Skip to content

3D magnetotelluric-derived electrical resistivity models of the Vulcan IOCG prospect #42

@bvkay

Description

@bvkay

-> submitter ORCID (or name)

0000-0002-9738-7277

-> slug

Kay-2024-Vulcan

-> license

CC-BY-4.0

-> alternative license URL

No response

-> model category

inverse model, model published in study

-> model status

completed

-> associated publication DOI

https://doi.org/10.1080/08123985.2024.2378132

-> model creators

0000-0002-9738-7277
0000-0001-7106-0789
0000-0002-7743-7812

-> title

3D magnetotelluric-derived electrical resistivity models of the Vulcan IOCG prospect under thick cover, South Australia

-> description

Three-dimensional electrical resistivity models derived from magnetotelluric (MT) data across the Vulcan iron-oxide copper–gold (IOCG) prospect in South Australia. The models image conductive sedimentary cover, a low-resistivity haematite breccia zone, and a deeper vertical conductive structure interpreted as a fluid pathway. Two inversion variants are provided: a smooth model and a structurally constrained “tear” model incorporating a known basement interface. These models provide constraints on subsurface structure and fluid pathways relevant to mineral exploration beneath thick conductive cover.

-> abstract

This dataset contains 3D electrical resistivity models derived from magnetotelluric (MT) data collected over the Vulcan iron-oxide copper–gold (IOCG) prospect, located ~30 km northeast of Olympic Dam, South Australia. The survey comprises 100 broadband MT sites on a 1 km grid across a 9 × 9 km area.

Inversion of MT responses resolves three primary domains: (1) conductive sedimentary cover (~850 m thick), (2) a low-resistivity zone associated with haematite breccia beneath the cover, and (3) a vertically extensive conductive structure extending to several kilometres depth, interpreted as a pathway for mineralising fluids.

Two model variants are included: a smooth inversion and a model incorporating a structural discontinuity (“tear”) at the base of the sedimentary cover, based on drillhole constraints. The models provide insight into lithospheric-scale fluid pathways and the geometry of IOCG mineral systems under conductive cover sequences.

-> scientific keywords

magnetotellurics, electrical resistivity, IOCG

-> funder

Geological Survey of South Australia, Accelerated Discovery Initiative (ADI) project RD02-260.
FMG Resources Pty Ltd, Accelerated Discovery Initiative (ADI) project RD02-260.

-> model embargo?

No response

-> include model code ?

  • yes

-> model code/inputs DOI

No response

-> model code/inputs notes

The models were generated using a commercial 3D magnetotelluric inversion algorithm (CGG RLM-3D). While the inversion software itself cannot be distributed, all relevant model inputs required to reproduce or reinterpret the results are provided.

Key inversion settings include:

  • Non-linear conjugate gradient optimisation with L-BFGS preconditioning
  • Frequency range: 0.01–100 Hz (4 points per decade)
  • Full impedance tensor inversion with per-site galvanic distortion estimation
  • Resistivity bounds constrained between 0.5–5000 Ωm
  • Smoothness-constrained inversion using a hybrid Laplacian regularisation

Full parameter files and input datasets are included, with:

-> include model output data?

  • yes

-> data creators

https://orcid.org/0000-0002-9738-7277
https://orcid.org/0000-0001-7106-0789
https://orcid.org/0000-0002-7743-7812

-> model output data DOI

No response

-> model output data notes

The output dataset includes both model results and predicted data products for two inversion variants:

Smooth model (uniform regularisation):

Tear model (structurally constrained at the base of sedimentary cover):

Predicted magnetotelluric responses are also provided as EDI files, enabling direct comparison between observed and modelled impedance tensors.

-> model output data size

23 Mb

-> software framework DOI/URI

No response

-> software framework source repository

No response

-> name of primary software framework (e.g. Underworld, ASPECT, Badlands, OpenFOAM)

CGG RLM-3D

-> software framework authors

No response

-> software & algorithm keywords

CGG RLM-3D, Viridien, NLCG, L-BFGS

-> computer URI/DOI

No response

-> add landing page image and caption

Image 3D perspective views of the smooth (top) and tear (bottom) resistivity models, highlighting differences in basement structure. Clipped resistivity values (30–60 Ωm) emphasise the vertical conductive feature and haematite breccia zone; the tear model better resolves structure at the top of basement consistent with drillhole and gravity constraints.

-> add an animation (if relevant)

No response

-> add a graphic abstract figure (if relevant)

Image Comparison of smooth and tear-constrained 3D resistivity models showing improved resolution of basement structures and strong spatial agreement with gravity and drillhole constraints, highlighting a conductive pathway linked to the mineral system.

-> add a model setup figure (if relevant)

Image Inversion mesh used for 3D MT modelling, including padding cells and a structural “tear” boundary at the base of the sedimentary cover. MT site locations are shown as black triangles; drillhole logs provide constraints on the tear depth. Mesh projection UTM Zone 54S.

-> add a description of your model setup

The 3D resistivity models were generated using a non-linear conjugate gradient (NLCG) inversion scheme with L-BFGS preconditioning, designed to produce smooth, data-consistent models. A maximum of 50 inversion iterations was applied, with adjoint updates performed periodically during optimisation.

The inversion utilised magnetotelluric impedance tensor data over a frequency range of 0.01-100 Hz, sampled at approximately four points per decade. All four complex impedance components were inverted, with simultaneous estimation of frequency-independent 2×2 galvanic distortion matrices at each site. A strong regularisation weight (1000) was applied to keep distortion matrices close to the identity.

Forward modelling and sensitivity calculations were controlled with solver tolerances of 1×10⁻⁶ and 1×10⁻⁵, respectively, with a maximum of 1000 iterations per linear solve. A 1D adjoint approximation was used to improve computational efficiency.

Model resistivity was bounded between 0.5 and 5000 Ωm. Regularisation followed a smoothest-model approach using a hybrid Laplacian formulation, with anisotropic smoothing applied via vertical (τ = 0.006) and horizontal (τ = 2) weighting parameters. Additional near-surface smoothing was imposed to a depth of 500 m to suppress inversion artefacts and reflect the expected lateral continuity of sedimentary cover.

Please provide any feedback on the model submission process?

Some file formats couldn't be uploaded, hence the zipped files.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions