Multi-stage enrichment processes for large gold-bearing ore deposits

Reviews of large gold deposits worldwide suggest that these deposits are not exceptional in their processes of formation but circumstances have combined to form a much larger deposit. Very large deposits are not far outliers but rather conform to the overall characteristics of deposits in their class worldwide. The largest deposits are unexceptional in the sense of their geology and geochemistry and are within the lognormal distribution of deposits in terms of grade and continue the power law distribution in terms of size.

However empirical observations suggest that the largest gold deposits tend to differ from smaller ones by having very complicated multistage ore paragenesis where multiple gold enrichment events and processes have been involved in the deposit generation. These observations suggest that multistage processes may even be a requirement for the formation of large deposits. In some deposits (e.g. Witwatersrand or Carlin) the different enrichment processes occur millions of years apart. In others, such as large porphyry or epithermal deposits, the different stages are much closer in time. In many deposits, particularly sedimentary-hosted deposits, early diffuse basin-wide enrichment occurs within a particular province that is then upgraded by more focused processes. The presence of this early diffuse basin-wide enrichment could explain the tendency for gold deposits to cluster into camps.

This model has important implications, as the presence or absence of multiple gold events could be used to discriminate, at the exploration and feasibility stages, between small deposits with single stage ore genesis and more complicated deposits with multistage enrichment and the potential for greater gold endowment. The exploration strategies that can be used to find new gold ores will necessarily differ depending on whether a single or multiple events are involved. For example, in the first instance, searches should concentrate on the area near a discrete source region using a simple source-pathway-trap model. In the second instance a much more probabilistic strategy should be employed, involving the identification of multiple ore systems that overlap in space.