Between Earth’s crust and core lies the mantle, a 2,900 km thick convecting layer of hot rock that is the engine driving our dynamic planet: it is the principal control on Earth’s thermal and chemical evolution and is, ultimately, responsible for almost all large-scale tectonic and geological activity. Despite this significance, we have little knowledge of the past structure and flow history of Earth’s mantle.

G-ADOPT is focused around the adoption, development and support of a computational modelling infrastructure for inverse geodynamics. It builds on several recent breakthroughs including a surge in accessible observational datasets; advances in inversion methods, using sophisticated adjoint techniques, that provide a mechanism for fusing these observations with dynamics, physics and chemistry; and two novel software libraries, Firedrake and dolfin-adjoint. When combined, these libraries provide a state-of-the-art finite element platform that offers a radical new approach for rigorously integrating data from fields such as tectonics, stratigraphy, geochronology, geochemistry, seismology and geodesy, with multi-resolution, time-dependent, geodynamical models, through high-performance computing.

This platform will enable robust reconstructions of the history of mantle convection and its impact at Earth’s surface, addressing a fundamental challenge central to the Earth sciences. G-ADOPT will facilitate the generation of unique 4-D datasets of Earth’s evolution, which will be of great value across the geoscientific community, with traceable provenance of input data and model configuration in full compliance with FAIR principles.

G-ADOPT will provide access to and support with next-generation finite element software libraries that rigorously integrate data, from fields such as tectonics, stratigraphy,geochronology, geochemistry, seismology and geodesy, with geodynamical models. The framework developed will be directly transferable to other disciplines.

Start date 16 February 2021
Expected completion date 30 June 2023
Investment by ARDC $767,567
Lead node
1 Community building and training
Workshops will be held with international partners, users and other stakeholders throughout the project to coordinate software development plans with international partners; refine data interfaces in conjunction with domain experts; train users; identify other applicable research areas; and finally engage the broader community and plan for the future.
2 High-performance Computational Modelling Platform for Inversion
The G-ADOPT platform will use HPC to provide access to state-of-the-art finite element software libraries that rigorously integrate geoscientific data with multi-resolution, time-dependent geodynamical models. Firedrake and dolfin-adjoint will be ported to national supercomputing facilities at NCI, then validated and verified against a suite of analytical and benchmark solutions and via synthetic experiments, with systematically increasing complexity. Functionality will be enhanced, as required. The efficiency of optimisation algorithms will be improved to reduce computational cost and extend the applicability of the platform.
3 Tools for Rapid Integration of Geoscientific Data
The necessary interfaces will be developed for integrating geoscience data with adjoint models. Data includes geophysical data from IRIS, ICGEM, GA's EFTF program and the ARDC cross-NCRIS funded National High-Resolution Geophysics Reference Collection, which constrain the present-day structure and flow-regime of Earth’s interior; plate tectonic reconstructions through GPlates/PyGplates, which reveal how Earth’s plates and their boundaries have evolved over geological time; and EarthChem and other community geochemical and geochronological databases (e.g. AusGeochem), which reveal critical aspects of the mantle to great depths and in deep time.
4 Extension to other fields of research
The applicability of the framework will be extended to other fields, through a series of proof-of-concept examples and associated cookbooks.
5 4-D datasets of Earth’s evolution
4-D datasets of Earth’s evolution that are generated during the project will be made FAIR.

Core features

G-ADOPT will drive a new class of interdisciplinary research and position Australian researchers at the frontier of this emerging discipline. The framework developed will be directly transferable to other disciplines.
Interdisciplinary collaboration
G-ADOPT will facilitate national and international collaboration, connecting researchers with data to those with models and driving a new class of interdisciplinary research.

Who is this project for?

  • Researchers
  • Research organisations
  • Infrastructure providers
  • Government (state and commonwealth)
  • Geophysicists
  • Data analysts

What does this project enable?

G-ADOPT will create a new paradigm for understanding Earth’s evolution. It will be central to ongoing efforts to reveal the geodynamic environments underpinning mineralisation, aligning with Australia’s research priority on Resources. G-ADOPT will convert the solid Earth sciences community from idealised forward models to data-driven simulation, facilitating the use of many geoscientific datasets in a geodynamical context, for the first time.

Handy resources

The University of SydneyVisit
Geoscience AustraliaVisit
Imperial College LondonVisit
Oxford UniversityVisit
University of MunichVisit