ASEG Extended Abstracts - 25th International Conference and Exhibition – Interpreting the Past, Discovering the Future, 2016

The Mt Magnet is one of Western Australia’s oldest gold mining districts with numerous mine workings and extensive past exploration. The extent of historic exploration has made developing new exploration concepts and targets increasing difficult. The aim of the study was to better understand the subsurface architecture of the region as well as identify previously unrecognised geological units which are considered prospective for gold mineralisation, namely banded iron formations or porphyry granite bodies. To achieve this aim a regional 3D geological model was constructed using geological interpretations and a constrained gravity inversion. The geological model was refined using the gravity inversion to create a model consistent with both geophysics and geological understandings. Ultimately from this work greater confidence was provided for drill targets in areas where previous exploration had not effectively tested these areas.

The paper describes the first field measurements using recently developed TranSIM instrumentation. The method is based on registering horizontal components of electric and magnetic field of distant lightning strokes (sferics). In fact it is an expansion of classical MT method in the VLF frequency range of 3 to 30 kHz thus providing information about conductivity distribution at the depths of approximately 50 to 300 meters.

Measurements are conducted atop basalt body in the vicinity of Ipswich (Queensland) over 16 sites on 2 parallel profiles. From 20 to 50 sferics were registered at every site. Data was processed using short time Fourier transform obtaining spectrograms of electric and magnetic field. Apparent resistivity was calculated and frequency pseudo sections are presented.

The apparent resistivity sections were matched with geological scheme of the site. Highly resistive zones correlate with high thickness of a basalt layer and conductive zones show where the underlying basalt sandstone is closest to the surface.

This paper highlights the complimentary potential field studies that have been done in parallel to the interpretation of the 13GA-EG1 Eucla-Gawler deep crustal reflection seismic line. Gravity and magnetic images have been interpreted and potential field data has been modelled using edge detection, forward modelling and inversions to pick out the main domains and structures. Seismic, MT and drill core analysis has been progressing in parallel to the potential field investigations. The different approach taken here was to allow more freedom and independence in the interpretations originating from the potential field studies, rather than constraining them with a predefined architecture from the seismic interpretation. Initial results show gravity and magnetic worms correlating with interpreted structures and domain boundaries. Inversions show the 3D distribution of magnetic susceptibility and densities associated with major features such as the Mundrabilla Shear Zone and folded feature seen in the Nawa Domain. This paper summarises the main findings from the potential field studies, which, in conjunction with the parallel studies, allows for a more robust understanding of the crustal architecture and assessment of the mineral potential of the region.

Geoscience Australia has been collecting Vibroseis deep crustal reflection seismic data since 1999. Since 2013, Geoscience Australia has also collected the uncorrelated single sweep data for each VP of each survey. For a typical survey using three 12 s sweeps, each with 20 s listening time, this means collecting 96 s + 20 s = 116 s of data per VP, instead of 20 s of data per VP. This is nearly a 6-fold increase in data volume, and has been made possible by the availability of high capacity USB data disks. Using extended cross-correlation on the uncorrelated record data, it is possible to image deeper into the earth than using standard vibroseis cross-correlation. Geoscience Australia applied the method of truncated cross-correlation to a deep crustal reflection seismic survey collected over the Yathong Trough section of the Darling Basin in central NSW in 2013. This area showed unusual reflectivity below the Moho in the mantle, stimulating this study. The extended correlation stack showed some faint reflectivity visible from 18s to 24 s which may link to a region of higher reflectivity visible at about 18s in the mantle. Also, the truncated extended cross-correlation method appeared to show improved reflectivity in the mid to lower crustal areas.

This work demonstrates how fully utilizing a modern reprocessing and depth imaging sequence significantly enhances seismic data quality, thereby extracting additional value from older seismic datasets. Using an example from the northern Browse basin, a case study is presented in which the combination of broadband processing, advanced demultiple techniques and anisotropic earth model building produce significant uplift in the imaging results from what is a historically challenging basin for successful seismic data.

The reprocessing sequence was undertaken in two phases. The first phase focused on lowering the noise levels and extending the useful bandwidth of the conventionally acquired seismic data using deghosting techniques. This was combined with 2D and 3D surface demultiple techniques to produce a dataset with low noise levels and a broad signal spectrum for migration. These data were subsequently input to the second phase of the reprocessing which focused on deriving a detailed and accurate earth model. Further anisotropy analysis and well calibration routines were performed to calibrate the earth model to the well data.

The final imaging was performed using TTI Kirchhoff prestack depth migration and comparisons were made to the previous time domain imaging results undertaken in 2010.

Inverse magnetotelluric (MT) problems are naturally ill-posed and smoothing criteria are typically added to stabilize the process. Smoothing and geo-electrical equivalency tend to produce unrealistic geological models. In reality the subsurface geology is differentiated by distinct rock units that are often better defined by boundaries rather diffuse or smooth boundaries. We present the application of fuzzy clustering as an added constraint within the inversion process to guide model updates toward earth models that resemble geological units. Fuzzy clustering divides the simulated model into clusters based on the similarity of model features. Moreover, fuzzy clustering enables the inclusion of additional prior information in the inversion process such as structural and/or petrophysical information. The inclusion of this information produces geo-electrical distributions that more closely reflect the true rock units and unit boundaries. This is demonstrated through several synthetic examples. The simulations show that by including prior petrophysical and/or boundary location information within the inversion the original conductivity distribution is well resolved.

The audio magnetotelluric (AMT) technique has been applied by Geoscience Australia to determine the nature and thickness of cover, plus the basement architecture in regions around Australia. The depth of cover derived from AMT data agrees with results obtained by other geophysical techniques and known information.

We present 1D anisotropic inversions of magnetotelluric data in two regions of the Otway Basin; Koroit, Victoria, and Penola, South Australia. In the Koroit region we have delineated an electrically anisotropic layer at approximately 2.5 to 3.5 km depth; this corresponds to the upper part of the Lower Cretaceous Crayfish Group, a known reservoir unit. The anisotropy strike is consistent between stations at approximately 160° east of north. We interpret the anisotropy at Koroit as resulting from pervasive NNW oriented, fluid-filled fractures, resulting in enhanced bulk electrical and hydraulic conductivity. This interpretation is consistent with permeability data from well formation tests. It is also consistent with the orientation of mapped faults in the area, which are favourably oriented for reactivation in the current stress field. In Penola, no persistent anisotropic layer has been defined even though the areas are geologically similar. The difference in the resistivity structure may reflect differences in the density of fractures or their fill material. Alternatively, it may reflect small differences in the amount by which the fractures are open, resulting from differences in the stress field and fracture orientation in each area.

The integrated interpretation of the airborne FALCON® Airborne Gravity Gradiometer (AGG) survey was designed to assist oil and gas producer Empire Oil and Gas in identifying target areas for hydrocarbon exploration in the Perth Basin. It was developed as a synthesis of the geological structure, tectonic evolution and principles of gravity and magnetic data behaviour. The survey identified areas containing large structural leads and trends as the target of future gas exploration activities, including infill 2D seismic acquisition. The prospectivity of some identified leads has been significantly increased by the gas discovery at wells Red Gully North-1 and Gingin-1, located in EP 389 in central Perth Basin

The need for a more effective tool to explore for channel hosted detrital iron ore deposits under post mineral cover has led to the development of a novel minimalistic seismic refraction survey method we call ‘Sparse Refraction’, to map the depth of cover over basement. The survey configuration employs a single source and a single receiver, at a fixed offset. The source-receiver pair are moved progressively along a traverse, and at each station a first break reading is recorded. Gravity data is also acquired. Combining the two datasets allows the target response to be isolated from the gravitational effects of palaeotopography - we call this residual gravity field the ‘Excess Mass’. The method allows large areas to be screened quickly and effectively, with low environmental impact.

This paper describes the journey from understanding the exploration problem, finding a solution, and the successful implementation of the new approach.

Geoscience Australia operates and maintains a state-of-the-art network of stations and sophisticated instrumentation that monitors natural and anthropogenic (human-made) hazards in Australia and around the globe through its Geophysical Networks.

Key responsibilities are to: operate and maintain the Australian National Seismograph Network (ANSN) and Urban Monitoring (UM) networks; operate and maintain Australian Comprehensive Nuclear-Test-Ban Treaty (CTBT) seismic, hydro-acoustic and infrasound technologies, as part of Australia’s commitment to support monitoring of worldwide nuclear testing; operate and maintain a national network of geomagnetic observatories which form part of a global observatory network; provide technical expertise and advice to Geoscience Australia projects, such as the National Geospatial Reference Systems, Hazard and Risk Infrastructure and Applications, Regional Development, Vulnerability, Resilience and Mitigation and the JATWS (Joint Australian Tsunami Warning System); and, provide technical and operational support for significant Australian earthquake events and aftershock deployment studies.

Geophysical data archives are stored on-site and can be freely downloaded from GA or international data centres. Seismic data can be accessed at GA and Incorporated Research Institutions for Seismology (IRIS) and geomagnetic data at INTERMAGNET.

Seismic data from Geoscience Australia’s Geophysical Networks feeds into important hazard maps including the probabilistic national earthquake hazard map and the probabilistic Tsunami hazard map. Geomagnetic data feeds into the International Geomagnetic Reference Field and has been used to develop the first 3-D conductivity map of Australia.

In this paper, we are proposing a new methodology of 2D seismic acquisition and processing that aims to improve imaging of complex 3D geological environments. The method requires a simultaneous acquisition along two parallel receiver lines. The adapted processing highlights locations of the reflectors that are out of vertical plane by filtering the data by the apparent angle of incidence. This filtering also produces static and residual corrections as a function of this angle. Another benefit of the proposed pre-stack plane filtering is producing 3D velocity model as well as set of individually filtered and migrated sections that can be distributed in 3D volume for the visualisation proposes and interpretation free of conflicting events that would be present in conventional 2D imaging.

With a goal of increasing the geological reliability of single and joint domain inversion, including those set in rugged terrain, we present an expanded scope for cross-gradient regularized inversion. The basic application covers the usual structural similarity objective - comparing the gradient fields of distinct property volumes derived from different geophysical domains - but a particular advantage comes when including gradients derived a priori, from geology or any ancillary property set, providing reference gradient control during single or joint inversions.

In the first example, straightforward cross-gradient joint inversion of synthetic datasets of airborne EM (AEM) and airborne gravity-gradiometry (AGG) improves both the lateral and vertical definition of closely adjacent but distinct bodies.

Two further AEM examples include surface and subsurface geology in the cross-gradient inversion; a) complex foothills setting, and b) buried massive sulphide. Both outputs demonstrate a marked improvement in interpretability over the standard smooth model approach.

The Australian National Ocean Bottom Seismograph Fleet is part of AuScope’s Australian Geophysical Observing System (AGOS) -an initiative of the Australian Government funded through the Education Investment Fund. These instruments will greatly contribute to the understanding of the crust beneath oceanic basins surrounding Australia. In 2014-15 the Australian National OBS Fleet was utilised by the petroleum industry on a number of seismic surveys. High-quality data were recorded at all OBS deployment sites, often to offsets sufficiently large to detect Pn phases - refractions from the upper mantle. Analysis of earthquake data recorded during marine seismic surveys suggests strong interaction between anthropogenic signals (airgun source, vessel noise) and the natural environment, and allows arguing that in some instances earthquake energy contaminates marine reflection data in the frequency pass-band needed for petroleum exploration. Recording earthquake and airgun signals at fixed locations opens up a completely new possibility for calibration and comparison of those signal strengths and spectral compositions.

The success of unconventional gas extraction is dependent on establishing sufficient permeability in otherwise low-porosity and low-permeability formations. In the case of shale gas, permeability can be established through hydraulic stimulation of deep formations, either through existing fracture networks or by creating new pathways for fluids to flow. Coal seam gas (CSG) permeability can be established through de-pressurisation of coal beds by extracting existing sub-surface fluids.

The primary geophysical technique for the monitoring of hydraulic stimulation and de-pressurisation has been microseismic, which measures small seismic events associated with rock fractures. The magnetotelluric method (MT) presents itself as an alternative geophysical approach for monitoring unconventional resource development. MT is directly sensitive to electrical resistivity with depth and orientation and could be used to infer fracture orientation, fluid migration and hydraulic conductivity.

We report on the first industrial MT field surveys for the spatial and temporal monitoring of fluid movement resulting from both hydraulic fracturing of a shale gas reservoir and de- pressurisation of a CSG formation. We show that increasing permeability enables conductive fluids to connect resulting in small drops in bulk resistivity. Such changes in resistivity can be mapped through modelling and inversion allowing a determination of areas with greater permeability and hence production capacity.

The advantages of 2.5D (2D geology, 3D source) airborne electromagnetic inversion in 3D geological mapping applications and the identification of conductive drilling targets compared to the more commonly used CDI transforms or simple 1D inversions are demonstrated using examples from different geological settings.

The 2.5D inversion application used in this work and described in Silic et al, 2015 is a substantially changed version of ArjunAir, Wilson et al., 2006, a product of CSIRO/AMIRA project P223F. The changes include a new forward model algorithm and a new inversion solver. The application enables the accurate simulation of 3D source excitation for full domain models inclusive of topography, non-conforming boundaries and very high resistivity contrasts. Solution is accurate for a geoelectrical cross-section which is relatively constant along a strike length that exceeds the AEM system footprint.

The major innovation includes a new inversion solver with adaptive regularisation which allows the incorporation of a misfit to the reference model and the model smoothness function. The regularisation parameter is chosen automatically and changed adaptively at each iteration, as the model, the sensitivity and the roughness matrices are changing, Silic et al, 2015.

Memory usage has been dramatically reduced and provides a usage estimate prior to execution. For speed the software has been parallelised using Intel MPI and can be used on standard computing hardware or computing clusters. Data from survey lines with lengths exceeding 30 kilometres can be inverted on high end laptop computers. The integrated software design allows the user to prepare a full survey inversion then execute this simply in a batch process. The user can visualise inversion progress at any time during process execution.

We allow flexibility in the selection of components and in the estimation of noise. A non-specialist can obtain a high value result from our 2.5D AEM inversion in terms of it achieving a more realistic geological section.

We show inversion examples from groundwater, minerals (VMS) and geological mapping AEM surveys projects and compare the results with known geology and drilling. We demonstrate the much improved mapping and target definition delivered by this inversion method when compared with the other more common transforms or inversion methods used on these projects.

It is usually assumed that the initial magnetisation curve for a rock, soil or ore sample is linear in the applied field, for fields much less than the coercivity of the magnetic minerals in the sample. This implies that the measured susceptibility, defined as the induced magnetisation divided by the inducing applied field, is independent of the field H that is used in the measurement and that the induced magnetisation of the rock unit in situ can be calculated, irrespective of the field used by the measuring instrument, by multiplying the measured susceptibility by the Earth’s field at the location of the rock unit. A better approximation for many materials that contain ferromagnetic (sensu lato) minerals is a quadratic dependence of the weak-field magnetisation on the applied field, given by Rayleigh’s Law, which yields a linear dependence of susceptibility on applied field. This field-dependent susceptibility is associated with hysteresis and a phase lag of magnetisation behind the applied field for AC measurements, which can masquerade as a phase lag produced by magnetic viscosity. Field-dependence of susceptibility is strongly affected by self-demagnetisation, so measurements of the Rayleigh coefficient η of strongly magnetic samples, as well as the initial susceptibility χ, must be corrected for self-demagnetisation in order to calculate intrinsic properties of the rock unit. Self-demagnetisation also largely explains why rocks containing low-Ti magnetite grains, which have high intrinsic susceptibility, exhibit only weak field-dependence of susceptibility, whereas rocks bearing titaniferous magnetite, monoclinic pyrrhotite or multidomain hematite exhibit relatively pronounced field-dependence of susceptibility. Under the conditions of the Neel approximation (ηH « χ), the Rayleigh laws are still obeyed even when self-demagnetisation is considered. However, considerable departures from the Rayleigh relations occur when ηH > χ This paper examines implications of field-dependent susceptibility for measurements of susceptibility and its anisotropy, and methods for correcting calculations of induced magnetisation.

This study uses broadband magnetotelluric (BBMT) and audio magnetotelluric (AMT) data to model the Eromanga and Georgina Basins in the Boulia region of western Queensland. Extensive data analysis to establish dimensionality, strike and the presence of galvanic distortion was conducted before inversion. The OCCAM 2D MT inversion code was used to produce conductivity sections for interpretation. The results of OCCAM inversions were compared the results of other inversion codes to ascertain the presence of any inconsistencies in the results. Several inversions were run for each profile to optimise inversion parameters with detailed inspection of data fit at each site to establish any systematic data misfits. Inversions were run on the full frequency of AMT data first as the resolution of the data was better. The broadband data was subset to frequencies above 0.4 Hz to focus the inversion on resolving the shallow features. A priori knowledge from the AMT inversions was very useful in interpreting the lower resolution BBMT data. Independent constraint in the form of drillhole and seismic data was used to aid interpretation of the inversions profiles. The BBMT inversions allowed the two-layer Georgina Basin signature evident in the south of the project area to be traced further north. They also delineated more complicated basin morphology in the west of the project area.

This study aims to clarify the tectonic evolution of the Lachlan Orogen by modelling the subsurface morphology of the Gilmore Fault Zone (GFZ). The GFZ marks a distinct geophysical contrast between (high gravity, low magnetic intensity) high-grade metamorphic rocks found in the Wagga metamorphic belt (WMB), to the west, and the (low gravity, uniformly high magnetic intensity) low-grade volcanic rocks found in the Macquarie Arc and Silurian rift basins to the east. Understanding the structure of this fault at depth should provide constraints on existing models for the tectonic evolution of the Lachlan Orogen.

Subsurface structure around the GFZ in the vicinity of Barmedman has been inverted by iterative 2.5D potential-field modelling of gravity and magnetics, constrained by pre-existing reflection seismic profiles, potential-field interpretations by previous workers, and physical properties data collected on representative lithologies.

Preliminary findings show that the surface structure mapped as the Gilmore Fault is an east-dipping, shallow thrust fault, and doesn’t correspond to the major crustal ‘suture’ envisaged in regional tectonic studies. It is a secondary antithetic structure off the main west-dipping, crustal penetrating fault that separates the Macquarie arc and the WMB. Another, steeper west-dipping fault cuts these structures and separates the WMB from the Silurian Tumut Trough. This larger structure defines the regionally extensive tectonic feature of the GFZ. The trace of the mappable Gilmore Fault (as opposed to the GFZ) is curved, and terminates abruptly to the north, indicating the Gilmore Fault is the base of a series of thrust flakes imposed on the pre-existing main fault in the GFZ.

West-dipping crustal-penetrating thrust faults east of the GFZ are indicative of successive collision, accretion and extension events. Thus far the modelled structure of the GFZ is not consistent with the terrane accretion model but is consistent with the accretionary orogen and orocline model.