Innovation

In order to meet the challenge of de-risking exploration undercover, new tools, platforms and workflows are required. Accordingly, multiple collaborations have been set up with world-leading researchers at Australian universities in order to maximise the potential and longevity of the datasets being acquired as part of the Exploring for the Future program. Key collaborations include:

  • Inversion and Data Analytics in Solid Earth Geophysics with Prof. Malcolm Sambridge at the Research School of Earth Sciences, Australian National University. This collaboration is focused on the implementation of new computer algorithms to produce more robust models of subsurface architecture from geophysical datasets such as AusAEM and AusArray. A central component of this project is the generation of robust uncertainty estimates for these models in order to clearly inform explorers of the robustness of subsurface estimates.
  • Lithosphere Imaging with Associate Prof. Juan Carlos Afonso at the Department of Earth and Planetary Sciences, Macquarie University. This collaboration builds on proprietary research computer codes to quantitatively integrate disparate geophysical datasets such as AusArray, AusLAMP and the South Nicholson Gravity Survey to map deep mineral distribution beneath the focus area between Tennant Creek and Mount Isa.
  • Deep learning study in collaboration with CSIRO Data61. Over the last two years Geoscience Australia and Data61 have developed the UncoverML machine-learning platform capable of analysing national datasets at 90 m cell sizes utilizing supercomputer facilities such as the National Computational Infrastructure (NCI). This collaboration is examining the utility of integrating deep learning algorithms with UncoverML for the purpose of predictive mapping of cover thickness and character.
  • National seamless coverage of Sentinel-2 satellite data and a Landsat TM data bare earth model with Dr. Dale Roberts at the College of Business and Economics, Australian National University.
  • Data enchanced Virtual Laboratories (DeVL) projects funded by NCRIS. The aim of this project is to unlock geophysical datasets for big data and high-performance computing by standardising and transforming datasets such as AusLAMP to be suitable in this high-performance environment. Geoscience Australia’s participation in this and other related projects ensures datasets collected as part of the Exploring for the Future program are in step with global data trends and demands.

These innovations require large amounts of research and testing; the products and codes derived from these activities will be available from mid to late 2018.