Higher Degree by Research Application Portal

Accurate regolith mapping underpins mineral exploration by understanding how our landscape has evolved, revealing weathering histories, erosion and deposition processes and providing important context for near surface geochemical signatures. In addition, regolith materials themselves can be important hosts for metal enrichment. One key aspect of regolith characterisation is the ability to distinguish between transported (by wind, water or gravity) and in situ relict materials. Changes in prolonged or rapid weathering and shifts in environmental conditions will fundamentally alter interpretation of surface geochemistry.

In this study we will use a multi-geophysics approach to assessing and classifying the regolith then to compare it to the underlying geology. Similar geometric and geochemical patterns likely indicate an in situ origin, while disparate patterns may indicate transported regolith. Detecting the appropriate geometric and geochemical patterns is yet to be solved. To do this, we will use recently developed image analysis workflow focussing of textures. The workflow combines ‘Haralick features’ with a machine learning technique ‘t-SNE feature embedding (Portes et al, submitted)’ to allow us to find textural correlations between different raster datasets.

The student should have a strong background in coding together with a grounding in a natural science. The student will have access to experts at both the University of Western Australia, and the CSIRO, Australia’s National Science Agency. Placement at both institutions will encouraged.

Students with a strong background in some area of Earth Sciences combined with good programming skills is essential.

Applicants should first contact Prof Mark Jessell to discuss the project.

Prof Mark Jessell Mark.Jessell@uwa.edu.au