Sub-solidus replacement of rapakivi textures during high-temperature potassium metasomatism of the Mannum Granite

Forster, Marnie1,2, Goswami, Naina1,2, Lister, Gordon1, Reid, Anthony3

1Research School of Earth Sciences, Australian National University, Canberra, Australia, 2MinEx CRC, Canberra, Australia, 3Geological Survey of South Australia, Adelaide, Australia

The Mannum Granite is a high-temperature A-type granitoid emplaced during the late stages of the Delamerian Orogeny in South Australia. It has attracted considerable research interest over past decades because it is a porphyritic A-type granite that displays well-developed rapakivi structures and has been linked to the process of magma-mingling above deep-seated mafic igneous bodies. It contains many mafic enclaves of varied sizes and textures including potassium feldspar and quartz porphyroblasts with disseminated sulphides. The alkali feldspar cores of the rapakivi textures are lobate and overgrow the plagioclase rims, suggesting alkali feldspars are younger.

We conducted 40Ar/39Ar geochronology in conjunction with ultra-high-vacuum (UHV) step heating 39Ar diffusion experiments on K-feldspar and biotite from the Mannum granite. Conjoint inversion of data from the UHV 39Ar diffusion experiments using the Wunderkind program applied to data from the K-feldspar experiment demonstrates the presence of highly retentive core domains capable of retaining radiogenic argon even to temperatures in excess of 600°C. Inversion of the geochronology data using the MacArgon program produces a temperature-time history that suggests that the K-feldspar in the core of the rapakivi texture formed as the result of sub-solidus solid state replacement and/or metasomatism at ~443 Ma, and then cooled rapidly at ~440 Ma. Since biotite spectra which define a plateau age at ~472 Ma are relatively undisturbed, the temperature excursion associated with the metasomatic event would have to have been limited in its magnitude (<~500°C) and duration (<< 1 Ma). 

The biotite age (~473 Ma) is close to the previously determined early Ordovician crystallisation ages for the granites in this belt. In addition, the biotite appears to be highly retentive, with closure temperature for cooling at 20°C/Ma determined as the result of the UHV 39Ar diffusion experiment at ~485°C. These are thus exceptionally retentive biotites, which correlate with the fact that they are iron-rich and fluorine-rich, and fluorine-rich biotite typically exhibits higher argon retentivity. Overall this short/sharp thermal pulse may have occurred in consequence of fluid movement associated with the Old Teal Flat Shear Zone to the east, which contains fabrics dated in this study at ~445 Ma.

Microstructural analysis shows sericite and calcite precipitated along grain boundaries, fractures and exsolution lamellae. The presence of interpenetrating grain boundaries, sericite, calcite and secondary K-feldspar is attributed to the infiltration of an acidic hydrothermal agent hydrolysing primary K-feldspar to produce secondary sericite and quartz, further reacting with Ca-rich plagioclase and quartz to produce secondary K-feldspar phase(s) and calcite. Microstructures thus confirm the presence of a potassium-rich metasomatic event at Mannum leading to fluid-assisted in-situ replacement of K-feldspars with concomitant precipitation of secondary mineral phases.


The work has been supported by the Mineral Exploration Cooperative Research Centre whose activities are funded by the Australian Government’s Cooperative Research Centre Program. This is MinEx CRC presentation.


Naina Goswami is a second year PhD candidate at The Australian National University (ANU). She finished her Masters (advanced) from ANU in 2018 and her B.Sc (Hons) Chemistry from University of Delhi 2016. She specialises in 40Ar/39Ar geochronology and geochemistry. Currently she is undertaking her PhD at ANU working on a MinEx-CRC project based in South Australia.

Timescale of events around the Cretaceous-Paleogene Boundary: Links between the Chicxulub impact, Deccan volcanism, and the Cretaceous-Paleogene mass extinction

Sprain, Courtney J.1, Renne, Paul R.2,3, Clemens, William A.3, Wilson, Gregory P.4, Self, Steve3, Vanderkluysen, Loyc5, Pande, Kanchan6, Fendley, Isabel7, and Mittal, Tushar8 

1University of Florida, Gainesville, United States, 2Berkeley Geochronology Center, Berkeley, USA, 3University of California-Berkeley, Berkeley, USA, 4University of Washington, Seattle, USA, 5Drexel University, Philadelphia, USA, 6Indian Institute of Technology, Mumbai, India, 7University of Oxford, Oxford, United Kingdom, 8Massachusetts Institute of Technology, Cambridge, USA

The Cretaceous-Paleogene boundary (KPB) mass extinction is one of the most important biotic turnover events in Earth history. This event is important to study for several reasons, the most relevant being its implications on our understanding of the effects of abrupt climate change. Although the temporal coincidence between the Chicxulub crater and the KPB has strongly implicated the impact as the main player in the mass extinction, the eruption of the Deccan Traps (DT) cannot be dismissed as a possible contributor. The timing of DT eruptions spans the KPB and, furthermore, the onset of DT volcanism roughly coincides with Late Cretaceous records of environmental change. Both the Chicxulub impact and DT volcanism have similar environmental forcing mechanisms, albeit acting on different timescales. Until recently, insufficient geochronology has made it difficult to tease apart effects from either agent. 

To better understand the effects of both the Chicxulub impact and the DT in the KPB crises, we developed a high-precision chronologic framework that outlines the temporal sequence of biotic and climatic changes, and proposed perturbations, around the KPB using 40Ar/39Ar geochronology and paleomagnetism. This work was primarily conducted in two areas: the Hell Creek region of NE Montana, USA and the Deccan Traps, India. The Hell Creek region is one of the best-studied terrestrial KPB sites in the world. We developed a high-precision chronostratigraphic framework for fluvial sediments within the Hell Creek, using 40Ar/39Ar dating, magnetostratigraphy, and chemical fingerprinting. This work constrained the timing of terrestrial faunal decline and recovery while calibrating North American Land Mammal Ages biostratigraphy. The coupling of our magnetostratigraphic sections and high-precision 40Ar/39Ar ages further allowed for calibration of the circum-KPB polarity chron (C29r) at unprecedented precision, enabling correlation of our record to other KPB records around the globe. To better understand the role of the DT in the KPB extinction, we developed high precision 40Ar/39Ar ages for >20 lavas ranging the entire DT stratigraphy.

Tying all of this work together, we are able to determine: 1) the decline in terrestrial faunas began between 400 ka and 150 ka pre-KPB, 2) terrestrial disaster faunas are constrained to the first ~25 ka of the Paleogene, and recovery occurred gradually over the next 850 ka, 3) over 90% of the DT volume was erupted in < 1 Ma, with 50-75% emplaced post-KPB, 4) the onset of volcanism is approximately coincident with the onset of pre-KPB warming, but despite this 5) pre-KPB records of climate change coincide temporally with the eruption of the smallest DT phases, suggesting that if the DT are the source of pre-KPB climate change, the release of climate-modifying gases cannot be directly related to eruptive volume as previously assumed. Overall, our new work highlights the close temporal relationship between the Chicxulub impact, Deccan volcanism, and the KPB. But more work is needed, specifically addressing Deccan volatile release and eruption tempo, in order to fully understand the impact of the DT on the Earth system and its role in the mass extinction.


Courtney Sprain is an Assistant Professor at the University of Florida. She received her Ph.D. in Earth and Planetary Science in 2017 from the University of California, Berkeley and her B.S. degrees in Geology and Geophysics from the University of Minnesota in 2012. She specializes in 40Ar/39Ar Geochronology and Paleomagnetism

Australian Drylands Depositional Environments: Local News, Globally Relevant

Wakelin-King, Gresley A.1

1Wakelin Associates, Melbourne, Australia

Landforms and sediments of continental drylands are relevant to the rock record and to sustainable land management. Globally, many strata were deposited in dryland settings, and modern landscapes are used as analogues for conceptual models of depositional processes. In the present day, ~40% of the world’s land is drylands, and it houses >30% of the people ; Australian drylands are ~80% of the main continent and contain communities, industries and biodiversity.  Understanding how drylands work is clearly desirable. Globally, drylands have highly variable rainfall, highly to extremely variable flow regimes, and a moisture deficit: attributes which govern biota life strategies, and create characteristic sedimentary deposits.

Previously, modern analogues focussed on northern hemisphere examples, typically from coarse-clastic, tectonically vigorous, high gradient catchments with flashy hydrology. Australia‘s context is different: low-relief, low-gradient, subtly expressed neotectonism, a blanket of regolith, and ephemeral rivers capable of big floods and sustained flows. Few Australian drylands sediments have been well-documented; other similar drylands (Thar desert, India; sub-Saharan Africa) may also be under-represented in (English-language) literature, or under-utilised as modern analogues.

This presentation journeys through some Australian continental sedimentary landscapes.

Mud-Aggregate Floodplain, Massive Mudrock

Mud aggregate floodplains are the modern analogue for massive mudrock. The current modern analogues (both located in Australia) are vertic soils transported as sand/ silt-sized bedload. Cooper Creek’s floodplain has black muds deposited/reworked in braid-like bars; Fowlers Creek is a cut-and-fill floodplain with vertically-accreted red muds.

Low-Angle Alluvial Fans

In Australia downstream-diverging fluvial networks are common, usually as low-angle alluvial fans. Channel systems range from coeval multithread to ~single channels sequentially moving across a depositional surface. Fine sediments transported by low-energy flows create broad low-gradient fans. Diverse topographic and climatic contexts lead to a range of sedimentary deposits, most of which are undocumented. Two examples are the mud-dominant Lodden Fan in semi-arid grassland, and the complex aeolian and fluvial sands in the Cooper Creek Fan (Strzelecki Desert).

Floodouts: a Fluvial Process

Floodouts are unchannelised river reaches (flow path is 100% floodplain). Declining discharge during development of Cainozoic aridity means that many Australian drylands rivers are underfit, or their flows no longer extend down the length of the network. Episodic and incomplete sediment transport promotes floodout formation, typically where flow emerges from lateral constriction and loses sediment transport capacity. Floodouts occur in rangefront plains, macrochannels, and valley confluences; they are valuable in land management. Some floodout bedding sequences could be identifiable in the rock record.

Flashy Flow Events, Froude Numbers, and Flat Bedding

Sedimentary fluvial rocks from a drylands setting may include planar-bedded sands. Comparison with the standard bedform diagram may lead to an interpretation of deposition during the high-energy (F = 1) flood peak of a “typical” flashy desert flow. However, bedforms in modern drylands rivers demonstrate more complex conditions including widespread subcritical (F <1) planar bedding, and low-energy gravel imbrication. Bedforms are governed by grain size and flow velocity, but also by flow depth and sediment composition; this leads to intriguing bedform combinations in rivers with rapid flood recession.


Dr. Wakelin-King is a geological geomorphologist, specialising in drylands rivers and sediments. Her professional practice aims at normalising geomorphology within sustainable landscape management. She researches fluvial processes in the Lake Eyre Basin and western NSW, and strongly advocates for fieldwork as a necessary component of remote-area landscape studies.

Orogenesis terminated by mafic underplate delamination at prior passive rift margins: The Delamerian-Ross example

Foden, John1, Tappert, Ralph1, Todd, Angas1, Segui, David1

1Department of Earth Sciences, University of Adelaide, Australia

Stretching from southern Africa to north east Australia the Late Neoproterozoic to Late Cambrian aged Delamerian – Ross Orogen formed at the rifted Rodinia break-up margin, facing the newly opened Pacific. The orogenic history of this margin reflected initiation of subduction of the Pacific plate. At the end of the Cambrian, along the entire belt, active convergent orogenesis was terminated abruptly by rapid exhumation, uplift and cooling. This event is recorded as a widespread regional Upper Cambrian unconformity from southern Africa across Antarctica and into Tasmania. Rapid erosion that resulted from this event produced latest Cambrian to Early Ordovician proximal and distal siliciclastic sediment deposits including conglomerates and fluvial sandstones as well as marine turbidites. These deposits include the South African Cape Supergroup, the Ross Orogen Carryer and Douglas Conglomerates, the Tasmanian Jukes and Owen Conglomerate and the western Victorian turbidites.

In South Australia Jurassic aged kimberlite intruded the Delamerian Orogen and transported an abundant population of mafic xenoliths ranging from garnet-pyroxenite (‘eclogite’) to pyroxenite and mafic granulite. Mineral assemblages include; Cpx-Gt-Rt, Cpx-Gt-Amp-Rt, CPX-Gt-Amp-Ky, Plag-Cpx-Gt± Amp, Ky, Il. These were sampled from lithospheric mantle recording pressures in the range 8 to 25 kbar. Exsolution of garnet and kyanite from Cpx provides evidence for cooling at constant or increasing pressure. Whole rock Nd isotopes yield an imprecise Late Neoproterozoic external isochron and their geochemical composition indicate that parental mafic magma was anorogenic rift-related tholeiite. Importantly the suite of samples forms clear compositional trends that show igneous ‘gabbroic’ pyroxene + plagioclase fractionation control even though many samples are now plagioclase-free.

The conclusion is that these were magmas produced during Rodinian rifting and breakup and formed underplated gabbros at Moho depth. Subsequent cooling to produce plagioclase-free, garnet and pargasite -bearing assemblages lead to increasing density and subsequent delamination resulting in buoyant crustal uplift and probably coupled with slab roll back led to orogenic termination. The common occurrence of high pressure pargasitic amphibole may implicate the role of hydrous flux from the subducting Pacific plate in catalysing high pressure cooling reactions in the mafic underplate. Critical to their density increase, P-T modelling of the pyroxenite bulk compositions indicates that at Moho depths (9kbar) cooling of the Neoproterozoic magmatic underplate would cross the garnet and pargasite-in reactions at 1000oC and the plagioclase-out reaction at 750oC. The time taken for initial mafic magmatic intrusions at 1350oC to cool to cross these important density increasing reactions at a Moho T of ~600oC is of the order of 5-15 m.y. Delamination may also promote local thermal convection leading to anomalous asthenospheric ascent and the production of post-tectonic magmas.

The style of orogenic termination described here forms a distinctive class and reflects subduction at the rift magma-rich margins of continental fragments formed during break-up of earlier continents. This orogenic style seems common to many belts formed during Gondwanan accretion.


Emeritus Professor John Foden is an igneous petrologist and geochemist with a lengthy history of teaching, research  and post-graduate supervision at the University of Adelaide . His specialist interests include; magma generation and differentiation, modern subduction magmatism in the Indonesian Sunda Arc, the orogenic history of the early Palaeozoic margin of Australia, the use of Fe-isotopes in the interpretation of high temperature processes and kimberlite and diamonds.

The status of geoheritage and geoconservation in Australia

Creswell, Ian1

1University of Newcastle, Callaghan, Australia

The protection of Australia’s natural heritage has been ongoing since the 1870s in every state and territory, however, efforts to identify and conserve important sites of geoheritage significance have had limited success.  This presentation reviews the policies and legislation governing geoheritage and geoconservation in all Australian jurisdictions and shows that there are inconsistencies and inadequacies in the processes to identify and protect areas of geoheritage significance.  Each state and territory has differing emphases on geoconservation and different degrees of success in achieving geoconservation goals.  In 2015 the Australian Government released the Australian Heritage Strategy as the overarching framework for the identification, management, and protection of Australia’s heritage across all levels of government and community.  While in recent years there have been a few ad hoc successes related to national heritage or to state heritage, it is not clear the strategy is working.  There is an urgent need for a nation-wide systematic approach to identifying representative geoheritage sites, and to enact processes for their protection.


Dr Cresswell is the co-Chair of the national State of the Environment report, and Chair of the Western Australian Biodiversity Science Institute. He has extensive experience working in environmental science in biodiversity conservation and discovery, oceans governance, fisheries management, wildlife regulation, and protected areas, including geoheritage. He has led major programs in CSIRO both in marine science and terrestrial and freshwater ecology science. Previous to this Ian worked within the Australian Government leading oceans management, sustainable fisheries assessments, international wildlife management, and the Australian Biological Resources Study. In collaboration with the V & C Semeniuk Research Group, Ian has studied coastal systems, including coastal dunes, estuaries, mangroves systems, and tidal flats, towards their management and assessing their geoheritage values.  In relation to geoheritage, Ian has a strong interest in geoconservation, policy, and legislation, and maintains an ongoing research interest in multiple-use management and coastal systems.

Extending Whole-Plate Tectonic Models into Deep Time: Linking the Neoproterozoic and the Phanerozoic

Merdith, Andrew1,*, Collins, Alan2, Williams, Simon3, Tetley, Michael1, Mulder, Jacob4, Blades, Morgan2, Young, Alexander5, Armistead, Sheree6, Cannon, John7, Zahirovic, Sabin7 and Müller, Dietmar7

1UnivLyon, Université Lyon 1, Ens de Lyon, CNRS, UMR 5276 LGL-TPE, F-69622, Villeurbanne, France. 2Tectonics and Earth Systems Group (TES), Dept of Earth Sciences, The University of Adelaide, Adelaide, South Australia 5005, Australia. 3Northwest University, Xi’an, China. 4School of Earth, Atmosphere and Environment, Monash University, Clayton, Victoria 3168, Australia. 5GeoQuEST Research Centre, School of Earth, Atmospheric and Life Sciences, University of Wollongong, Northfields Avenue, NSW 2522, Australia. 6Geological Survey of Canada, 601 Booth Street, Ottawa, Ontario, Canada & Metal Earth, Harquail School of Earth Sciences, Laurentian University, Sudbury, Ontario, Canada. 7Earthbyte Group, School of Geosciences, University of Sydney, Sydney, New South Wales, 2006, Australia

Recent progress in plate tectonic reconstructions has seen models move beyond the classical idea of continental drift by attempting to reconstruct the full evolving configuration of tectonic plates and plate boundaries through time. These advances are an essential step in in quantifying the role plate tectonics has had in the evolution of Earth-surface systems, including the biosphere, atmosphere and hydrosphere, as well as palaeogeographies and the evolving shape of the Earth surface (palaeobathymetry and topography). Previous work has resulted in a number of full-plate reconstructions spanning the last 1 Ga. However, so far these models cover discrete time periods, meaning that a complete model with a consistent set of plate motions and boundaries is not yet available to the Earth Science community. This is a particular problem for the Neoproterozoic and Cambrian, as it means that many existing interpretations of geological and palaeomagnetic data have remained disconnected from younger, better-constrained periods in Earth history. Here we present a continuous full-plate model spanning 1 Ga to the present-day, that is focused on a revised and improved model for the Neoproterozoic–Cambrian (1000–520 Ma), but connects with models of the Phanerozoic and opens up pre-Gondwana times for quantitative analysis and further regional refinements. The model is presented in a purely palaeomagnetic reference frame, and is otherwise geologically-derived, based on preserved data from past-plate boundaries. This is a first step in the direction of a detailed and self-consistent tectonic reconstruction for the last billion years of Earth history.


Andrew completed his PhD at the University of Sydney in 2017, investigating the kinematic and tectonic history of the Neoproterozoic Earth. He has since been a Post-doctoral Fellow at Université Claude Bernard Lyon 1 since 2018, constructing tectonic models of serpentinisation in oceanic lithosphere.

New constraints on the tectonic and metallogenic history of Lachlan Orogen

Habib, Umer1, Leslie, Chris1, Meffre, A/Prof. Sebastien1, Schaap, Thomas1, Wells, Tristan1

1CODES, University of Tasmania, Hobart, Australia

The Lachlan Orogen has a long and complicated geological history involving:

  • Cambrian collision and orogenesis
  • Ordovician continental-derived sediment deposition
  • Ordovician island arc development formation of porphyry and epithermal deposits
  • Silurian orogenesis and extension and formation of volcanic-hosted massive sulphide deposits
  • Devonian orogenesis and formation of granite-related deposits

This geological framework is well-supported by data from many previous studies. However, the exact plate configurations responsible for these geological events remain poorly constrained. These constraints are required to make predictions about the location of ore deposits and to gain a better understanding of the structure and composition of the continental crust.

New data acquired throughout the Lachlan orogen over recent years combined with data from previous studies have helped to improve scientific knowledge showing that: 

  • During the earliest geological history of the Lachlan Orogen there was at least 2 Cambrian arcs: one continental and the other oceanic. The continental arc extended along the western edge of the orogen. The intra-oceanic island was dismembered after it collided with the Selwyn Block and Tasmania.
  • Geochemical and geochronology data from the Melbourne zone and the Selwyn block in Victoria show that most of this zone is likely underlain by crust containing juvenile components rather than thick Mesoproterozoic continental crust.
  • The intra-oceanic Macquarie Arc began approximately 10 Ma after the end of the Cambrian magmatism and was active for a further 40 million years.
  • Magmatism in Macquarie Arc began to be contaminated by continental material starting at 450 Ma in the Molong area. Continental contamination and porphyry development occurred at different times within the magmatic history of the arc.
  • The Lachlan Orocline model explains much of the tectonic evolution of the area from the latest Ordovician through to Devonian periods.

Although these constraints are useful in refining the tectonic models, many details remain unresolved and uncertain.


Sebastien Meffre is Associate Professor at the University of Tasmania. He is the head of Earth Sciences and also works within the Centre for Ore Deposit and Earth Sciences (CODES). His current research interests include: the U-Pb isotopic system, understanding the plate tectonic processes and interactions and geochemistry.

Extracting more from exploration soil samples. The evolution of UltraFine+ and next generation analytics

Noble, Ryan R.P.1, Cole, David T.1, Williams, Morgan J.1, Lau, Ian C.1, Anand, Ravi R.1

1CSIRO Mineral Resources, Kensington, Perth, Western Australia

Continued exploration success requires consistent innovation. While large geochemical surveys conducted by mining companies are common, the suite of data we collect beyond standard soil chemistry has remained stagnant. A novel integration of analytical  methods known as UltraFine+™ extracts the “standard soil chemistry” of the <2 µm soil clay fraction, which is combined with soil VIS-NIR and FTIR spectral mineralogy proxies and physicochemical properties to improve interpretation of soil chemistry leading to better targeting for gold and base-metal exploration.

At numerous study sites across Australia, we demonstrate how the integration of spectral mineralogy proxies and particle-size variation can assist in understanding landscape processes and anomaly formation, and in some settings provide explanations for false positives. Through providing uncertainty estimates on spatial geochemical predictions and reducing the influence of explainable false positives, key information is communicated to decision makers for more confident targeting. A marked decrease in censored results using UltraFine+TM for gold (from 63% to 10% below detection limit) is a major improvement over historical techniques. Automated analytics pipelines using a variety of unsupervised machine learning techniques (e.g. dimensionality reduction, clustering) ensure the rapid interpretation of survey datasets, and accelerate the path to discovery.


Ryan is a principal research scientist with CSIRO working in soil and groundwater chemistry applied to mineral exploration

There are two sides to every trough – the story of young plate collision and its shadow

Keep, M1., Haig, D. W.2

1School of Earth Sciences, The University of Western Australia, 2Oceans Institute, University of Western Australia

As the orogenic product of the collision between the Australian Plate and Banda Arc, the Timor Orogen provides a unique insight into the processes and dynamics of the early stages of plate collision. These early stages, overprinted during ongoing orogenesis, provide information and constraints as to the timing of elements of subduction, the episodic nature of collision, and the resulting regional geodynamics.  Investigating the tectonic, kinematic and geodynamic history of an orogen requires integration of a broad range of field and analytical techniques, the most important element of which must be that any models are consistent with the field geology. That requires a robust understanding of the age and nature of the exposures.

The collisional front of the Timor Orogen includes a severely dismembered sequence of stratigraphic units, which include a number of different pre-collision stratigraphic associations, as well as a structural melange and a synorogenic phase. Juxtaposition of different elements along young, high-angle strike-slip faults means that stratigraphic associations are often juxtaposed across vertical boundaries, commonly eroded. Most original fold-and-thrust geometries are largely dismembered. Biostratigraphy, used in the field as the main determinant of stratigraphic age and association, is perhaps the most important and robust tool in the geodynamic toolbox, although largely overlooked due a lack of familiarity. Painstaking and robust stratigraphic analysis of thousands of samples from across East Timor over the last two decades has facilitated reconstruction of the stratigraphy of the island, which is not easily recognised through lithostratigraphic approaches, allowing interpretations of original formation and subsequent deformation at the collisional front.

In addition, broad regional interpretation of extensive 2D and 3D seismic data from Australia’s North West Shelf over the last 25 years, tied closely to well data (biostratigraphic ages), has allowed correlation of regional events from the Timor Sea westwards to the Carnarvon Basin. These events, marked by uplifts, unconformities, inversion episodes and stratigraphic responses, relate closely to events unfolding at the collisional front. With strain partitioned more strongly at the leading edge, more subtle responses in the strain shadow are not overprinted, and less prone to misidentification. These events corroborate the timing of events seen at the collisional front, occur over a wide area, and tell a different part of the collisional story.

Reconstructing this young and complex orogen requires data from both sides of the Timor Trough.


Myra Keep and David Haig have been working in East Timor since early 2003, making many dozens of visits and collecting thousands of samples for dating. Together they have over 65 years of work history on the Australian NWS, from PNG to the Carnarvon and Perth basins.

Random Forest Based Mineral Potential Mapping for Porphyry Cu-Au Mineralisation in the Eastern Lachlan Orogen

Ford, Arianne1

1Kenex Ltd., Lower Hutt, New Zealand

Random forests represent a machine learning implementation of a decision-tree algorithm that can be applied to data-driven mineral potential mapping. Most published studies using random forests include relatively small numbers of input maps that are typically pre-classified by an expert familiar with the mineral system being targeted. The aim of this study was to investigate how random forests performed using different input parameters in terms of the individual predictive maps and training data. Four different implementations of the random forest algorithm were produced based on a case study using data from the eastern Lachlan Orogen in NSW for the purposes of targeting porphyry Cu-Au mineralisation related to the Macquarie Arc: (1) using a large number of multi-class categorical or non-thresholded predictive maps that have had no favourability criteria applied; (2) using a large number of binary predictive maps that have had statistically valid and geologically meaningful thresholds determined through weights of evidence analysis and expert review; (3) using a subset of the binary predictive maps that were used in a weights of evidence mineral potential mapping study; and (4) using this same subset of binary predictive maps with weighted training data. These results were then compared to the results of an existing weights of evidence mineral potential mapping study.

The results of the random forest analysis demonstrate how both the ranking of the input maps and subsequent mineral potential varies considerably depending on the degree of intervention from an expert in the modelling process. The first approach produced a prospective area that covered 47.7% of the study area, the second approach 6.5%, the third approach 23.4%, and the final approach with the weighted training data 40.4%. In comparison, the weights of evidence study produced a prospective area that covered 15.2% of the study area, however failed to predict one of the training points within this prospective area. Increasing the complexity of the input data improved the predictive capacity of the mineral potential maps for targeting the porphyry Cu-Au mineralisation when expert review was used to determine meaningful thresholds and classifications for the input predictive maps. However, when a large number of multi-class categorical or non-thresholded predictive maps were used as input to the random forest (i.e. no favourability criteria were applied, so the algorithm determined the thresholds rather than an expert), a poor result was obtained. The results also highlight that the main limitation of using random forests (and other machine learning approaches) for mineral potential mapping is the lack of a sufficient number of economically significant deposits which can be used to train a large number of input predictive maps.

The random forest study clearly demonstrates that the use of predictive maps that have statistically valid, geologically meaningful, and practically useful thresholds and reclassifications assigned produce more robust mineral potential maps that can be used for exploration targeting.


Arianne is a spatial data analyst whose focus is on the use of mineral potential mapping and spatial statistics for mineral exploration. She spent more than 10 years as an academic before moving to industry to work on delivering value-added geoscience data to both government organisations and the exploration industry.

About the GSA

The Geological Society of Australia was established as a non-profit organisation in 1952 to promote, advance and support Earth sciences in Australia.

As a broadly based professional society that aims to represent all Earth Science disciplines, the GSA attracts a wide diversity of members working in a similarly broad range of industries.

Conference Managers

Please contact the team at Conference Design with any questions regarding the conference.
© 2020 Conference Design Pty Ltd