Linking Antarctic geological observations and geophysical data in a probabilistic space

To understand the development of the Antarctic continent, and study properties of its crust and lithosphere, we have access to mainly sparse geological observations and extensive, but low resolution, geophysical data. Early models are often based on only one or a few datasets, and interpretations can be non-unique. With a multivariate and stochastic model, we can better constrain ambiguities and depict interpretations of the Antarctic crust and lithosphere robustly and in a repeatable, shareable, way.

Recently, a number of improved geophysical datasets have been published. Data includes gravity acceleration from satellites, airborne measurements of the magnetic field, and maps of subglacial topography. Seismic models of the crust and lithosphere have also been refined, with new data and improved processing methods. Similar progress is seen in geological studies: new geological data have been acquired, and older data are reviewed and compiled. Observations from outcrops have been extended by marine core data and studies of glacial erratics to suggest properties of the subglacial terranes. Interpretations are supported by tectonic reconstructions of the East Antarctic continental margin.

We present an example of Antarctic basement/lithospheric terranes interpreted by linking geological observations with geophysical data utilising a probabilistic and multidimensional grid model. We use these domains to generate subglacial heat flow maps of the catchment area of the Aurora Basin for wider interdisciplinary use. The models are based on age-constrained crustal heat production, and we also include thermal properties from observations in adjoint Gondwanan margins, where available. We populate the domains with properties as age and provenance from geological data and use geophysics to extrapolate domain boundaries into the Antarctic interior.

The probabilistic approach illustrated in this presentation provides a robust and repeatable workflow. Our results and process are shareable with the broader community to use for interdisciplinary studies, and as a platform that will allow ongoing refinement.