Insights from the Cobar AEM Dataset

Artifacts in geophysical models can mislead exploration efforts, resulting in wasted resources and missed opportunities. In this article, we demonstrate how our 2.5D inversion software mitigates artifacts inherent in 1D inversion methods, using line L10390 from the Cobar AEM survey. This example illustrates the reliability of our approach in complex geological settings, enabling explorers to separate real geological features from inversion-induced anomalies.

Background

The Cobar AEM survey covers a geologically complex region, with steeply dipping structures and heterogeneous regolith layers. The Woorara Fault section, spanning lines L10380 to L10420, posed significant challenges for traditional 1D inversion techniques. Known for producing false anomalies—such as dipping conductors or misplaced conductive zones—1D methods often misrepresent such environments, complicating geological interpretation.

The Challenge

The primary difficulty in the Woorara Fault dataset arises from sharp conductivity contrasts at geological boundaries and topographic variations. These conditions often lead 1D inversions to create:

False Conductors: Artificially dipping anomalies at the edges of conductive slabs.
Misplaced Features: Misaligned or exaggerated geological structures.

These artifacts not only obscure real targets but can also result in costly exploratory efforts in non-prospective areas.

The above inversion result comes from the inversions produced by Geoscience Australia using the GA-LEI software.

Our Solution

Our 2.5D inversion software addresses these issues by:

Accounting for Geological Complexity: It models subsurface conductivity variations in two dimensions while incorporating the three-dimensional footprint of the AEM system.
Reducing Misfits: Using a numerical solution to Maxwell’s equations, our software minimizes misfits between observed data and the modeled response, as demonstrated in convergence GIFs.
Interactive Analysis: Users can visually validate the absence of artifacts through dynamic misfit plots and overlay comparisons with known geological features.

When applied to the L10380–10420 dataset, our inversion successfully differentiated real geological features from artifacts, resolving discrepancies present in the 1D results.

Results and Interpretation

Key improvements in our 2.5D inversion results include:

Artifact-Free Models: Dipping conductors falsely predicted by 1D inversion were absent, confirming their non-geological origin.
Enhanced Feature Resolution: The Woorara Fault’s position and associated conductive zones were accurately delineated, consistent with known geological data.

The convergence animations provided additional evidence of our model’s accuracy, demonstrating how iterative adjustments refined the solution to align with the observed data.

Note: this is very similar to the findings of the inversion contractor hired by the NSW government.

Conclusion

Line L10390 from the Cobar AEM survey illustrates how our 2.5D inversion software overcomes the limitations of 1D methods. By eliminating artifacts and producing geologically consistent models, our technology provides a critical advantage in mineral exploration. For regions like Cobar, where geological complexity demands precision, our approach delivers unmatched reliability and accuracy.

For more information about this survey and interpretation please refer to the Cobar AEM Survey interpretation report that was completed by Geoscience Australia.