Assessing the variability of expert estimates in the USGS Three-part Undiscovered Mineral Resource Assessment methodology: A call for increased skill diversity and scenario-based training

Future mineral exploration will necessarily be conducted in increasingly challenging and uncertain search spaces as near-surface, high-quality ore deposits are progressively depleted. Faced with this increase in task complexity, an important consideration from an exploration management perspective is the behavioural aspect of information interpretation and decision-making.

One such challenging search space is the Sandstone Greenstone Belt, Western Australia, covering an area of approximately 920 km², that is deemed prospective for the discovery of archean orogenic gold deposits, with mined (historic production), inferred and indicated resources (JORC 2004 and 2012) totalling 54 t Au. Gold endowment estimates made by geoscience experts, during an exploration project evaluation workshop, were compared with estimates from a group of non-geoscientists, made during a separate but identical workshop. Significant differences were identified between the estimates of the expert geoscientists and the non-geoscience expert group, with the latter proving more conservative. However, a portion of the geoscience experts (N = 11) group produced conservative estimates, comparable to a non-geoscience expert group (N = 10), with both suggesting the existence of additional gold deposits of similar size and quality to known resources (with group estimates for median total endowment of 99 t and 120 t Au, respectively). The remaining geoscience experts (N = 11) presented significantly more optimistic, albeit inconsistent, estimates for the gold endowment of the project area, predicting the existence of undiscovered deposits significantly larger than those already defined in the belt (with a group estimate for median total endowment of 350 t Au).

Although the true undiscovered gold endowment within the project area remains uncertain, several possible factors can explain the variations in estimates. These include the application of contrasting strategies, with participants opting to apply more empirical or conceptual methods, and to differences in background experience, resulting in distinct skillsets and varying ability to estimate uncertainty. To improve the quality of expert estimates, it is suggested that individual expertise and appropriate assessment strategies can be developed through scenario-based training courses, and that greater skill and experience diversity within exploration teams is desirable, leading to more balanced aggregate estimates. Further research is warranted to determine which, if any, of the proposed factors account for these disparities. This research could be used to adapt the composition of exploration teams and develop training programs to promote the development of expertise in predictive exploration targeting, in order to promote discovery of future mineral resources.