<div>A keynote talk talk given at Uncover Curnamona 2022 by Angela O'Rourke outlining the rationale, work program and new data acquisition for Geoscience Australia's Darling-Curnamona-Delamerian Project within Exploring for the Future</div> This presentation was given to the 2022 Uncover Curnamona 2022 Conference 31 May - 2 June:<br>(https://www.gsa.org.au/common/Uploaded%20files/Events/Uncover%20Curnamona%202021/UC2022_short_program_A4_web%20(003).pdf)

Building on newly acquired airborne electromagnetic and seismic reflection data during the Exploring for the Future (EFTF) program, Geoscience Australia (GA) generated a cover model across the Northern Territory and Queensland, in the Tennant Creek – Mount Isa (TISA) area (Figure 1; between 13.5 and 24.5⁰ S of latitude and 131.5 and 145⁰ E of longitude) (Bonnardot et al., 2020). The cover model provides depth estimates to chronostratigraphic layers, including: Base Cenozoic, Base Mesozoic, Base Paleozoic and Base Neoproterozoic. The depth estimates are based on the interpretation, compilation and integration of borehole, solid geology, reflection seismic, and airborne electromagnetic data, as well as depth to magnetic source estimates. These depth estimates in metres below the surface (relative to the Australian Height Datum) are consistently stored as points in the Estimates of Geophysical and Geological Surfaces (EGGS) database (Matthews et al., 2020). The data points compiled in this data package were extracted from the EGGS database. Preferred depth estimates were selected to ensure regional data consistency and aid the gridding. Two sets of cover depth surfaces (Bonnardot et al., 2020) were generated using different approaches to map megasequence boundaries associated with the Era unconformities: 1) Standard interpolation using a minimum-curvature gridding algorithm that provides minimum misfit where data points exist, and 2) Machine learning approach (Uncover-ML, Wilford et al., 2020) that allows to learn about relationships between datasets and therefore can provide better depth estimates in areas of sparse data points distribution and assess uncertainties. This data package includes the depth estimates data points compiled and used for gridding each surface, for the Base Cenozoic, Base Mesozoic, Base Paleozoic and Base Neoproterozoic (Figure 1). To provide indicative trends between the depth data points, regional interpolated depth surface grids are also provided for the Base Cenozoic, Base Mesozoic, Base Paleozoic and Base Neoproterozoic. The grids were generated with a standard interpolation algorithm, i.e. minimum-curvature interpolation method. Refined gridding method will be necessary to take into account uncertainties between the various datasets and variable distances between the points. These surfaces provide a framework to assess the depth and possible spatial extent of resources, including basin-hosted mineral resources, basement-hosted mineral resources, hydrocarbons and groundwater, as well as an input to economic models of the viability of potential resource development.

To meet the increasing demand for natural resources globally, industry faces the challenge of exploring new frontier areas that lie deeper undercover. Here, we present an approach to, and initial results of, modelling the depth of four key chronostratigraphic packages that obscure or host mineral, energy and groundwater resources. Our models are underpinned by the compilation and integration of ~200 000 estimates of the depth of these interfaces. Estimates are derived from interpretations of newly acquired airborne electromagnetic and seismic reflection data, along with boreholes, surface and solid geology, and depth to magnetic source investigations. Our curated estimates are stored in a consistent subsurface data repository. We use interpolation and machine learning algorithms to predict the distribution of these four packages away from the control points. Specifically, we focus on modelling the distribution of the base of Cenozoic-, Mesozoic-, Paleozoic- and Neoproterozoic-age stratigraphic units across an area of ~1.5 million km2 spanning the Queensland and Northern Territory border. Our repeatable and updatable approach to mapping these surfaces, together with the underlying datasets and resulting models, provides a semi-national geometric framework for resource assessment and exploration. <b>Citation:</b> Bonnardot, M.-A., Wilford, J., Rollet, N., Moushall, B., Czarnota, K., Wong, S.C.T. and Nicoll, M.G., 2020. Mapping the cover in northern Australia: towards a unified national 3D geological model. In: Czarnota, K., Roach, I., Abbott, S., Haynes, M., Kositcin, N., Ray, A. and Slatter, E. (eds.) Exploring for the Future: Extended Abstracts, Geoscience Australia, Canberra, 1–4.

The world is turning to the minerals sector to meet sustainable development goals on the path to net zero emissions, buoyed by modern manufacturing. Discovery and development of new and varied mineral deposits is essential to reach these goals. However, despite concerted efforts, exploration success rates are in decline globally. To provide an advantage to Australia’s mineral sector, the Australian Government has significantly invested in precompetitive geoscience to unlock both geographic and conceptual frontiers for further exploration and discovery by private industry. Over the last decade, Geoscience Australia, in collaboration with state/territory geological surveys and academia, has undertaken geoscience data acquisition and analysis at an unprecedented scale aligned with UNCOVER initiative through programs like Exploring for the Future. This strategic move has reversed Australia’s declining market share of global exploration investment, stimulated new minerals industries, led to the discovery of world-class mineral deposits, and opened new undercover provinces for exploration. Here, I highlight some key successes, consider some key challenges, and suggest a future direction for precompetitive geoscience. Australia is at the forefront of mineral systems science underpinned by world-leading standardised national geological and geophysical (i.e. potential field) data coverages. Acquired at increasing resolution over decades, they have been at the vanguard of mineral exploration as they effectively map lateral geological changes yet provide limited and non-unique insights with depth. Recognising mineral deposits are the consequence of large geological systems, a critical step change in the last decade has been a focus on extensive first-pass or framework 3D imaging of the Australian continent through the systematic collection of magnetotelluric (AusLAMP), passive seismic (AusArray) and airborne electromagnetic (AusAEM) data, supplemented by higher fidelity deep reflection seismic profiles. Aided by significant advances in geophysical processing, Bayesian inference and big data analytics, when integrated with classic geoscience these datasets are revealing new first-order controls on mineralisation and identifying new exploration opportunities. Examples include discovery of lithospheric thickness controls on sediment-hosted base-metal deposits, clear scale reduction approaches to targeting iron oxide-copper-gold systems using electrical methods and mapping source rocks of hydrothermal systems. Using statistical modelling, the predictive power of each dataset or derivative can be assessed allowing an unbiased national view of Australia’s mineral potential to emerge. Importantly, these advances are coupled with recommencement of stratigraphic drilling programs to test inference and demonstrably reduce risk of exploring in frontier regions. Systematic quantitative mineral potential analysis rapidly highlights the importance of data consistency, completeness, and the robustness of validation datasets and in so doing reaffirms the critical role geological surveys play as custodians of this information. The diversification of mineral demand to include critical minerals has both leveraged this information to identify new types of mineral deposits but also highlights the youthfulness of mineral systems science. In response there are growing international efforts to grow understanding of minerals systems science for all elements to enable exploration for critical minerals and realise secondary prospectivity of mine waste. The wave of 3D imaging of Australia is developing a framework 3D digital twin and national scale mineral potential models are emerging. The challenge for precompetitive geoscience is to strategically infill this coverage to further accelerate exploration and development by industry. However, given competing land use claims and increasing environmental, social and governance (ESG) requirements on the minerals sector, success requires a common understanding of subsurface geology across minerals, energy and groundwater industries, which dovetails with surficial, social and governance datasets. Delivery of such integrated subsurface understanding is an exciting and vital challenge for geological surveys and their collaborators.