Search

Recent decades have seen an exponential rise in the application of machine learning in geoscience. Fundamental differences distinguish geoscience data from most other data types. Geoscience datasets are typically multi-dimensional, and contain 1-D (drillholes), 2-D (maps or cross-sections), and 3-D volumetric and point data (models/voxels). Geoscience data quality is a product of its resolution and the precision of the methods used to acquire it. The dimensionality, resolution, and precision of each layer within a geoscience dataset translates to limitations in spatiality, scale and uncertainty of resulting interpretations. Historically, geoscience datasets were overlaid cartographically, to incorporate subjective, experience-driven knowledge, and variances in scale, and resolution. The nuances and limitations that underpin the reliability of automated interpretation are well understood by geoscientists, but are rarely appropriately transferred to data science. However, for true integration of geoscience data, such issues cannot be overlooked without consequence. To apply data analytics to complex geoscience data (e.g., hydrothermal mineral systems) effectively, methodologies must be used that characterise the system quantitatively, using collocated analyses, at a common scale. This paper provides research and exploration insights from an innovative district-wide, scale-integrated, geoscience data project, which analysed 1,590 samples from 23 mineral deposits and prospects across the Cloncurry District, Queensland, Australia. Ten different analytical techniques, including density, magnetic susceptibility, remanent magnetisation, anisotropy of magnetic susceptibility, radiometrics, conductivity, scanning electron microscopy (SEM)-based automated mineralogy, geochemistry, and short-wave infrared (SWIR) hyperspectral data with 561 columns of scale-integrated data (+2151 columns of SWIR). All data were collected on 2 cm x 2.5 cm sample cylinders; a scale at which the confidence in coupling of data from techniques can be high. These data are integrated by design, to eliminate the need to downscale coarser measurements via assumptions, inferences, inversions, and interpolations. This scale-consistent approach is critical to the quantitative characterisation of mineral systems and has numerous applications to mineral exploration, such as linking alteration paragenesis with structural controls and petrophysical zonation. [ABSTRACT FROM AUTHOR]

Abstract: Geological logging by mine and exploration geologists involves the application of a classification code that requires a geologist to make a subjective decision to classify a particular rock unit (e.g. rock type, alteration intensity, mineralisation content). Where multiple geologists carry out logging this approach will most likely result in inconsistencies that are difficult to monitor or assess over time. A potential solution to this problem is to have a method to determine the nature of a given sample based on a set of fixed and quantifiable parameters. Plutonic Gold Mine (Plutonic), in Western Australia, has a growing dataset of more than 150 000 portable X-ray fluorescence (pXRF) multi-element analyses from core and underground face samples, reconciled to Au fire-assay concentration, and logging codes. This study analyses and validates the consistency with which Plutonic's logging codes have been applied from 2009 to 2013. In doing so, we have developed a schema that allows objective logging by utilising elemental data from pXRF analyses to allocate a mineralisation code. A K/V ratio is used to determine the degree of alteration, while As+12Cu concentrations are used to determine how sulphide-rich a sample is. Thresholds were defined to maintain consistency with pre-existing logged dataset. While far from replacing the geologist, this method is complementary that adds significant value to a dataset. [Copyright &y& Elsevier]