b'AEGC 2023Short abstractsrelatively uniform grain-size and composition, and sit withinintegration of geoscience data such issues cannot be a background of uniform density, linear relationships existoverlooked without consequence. To apply data analytics between magnetite and iron content, density, and magneticto complex geoscience data effectively, we need to develop susceptibility, which can be integrated with geophysicalmethods that characterise mineral systems quantitatively at a modelling to infer Iron content in magnetite resources. common scale.Although simple in theory, inferring an iron resource basedHere we present insights from an innovative district-wide, on these principles is not straightforward. First, for readilyscale integrated, geoscience data project, which analysed integrable data, petrophysical properties and mineralogy1590 samples from 23 ore deposits and prospects across must be measured on identical volumes. Second, weaklythe Cloncurry District, Queensland. Ten different analytical susceptible rocks commonly hold iron in weakly-magnetictechniques, including density, magnetic susceptibility, silicates, whereas high susceptibility samples are affected byremanent magnetisation, anisotropy of magnetic susceptibility, self-demagnetisation. Therefore, one end of our theoreticallyradiometrics, conductivity, modal mineralogy from SEM-TIMA, linear function is non-linear due to the presence of Fe-silicates,geochemistry, and short-wave infrared (SWIR) hyperspectral and the other non-linear due to self-demagnetisation. Thisdata resulted in 561 columns of scale-integrated data (+2151 is overcome by incorporating geochemical data, modifyingcolumns of SWIR). All data were collected on 2x2.5 cm sized workflows to estimate true susceptibilities, and incorporatingsample cylinders, a scale at which we can be confident of the self-demagnetisation into modelling. Lastly, non-uniqueness inspatial coupling of the techniques. This data is integrated geophysical modelling is omnipresent, but can be addressedby design, eliminating the need to downscale coarser by utilising petrophysics and incorporating depth constraintsmeasurements using assumptions, inferences, inversions, from drilling and test modelling. Provided magnetic models areand interpolations. This scale consistent approach is critical within error of geological reality (particularly depth) we ensureto quantitative characterisation of mineral systems and the modelled iron resource accurately approximates the in-situhas numerous applications to mineral exploration, such as resource. linking alteration paragenesis with structural controls and The great advantage of the MagResource method is that we usepetrophysical zonation.physics to invert a magnetic field that is scalar to the resource. Whilst high resolution magnetic data is requisite, spatiallyOnshore Basin Inventories: Building foundational representative sampling is not, because our geophysical inversion is already spatially representative. Samples whichunderstanding of energy resources for Australias span the range of susceptibilities and magnetite contents arefuture.used to define the gradient of magnetic susceptibility: IronAdam Bailey, Lidena Carr, Susannah MacFarlane, Dianne content, which converts a modelled magnetic moment intoEdwards, Chris Boreham, Kamal Khider, Russell Korsch, Liuqi contained iron. By quantifying iron resources using 99% non- Wang, Emma Grosjean and Jade Andersoninvasive technology we can drastically reduce early exploration expenditure and define resources faster without the culturalGeoscience Australiaand environmental impacts.Geoscience Australias Onshore Basin Inventories programme provides a whole-of-basin inventory of geology, petroleum Integration by design: A novel approach to thesystems, exploration status and data coverage of hydrocarbon-multimodal Big Data conundrum in geoscience. prone onshore Australian basins. Volume 1 of the inventory covers the McArthur, South Nicholson, Georgina, Wiso, James Austin 1, Morgan Williams, Renee Birchall, Ben Patterson2,Amadeus, Warburton, Cooper and Galilee basins and Volume2 Jessica Stromberg1, Tina Shelton, Andreas Bjork1, Courteneyexpands this list to include the Officer, Perth and onshore Dhanaram3, Michael Gazley1 and Vladimir Lisitsin3 Canning basins. They provide a single point of reference and create a standardised national inventory of onshore basins. In 1 CSIRO Mineral Resourcesaddition to summarising the current state of knowledge within 2 Ex- CSIRO Mineral Resources 3 Geological Survey of Queensland each basin, the Onshore Basin Inventory reports identify critical science questions and key exploration uncertainties, as relevant Recent decades have seen an exponential rise in theto both conventional and unconventional hydrocarbons, application of machine learning to geoscience. Importantly,that may help inform future work programme planning and fundamental differences distinguish geoscience data fromaid in decision making for both government and industry most other data types. Geoscience datasets are spatiallyorganisations. Under Geoscience Australias Exploring for the multi-dimensional, containing 1D (drill holes), 2D (plans/ Future (EFTF) programme, six new Onshore Basin Inventory cross-sections), and 3D volumetric and point data (models/ reports will be delivered.voxels). Geoscience data quality is a product of its resolution and the precision of the methods used to acquire it. TheThese reports will be supported by selected value-add dimensionality, resolution, and precision of each layer withinproducts that aim to address identified data gaps and evolve a geoscience translate to limitations in spatiality, scaleregional understanding of basin evolution and prospectivity. and uncertainty of resulting interpretations. Historically,Petroleum system modelling is being undertaken in selected geoscience datasets were overlaid cartographically,basins to highlight the hydrocarbon potential in underexplored incorporating subjective, experience-driven knowledge, andprovinces, and seismic reprocessing and regional geochemical variances in scale, resolution. The nuances and limitationsstudies are underway to increase the impact of existing that underpin the reliability of automated interpretation aredatasets. The inventories are supported by the ongoing well understood by geoscientists but rarely appropriatelydevelopment of the nationwide source rock and fluids atlas, transferred to data-science. However, for true mathematicalaccessed through Geoscience Australias EFTF Data Discovery 81 PREVIEW FEBRUARY 2023'