ASEG Extended Abstracts - ASEG2013 - 23rd Geophysical Conference, 2013

Hood 10 is a small polymetallic (Cu - Zn) VHMS mineral deposit located in the Archean aged Slave Craton in arctic Canada, and is 100% owned by MMG Ltd. The deposit is hosted in the Napaktulik volcanic belt, a bimodal greenstone belt of approximately 2.67 Ga age. The deposit was initially discovered in the early 1970’s by Texas Gulf and covered with airborne and ground frequency domain EM (FDEM) and magnetics. The property has changed hands several times over the intervening years, with drill programs in the mid 1970’s, early 1980’s and early 1990’s resulting in a non-JORC compliant resource of 1.2 million tonnes of 4.4% zinc and 4.1% copper, and defining the ore body along a 250 m strike length and to a depth of 220 m. The property was acquired by MMG Ltd in 2009. In 2012, MMG Ltd conducted significant geophysical exploration on the property for the first time, acquiring surface and airborne time domain EM (TDEM) coverage of the deposit, as well as TDEM on 11 of 12 holes drilled into Hood 10 in 2012. New geophysical modelling of the TDEM results, drilling results, recent surface geological mapping and structural studies have significantly increased the understanding of the deposit. The current non-JORC compliant resource stands at 2.4 million tonnes of 3.5% zinc and 4.5% copper.

We will present the progress made on the development of a computational algorithm to model 3D Magnetotelluric data using Vector Finite Element Method (VFEM). The differential equations to be solved are the decoupled Helmholtz equations for the secondary electric field, or the secondary magnetic field, with a symmetric conductivity tensor. These equations are modified to include anisotropic earth and complex geometry (such as surface topography, and subsurface interfaces). The primary field is the solution of an air domain, homogeneous half-space or layered earth.

This study will compare the application of two boundary conditions, the Generalize Perfect Matched Layers method (GPML) versus Dirichlet boundaries. Dirichlet boundary conditions are applied on the tangential fields, assuming that the boundaries lie far away from the inhomogeneous model. The GPML scheme defines an artificial boundary zone that absorbs the propagating and evanescent electromagnetic fields, to remove boundary effects (Fang, 1996).

In this algorithm, high order edge elements are defined based on covariant projections for hexahedral elements (Crowley, et al., 1988). The vector basis functions are defined for the 12 edges (linear) element, 24 edges (quadratic) element, and 48 edges (cubic) element. By this definition, the vector basis will have zero divergence in the case of rectangular elements and relatively small divergence in the case of distorted elements. They are defined to study their numerical accuracy and speed, and to see if the divergence correction is automatically satisfied.

Mining and energy development in South Australia’s far north is set to have significant consequences for the water resources of the region. These sectors generate significant economic value to the State and their support remains a priority for the government. The scale of the planned developments and the potential from current exploration programs facilitated by the South Australian Government’s PACE Program will result in an increase in infrastructure requirements, including access to water resources and Aboriginal lands for potential mine development. Increased demand for water and in particular groundwater is compromised by the limited information we have about these resources. There is a recognised need to develop this knowledge so that water availability is not a limiting factor to development. The Goyder Institute’s Long-Term Outback Water Solutions (G-FLOWS) Project was established to help address this. Particular reference is made to work completed in the Musgrave Province. It illustrates the role of local scale AEM, acquired for exploration, and regional scale airborne magnetics and terrain data in helping develop a hydrogeological conceptual model for the Province. The AEM data reveal a complex and extensive inset palaeovalley system which contains groundwater of variable quality (2000 - 4500 mg/L TDS). Examination of their location against the regional magnetics indicates a strong litho-structural control on their orientation. If mineral resources were to be developed in the area, these groundwater systems would represent the best option for water supply. A regional scale water resource map, based on information gleaned from the geophysics, existing hydrogeological and digital elevation data, is presented that provides a framework for groundwater resource determination when/if mineral deposits were to be mined in the region.

Artesian mound springs occur along the south-western edge of the Great Artesian Basin, in northern South Australia, but their underground structure and relationship to faulting is not well understood. We have performed geophysical surveys over three different systems using a range of techniques: early-time TEM, self-potential, and magnetotellurics.

The self-potential data contains a local response due to specific spring vents, and also a broader stronger response due to laterally extensive upwelling in the lower part of the Bulldog Shale, at depths of approximately 100 m. Modelling of TEM and magnetotelluric data show that the confining Bulldog Shale, which is generally very conductive, contains resistive areas underneath springs and spring complexes which are believed to be related to spring-related carbonate deposition. Magnetotelluric modelling in particular indicates that anisotropic resistivity in the form of vertical sheets at a depth of 100 to 200 m, can explain the observations more readily than a conductive 2D feature, suggesting that the structures underlying the springs are sets of closely-spaced faults. The orientation of this anisotropy matches the regional NNW/SSE orientation of spring complexes.

Predictive mapping of mineral systems represents an important tool for assessing the potential for undiscovered mineral resources. Recently, assessments for a range of uranium mineralisation styles have been performed across three regional studies in Queensland, South Australia and the Northern Territory. These investigations have been undertaken using a mineral systems framework which considers key system components including sources, fluid-flow drivers and pathways, and depositional mechanisms.

This method places a strong emphasis on identifying important processes leading to ore formation, which are then translated into mappable geological proxies using a range of input datasets and derivatives. In areas of outcrop or shallow cover, these processes may be mapped using geochemical and observational geological data. However, deeper-buried terranes, such as those dominating most of the Australian continent, require the use of geophysical data to generate proxies for targeting the desired processes.

Importantly, and unlike many other available techniques, the method employed does not rely on the locations of known mineralisation to generate maps of mineral potential. This allows assessment of mineral potential in greenfield regions of Australia, including those beneath significant volumes of cover. Results from the regional studies completed to date successfully reproduce the locations of known mineralisation and highlight potential in areas not currently recognised as mineralised. Such mineral system analyses provide predictive models which may be the focus for follow-up investigation, including drilling.

Geoscience Australia has been acquiring both broadband and long-period magnetotelluric (MT) data over the last few years along deep seismic reflection survey lines across Australia, often in collaboration with the States/Territory geological surveys and the University of Adelaide.

Recently, new three-dimensional (3D) inversion code has become available from Oregon State University. This code is parallelised and has been compiled on the NCI supercomputer at the Australian National University.

Much of the structure of the Earth in the regions of the seismic surveys is complex and 3D, and MT data acquired along profiles in such regions are better imaged by using 3D inversion code rather than 1D or 2D code.

Preliminary conductivity models produced from the Youanmi MT survey in Western Australia correlate well with interpreted seismic structures and contain more geological information than previous 2D models. GA has commenced a program to re-model with the new code MT data previously acquired to provide more robust information on the conductivity structure of the shallow to deep Earth in the vicinity of the seismic transects.

The Late Cretaceous Belfast Formation is a fine grained, regionally extensive mudstone unit that provides the regional seal for a number of gas discoveries in the Otway Basin. A distinct decrease in sonic velocity and resistivity through the Belfast Formation is observed in wells on the western (down thrown) side of the Sorell Fault Zone. The velocities are anomalously low relative to expected velocities based on depth trends and/or petrophysically estimated velocities. Analysis of conventional log data in conjunction with bulk rock x-ray diffraction (XRD) data from rock chips through the interval illustrate that the velocity reversal is not a function of changes in lithology or porosity. We interpret the observed log response to be a function of increased pore pressure gradient, overpressure, through the Belfast Formation.

Increased drilling mud weights through this interval in a number of wells supports this conclusion. There have been no attempts to pressure test the Belfast Formation, however, there are abundant pressure test data for the underlying reservoir rocks. Pressure tests do show evidence of overpressure in reservoir intervals in Triton-1. It is not surprising that higher overpressures are not observed and/or recorded in the reservoir interval as the overpressure in the Belfast Formation is likely a consequence of disequilibrium compaction. Such overpressure has limited opportunity to create overpressure in adjacent formations. The disequilibrium compaction is likely caused by rapid sedimentation during the Late Cretaceous.

Using the Eaton Method (Eaton, 1975) and sonic log data, a maximum pore-pressure approximately 30% higher than hydrostatic is calculated in the shale. This prediction, however, cannot be made without an implicit assumption regarding a ‘normal’ compaction trend; we utilise a Gardner density-derived velocity prediction in addition to a standard depth trend extrapolation method. Observations in this paper are consistent with recently published evidence for overpressure in Otway Basin west of the Shipwreck Trough.

The occurrence of overpressure has implications for velocity analysis of seismic data and depth conversion in and around the Shipwreck Trough. Overpressure effects on velocity are not readily observable during velocity analysis of CMP gathers, which is partly due to the lack of reflectivity through the Belfast Formation and partly as the velocity of any reflector in this shale section will have a slower than expected velocity and can be potentially dismissed as multiple energy. This results in stacking velocities that are over-predicted through the shale, which can limit their use in detecting overpressure and cause depth conversion errors.

Government Geological Surveys study the Earth at the regional, province or national scale, and acquire vast volumes of technically complex data. Clients are increasingly requesting not just the basic national data, but also processed and derived products that enhance the geological information contained in the data. High performance computers facilitate a step-change in advanced processing and modelling of large, complex data, and will help Government Agencies deliver more sophisticated products to industry and researchers. Data processing and inversion are no-longer limited by the computational effort required.

Geoscience Australia is collaborating with the National Computational Infrastructure facility (NCI) at the Australian National University to develop advanced methods for extracting the maximum geological information from large data volumes. The new methods include: Modelling of potential field data in spherical coordinates; Inversion of magnetotelluric tensor data to a full 3D mesh of resistivities, and; Monte Carlo inversion of AEM responses. These algorithms are being implemented in a new Virtual Geophysical Laboratory in which government data and advanced processing methods are brought together in a single high performance computer environment.

A 2D land seismic reflection line was acquired 14-17 June 2011 in the Surat Basin, Queensland. Two source types were utilised; a heavy Vibroseis (AHV-IV 60,000 lb), and an impulsive surface source (Geokinetics proprietary Dual Synchronised Electrical Impulsive Source “onSEIS”). The recording system was OYO Geospace Seismic Recorder “GSR” cable-free nodes. The main objectives of the study were to (1) evaluate the response of the different seismic sources on the local geological framework; (2) to establish the optimal source effort for a planned 2D survey within the area. Vintage dynamite data acquired along the same line in 1984 was also used in the source effort analyses.

Qualitative analyses included visual inspection of raw shot gathers and processed brute stack sections; reflectors coherency, signal-to-noise ratio, spatial distribution, and signal penetration, were determined on raw displays. Quantitative analyse included amplitude spectra on Brute stack sections. Frequency content and signal bandwidth were critically evaluated to define which source performs better in terms of seismic resolution.

The dynamite single records showed good S/N ratio with clear reflectors at the target depths down to 1800 ms. However, the processed stack showed less bandwidth and lower imaging resolution due to the coarse source and receiver sampling during data acquisition. The onSEIS and Vibroseis signatures were comparable to the dynamite shots at the deep targets. However, the higher trace fold, longer offsets, and tighter spatial sampling for the onSEIS and Vibroseis, provided higher imaging resolutions that illuminated the entire section.

The Curie depth is the depth at which the crust or upper mantle ceases being magnetic due to temperature effects. Although there are several methods available to map this depth, magnetic methods are most often used. These methods, however, introduce uncertainty into the depth estimates and this uncertainty has not been adequately addressed in previous studies.

In this study, we have used magnetic data, at multiple scales, in combination with Monte Carlo techniques to evaluate both the Curie depth and its uncertainty for the Australian continent. Variations in the Curie depth for Australia are related to differences in mineralogy and thermal regimes across differing provinces of Australia, and may also be used to further our knowledge of crustal geothermal gradients. Increasing our knowledge in these areas will advance our understanding of uranium, geothermal and hydrocarbon systems in Australia.

The Virtual Geophysics Laboratory (VGL) is an environment that was developed as a data discovery and delivery facility, incorporating software and computing facilities. This design enables geoscientists to store, discover, retrieve and process datasets. Recent developments are expanding the VGL to incorporate the functionality of the Underworld software. Underworld is open source, parallelised software capable of calculating the 3D temperature distribution in the crust.

Numerical modelling of temperature is a tool that can be used to predict the temperature distribution at depth between and beneath measurement points based on a 3D geological map. Due to their size, temperature simulations performed on some of these 3D maps can have large memory requirements. To address these requirements and reduce computation time, cluster computing facilities (which have multiple individual processors) can be utilised. In order to analyse uncertainty quantitatively, hundreds to thousands of individual runs are often needed and therefore cluster computing facilities are necessary to complete the runs in a reasonable timeframe.

The new developments to VGL will facilitate the discovery and access to 3D geological maps. They will also provide easier access to the Underworld software, and will provide the high performance computing facilities (hosted at the National Computational Infrastructure and elsewhere) required to run large models. The metadata associated with each run performed using VGL is automatically stored, and therefore runs completed on VGL will be repeatable and testable.

A modified inversion approach is presented for the effective separation of sources in marine simultaneous shooting acquisition. The method aims to distribute all energy in the simultaneous shot records by reconstructing the individual shot records at their respective locations. The method is applied to a simulated simultaneous long offset data set, where two sources are used to acquire long offsets with conventional cables. In the second example, the performance is investigated on a data set from Western Australia, where two sources where located within close proximity, with only a small cross line distance between them. Results demonstrate that the individual sources can be separated satisfactory for both simultaneous source configurations.

A magnetic body has a resultant magnetisation that is the vector sum of its induced and remanent components. Ignoring the role of remanence may lead to erroneous resultant magnetisation direction and hence inaccurate geophysical models. Determination of this resultant magnetisation direction may be obtained from total magnetic intensity data for an isolated magnetic anomaly. We present a new method to recover the resultant magnetisation direction that proceeds by iteratively calculating an approximate source layer, and cross correlating trial magnetisation directions. This method is tested with a number of remanent component orientations for compact and elongate sources. For a compact source, the accuracy to which the correct resultant magnetisation direction can be recovered is generally found to be less than 5°. The method retains accuracy for low inclination resultant magnetisation directions, however, begins to lose sensitivity as the resultant inclination becomes steep. Application of the method to a case study of the Black Hill Norite recovered resultant magnetisation directions in agreement with paleomagnetic results. Here the resultant directions from the two methods are found to be consistent, with minor differences possibly due to limited paleomagnetic sampling of magnetisation directions that changed as the intrusion cooled. A higher bulk Q value, than found from the limited sampling, is suggested for one of the three anomalies studied, as supported by the results of other authors.

Remotely sensed geoscience data can assist detailed geological field mapping in areas of thick vegetation and poor outcrop. However, the potentially high dimensionality of these data makes it difficult to visually interpret and fully comprehend. Machine learning algorithms provide an efficient semi-automated means of recognising and identifying patterns in data. We use Random Forests for supervised classification of geologic units from airborne geophysical and soil geochemical data in the economically significant Hellyer - Mt Charter region of western Tasmania. A backward-recursive variable selection method is used to select the most relevant and useful data for this problem. This reduces computation cost and enhances interpretation of results without significantly affecting prediction accuracy. Random Forests generates accurate predictions of the spatial distribution of surface geologic units from these data. An example is provided regarding the use of Self- Organising Maps, an unsupervised clustering algorithm, to identify distinct but spatially contiguous clusters within a geologic unit. By visualising cluster spatial distribution and identifying key variable contributions to cluster differences, we interpret the geological significance of intra-class variability.

We investigate elements of a framework to efficiently map hydraulic conductivity (K) over wide areas by integrating surface and logging NMR data. Central to the broader framework is the need to characterize site- and measurement-specific relationships between surface NMR, logging NMR, and hydraulic conductivity data sets. We outline progress toward this goal focusing separately on the relationship between each of the three data types, emphasizing most recent findings. We identify well-development as an important factor affecting the relationship between NMR logging data and K. We also present results showing that multi-pulse surface NMR sequences provide access to NMR parameters more closely related to K. Finally, we present data examples illustrating extension of the framework to wide-scale acquisition of 2D surface NMR data.

Tectonism is believed to have exerted a strong control on the current disposition of sedimentary hosted uranium mineral deposits in the Frome Embayment, in South Australia. However the extent of this control has not been well understood, nor documented. Here, we examine the combined use of regional and finer-scale TEMPEST AEM data sets, linked to a structural interpretation of airborne magnetics and ground gravity, to extend our understanding of the evolution, geometry and variability of sediment packages associated with sedimentary uranium mineralisation in the Curnamona and Billeroo Palaeochannel systems. Through the analysis of both smooth and blocky model LEI inversions of these AEM data, we contend that structural control was critical in determining the initial orientation of the palaeovalleys and the location of basal sequences of the Eyre Formation, the host to known uranium mineralisation. We examine this influence in the context of the Goulds Dam uranium deposit. The presence of reactivated basement faults which controlled the initial orientation of the palaeovalley systems may also have a role in providing the loci for mobile reductants from underlying basin sequences, although this requires further investigation.

In 2011, Geoscience Australia collected 484 km of deep crustal seismic reflection data. The survey traverses the northeast Yilgarn Craton, western Officer Basin and the western Musgrave Province. The aim of the survey was to obtain structural information about a relatively unexplored region of the Officer Basin, as well as provide insights into the crustal-scale architecture of the surrounding region. The broad scale architecture is of particular interest because it delineates the boundary between two major continental blocks; the Musgrave Province and the Yilgarn Craton.

This study presents a geological framework for the region using a three-dimensional model as a medium for visualisation and is a virtual representation of the geology of the region. It derives from interpretations made by integrating spatial geoscience datasets collected at the Earth’s surface. The most significant constraint for the subsurface geology comes from the deep crustal seismic reflection survey, 11GA-YO1.

The resulting model is a georeferenced, closed volume that represents a geological framework to be tested and understood within the context of geodynamic models for the region.

The Bryah Basin is part of the Capricorn Orogen, a collision zone between the Archaean Pilbara and Yilgarn Cratons in western Australia. The Basin is host to significant mineralisation, including mesothermal orogenic gold, copper-gold volcanogenic massive sulphides. Among the challenges in the exploration for these mineral systems is the paucity of outcrop and the extent and variability of a complex regolith cover. To better understand this regolith, a reconnaissance, regional-scale, fixed-wing time domain AEM survey was undertaken over the Bryah Basin in 2012. The resulting data were inverted using a smooth model layered earth inversion. In this paper we compare results on mapping regolith variability obtained from the full inversion of the AEM data against that defined from the fast approximate transform of the same data set. The inverted data show the most dominant regolith features are associated with sediment filled palaeovalleys. The regional regolith framework determined from this study provides a basis for better understanding and interpreting an extensive regolith geochemical data set with respect to metalloid anomalies linked to buried Cu-Au mineral systems

The inversion of large-scale magnetic data sets has historically been successfully achieved through integral transforms of the large, dense sensitivity matrix. Two well-known transforms, the discrete Fourier and multi- dimensional wavelet, reduce the required storage and ultimately speed of the inversion by storing only the necessary transform coefficients without losing accuracy. The main drawback of the approaches is the required calculation of the entire dense sensitivity matrix prior to the transform. This process can be much more costly than the inversion itself. We solve the magnetostatic Maxwell’s equation using a finite volume technique on an ocTree-based mesh. The ocTree mesh greatly reduces the time required for the inversion process. When working in the differential equation domain it is not necessary to explicitly form the sensitivity matrix; this decreases the storage requirement of the problem and increases the overall speed of the inversion. The principal mesh is broken up into sub-domain ocTree grids to further enable parallelization of the forward problem. These grids extend the entire domain of the principal mesh to include large regional features that may influence the data. We present the discretization of the equations and verify the accuracy of the modelling both with the principal mesh and with multiple sub-domains. We show a synthetic example and a large field example consisting of over 4 million data and 5 million model cells that was inverted on a desktop computer.