Commentary and critique of “accounting for mineral resources: Issues and BEA's initial estimates”

The Bureau of Economic Analysis (BEA), U.S. Department of Commerce, in keeping with broad international efforts to green national gross domestic product (GDP) accounts, has attempted to create satellite accounts for mineral depletion. The present paper criticizes the BEA's unit rent measure of resource depletion, which is based upon Hotelling's pure theory of exhaustion. Following a thorough criticism of Hotelling's theory as a real-world model of mineral prices, practical issues related to the treatment of recycling, exploration, definition of reserves/resources, and their valuation/exploitation are introduced. For crude oil and natural gas and the nonfuel minerals, specific application problems of joint products and residuals complicate BEA's assumptions and efforts.     

Intrinsic sample methods for mineral exploration and resource assessment

We report on an objective methodology, referred to as intrinsic sample methodology, for the delineation of exploration target areas or resource areas for assessment. Important features of the methodology include (1) identification of recognition criteria for critical genetic factors, (2) synthesis of new variables from enhanced geodata, (3) estimation of logit probability models, and (4) cutting of estimated logit probabilities to delineate exploration targets or resource areas. The methodology is demonstrated on the Walker Lake quadrangle of Nevada and California.     

A comparison of state space and multiple regression for monthly forecasts: U.S. Fuel consumption

Monthly consumption forecasts for U.S. oil, natural gas, and coal are made using state space and multiple regression applied to the same data. These forecasts are compared with actual consumption for a test period. The forecasts made using state space are preferred to those made using multiple regression models for both expost and exante cases. The state space forecasts track data cycles better than do the regression forecasts. Average absolute forecast errors are less for the state space models than they are for the multiple regression models.     

Estimation of uranium endowment by subjective geological analysis—a comparison of methods and estimates for the San Juan Basin,New Mexico

This study investigates how estimates of uranium endowment made by a geologist using an appraisal system that is based upon a formalization of geoscience and decision rules compare with estimates made by informal and unconstrained intuitive processes. The motivation for this study derives from the premise that formalization of decisions would mitigate the heuristic biases and hedging that may result from the use of unconstrained intuitive processes. Estimates of the uranium endowment of the San Juan Basin of New Mexico by four methodologies are compared in this study. These methods, ranked from top to bottom by degree of decomposition (mitigating of heuristic bias)and control on hedging, are as follows Implicit 2 1.5 × 10⁶ s.t. of U₃O₈ Implicit 1 1.6 × 10⁶ s.t. of U₃O₈ NURE (1980) 2.4 × 10⁶ s.t. of U₃O₈ Appraisal system 3.9 × 10⁶ s.t. of U₃O₈ The magnitude of expected uranium endowment estimated by these methods, ranked from smallest to largest, is in this same order. With the exception of the NURE estimates, the magnitude of the variance (uncertainty)of uranium endowment, ranked from smallest to largest, also is in this same order. These results prompt the suggestion that the more decomposed and formalized the estimation procedure, the greater the expected value and the variance of uranium endowment. Equivalently, predicating U ₃ O ₈ endowment estimation strictly upon that part of the geologist's geoscience that is useful in making U ₃ O ₈ endowment estimates and upon his understanding of the region's history produced larger estimates than have previously been reported. However, this method of estimation also shows that uncertainty about the actual state of U ₃ O ₈ endowment is much greater than previously described.     

Estimation of exploration potential of a metallogenic unit by parametric modeling of the distribution of mineral occurrences when exploration is incomplete. Case study: Walker Lake quadrangle of Nevada and California

Analysis of the distribution of ore-bodies in space can be useful generally in the estimation of mineral resources or the management of exploration. This study is directed to analysis of the undiscovered potential of well known areas of mining district size, referred to as metallogenic units (MUs). The analysis employs an effort-adjusted and truncated probability model for number of occurrences within a subdivision (quadrat) of a MU and Monte Carlo sampling to generate an approximation to the probability distribution for number of occurrences and number of mines within an MU when it is totally explored. Exploration potential for Monitor, Bodie, Aurora, and Camp Douglas MUs (Walker Lake quadrangle of Nevada and California) are estimated to be 9, 4, 7, and 4 mines, respectively.     

Air quality and view degradations due to copper mining and milling: Preliminary analysis and cost estimates for Green Valley,Arizona

Green Valley, Arizona is a retirement community located near major open-pit copper mines. Mining and milling activities create dust, which degrades air quality, and mine dumps and tailings banks, which degrade the viewscape. Although the tailings banks are highly visible, this study found that consumer surplus lost from dust-polluted air is more than twice that from degradation of viewscape. Consumer surplus lost from both air quality and viewscape degradation is estimated to be between $116,000,000 and $169,000,000. These preliminary estimates are based upon two studies: hedonic property values, and a contingent valuation survey of willingness to pay.     

Mineral Favorability Mapping: A Comparison of Artificial Neural Networks,Logistic Regression,and Discriminant Analysis

A Probabilistic Neural Network (PNN) was trained to classify mineralized and nonmineralized cells using eight geological, geochemical, and geophysical variables. When applied to a second (validation) set of well-explored cells that had been excluded from the training set, the trained PNN generalized well, giving true positive percentages of 86.7 and 93.8 for the mineralized and nonmineralized cells, respectively. All artifical neural networks and statistical models were analyzed and compared by the percentages of mineralized cells and barren cells that would be retained and rejected correctly respectively, for specified cutoff probabilities for mineralization. For example, a cutoff probability for mineralization of 0.5 applied to the PNN probabilities would have retained correctly 87.66% of the mineralized cells and correctly rejected 93.25% of the barren cells of the validation set. Nonparametric discriminant analysis, based upon the Epanechnikov Kernel performed better than logistic regression or parametric discriminant analysis. Moreover, it generalized well to the validation set of well-explored cells, particularly to those cells that were mineralized. However, PNN performed better overall than nonparametric discriminant analysis in that it achieved higher percentages of correct retention and correct rejection of mineralized and barren cells, respectively. Although the generalized regression neural network (GRNN) is not designed for a binary—presence or absence of mineralization— dependent variable, it also performed well in mapping favorability by an index valued on the interval [0, 1]. However, PNN outperformed GRNN in correctly retaining mineralized cells and rejecting barren cells of the validation set.     

A Comparative Analysis of Favorability Mappings by Weights of Evidence,Probabilistic Neural Networks,Discriminant Analysis,and Logistic Regression

This study compares the performance of favorability mappings by weights of evidence (WOE), probabilistic neural networks (PNN), logistic regression (LR), and discriminant analysis (DA). Comparisons are made by an objective measure of performance that is based on statistical decision theory. The study further emphasizes out-of-sample inference, and quantifies the extent to which outcome is influenced by optimum variable discretization with classification and regression trees (CARTS).Favorability mapping methodologies are evaluated systematically across three case studies with contrasting scale and geologic information:

Econometric and learning curve estimation of U.S. potential oil supply

This study employs (1) a simple econometric model to generate a time series of drilling footage to the year 2040 and (2) learning models to estimate the oil reserve additions from that drilling, given scenarios of oil price and projected U.S. population. Reserve additions are estimated separately for the lower 48 states and Alaska regions by estimating separate drilling footage and learning models for each region. Generally, the estimates of potential supply from undiscovered oil fields and from extensions of known fields are more optimistic than recent estimates by others. For a $1989 price of about $20/barrel (bbl), which is similar to recent prices, the potential supply of oil is estimated to be approximately 60.7 billion bbl, with 95-percent confidence bounds of 54.3 and 67.1 billion bbl. For a price of $25.50/bbl, potential supply is estimated to be approximately 82 billion bbl, with 95-percent confidence bounds of 74.5 and 89.5 billion bbl. Although estimates of potential oil supply for the entire United States are more optimistic than other recent estimates, the part of that supply estimated to be forthcoming from Alaska is smaller than other recent estimates: 2.3 and 3.3 billion bbl for prices of about $20 and $25.50 per barrel, respectively. Thus, reserve additions from the lower 48 states through development drilling and through improved recovery and production technologies will become increasingly important to future U.S. oil supply.