Search-Oriented Technique for the Numerical Forecasting of Gold Placers: Evidence from the Vagran Placer District

Abstract

The field research and analysis of data from unpublished reports made it possible to identify two placer gold types within the Vagran placer cluster (northern Urals) that are indicators of the primary gold–sulfide–quartz and hypogenic-hypergenic mineralization. They are used as references for developing a digital system for predicting parameters and localizing primary placer gold sources. In the present work, the typomorphic characteristics of placer gold (grain size, roundness, sorting and fineness, as well as trace element content) are formalized as a quantitative assessment. These data were processed using the methods of multiplicative indicators, linear regression, and random regression tree algorithm, which make it possible to forecast the composition and localization of the primary mineralization with greater confidence and stability than ordinary parameters separately. Since the data needed for such an assessment are sufficient in unpublished reports on the sampling of heavy mineral concentrates, additional field and high-cost laboratory studies are optional. Analysis of the correlation matrix allowed to identify characteristic indicators for the primary gold–sulfide–quartz and hypogenic-hypergenic mineralization, as well as to propose recommendations for the prospecting and exploration of primary Au potential in the ore cluster. All three methods provide consistent results, making it possible to predict the parameters of primary Au content. The multiplicative coefficient calculation method, which yielded a more contrasting (prominent) result, seems to be the simplest and most generalized tool suitable for any set of consistent data, whereas other methods require a more in-depth analysis of input parameters. The proposed forecast estimation method promotes the efficiency and partial automatization of the determination of primary mineralization potential in regions.