another data source, and each record was treated as the
authoritative record against which ULD mines found to be nearby and named similarly, were
matched. The minimum distance between the matched ULD and RRC mines was just under 5 meters,
from ULD duplicate to RRC, and the maximum was approximately 8500 meters. Both minimum and
maximum distance values represent outliers, although with or without outliers, the average spatial
offset remains in vicinity of 1100-1200 meters.

The  accuracy range represented in the composite database is generally between 0 and 1500 meters for
database records compared to mines on USGS topographic maps. While this range  is broad, it
represents an average discrepancy between coordinate  information available in the database and the
USGS map location where the deposit in question may be found. The ends of this range were
determined by comparing a small sample of ULD mine records (chosen  for being located within the
bounds of USGS 1:24,000 scale maps that contained the most mines) to  mines  represented on USGS
24K maps.  The USGS mine locations were deemed authoritative since they were identified from
1:24,000 scale maps (or better) and sometimes confirmed and labeled with mine names using  local
authoritative data when available.
                                            10

-------
MAS/MILS and MRDS are the uranium location data sources relied upon to provide location
information. These data sets have no reported accuracy but were both judged to be off by an average
of 200 meters when compared to USGS 24K mines in the quality assurance steps performed on the
ULD mentioned below.
Duplicate Record Identification and Reliability Value Assignment

Due to the various overlapping data sets used to compile this composite database, many duplicate
records exist. The reliability characteristics of the individual records were assessed so that, based on
a single reliability value, the least reliable location records might be identified.  In addition to this
process, duplicate records were identified based on proximity of coordinates and approximate match
of mine name.  Since mine names may differ due to user entry differences but still represent the same
mines, mine names were manually reviewed for likeness. Additionally, since a single mine or
uranium activity could have many coordinate pairs associated with it, a buffer of 2400 meters
(~1.5 miles) was ultimately used to consider the nearby mines  for duplication. Additionally, when
comparing mine names and assigning Dup_MatchName levels, the operators used a conservative
measure in order that mine records with even the smallest chance of being unique were left in the
database. As a result, duplicates remain, but none that could be removed beyond the  shadow of a
doubt.

It should also be noted that in response to reviewer comments, the RRC database of Texas mines, and
MAS/MILS mines for the rest of the U.S. (apart from the original western study area) were added to
the  ULD at the end of the contract period. As a result, the reliability and duplicate record codes are
not populated for these records.

A program was  written to assist an operator in tagging likely duplicate records, based on name
similarity and proximity, with the following fields of information: Dup_MatchName, Dup_MatchID,
and Dup_Dist.  The likelihood of being a duplicate record, combined with the reliability information,
allows for the removal from the ULD of the least reliable records or the most likely to be a duplicate
location.

Each of the following fields of information was added based on assumptions about what constitutes a
reliable record of information. Since few fields were reliably populated other than LATITUDE,
LONGITUDE, and MINE_NAME - these fields formed the basis for the decision-making regarding
which records to keep. The reliability fields (see below) were  populated in such a way that their
content could be summed into a single reliability score.  Since  there are 10 reliability fields, the
maximum reliability score is theoretically "10", and each field is either populated with a "1" or "0"
(or  blank) for each record.

The basic premise in the reliability scoring is that a record is worth keeping if it is either completely
unique (by way of name and location) or if it cannot reliably be equated to any other record due to
subtle or great differences in the mine name.
                                            11

-------
Reliability Value Assignment

Below is a listing of all fields populated used to establish each uranium location record's relative
reliability:

BLK UNIQLL
A value of "1" in this field indicates that the record has a non-blank mine name or that it has a non-
blank mine name AND a unique coordinate pair.  [Note: Records with blank mine names, that shared
coordinate pairs with records with non-blank mine names, were considered to be marked for deletion
during the duplicate identification process.]

GEN_UNIQLL
A value of "1" in this field indicates that the record has a non-generic mine name or that it has a non-
generic mine name AND a unique coordinate pair.  [Note: Records with generic mine names but that
shared coordinate pairs with records with non-generic mine names, were considered to be marked
for deletion during the duplicate identification process.]

IDENT  WI
A value of "1" in this field indicates that the record contains a unique coordinate pair as compared to
all other records in the source database, e.g., in DB1.

IDENT  WO
A value of "1" in this field indicates that the record contains a unique coordinate pair as compared to
all other records across all source databases, e.g., a coordinate pair not found elsewhere in DB1-
DB24.

SPAT ISOL
A value of "1" in this field indicates that the record is spatially isolated.  Spatial isolation is defined
here as a uranium activity with no other uranium  activity within 2400 meters (1.5 miles).

DOCS
A value of "1" indicates that the data source has documentation.

IDENT  NMAEC
A value of "1" indicates that the uranium activity shares the precise name of a uranium activity listed
in the authoritative U.S. Atomic Energy Commission's Uranium Mine and Properties Database
(UMPD) within the same State.  [Note: Tthe UMPD mine list only covers uranium mine activities in
the 4 Colorado Plateau states.]

IDENT  IDMILS
A "1" indicates that the record is not known to have originated from MAS/MILS. [Note: For
example, all but approximately one dozen records from DB12 (State of Arizona) were tagged with a
MILS ID, rendering nearly the entire data source a duplicate of records available from DB5
(MAS/MILS).]
                                             12

-------
QA FLAG
A "1" indicates that the record was not identified as a spatial outlier based on comparing the State
attribute field contents with the actual State in which a point-in-polygon procedure locates the
coordinate pair.  For example, the coordinates for a few records from DB18 (New Mexico Mines
Database) that are known uranium activities in New Mexico, actually appeared in neighboring States.
This field was de-populated and re-populated with  "9"s where the location placed the record outside
of the states.

IDENT COORD
A "1" indicates that the record does not share identical coordinates with any other record in the entire
database.  [Note: A non-unique coordinate pair may be an indication of either a duplicate record OR
of unreliable location information in the case where a Public Land Survey System (PLSS) section
centroid is substituted for multiple uranium activity occurrences in the absence of true coordinate
information. A repeated coordinate pair may also  be a sign that an approximate central location
(such as the entrance to a uranium deposit or area) was recorded for multiple physically unique mine
features (such as shafts, openings, adits, etc.).  Such a location may be less reliable when trying to
identify the size of nearby at-risk populations, as the coordinate pair is an approximation.]

Not all records were considered in the duplicate identification process.  Records that would be kept
without needing to be run through the name and proximity match process, include:

    •   Spatial outliers (where SPATJSOL = 1)
    •   The "unmatchables" (where BLKJJNIQLL = 0 or GENJJNIQLL = 0)
    •   The "already matched" (where IDENT IDMILS = 0)
    •   Supposed mine features (all but one record from a group of records originating from a single
        data source with identical names  and coordinates: IDENT_WI = 0) [Note:  The thought
        process here is that records with identical coordinate information do not contribute to a
        better estimate of impacted populations and environmental resources. Furthermore, the
        premise is that multiple identical records that originate from the same data source are likely
        to represent physically distinct features of the same uranium activity.]
    •   Records added after the duplicate process had already  been performed.
Duplicate Records

Below is a listing of fields that were populated in a partially automated procedure to identify those
records that are likely to be duplicates or provide redundant uranium activity information. Some
terminology used below:

    •  keeper - this term refers to uranium activity records that will be kept in the final ULD.
    •  duplicate - this term refers to uranium activity records that have similar or identical names
       AND that are within 2400 meters (~1.5 miles) of each other.

Dup  MatchID
This field reports the ICF_ID of the duplicate record that will be kept in the database - the keeper ID.
If the ICF_ID and Dup_MatchID are the same it means that the record is unique, spatially isolated,
does not contain a name or proximity match with any other record in the database, or, for other
reasons, is considered a "keeper." A sort on this field will show which records were found to be alike
and considered duplicates.
                                             13

-------
Dup_MatchName
This field records the quality of the name match with the keeper by assigning a value of 1-4, where:
    1 = the names are a perfect match;
    2 = the names are slightly off (e.g., Mary vs. Mary Mine; Mary Prospect vs. Mary Mine; Margie
       2 vs. Margie #2; Buck Shot vs. Buckshot; Tramp #2 Group vs. Tramp #2 Mine Group);
    3 = the names are more than slightly off but still considered a match (e.g., Buffalo Head Mining
       Co. Claim # vs. Buffalo Head Mine; Mammoth-Lincoln vs. Mammoth Mine); and
    4 = the name of one record, that shares identical coordinates with another record, is a generic
       name (e.g., Lucky Strike versus Uranium Prospect) or blank.

Dup  Dist
This field records the distance (meters) between a duplicate and its keeper.  For keepers, the value of
this field is zero.
Other Criteria Applied When Tagging Duplicate Records

Records with names that include the word "Group" were not considered duplicates of other records
with identical names excluding the word "Group" in case they represent distinct mine features (e.g.,
Dorothy Mine vs. Dorothy Mine Group).  [Note: This criterion was observed to leave in many
possible duplicates.]

Records with names that include a number were not considered duplicates of other records with
identical names excluding the number in case they represent distinct mine features (e.g., Dorothy
Mine vs. Dorothy Mine #4).
Originating Data Source Reliability

For each state, all contributing data sources were ranked so that when duplicates are identified, the
record from the most reliable data source might be designated the "keeper" (see ranking list below).
It was furthermore decided that records with either identical or approximate mine names AND
approximate (but not identical) coordinates, originating from the same data source, would be kept in
the likelihood that they represented true geographically distinct mine feature locations. In Table 4
below, the data sources are listed for each State in order of most to least reliable, moving left to right.
The least reliable data sources are those for which documentation was poor or unavailable, attribute
information was  limited or mostly unpopulated, and the spatial accuracy of the locations associated
with the data source, relative to the USGS mine locations, was low.
                                             14

-------
                          Table 4. Data Source Ranking By State
                         (Numbers refer to numbers assigned to each
                        original data source and recorded in the each
                       uranium location record in the DB ALIAS field.)
State
AZ
CA
CO
MT
ND
NM
NV
SD
TX
UT
WA
WY
Database Number,
in order of most to least reliable from left to right
12,5,4,13,11,18
16,5,4,13
2,5,4, 13,3 (DB1
removed entirely for spatial unreliability)
23,5,4,22
5, 4, 22
18,5,4,13,11
5, 4, 20
15,5,4,22
5,24,17,13,4,21
2,7
5,4
2,19,5,4,15
Using the Reliability Factors and Identified Duplicates to Remove Duplicates

Once the potential duplicates, based on name match and proximity are identified, the records are
tagged with the above reliability factors.  No records were ultimately removed based on a low
reliability score, however many were removed based on the findings of the duplicate removal process.
Table 5 lists the reliability score distribution for ULD records.

                         Table 5. Distribution of Reliability Values
Reliability Score
4
5
6
7
8
9
10
Record Count
54
874
1548
5567
8159
6221
176
                                           15

-------
The first step was to separate out the keepers and remove the duplicates from the subset of records
included in the duplicate identification process.  This step could theoretically have been modified by
only removing those duplicates that also had low reliability scores.  Also, only those records that were
within a tighter Dup_Dist could have been removed (such as closer than 1000 meters instead of 2400
meters). However, it was decided here to remove all duplicates.
Data Sources

More than 20 data sources were collected to compile the ULD. These data sources are discussed
below. They were combined into a composite Microsoft Access 2000 database, Master03.mdb and
also into shape files (e.g., ULD_albers.shp) that can be readily used by ESRI ArcGIS software.

The individual databases are numbered 1-253 and this ID is recorded in the DB_ALIAS field of the
ULD. The field ICF_ID records a unique ULD record number in addition to the data sources'
inherent unique ID that is recorded in the DBUNIQUE field. Here are brief descriptions of the data
sources:

DB1: BRASSCAP
This database is managed by the Colorado Department of Natural Resources, Division of Minerals
and Geology. This database contains 3491 uranium sites but only 420 of these sites contain
geographic coordinates. Geographic coordinates are in latitude and longitude, NAD27 Datum. This
source has been eliminated from the final data set due to the uncertainty of its reliability.

DB2: Colorado (BLM) Abandoned Mine Land Inventory (AML)
The Colorado BLM AML database is managed by the BLM Colorado State Office. The Bureau of
Land Management's National Database was the source for these records. The original projection of
the data was UTM.

DB3: Colorado (FS) Abandoned Mine Land Database
The Colorado-Forest Service-AML database is managed by the Forest Service, Region 2. All data
were collected under contract by the Colorado Department of Natural Resources, Colorado
Geological Survey. The contractor used Arc/Info to create a GIS layer for each Forest Service
Ranger District. Information in this database is for the Creede, Cebolla, and Norwood Ranger
Districts. Only 13 records of this database are in the composite shape file  (ULD_albers.shp).

DB4: U.S. Geological Survey Mineral Resources Data System (MRDS) database (McFaul et al,
2000)
While the original MRDS  database is global in geographic extent, a United States only Access
version of the database was filtered for only uranium commodity mines in the states of interest.  A
spatial database was created from the provided latitude/longitude information, projected to an Albers
projection, and the records were tagged with ICF IDs. Spatial correlation with USGS mines is 100-
200 meters based on small comparison sample.
3 Four sources, i.e., DB1, and DB8-DB10, were eliminated from the master database due to concerns about
  reliability.
                                            16

-------
DBS: U.S. Geological Survey Minerals Availability System/Minerals Industry Location System
(MAS/MILS) database (McFaul et al, 2000)
Spatial data were created from the latitude/longitude coordinates provided, projected to the Albers
projection, and then tagged with ICF_IDs. Additional information on ownership and production are
available for these records and can be joined from the DBF files provided (DB5_Owners,
DB5_Production)4. The MAS/MILS data contained in the ULD may show different numbers of
locations than the BASINS repository for each state since the ULD data have been filtered in the
following manner: (1) only records with Uranium in the commodity field were included, (2) only
records with both required coordinates were included, and (3) some mine records from DBS have
been replaced by mines from more authoritative data sets.  Spatial correlation with mines noted on
USGS topographic maps is 100-200 meters based on a small comparison sample.

DB6: Utah (BLM) Abandoned/inactive Mine Land (AML) Inventory
The Utah BLM AML database is managed by the Bureau of Land Management, Utah State Office.
Spatial correlation with USGS mines is within 150 meters based on comparison sample.

DB7: Utah Abandoned Mine Reclamation (AMR) Database
The Utah Abandoned Mine Reclamation Database is managed by the Utah Department of Natural
Resources, Division of Oil, Gas and Mining. Spatial correlation with USGS mines is within 50
meters or less based on comparison sample.

DB11: Navajo Lands Project
An Arc View shape file containing nearly one thousand mine locations, but unsupported by any
metadata, was provided via the CD ROM series "Abandoned Uranium Mines Project Arizona, New
Mexico, Utah - Navajo Lands Project Atlas, 1994-2000."  These records were tagged with ICF_IDs
ranging from 10,413 to 11,299. Since no projection information was available, it was assumed to be
geographic with North American Datum 1927, and was projected to Albers. The location shape file
was culled from dozens of maps and mining claim documents, and these locations  are approximate.
The data are of varying accuracy and were assembled to obtain a general idea of mine concentrations
to target remote sensing forays of the Navajo Study. Updated information related to this project
became available too late for inclusion into this database effort (NAMLRP, 2004).

DB12: State of Arizona
These data were also provided undocumented and in the form of an ESRI shape file.

DB13: Forest Service
These data were received in shape file format with an accompanying data dictionary and Users Guide.
Records were projected to Albers. Approximately 15% of the records had latitude/longitudes that
placed the locations outside of Arizona and New Mexico.  Only nine records of this database are in
the composite shape file (ULD_albers.shp).
4 Note that ownership and production information is only available for 11 western states included in this
  compilation, however, such information is available in MAS/MILS and can be extracted by linking to the
  original MAS/MILS data sets using the DBUNIQUE field which corresponds to the MAS/MILS.SEQ field.
                                            17

-------
DB14: Bureau of Land Management
These data were received in shape file format but without accompanying projection information or
other documentation. However, the data do have information for numerous fields. The records were
tagged with ICF_IDs ranging from 14,987 to  16,935, and projected to Albers, assuming they were
unprojected to a datum of NAD27. Thirty-one records appeared outside of the United States. These
data were removed from the single data set as they appear to represent mines other than uranium and
a commodity field is not available for filtering purposes.

DB15: South Dakota Abandoned Mine Lands Inventory
These data were received as part of a Microsoft Access 2.0 and Microsoft Access 95 application that
allow users to query this relational locations database.  Uranium locations were not flagged such that
they could be identified, however, a contributing member of the team that developed the database
reported that all locations in Fall River County, South Dakota,  unless otherwise identified, were
uranium locations. These locations, and their associated data from the related data tables, were
exported from the Access application.  Some export difficulties regarding MEMO fields required that
these larger fields were not brought along into this compendium.

A report entitled "Comprehensive Inventory of Known Abandoned Mine Lands in the Black Hills of
South Dakota" from January 26, 1998, accompanied the application. This report identified three
input sources, including (1) USGS database for Metallic Mineral Districts and Mines in the Black
Hills, (2) USFS inventory of sites  on National Forest lands, and (3) a South Dakota database
compiled from literature search of known AML sites in the Black Hills. The final database had single
records for locations and removed any duplication. All USFS locations were field verified.  About 70
of the remaining non-USFS locations were  field verified with a GPS unit. The report stated that none
of the locations visited were deemed to be in need of remediation, however some seemed to pose
"considerable" environmental risk. Apart from this report an install manual was made available.
Latitude and longitude information were available (based on the nearest one quarter of a quarter
section to the location) and were assumed to be NAD83.  A GIS file and a DBF file were generated in
Maplnfo. ICF_Ids and the DB_ALIAS field were added in Excel.  There was no apparent match to
USGS  mines.

DB16: California (FS) Mines
These data were received partly in the body of an email and partly as an Excel spreadsheet. Spatial
files were generated in Maplnfo from provided coordinates; datum was assumed to be NAD83.
There were no apparent matches to USGS mines. Only 17 records of this database are in the
composite shape file (ULD_albers.shp).

DB17: Texas Department of Health
Twenty-seven locations with driving directions to the sites were received via fax.  Delorme Streetmap
Version 9 was used to locate the locations by  placename and then by major route intersection. A staff
member of the Texas Department  of Health assisted our effort  by reviewing the mine locations for
about a half dozen mines and sending in corrections. Most mines had been placed within one-tenth of
a mile  (160 m) of the actual location. However, one was off by over a half-mile (800 m). There was
a limited match to USGS mine locations. None of these are included in the composite shape file
(ULD_albers.shp).
                                            18

-------
DB18: New Mexico Mines Database
EPA's Office of Radiation and Indoor Air (ORIA) funded the digitization of this collection of
location data that, in part, dates back to the DOE NURE Program in the 1970s. Paper files were
converted to a relational Access database by staff at the New Mexico Bureau of Geology and Mineral
Resources, a division of the New Mexico Institute of Mining and Technology.  Full documentation
was received with this data set. The Access tables were converted to GIS format in Maplnfo and then
converted to DBF in Excel. Perfect spatial alignment was observed with USGS mine locations.

DB19: Wyoming Abandoned Mine Land
No documentation was received with this data set.  There was no apparent match with mine names on
USGS maps, or with mine names.

DB20: Nevada Bureau of Mines and Geology
These locations were digitized from a book that contained a mine location map. Names were hand-
entered from entries in the book. Only those locations that did not appear in other data sources were
digitized. This was determined by comparing the hardcopy map to a GIS map of Nevada locations
from MAS/MILS and MRDS. Some names and locations matched USGS mine locations.

DB21: Texas Mines from Adams & Smith Report.
These locations were digitized from a map contained in the report. No documentation or location
information was included with the map. No apparent match was found with USGS mines for either
names or locations.

DB22: Dakotas Mines from USAEC Map 1967
Locations were digitized from this map. No documentation was available, apart from a brief legend
and symbolization that identified two types of uranium available from the mines. There was no
apparent match with USGS mine locations.

DB23: Montana State Library (MILS)
These location data were accompanied by full metadata detailing accuracy mine location information.
It appears however that these locations may have been drawn from MAS/MILS as their spatial
locations are highly correlated. Some name matches were found for USGS mines; locations were off
by 0-1000 m.  Only eight records of this database are in the composite shape file (ULD_albers.shp).

DB24: Inactive Mineral Production Sites - University of Texas
These locations were hand-entered from a fax. No documentation accompanied this data source. No
USGS mine correlations were found. Only seven records of this database are in the composite shape
file (ULD_albers.shp).

DB25: Railroad Commission of Texas (RRC) Uranium Mines
No metadata were  available as it was a work in progress, however information about how the
locations were verified was available. Local area specialists combined aerial photography with some
field-truthing in order to verify the mine locations.  This is the most authoritative collection of
uranium mines currently available in Texas, although it appears to be incomplete from the  perspective
of uranium occurrences. Sources at the USGS report NURE (National Uranium Resource
Evaluation) folio evidence of uranium production from additional locations in Texas. However that
data was not available at the time of this compilation.
                                           19

-------
  Figure 1. Western Uranium Locations From the EPA Uranium Location Database
Legend

  o   EPA Identified Uranium Locations
        Miles
       ^^M
0  75150    300   450
   f.

" V "
                                    20

-------
      Figure 2. Western Uranium Locations from the USGS MRDS Database
Legend

  *   MRDS locations
       Miles
       ^^M
0 75150   300   450
                                  21

-------
           Figure 3. Density of Western Uranium Mines
Using the MAS/MILS Database Portion of the Uranium Location Database
                                     Legend
                                     Western Uranium Mine
                                             Density By
                                       Hydrologic Unit  Code
                                           HUCs>100 U mines

                                           HUCs 51-100 U mines

                                           HUCs 11-SOU mines

                                           HUCs 6-1OU mines

                                           HUCs 1-5 U mines
                                       In the MAS/MILS Database the
                                       Upper Dolores (CO). San Miguel (CO),
                                       and Lower Dolores (CO) Hydrologic
                                       Unit Codes (HUCs) Each
                                       Have > 300 Uranium Mines

-------
References

CRCPD 2004—Conference of Radiation Control Program Directors, Part N, Regulation and
Licensing of Technologically Enhanced Naturally Occurring Radioactive Material (TENORM).
Suggested State Radiation Control Regulations Volume I. Frankfort, KY: CRCPD. April 2004.

IAEA 2003—International Atomic Energy Agency. Technical Report Series No. 419. Extent of
Environmental Contamination by Naturally Occurring Radioactive Material (NORM) and
Technological Options for Mitigation. Vienna, Austria. 2003.

McFaul, E.J., G.T. Mason Jr., W. B. Ferguson, and Lipin, B. R. U.S. Geological Survey. U.S.
Geological Survey Mineral Databases—MRDS and MAS/MILS by McFaul et al, U.S. Geological
Survey Digital Data Series DDS-52 Two Discs. 2000.

NAMLRP 2004.  Geographic Information System Database Development for the Navajo Abandoned
Mine Lands Reclamation Program, September 2004.

NAS 1999a—National Academy of Sciences. Evaluation of Guidelines for Exposures to
Technologically Enhanced Naturally Occurring Radioactive Materials. Washington, DC: National
Academy Press, 1999.

U.S. DOE/EIA 2005—U.S. Department of Energy, Energy Information Administration. Domestic
Uranium Production Report, 2nd Quarter 2005 Production. Washington, DC: U.S. DOE. August
2005. http://www.eia.doe.gov/cneaf/nuclear/dupr/dupr.html

U.S. DOE/EIA 2005b—U.S. Department of Energy, Energy Information Administration. Domestic
Uranium Production Report, Bra4 Quarter 2005 Production. Washington, DC:  U.S. DOE. August
2005. http://www.eia.doe.gov/cneaf/nuclear/dupr/dupr.html

U.S. DOE/EIA 2003—U.S. Department of Energy, Energy Information Administration. Uranium
Industry Annual:  2002. DOE/EIA-0478(2002). Washington, DC: U.S. DOE. May 2003.

U.S. DOE/EIA 2003b—U.S. Department of Energy, Energy Information Administration. Domestic
Uranium Production Report. Form EIA-851. February 2003.
http://www.eia.doe.gov/cneaf/nuclear/page/at a glance/qtr upd/qupd.html

U.S. DOE/EIA 1992—U.S. Department of Energy, Energy Information Administration. Domestic
Uranium Mining and Milling Industry: 1991 Variability Assessment. DOE/EIA-0477(91).
Washington, DC: U.S. DOE. December 1992.

U.S. EPA 2006a—U.S. Environmental Protection Agency. Technologically Enhanced Naturally
Occurring Radioactive Materials (TENORM) from Uranium Mining: Volume I: Mining and
Reclamation Background. EPA 402-R-05-007. Washington, DC: U.S. EPA. 2006.

U.S. EPA 2006b—U.S. Environmental Protection Agency. Technologically Enhanced Naturally
Occurring Radioactive Materials (TENORM) from Uranium Mining: Volume II: Investigation of
Potential Health,  Geographic and Environmental Issues of Abandoned Uranium Mines. (In press).
Washington, DC: U.S. EPA. 2006.
                                           23

-------
U.S. EPA 2003—U.S. Environmental Protection Agency. Guidance—Potential for Radiation
Contamination Associated with Mineral and Resource Extraction Industries. Washington DC: U.S.
EPA, Office of Radiation and Indoor Air. April 15, 2003.

U.S. EPA 2001—U.S. Environmental Protection Agency, Science Advisory Board, TENORM—
Evaluating Occurrence and Risks, an SAB Advisory; A Science Advisory Board Advisory on EPA 's
Approach for Evaluating Occurrence and Risks of Technologically Enhanced Naturally Occurring
Radioactive Materials (TENORM). EPA-SAB-RAC-ADV-01-001. Washington, DC: U.S. EPA.
February 2001.

U.S. EPA 2000—U.S. Environmental Protection Agency. Report to Congress on EPA 's Guidances on
Technologically Enhanced Naturally Occurring Radioactive Materials. Washington, DC: U.S. EPA.
June 2000.

U.S. EPA 1994—U.S. Environmental Protection Agency, Science Advisory Board. An SAB
Report: Review of Diffuse NORM Draft Scoping Document. EPA-SAB-RAC-94-013. Washington,
DC: U.S. EPA. May 1994.

U.S. EPA 1995—U.S. Environmental Protection Agency. Extraction and Beneficiation of Ores and
Minerals: Uranium. Vol. 5. Technical Resource Document. EPA 530-R-94-032. Washington, DC:
U.S. EPA, Office of Solid Waste. January 1995.

U.S. EPA 1985—U.S. Environmental Protection Agency. Report to Congress on Wastes from the
Extraction and Beneficiation of Metallic  Ores, Phosphate Rock, Asbestos, Overburden from Uranium
Mining, and Oil Shale. EPA 530/SW-85-003. Washington, DC: U.S. EPA. December 1985.

U.S. EPA 1983a—U.S. Environmental Protection Agency. Report to Congress on the Potential
Health and Environmental Hazards of Uranium Mine Wastes. Vol. 1. EPA 520/1-83-007.
Washington, DC: U.S. EPA. June 1983.

U.S. EPA 1983b—U.S. Environmental Protection Agency. Report to  Congress on the Potential
Health and Environmental Hazards of Uranium Mine Wastes. Vol. 2. EPA 520/1-83-007.
Washington, DC: U.S. EPA, June 1983.

U.S. EPA 1983c—U.S. Environmental Protection Agency. Report to Congress on the Potential
Health and Environmental Hazards ofUraniumMine Wastes. Vol. 3. EPA 520/1-83-007.
Washington, DC: U.S. EPA. June 1983.
                                           24

-------