USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007-5039

Back to Table of Contents

Probabilistic Assessment Methodology

The methodology used for this Bay RMP area study was the three-part assessment commonly used by the U.S. Geological Survey and other entities for mineral assessments (Singer, 1993; Drew and others, 1999). In the first part of the assessment, tracts are delineated as permissive for the occurrence of specific mineral deposit types (models) based on similar geological variables (host rocks, structural setting, etc.) in the study area. In the second part, grade and tonnage models are selected that best reflect the descriptive mineral-deposit models selected and that are consistent with known deposits and occurrences in and near the area being assessed. These grade and tonnage models describe the distributions of grades (metal contents) and sizes of known examples of the deposit model, and therefore indicate the grades and tonnages likely to occur in the area being assessed.

The third part of the assessment estimates the number of undiscovered deposits of each type, consistent with the descriptive and grade and tonnage models. Mineral deposits known to exist within the tract at the time the assessment is carried out are specifically excluded from inclusion in these “undiscovered” estimates. Each estimate is made such that approximately one-half of the estimated number of undiscovered deposits have tonnages larger than the median tonnage, and such that one-half of the grades in the undiscovered deposits are greater than the median grade in the grade and tonnage model identified for the deposit type. The quantitative estimates produced here for the Bay RMP area indicate the likely number of undiscovered deposits of a median grade and median tonnage for the deposit model in question, at the 90^th, 50^th and 10^th percentile levels. Uncertainties in the estimates are accounted for by the spread in number of deposits between the 90^th and 10^th percentile (probability) levels. Estimates of the number of deposits are based on several factors, including (1) evaluation of deposit densities from well-explored areas of similar geology (Singer and others, 2001), (2) extrapolation from known deposits or the frequency of related deposit types in the region, (3) identifying geochemical anomalies of elements associated with the deposit type, (4) identifying areas of similar geologic settings or processes to known deposits in or near the region, (5) identifying geophysical signatures similar to those of the deposit model, and (6) statistical guides of the range of uncertainties in the distribution of data that can result in a given mean number of deposits (Singer, 1993, 1994; Singer and Menzie, 2005).

Following the three-part assessment, a simulation analysis was used to estimate the total mineral endowment (metal content) represented by undiscovered deposits of each type. Probability distributions for the total contained mineralized rock and metals are estimated using the USGS EMINERS software program (Duval, 2004), derived from a Monte Carlo simulator designed by Root and others (1992). The EMINERS program uses piecewise linear approximations of the number of deposits and the tonnages and grades of metals in the simulations to avoid the effects of high values due to skewed probability distributions. A random number generator is used to sample distributions similar to the grade and tonnage distributions for each deposit type and estimate of number of deposits. The simulation randomly samples the distributions 4,999 times and calculates an estimate of the contained metals expected in the undiscovered deposits.

This three-part assessment methodology yields results that include probabilistic expressions of uncertainty. To emphasize the extent of this uncertainty, results reported here include the 95^th and 5^th percentiles (probabilities) for contained metals, in addition to estimated mean values. The 95^th percentile probability indicates 19 in 20 chances, while the 5^th percentile level refers to a 1 in 20 chance that the amounts shown will be at least that large. The 95^th and 5^th percentiles are considered reasonable minimum and maximum values, and the mean is the average, or expected value.

Back to Table of Contents