A computerized system to estimate potential uranium resources

Resources

and Energy

13 (1991) 201-215.

North-Holland

A computerized system to estimate potential uranium resources Sujit Das and Russell Lee* Oak Ridge National Laboratory, Oak Ridge, TN 378314205, Received

September

1989, final version

received

October

USA

1990

The URAD system is the only comprehensive computerized database for information and estimates of the quantity of undiscovered uranium resources in the United States. The database includes subjective estimates of uranium endowment and geographic and geologic descriptions for approximately 700 area-specific assessments. The estimated quantity of undiscovered resources are represented in the form of probability distributions. This probabilistic approach makes the uncertainty about these resources explicit.

1. Introduction Uranium is used to fuel nuclear reactors that generate electric power. Discussions of the uranium industry and market are provided in Ahmed (1979), Energy Information Administration (EIA) (1983, 1984a, b, 1988a, b), Neff (1984), and Taylor and Yokel1 (1979). Estimates of the potential uranium resources in the U.S. are used to assess the economic viability of the U.S. uranium mining and milling industry. In particular, ‘resource capability’ is one of the criteria used by the U.S. Secretary of Energy in an annual determination of industry viability [EIA (1988a)]. Methods for estimating undiscovered uranium resources have received attention since 1975, mainly as part of the National Uranium Resource Evaluation (NURE) Program of the Energy Information Administration (EIA) of the U.S. Department of Energy (DOE) [Harris (1976, 1977, 1984), U.S. Department of Energy (1980), Harris and Agterberg (1981), Harris and Carrigan (1981)]. *Juanita Hunt was the key word processing person. Her contributions are gratefully acknowledged. The authors are also indebted to two anonymous reviewers at Oak Ridge National Laboratory; to WI. Finch of the U.S. Geological Survey, whose review and comments on an earlier draft led to improvements in the exposition of the paper; to Matthew Kahn, who made several suggestions on having more of an economic context to the paper; and to a reviewer who made a number of critical, helpful comments. The research upon which this paper is based was part of the Uranium Resources Assessment Data System Maintenance project at Oak Ridge National Laboratory. The project was supportedby the Energy Information Administration of the U.S. Department of Energy under the guidance of Luther Smith, who provided information on uranium data collection procedures. 01654572/91/$03.50

0

1991-Elsevier

Science Publishers

B.V. (North-Holland)

202

S. Das and R. Lee, Estimation of potential uranium resources

The purpose of this paper is to provide a brief description of the Uranium Resources Assessment Data (hereafter referred to as URAD) system developed under the NURE Program, which is used for the estimation of U.S. potential uranium resources. These estimates are used in the EIA publication ‘Uranium Industry Annual’ and in the EIA’s submission for the biennial report ‘Uranium Resources, Production, and Demand’, which is published jointly by the OECD Nuclear Energy Agency and the International Atomic Energy Agency. 2. Overview of URAD system The URAD system contains subjective geologic information about the total uranium resources in approximately 700 areas within twenty regions of the United States. Some of the geologic information is in the form of probability distributions. Numerical procedures are used on the probability distributions to compute estimates of resource quantities from these geologic data. Estimates are produced for environments that have geological characteristics favorable for the occurrence of uranium deposits. The estimates are classified into the categories of Estimated Additional Resources (EAR) and Speculative Resources (SR). Engineering-economic cost estimates are a key ingredient of the calculations. The resources are allocated among Forward Cost (i.e., costs that have not yet been incurred) Categories of $15, $30, $50, and $100 per pound U,O,. EAR is based on direct geological evidence of occurrence and on estimates of the uranium that can be recovered within the given cost ranges, whereas SR is estimated from indirect evidence and geological extrapolations. The estimated quantities of uranium that would be available at different forward costs or prices can be used to construct longrun supply curves and to model the future economic competitiveness of the U.S. uranium industry. The mathematical procedures used in this system were developed by Oak Ridge National Laboratory (ORNL) personnel, principally Ford and McLaren (1980) with the assistance of others, under work sponsored by the Grand Junction Area Office of the Department of Energy. By treating the resource quantity as a random variable and by incorporating expert opinion expressed as subjective probability distributions, the method is implicitly Bayesian. In that respect, the spirit of this approach is similar to that in Solow and Broadus (1989). The URAD system can be thought of as an explicit method of deriving a ‘prior’ probability distribution for the quantity of resources. A ‘posterior’ distribution would be estimated as more detailed data are gathered on individual properties. The URAD system is run under Digital Equipment’s (DEC’s) TOPS-10 operating system. The system utilizes FORTRAN subroutines for the calculation of endowment and economic potential resources, and employs

S. Das and R. Lee, Estimation

of potential

uranium resources

203

System 1022 for database management and some report generation. MAPPER, an internally written Oak Ridge National Laboratory software system, has been used where graphical displays have been required. MAPPER is a front end preprocessor for the Tell-A-Graf graphics package, which is used for creating graphic outputs. URAD has also been installed in an IBM-compatible PC environment using a combination of FORTRAN and ORACLE information management software.’ The flow of data in the system is shown in fig. 1. Reference Information are System 1022 files containing geographic and geologic names; reference codes for location identification; and information about its geologic characteristics (recognition criteria), uranium reserves, potential resources and production. The reserves data are proprietary; hence only aggregate information on the probability distributions for the twenty resource regions of the United States are maintained. The resource data include subjective uranium endowment estimates and geographic and geologic descriptions for approximately 700 area-specific assessments spanning the two potential classes: EAR and SR. Geologists’ judgments account for 95 of the 156 attributes stored for each assessed area in the Master Data File; for example, size, thickness and oregrade distribution are a few of the major attributes which require geologists’ judgments. These data were collected through national surveys using the DOE’s former GJ-85 form. The remaining values are estimated by the URAD System - for example, the capital and operating costs, cutoff and average grades and the probability distribution of resource estimates in terms of mean, variance, and third moment. The Master Data File serves as both an input and output for potential resources data. The Endowment Model uses the subjective geologic judgments of three random variables to describe the fraction of the area underlain with ore, the tonnage of ore associated with that fraction and the grade of the endowed rock. These data are from the Geologists’ Judgments and the Reference Information databases (refer to fig. 1). The uranium endowment distribution function is then computed by multiplying the above three random variables together times the area being assessed (usually measured in square miles). The estimation methodology is discussed further in section 3. Mining Engineers’ and Mineral Economists’ Judgments concerning the costs of finding, mining and milling the undiscovered uranium ore are used to develop a Cost Model. Geologists’ subjective information such as mine type and depth interval, together with the computed endowment, are utilized to obtain eleven separate cost factors in the cost model. These cost factors are various components of capital and operating costs. The relationships between the cost factors, the geologists’ subjective information, and the cost escalation ‘At the time of development, there were neither systems similar to the URAD system nor standard mathematical routines. Consequently, the software was custom developed. The FORTRAN language was chosen for its strength in making scientific calculations. MAPPER or any other graphics software could be used for the graphical displays.

S. Das and R. Lee, Estimation of potential uranium resources

ENDOWMENT

GEOGRAPHIC, + CONTROL AREA

MASTER DATA FILE

COST

MODEL MODEL

ECONOMIC POTENTIAL RESOURCES

(700 ASSESSMENTS)

MODEL

($15,30,5O,lOO/LB)

nureflow.dnv Fig. 1. URAD

system information

flow.

indices that are used to estimate extraction and recovery costs, have been reviewed and updated annually since the model was first used in 1980. The Marshall/Swift Price Index, Wholesale Price Index and Chemical Engineering Price Index are cost indices that are used to take into account the escalation of mining production costs over time. The Economic Resources Potential Model uses the estimated endowment and cost factors to develop probability distributions for the potential resources that are expected to be ultimately discovered and produced at current dollar

S. Das and R. Lee, Estimation of potential uranium resources

205

distribution costs of $15, $30, $50 and $100 per pound U30,. 2 The probability is described in terms of its moments. Normally, execution of the model results in a printed assessment report about each favorable area of interest and the creation or update of records of those areas on the 1022 Master Data File. The favorable areas are defined a priori based on their geologic structure. Once individual resource assessments are created or updated, other summary reports may then be generated. Summary reports, where various individual assessments may be added together to form regional or national totals, are normally generated from the data stored in the 1022 Master Data File. FORTRAN, 1022 or MAPPER utility programs may be used. Das et al. (1988) provide an operations manual for this system. The following discussion describes the URAD estimation methodology in greater detail.

3. Estimation methodology The URAD system uses the NURE assessment methodology which uses the geologic analogy concept (i.e., if one area has known ore deposits of certain sizes and grades, other areas with very similar geologic settings may have ore deposits of similar sizes and grades). Guidelines or criteria to recognize areas having similar geologic, geochemical and geophysical characteristics have been established for the world-class and other important deposits. These criteria are reported as the conclusions of hundreds of reports on areas studied during the past 30 years, and several summaries of recognition criteria are available for use in resource assessments [Mickle and Mathews (1978), Mathews et al. (1979)]. The information required to estimate geologic endowment in the NURE method is obtained from subjective judgments. The reliability, or repeatability, of the assessments is reasonably guaranteed in this method by means of elicitation. Elicitation of the undiscovered uranium endowment involves formalized discussion, among the principal scientist and the team of experts, to derive the factors needed to calculate the uranium endowment of a favorable area. Elicitation is a convenient, relatively fast, thorough, complete, and continuous process leading from information on known deposits to estimates of the undiscovered deposits. It is the key element in this method, and the credibility of the estimate can be judged on the basis of the audit trail developed during the process. McCammon et al. (1986) have attempted to reduce the subjectivity in this approach by integrating large data sets with genetic models in a relatively well-explored basin in New Mexico. The deposit-size-frequency (DSF) 2The dollar categories are arbitrary. under $15 per pound U308.

Others

can be used. The current

spot price is somewhat

206

S. Das and R. Lee, Estimation

of potential uranium resources

description of the endowment (discussed in section 3.1) also permits the utilization of geological knowledge in a less subjective manner. In the NURE method, uranium resource estimates are detined as random variables, recognizing that the quantity of uranium to be found and recovered from any particular region is probabilistic and inherently unknowable. URAD attempts to quantify this uncertainty and thus to eliminate simplifying assumptions. The method of moments is used to estimate the probability distribution. Piepel et al. (1981) have also made probabilistic estimates of U.S. uranium supply based on the DOE-GJO database using this methodology.3 The following discussion describes the major computational components of URAD in greater detail.

3.1. Endowment Endowment is defined as the quantity of uranium, irrespective of economic considerations, contained in undiscovered deposits at a grade of 0.01 percent U,O, and higher. The conditional uranium endowment of a given area (U,) is calculated as the product of four statistically independent random variables [i.e., projected surface area of the favorable area (A), the fraction (F) of A that is underlain by endowment, tons (T) of endowed rock per square mile within A x F, and average grade of endowment, in decimal fraction form (C)l. Mathematically, U,=AxFxTxG. The above factors are sequentially estimated based on geologists’ judgments with F, T and G being treated as random variables from a lognormal probability distribution, that is described by its high, low and most likely values. These three values of the random variables correspond to the 95th percentile, 5th percentile and mode, respectively, of a probability distribution function. The lognormal distribution has a rather general form and is commonly used to describe the distribution of geologic phenomena. The endowment is expressed in terms of the probability of occurrence of deposits. First, a three-parameter lognormal distribution is determined for each of these random variables (i.e., F, T and G) based on the three values estimated 3Mckelvey’s reserve-abundance relationship is another method commonly used for resource estimation [Harris (1977)]. In this method, appraisal of endowment consists of examining the relationship of the reserves of metals at a particulat instant in time to their crustal abundance in the earth. Thus, the relationship across metals of a time-varying measure (i.e., known economic reserves) to a time-invariant measure (i.e., crustal abundance) is used as a basis for estimating the unknown economic resources of a region. Even if the physical underpinnings of this approach were correct, an assessment of geologic resources using this relationship would be relevant only for economic conditions at a particular instant in time. Further discussions of methods for estimating mineral resources are in Lee and Thomas (1983).

S. Das and R. Lee, Estimation

of potential

uranium resources

207

by the geologist for each of these three random variables4 Next, the three moments (i.e., the mean, variance, and the third moment about the mean) are obtained for the distributions of each of the random variables.5 Finally, the distribution function for U, is determined from the resulting moments of each of the random variables.6 Depending on the moments, the distribution function may include the lognormal, normal, or Student’s t distributions as special cases. A detailed discussion of this computational procedure can be found in Ford et al. (1980). Finch and McCammon (1987) have recently suggested replacing the factors F x T (i.e., the fraction the surface area underlain by endowment and the tonnage of uranium ore in each unit area of endowment) by a depositsize-frequency (DSF) description of the endowment. The DSF method uses an estimate of the number of deposits in each size (i.e., tonnage) category. This deposit-size-frequency distribution is estimated from a number of geologic attributes of the area which are known to give indications of the number and sizes of deposits. Finch and McCammon suggested that the DSF method describes the endowed area in greater detail by considering individual deposits within the area, and by relating the number and sizes of the deposits to specific geologic knowledge. The DSF method merits consideration in any future enhancements of the URAD system, although it would necessitate more research to calibrate the relationships between specific geologic attributes and the deposit-size frequency distribution. Also more geologic input data would be required for the enhanced URAD system.

3.2. Economic

potential

resources

Economic potential resources are based on economic considerations, i.e., the revenue from the ore should at least cover its extraction cost. These resources are calculated for four different cost categories: $15/lb, $30/lb, 4A three-parameter

lognormal

distribution

f(x)=exp[-{(ln(x-d)-a)/bJ2/2]/d2n[b(x-d)] 0, Three parameters values as follows:

a, b and d are calculated

f(x) with parameters for for

a, b and d is defined as x>d. xsd.

from the high (x1), low (xs) and most

likely (M)

x1 =d+exp(a+ 1.645b), x,=d+exp(a-l.645b) and M=d+exp(a-b*). ‘The mean (m) of a lognormal distribution is defined as: d+exp(a + b*/2). The other two moments (i.e., variance and the third moment) are calculated as the expected values of (I-m)’ and (x - m)3, respectively. 6For example, for three random variables x, y and z with means as m,, my and m, respectively, the three resulting moments can be calculated as the expected values of (xyz-mlmym,), and (~y~-rn~rn,rn~)~, respectively. The resulting endowment distribution (as (xyz-m,m,m,)z, defined by the parameters a, b and d) is obtained by solving equations represented by the above three variables.

S. Das and R. Lee, Estimation of potential uranium resources

208

$50/lb and $lOO/lb for each of the 700 assessment areas, for which endowment calculations were made. The equation for economically potential resources is very similar to eq. (1) for the estimated endowment: U,i= A x F x T x F(c) x G(c)

(potential at cost category Ci).

(2)

The only difference from eq. (1) lies in incorporating F(c) (the fraction of rock having grades exceeding the required cutoff grade) and substituting G(c) (the average grade of this fraction) for the average grade of endowment, G, in eq. (1). These terms are the ones which take economic factors into consideration. The cutoff and average grades are dictated by economic operating and capital cost criteria as follows:

Cutoff grade

Operating Costs ($/ton of ore) y0 u,O, =20 x Cost Category ($/lb of U30s) x Mill Recovery’ ( ore1

Average grade

% !!$ ( 1

(3)

Operating + Capital Costs ($/ton of ore) =20 x Cost Category ($/lb of U,O,) x Mill Recovery’ (4)

Mill recovery is expressed as a fraction between 0 and 1; whereas cutoff and average grades are expressed as %(U,O,/ore). The fraction of rock having grades exceeding the required cutoff grade c, F(c), is calculated as the ratio of rock above the cutoff grade to rock of all grades. The distribution of ore grade is described in terms of a probability distribution, viz., a grade-tonnage curve. In discrete terms, this grade-tonnage curve can be defined by a histogram. F(c) is obtained from the area of the histogram of tonnage versus grade increment to the right of the cutoff grade, divided by the area of the entire curve. The main problem lies in determining this histogram of grade versus tonnage for undiscovered deposits. First, G(c) the average grade is assumed to be a linear function of cutoff grade (c) based on DOE mineral inventory analysis of already similar discovered deposits. F(c) is expressed in terms of this linear function as follows:

j

G’(x)/[G(x)-x]dx

0.01%

The cutoff grade (c) required for obtaining F(c) is defined as the larger of (3)

S. Das and R. Lee, Estimation

of potential

uranium

resources

209

or the value obtained after replacing average grade in (4) by the linear average grade vs. cutoff grade function in terms of cutoff grade. For economic potential resources, F(c) and G(c) are random variables and their probability distributions are determined in a similar way to that for the endowment. 3.3. Cost calculations The estimate of economic potential resources for a given forward cost category is a function of cutoff and average grade as shown above in eq. (2). Economic factors such as operating and capital costs determine which grades of material can be recovered at a desired cost per pound as shown above in eqs. (3) and (4). Thus, in the estimation of undiscovered resources, economic factors for discovering, mining, and milling the undiscovered deposits in the favorable area are determined, and the costs are computed considering information about deposit location, depth and other parameters. The total capital costs are calculated as the sum of acquisition costs, exploration costs, development drilling costs and mining and milling capital costs. The operating costs include haulage costs, royalty costs, ad valorem taxes, severance taxes, labor, and other mining and milling operating costs. These production cost components largely depend on the geologic characteristics of the ore bodies. Table 1 provides a tabulation of the cost equations and the relationships between the geologists’ judgments, the internal model variables, and the computed costs as given in Blanchfield (1980). The mining and milling cost models include: open pit, underground and solution. Royalty cost models are disaggregated by different land types. A review of the mathematical expressions used in these cost models can be found in Blanchfield (1980), Golembiewski (1987), and Szymanski et al. (1989). 4. Application URAD produces several summary reports, where various individual assessments are added together to form regional or national totals. Table 2 contains a list of the available summary reports. A total of nine different graphs of cumulative distributions of regional and national uranium resources can also be generated. Fig. 2 shows the cumulative distribution for $30/lb uranium resources in the United States as a whole. The $30/lb category denotes those resources that can be recovered at a forward cost of $30/lb or less. In the figure, for Speculative Resources, there is a very high probability (almost 1) that they exceed 2.5 x lo5 tons of U,O, and a very low probability (almost 0) that they exceed 10 x lo5 tons. The expected value (i.e., the mean value of the probability distribution) is about 5.3 x 10’ tons. The other two probability distributions are for Estimated Additional

costs

X X X

X

Mine loss

X

(1)

Mine operating costs Haulage costs Mill operating costs Royalty costs Ad valorem taxes Severance taxes

Total operating costs

Acquisition costs Exploration costs Development drilling costs Mine operating costs Mill operating costs

Total capital costs

Calculated

Region

Independent

type (2)

Mine

X

X X X

X X X

Land type (5)

determinants.

Mill size (4)

of cost factor

Depth interval (3)

variables

Summary

X X X

State (6) X X X

A.F.T. (7) X

Resource class (8)

X X X

X

X

cost category (9)

S. Das and R. Lee, Estimation of potential uranium resources

211

Table 2 Description

of summary

reports

available

from URAD

system.

Task

Description

1

ADDF85

2 3 4

AGEHRU AGEREP AREREP

5 6 7 8

CAREP CCREP DEPREP ESOREP

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

GEOREP GTREP HRUREP INDEX IRAREP JAPREP LTREP PDVREP PIREP QDREP REGREP RLSREP SLSREP STAREP SUM22

24

YLTREP

From Form GJ-85 input, create a new 1022 dataset or add records to an existing one Summary of host rock-unit codes broken down by Geologic Age Summary of potential by geologic age Summary of potential by class, resource region, exploration area, ore reserve area and locality Summary of potential by control areas and class Ranking of potential assessments with cost category and class Summary of potential by deposit type Descending ranking of assessments by unconditional mean of $50/lb endowments Master list of geographic codes Summary of potential resources by geologic type Summary of potential resources by host-rock-unit List of localities with index to endowment and $5O/lb rankings Produce individual resource assessment reports Produce regional and national uranium resource totals and percentiles Summary of potential resources by years-lead-time Conditional means and 25th and 75th percentiles for probable potential Summary of potential resources by P.I./Assessor Summary of potential resources by quadrangle Summary of potential resources by class and region Summary of potential by class, region and land-status-type Summary of potential by class, state and land-status-type Summary of potential by state and class Produce cumulative distributions of regional and national uranium resources, and Iiles to produce MAPPER graphs Summary of potential resources by years-lead-time

1. RESERVES-MEAN:

0

66830

4

2

6

8

IO (x 105)

U, TONS U308

Fig. 2. Illustration

of URAD

output

providing

estimate

of $30/lb uranium

resources.

212

S. Das and R. Lee, Estimation

of potential uranium resources

Table 3 Example additional

of estimates of uranium resources computed by the URAD system. Estimated resources (EAR) and speculative resources (SR) in the SfiO-per-pound forward-cost category by land status at the end of 1987. Estimated additional

resources”

Million pounds Land status

U,Ds

Public Lands Bureau of Land Management and Forest Service Lands 530 b Bureau of Reclamation Wilderness Areas 20 National Park Service Lands 30 b Wildlife Refuges DOE-Administered 10 Indian Lands 150 State Lands IO Private Fee Landsd 1,450 Other (Military Reservations, Waterways, Reclamation Projects, Proposed Withdrawals, etc.) 70 Total

2,330

_ Percent of total EAR

Speculative

resourcesa

Million pounds U,Os

Percent of total SR

22.8 C

280 b

0.7 1.3 C

20 10

14.2 0.2 LO 0.5 0.1 E

230 70 1,290

11.9 3.7 65.5

0.3 6.6 3.2 62.1

b b

3.0

60

2.9

100.0

1,960

100.0

“Values shown are the mean values for the distributions of estimates of EAR and SR, rounded to the nearest 10 million pounds U,Os. bValue is less than 10 million pounds U,Os. ‘Value is less than 0.1 percent. dIncludes railroad lands and patented claims. Note: Totals may not equai sum of components because of independent rounding. Source: Prepared by staff of the Nuclear and Alternate Fuels Division, Office of Coal, Nuclear, Electric and Alternate Fuels, Energy Information Administration, based on uranium resources data developed under the DOE National Uranium Resource Evaluation (NURE) program, using methodology described in An Assessment Report on Uranium in the United States of America (October 1980), and in U.S. Department of Energy, Uranium Industry Seminar (October 1980).

Resources and Reserves, and are interpreted Resources.

similarly to that for Speculative

Information on the projected domestic uranium resource capability under various market assumptions are required by the EIA for its annual assessment of the viability of the domestic uranium mining and milling inustry. The EM’s annual assessment is based on the resource estimates of URAD and are reported in Uranium Zn~ustry Annual, an EIA publication. The estimates are presented by resource category [i.e., Reserves, Estimated Additional Resources (EAR), Speculative Resources (SR) and Endowment], forward cost categories, resource region and land status. An example is in Table 3. As discussed earlier, the estimation of economic potential resources takes

S. Das and R. Lee, Estimation of potential uranium resources

213

Table 4 Estimated

Resource

resources (tons of U,Os) in the SSOjlb forward-cost category for uranium resource regions for two different sets of cost indices.

region

Alaska Pacific Coast Basin and Range Columbia Plateaus Colorado Plateaus Central Lowlands Appalachian Highlands

selected

U.S.

Estimated additional resources (EAR)

Speculative resources (SR)

Total resources

Cost index set

Cost index set

Cost index set

(1)”

(2Y

(1)”

(2Y

(1)”

(bY

2,419 7,313 94,930 0 427,570 0 35,130

2,414 7,183 90,818 0 417,881 0 34,612

183 3,189 97,498 7,322 239,879 34,046 164,844

175 3,116 94,540 7,032 231,819 33,332 160,918

2,602 10,502 192,428 7,322 667,449 34,046 199,974

2,589 10,299 185,358 7,032 649,700 33,332 195,530

“(1): CEP (1980)=1.219; MS1 (1980)=1.209; (1981) = 1.027; WPI (1982) = 1.026. ‘(2): CEP (1980)=1.311; MS1 (1980)=1.292; (1981)= 1.091; WPI (1982)=1.063.

MS1 (1982)=1.069;

WPI

(1980)=1.136;

MS1 (1982)=1.143;

WPI

(1980)=

WPI

1.208; WPI

into account economic factors for discovering, mining, and milling the undiscovered deposits. Cost indices, such as Wholesale Price Index-Industrial Commodities (WPI), the Marshall and Swift Mining-Milling Equipment Cost Index (MSI), and the Chemical Engineering Plant Cost Index (CEP) are used in cost equations and are annually revised.’ Table 4 shows the resource estimates in the $50/lb forward cost category for selected U.S. uranium resource regions* for two different sets of cost indices. Larger economic index values result in increased costs to extract U,O,, thereby reducing economic resource potential. The percentage decreases in the estimate of resource potential are in the range of 0.2’%4.4%. For example, in the Columbia Plateaus the decrease in resource potential is identical in both resource categories, i.e., 4.0%. The URAD system is continually updated as more information is available from USGS assessors. Recently, the undiscovered uranium endowment estimates for collapse-breccia pipes in the Grand Canyon region of Arizona and Utah have been revised to approximately 1.3 million tons of U,O,, instead of an earlier estimate of 155,000 tons of U,O,. The revised endowment estimate is considered more realistic because additional production and exploration data on collapse-breccia deposits were made available to the USGS assessors. This revised estimate caused an increase in the esti‘These cost indices are used relative to the cost Various indices are used relative to different years (1980, 1981, 1982). ‘The U.S. uranium resource regions are classified of these U.S. resource regions are according to U.S.

indices for the years 1980, 1981, and 1982. - CEP (1980), MS1 (1980, 1982) and WPI into thirteen different regions. The definition Department of Energy (1980).

214

S. Das and R. Lee, Estimation

of potential uranium resources

mates of uranium resources in this region in the $30/lb, $50/lb, and $lOO/lb forward-cost categories by 38, 11, and 9 times, respectively, compared to the earlier estimates.

References Ahmed, S.B., 1979, Nuclear fuel and energy policy (Lexington Books, Lexington, MA). Blanchfield, D.M., 1980, Methodology for uranium resource estimates and reliability, Paper presented at Grand Junction Uranium Symposium, Grand Junction, CO. Das, S., C.E. Ford and C.T. Carter, 1988, The uranium resources assesment data (URAD) system: Operations manual, Report no. ORNL/CF-89/60 (Oak Ridge National Laboratory, Oak Ridge, TN). Energy Information Administration, 1984a, United States uranium mining and milling industry: A comprehensive review, A report to Congress by the President of the United States. Report DOE/S-O028 (U.S. Department of Energy, Washington, DC). Energy Information Administration, 1984b, Survey of United States uranium marketing activity 1983. Report DOE/EIA-0403(83) (U.S. Department of Energy, Washington, DC). Energy Information Administration, 1988a, Domestic uranium mining and milling industry: 1987 viability assessment. Report DOE/EIA-0477(87) (Department of Energy, Washington, DC). Energy Information Administration, 1988b, Uranium industry annual 1987. Report DOE/EIA047(87) (Department of Energy, Washington, DC). Finch, W.I. and R.B. McCammon, 1987, Uranium resource assessment by the geological survey: Methodology and plan to update the National Resource Base, U.S. Geological Survey Circular 994 (U.S. Department of the Interior, Washington, DC). Ford, C.E. and R.A. McLaren, 1980, Methods for obtaining distributions of uranium occurrence from estimates of geologic features. K/CSD-13 (Oak Ridge National Laboratory, Oak Ridge, TN). Golembiewski, Gary, 1987, Review of the cost factor generator algorithms, Draft report (Science Applications, Inc., Oak Ridge, TN). Harris, D.P., 1976, A survey and critique of quantitative methods for the appraisal of mineral resources. Report GJBX-12(76) (U.S. Energy Research and Development Administration, Washington, DC). Harris, D.P., 1977, Quantitative methods for the appraisal of mineral resources. Report GJBX14(77) (U.S. Energy Research and Development Administration, Washington, DC). Harris, D.P., 1984, Mineral resources appraisal, mineral endowment, resources, and potential supply: Concepts, methods and cases (Clarendon Press, Oxford). Harris, D.P. and F.P. Agterberg, 1981, The appraisal of mineral resources, Economic Geology, 75th Anniversary Volume. Harris, D.P. and F.J. Carrigan, 1981, Estimation of uranium endowment by subjective geological analysis ~ A comparison of methods and estimates for the San Juan Basin, New Mexico, Economic Geology 76, no. 5, 1032-1055. Lee, R. and C.R. Thomas, 1983, Recommendations for the analysis of nonfuel mineral supply. Mineral data analysis contract no. JO1 13112 (U.S. Bureau of Mines, Washington, DC). Mathews, G.H. et al., 1979, Preliminary recognition criteria for uranium occurrences: A field guide. Report GJBX-32(79) (U.S. Department of Energy, Washington, DC). McCammon, R.B. et al., 1986, Estimation of uranium endowment in the Westwater Canyon Member, Morrison formation, San Juan basin, New Mexico, using a data-directed numerical method, in: C.E. Turner-Peterson et al., eds., A basin analysis case study - The Morrison formation, grants uranium region, New Mexico: Studies in Geology no. 22 (American Association of Petroleum Geologists, Tulsa, OK). Mickle, D.G. and G.W. Mathews, 1978, Geologic characteristics of environments favorable for uranium deposits. Report GJBX-67(78) (U.S. Department of Energy, Washington, DC). Neff, T.L., 1984, The international uranium market (Ballinger, Cambridge, MA). Piepel, G.F., L.F. Long, R.A. McLaren and C.E. Ford, 1981, Probabilistic estimates of U.S. uranium supply. Report no. PNL-3595 (Pacific Northwest Laboratory, Richland, WA).

S. Das and R. Lee, Estimation of potential uranium resources

215

Solow, A.R. and J.M. Broadus, 1989, Loss functions in estimating offshore oil resources, Resources and Energy 11, 81-92. Szymanski, W.N., R. Mojib and R. Lee, 1989, Sensitivity analysis of mining, milling and processing cost equations in the uranium resources data system, Draft report (Oak Ridge National Laboratory, Oak Ridge, TN). Taylor, J.H. and M.D. Yokell, 1979, Yellowcake: The international uranium cartel (Pergamon, New York). U.S. Department of Energy, 1980, An assessment report on uranium in the United States of America. Report GJO-11 l(80) (U.S. Department of Energy, Washington, DC).