Michael Roach's research while affiliated with University of Tasmania and other places

In this study, a regional model that defines the three-dimensional geometry of the subsurface geology beneath the complex, prospective northwestern Tasmania has been developed. This has been achieved using a series of potential field inversions constrained by surface geology, geological sections, seismic interpretations and a newly extended petrophysical dataset. Three major episodes of granitic magmatism are preserved in Tasmania: in the Neoproterozoic, Cambrian and Devonian. Granite bodies are hence considered important indicators of mineralization for explorers in an area of challenging vegetation, topography and cover sequences. Forward modelling and property-based inversions of the pre-existing geological model show that the previously interpreted subsurface geometry is not compatible with potential field data. Four sub-regions displayed a large discrepancy between calculated and observed data. This study redefines the subsurface geometries of these sub-regions through individual geometry inversions. The density and magnetic susceptibility ranges of units are further refined through property inversions. The modified geometry of the Devonian granites in the four sub-regions may be summarized as follows: 1) the Housetop Granite is relatively thin (≤5 km thickness), whereas 2) the Heemskirk and Meredith Granites are very thick and granite extends to a shallower depth between these bodies than previously interpreted. This region between plutons is thus a more prospective region than previously thought. 3) For the first time, an intrusive body underlying the eastern part of the Rocky Cape Group has been identified. Its petrophysical properties are similar to that of a granite, and its top is interpreted at a depth of >3 km. This interpreted low density (granitic) unit may be either Neoproterozoic or Devonian. 4) A new non-magnetic, low density Cambrian granite, with a minimum burial depth of 1 km, is also modelled in 3D, within the Mount Read Volcanics, in the south of the study area. Our approach, whereby sub-regions are identified for more detailed modelling, enables new constraints to be introduced in a computationally efficient way, and has general application to refining the geometry of key structures in prospective regions.

The Heazlewood-Luina-Waratah area is a prospective region for minerals in northwest Tasmania, Australia, associated with historically important ore deposits related to the emplacement of granite intrusions and/or ultramafic complexes. The geology of the area is poorly understood due to the difficult terrain and dense vegetation. We construct an initial high-resolution 3D geological model of this area using constraints from geological maps, and geological and geophysical cross sections. This initial model is improved upon by integrating results from 3D geometry and physical property inversion of potential field (gravity and magnetic) data, petrophysical measurements, and updated field mapping. Geometry inversion reveals that the Devonian granites in the south are thicker than previously thought, possibly connecting to deep sources of mineralization. In addition, we identified gravity anomalies to the northeast that could be caused by near-surface granite cupolas. A newly discovered ultramafic complex linking the Heazlewood and Mount Stewart Ultramafic Complexes in the southwest has also been modeled. This implies a greater volume of ultramafic material in the Cambrian successions and points to a larger obducted component than previously thought. The newly inferred granite cupolas and ultramafic complexes are targets for future mineral exploration. Petrophysical property inversion reveals a high degree of variation in these properties within the ultramafic complexes indicating a variable degree of serpentinization. Sensitivity tests suggest maximum depths of 2-3 km for the contact aureole that surrounds major granitic intrusions in the southeast, while the Heazlewood River complex is likely to have a deeper source up to 4 km. Our case study illustrates the value of adding geological and petrophysical constraints to 3D modeling for the purpose of guiding mineral exploration. This is particularly important for the refinement of geological structure in tectonically complex areas that have lithology units with contrasting magnetic and density characteristics.

The Lachlan Orogen’s mineral wealth is a direct result of tectonic processes that took place in the early Palaeozoic, but the exact nature and timing of events is widely contested. Here, we apply new methods of deforming tectonic reconstruction modelling to the area. The resulting reconstructions enable us to consistently compare alternative, previously-proposed models and test them against new and old data. This approach highlights model self-inconsistencies and incompatibilities with available data. We adopted an approach where the most valid components of individual tectonic reconstructions were combined to produce a new reconstruction model constrained by the most recent data. The new model invokes two concurrent subduction zones from the Early Cambrian to the Late Ordovician. It includes a consistent continent-dipping subduction at the Eastern Gondwanan margin, and an outboard subduction complex, which experiences multiple reversals. These are responsible for an unnamed Cambrian Arc and its obduction in Tasmania, which is part of the microcontinent VanDieland before accretion to Gondwana in the late Cambrian. The Macquarie Arc later develops in the Ordovician over the Cambrian Arc. A single continent-dipping system then resumes following the Benambran Orogeny, when oroclinal folding occurs across south-eastern Australia followed by east-west shortening of the Tabberabberan Orogeny.

Mineral exploration and geological mapping of highly prospective areas in western Tasmania, southern Australia, is challenging due to steep topography, dense vegetation, and limited outcrop. Synthetic aperture radar (SAR) can potentially penetrate vegetation canopies and assist geological mapping in this environment. This study applies manual and automated lithological classification methods to airborne polarimetric TopSAR and geophysical data in the Heazlewood region, western Tasmania. Major discrepancies between classification results and the existing geological map generated fieldwork targets that led to the discovery of previously unmapped rock units. Manual analysis of radar image texture was essential for the identification of lithological boundaries. Automated pixel-based classification of radar data using Random Forests achieved poor results despite the inclusion of textural information derived from gray level co-occurrence matrices. This is because the majority of manually identified features within the radar imagery result from geobotanical and geomorphological relationships, rather than direct imaging of surficial lithological variations. Inconsistent relationships between geology and vegetation or geology and topography limit the reliability of TopSAR interpretations for geological mapping in this environment. However, Random Forest classifications, based on geophysical data and validated against manual interpretations, were accurate (∼90%) even when using limited training data (∼0.15% of total data). These classifications identified a previously unmapped region of mafic–ultramafic rocks, the presence of which was verified through fieldwork. This study validates the application of machine learning for geological mapping in remote and inaccessible localities but also highlights the limitations of SAR data in thickly vegetated terrain.

Tasmania in southeast Australia is underlain by basement rocks dating from the Precambrian to the Cretaceous. One important event in the tectonic evolution of Tasmania is the intrusion of Devonian Granites which has resulted in metamorphism and mineralization in some regions. We construct a regional-detailed 3D model for northwest Tasmania including major geological units and estimate density values of each unit to investigate the tectonic setting and geometry of Devonian Granites. Our model contains 20 units with different density properties. The residual gravity data have been used to model the properties of these units and consequently the geometry of Devonian Granites. The new inversion refines the geometry of the Devonian Granites with respect to initial models. A new granite intrusion is also revealed by this study which is the subject of further investigation. Presentation Date: Wednesday, October 19, 2016 Start Time: 8:50:00 AM Location: Lobby D/C Presentation Type: POSTER

Geophysical and remote sensing methods are routinely used to supplement geological observations. They are also of value where geological observations are not available due to a thick regolith profile or where ground access is restricted. Dealing with these large, disparate datasets poses an increasing challenge and opportunity, to the mineral exploration industry. Machine learning algorithms present an efficient, data-driven means of adding value to these data through semi-automated lithological mapping. Recent research has demonstrated that the combination of the Random Forests™ and Self Organizing Maps classification methods significantly improved lithological mapping in a VHMS setting. This study applies these algorithms to the orogenic gold setting, using an exemplar case study located in Australia's Eastern Gold Fields. The study area is characterized by Archean stratigraphy overlain by regolith comprising up to 80 m of in situ saprolite and Tertiary transported cover. We address the capacity of these algorithms to improve lithological mapping in the study area and which data are required for the best results. We examine the value of implementing this method at various stages of a projects life cycle. Using Random Forests and Self Organizing maps, we are able to produce a data-driven lithological map of the study area. This has been achieved using a desktop computer and an interoperable combination of industry standard and freely available open source software. Quantifying uncertainty in the classification allows us to better identify areas where the available data adequately facilitates mapping and targeting; and those areas of high complexity where further exploration effort is required.

The continental crust of southeast Australia is a complex and highly prospective area. Southeast Australia comprises the Delamerian and Lachlan Orogenies which, together with the Eastern Tasmania Terrane, are understood to have Phanerozoic basement. In contrast, the Western Tasmanian Terrane comprises areas of exposed Neoproterozoic basement which were assembled along the proto-Pacific margin of East Gondwana. In this study, the crustal structure across southeast Australia and Tasmania is considered using seismic and aeromagnetic methods. We use previous passive seismic results and present a new analysis of magnetic data. The Curie temperature, the temperature at which magnetic rocks lose their magnetisation, is investigated using spectral analysis of aeromagnetic data and the Curie point depth (CPD) is consequently determined. CPD is compared to the depth of the seismic Moho discontinuity throughout the study area. The Moho depth and newly calculated CPD throughout the study area vary from ~20 to >38 km and ~25 to >45 km, respectively. The CPD is slightly shallower than the Moho across the study area. The Delamerian and Lachlan Orogenies are underlain by a 30-35 km and ~40- 50 km deep Moho respectively, while average CPD depths are ~30 and ~28 km for these regions. A relatively shallow CPD is observed in the northeast of the study area and corresponds to Cainozoic volcanism in eastern Australia. The shallow Moho beneath Tasmania supports the idea of crustal thinning during Gondwana breakup. In Tasmania, CPD increases in depth from ~21 km in the northwest to >31 km in north. This is consistent with variations in the depth of the Moho from 25 km in the northwest to 37 km in the north.

A high-resolution digital elevation model (DEM), generated from airborne light detection and ranging (LiDAR) remote sensing data, is used here to estimate the 3-D orientation of bedding planes. Methods for enhancement, manual identification and extraction of lineaments, and estimation of best fit planes representing bedding are presented and evaluated for a study area in folded metasedimentary rocks in northeast Tasmania, Australia. Estimated bedding plane dip directions are shown to be accurate and reliable when compared with field-based observations. The same cannot be said for dip angle estimates. It is likely that small errors in the location of a manually digitized lineament will affect dip estimation more than dip direction estimation, particularly for steeply dipping structures. Fold axis orientations calculated from the stereographic analysis of estimated bedding closely correspond to orientations determined from field data. The mean absolute differences $pm$ standard error for 12 of the 14 regularly spaced domains located within the study area were $8.7^{circ} pm 1.2^{circ}$ for the fold plunge and $4.9^{circ} pm 0.9^{circ}$ for the fold trend. The techniques described here for the extraction of bedding plane orientations from high-resolution DEMs complement field-based geological mapping and can assist structural interpretations.