ASEG Extended Abstracts - ASEG2009 - 20th Geophysical Conference, 2009

The geology and petroleum potential of the western Tasmanian offshore basins is poorly understood. As part of a strategy to improve the understanding of these basins, aeromagnetic data was acquired by Geoscience Australia and Mineral Resources Tasmania under a National Geoscience Agreement and partly funded by the Commonwealth Government's Offshore Energy Security Program. The survey acquired 141,234 line km of high quality data with a line spacing of 800 m across the Bass, southern Otway and Sorell basins and Torquay Sub-basin (Figure 1). The aim of this survey was to acquire new aeromagnetic data to help delineate the structural architecture of the basins and underlying basement and the distribution of igneous rocks. The data fill a gap in the existing aeromagnetic coverage between Tasmania and mainland Australia and provide fresh insights into basement structure and its control on basin architecture and sedimentation patterns during the breakup of Gondwana and separation of Australia from Antarctica.

We present an example of the use of a software framework that combines 3D geological mapping and geophysical modelling. This approach would be relevant in a wide range of applications, from regional geological mapping, groundwater studies, through to mineral or geothermal prospecting. In this instance, we illustrate a typical workflow with a case study involving the San Nicolas VMS deposit and surrounding region. An initial 3D geological map, based on sparse surface geological observations, is progressively refined using AEM and potential field modelling results. At each stage, the geological map is used to capture and communicate the inferred distribution of map units. During the latter stages of interpretation, the geological map is also used to provide constraints for potential field modelling. In addition to the opportunity for direct detection of highly conductive basement-hosted sulphide mineralisation, the availability of AEM data proved crucial in the San Nicolas example in that we were able to use these data to determine the thickness of a moderately conductive transported cover unit. This significantly reduced the ambiguity of subsequent potential field interpretations.

To be reliable, Earth models used for mineral exploration should be consistent with all available geologic information. Furthermore, combining several complimentary types of geophysical data collected over the same Earth region can further reduce ambiguity and enhance inversion results. Phillips (2001), Williams (2006) and Farquharson et al (2008) provide examples of incorporating some types of geologic information to dramatically improve inversion results. We expand the types of geologic information that can be incorporated and present a cooperative inversion strategy for inverting multiple data types.

An emerging field in geophysics is joint inversions that aim to improve the reliability of subsurface modelling (Haber & Oldenberg 1997, Gallardo & Meju 2004 ). Such inversions are where two complementary geophysical data sets are simultaneously inverted to yield a unified model. Individual inversions can produce highly varied models that fit the data equally well. Joint inversions may help to reduce this ambiguity without having to introduce external constraints. The integrated models also allow for more rigorous interpretations.

The difficulty with joint inversion is how to link different data sets. There are two approaches to linking the different data sets, the petrophysical approach and the structural approach. The petrophysical approach links geophysical parameters through rock properties such as porosity. The structural approach is based on the premise that the geological conditions that control changes in one parameter will also affect the other parameter. This means changes in one parameter should coincide with changes in the other parameter.

We look to explore both joint inversion approaches in the first instance by combining gravity and magnetotelluric (MT) data. Both techniques offer cost effective ways of determining the Earth’s subsurface properties. Gravity is already widely accessible and ATT data is being increasingly acquired. This creates many opportunities to apply the inversions.

Building models of a geological system using potential field is almost a routine exercise, where both gravity and magnetic data are used to interpret the architectural framework of a geological setting. Although nonuniqueness in potential field interpretation is inevitably a major issue (Skeels, 1947, Roy, 1962), it is argued by many authors (Pedersen, 1979; Bosch et al., 2006) that the use of data redundancy, for example, joint use of gravity and magnetic data as well as appropriate constraints in the interpretation/inverse modelling overcomes the ambiguity limitation of potential field interpretation.

However, joint use of gravity and magnetic data especially for geometric inversion is a nontrivial task. By geometric inversion of potential field data we mean determining the geometric configuration of the modelled body whose physical properties, such as density/susceptibility are known. Such potential field inversions are inherently nonlinear. The presence of nonlinearity between the data and model space often results in an apparent paradoxical situation, offering a less stable and thus a less reliable inverse model while exercising the data redundancy by joint use of gravity and magnetic data. We address the issue of producing a stable solution for a 2D joint geometric inversion of gravity and magnetic data via the “Pareto optimal” criterion. To this end, we state that the joint inversion of gravity and magnetic data can be realised via a multi-objective optimisation problem (MOP), where the objective functions related to the best fitting gravity and magnetic data are satisfied simultaneously. Very recently, MOP has been introduced by Kozlovskaya et al. (2007) in seismic anisotropy studies of lithosphere by jointly inverting shear wave splitting and P-wave residual data. With regard to the MOP our goal is not to determine a unique, possibly globally optimal, solution, but to determine a class of optimal solutions defining a compact set. Such a class of solutions is known as a Pareto set or Pareto front, where each solution is optimal in a Pareto sense; meaning there exists no other solution which will improve one objective function without degrading the other objective function (Coello Coello et al., 2002). To determine a Pareto front we use the ε-constraint method, where one objective is minimised while the other one is constrained with a bound by a pre-selected value of ε. To minimise the objective function we use a Particle Swarm Optimization (PSO) scheme, which is a global stochastic optimisation scheme. The rationale in selecting PSO is that it is easy to implement and is computationally inexpensive since its memory and its CPU speed requirements are low (Jones, 2006). PSO has been used for both continuous and discrete optimisation problems (Clerc and Kennedy, 2002). The application of PSO in geophysical inversion is demonstrated by Shaw and Shrivastava, (2007). In this study we discuss the application of MOP and PSO in the joint inversion of gravity/magnetic data as

well as testing the method through numerical experiment using synthetically generated gravity and magnetic data for a simple but realistic geological model.

Magnetic instruments such as cross-wing gradiometers, vertical gradiometers and full tensor SQUID magnetometers presented challenges for geological interpretation and geophysical inversion. In particular, the full tensor magnetometer presents many new challenges for an interpreter where only the vertical derivative of the vertical magnetic component presents a useful geological analogue for visual interpretation. With six channels of information how do we make practical use of the other five channels which implicitly contain useful information about the 3D distribution of magnetic properties?

Joint inversion of all six channels is the logical solution whereby the data is inverted directly to a 3D magnetic susceptibility model. When compared with the scalar amplitude of the total magnetic intensity measurement, the magnetic tensor has valuable 3D information. For example just a few samples can provide sufficient information to immediately determine if an igneous pipe is on the left or right side of the flight line. A few more samples can locate the position and depth of a pipe that is off to the side of a flight line.

Joint inversion can be used with various combinations of sensors and derived parameters. For example a cross-wing total magnetic field gradiometer can be used with the centre point total field value to derive important off-line geological information. The first vertical derivative derived from gridded data can be combined with total magnetic intensity measurements for two channel joint inversion to optimise the quality of depth, width, dip and depth extent inversions.

Examples are provided to illustrate the improvement in geological information extraction when compared with single channel inversion of total magnetic intensity data. The methods provide new opportunities to look at the latest generation of instruments and new ways to look at old surveys.

Inversion of geophysical data seeks to extract a model, or suite of models, representing the subsurface physical properties that can explain an observed geophysical dataset. Due to the inherent non-uniqueness of inversion, any recovered property distribution is only one of an infinite number of possible distributions that could explain the observed data. The most desirable solutions are those that can explain the observed geophysical data and also reproduce known geological features; a goal that can only be achieved by including any available geological information into the inversions as constraints.

One approach to achieving this goal of integration is to supply a full 3D model of geological observations and interpretations to the inversion and test the hypothesis that those interpretations are consistent with the geophysical data (McGaughey, 2007; Mclnerney et al., 2007; Oldenburg and Pratt, 2007). However, in greenfields mineral exploration where limited geological knowledge exists, it may be impossible to define such a 3D model everywhere in the region of interest. An alternate approach is to supply only the available sparse geological observations to the inversion to recover a prediction about the subsurface distribution of geological features that may be required to satisfy both the known geological constraints and the observed geophysical data. This postpones much of the geological interpretation until after the inversions have been performed and reduces the lead time to recover an inversion result and enable the results of inversions to be used in decisions to acquire further geological and geophysical data or to assist with geological interpretation.

We describe a new method for preparing the geological constraints required for this sparse data approach. It is specifically targeted for use with the University of British Columbia - Geophysical Inversion Facility (UBC-GIF) GRAV3D and MAG3D gravity and magnetic inversion programs (Li and Oldenburg, 1996, 1998). The UBC-GIF inversion approach allows constraints to be assigned to each cell using four sets of parameters:

A reference physical property which provides the best estimate of the arithmetic mean physical property in the cell.

A smallness weight which provides an estimate of the reliability of the assigned reference physical property. The weight is a unitless value ≥ 1 with increasing values indicating higher confidence.

Lower and upper physical property bounds indicating the absolute limits on the property range that can be assigned to the cell. These effectively represent a confidence interval on the supplied reference property.

Smoothness weights controlling the variation in properties between each adjacent cell in each direction. Values > 1 promote smoother property variations between cells. Values < 1 (but > 0) promote discontinuities in properties between cells.

The inversion will recover a physical property model with properties for each cell that lie between the defined bounds and are as close as possible to the supplied reference physical

properties, while still reproducing the observed geophysical data. If possible, the reference physical properties will be matched more closely in those cells that have the highest smallness weights.

A number of helicopter AEM systems are discussed including AeroTEM, HeliGeotem, RepTEM. SkyTEM and VTEM. Field data are shown from a range of exploration environments, including AeroTEM data across an oil sands prospect in Alberta, Canada, HeliGEOTEM mineral exploration data from Labrador, Canada, RepTEM uranium exploration data from South Australia, SkyTEM salinity mapping data from South Australia and VTEM data across a massive sulphide in Western Australia. The data were modelled with layered-earth inversions and selected anomalies were modelled successfully with conductive plates in free-space.

The fixed-wing MEGATEM airborne transient electromagnetic (EM) system was introduced in 1998 (Smith et al., 2003) as a four-engine version of the GEOTEM system (.Annan and Lockwood, 1991). The four engines were required so that the system could fly safely at high altitudes in the Andes Mountains (Smith et al., 2003). The larger aircraft required to carry four engines was also able to carry a larger transmitter loop, so the dipole moment of the system increased to more than one million Am² - hence the name MEGATEM. However, the aircraft was also able to draw power from more generators and carry heavier more powerful transmitter electronics, so a further increase was achieved in 2001 so that the dipole moment became more than two million Am² (Smith et al., 2003). As a consequence, the system name was modified to MEGATEM^II. Figure 1 shows the MEGATEM system with the EM transmitter and receiver labeled. The system also flies with a magnetometer for measuring the intensity of the earth’s field.

The ability to detect and characterize targets of high conductance (>2 000 S) has become a primary technical criteria for airborne EM surveys being flown to explore for nickel sulfide deposits. A variety of commercial airborne EM systems have flown over the same high conductance target as well as a prototype system These results show what the current capabilities of the various systems and as well, what type of future technology is required to significantly improve the high end conductance aperture of airborne EM technology.

As a check on direct current measurements of the transmitted waveforms as measured with a wideband transducer, we monitored the EM field of the SeaTEM system during several flights over a ground loop. The ground loop was laid out over very resistive granite in Western Australia. Ground measurements were made of the current and voltage induced in the ground loop. For the first time, we also collected point-sampled total field and component magnetic fields using commercial sensor systems to investigate the effectiveness of direct B field rather than dB/dt measurements.

The AeroTEM system’s rigid structure and bucking coil makes the recording of interpretable on-time data possible. When processed using standard methods the on-time dB/dt response is comparable in conductivity aperture to the off-time calculated B-field response. However, standard primary field removal methods distort both on-time dB/dt and calculated B-field responses. By modifying primary field removal routines to incorporate the rigidity of the system it is possible to minimize these distortions. While this directly improves dB/dt and calculated B-field responses, it opens the door to other advanced processing techniques that require the accurate removal of the primary field such as the detection of perfect conductors. Initial results from synthetic and field data are used to illustrate the potential of these routines.

A series of demonstration tests were conducted using the Z-TEM, airborne AFMAG system over deep targets in the Athabasca Basin of Saskatchewan, Canada. These tests were conducted in mid-2008 and were flown to test Z-TEM’s ability to detect large conductive targets at depth; deeper than conventional airborne EM methods. Data are presented over areas where the conductors are located 800 metres beneath the surface. As well, a case of Z-TEM following the plunge of a conductor to over 800 metres depth is shown.

TEM surveys typically operate with a simple transmitter waveform, such as a 50% duty cycle alternating square wave. The frequency of transmission and the duration of measurement is decided before the survey or, frequently in the case of ground TEM surveys, adapted during the survey by the operator to suit the conditions.

With some sensor types in particular, achieving good quality data throughout all parts of the decay is difficult. Additionally, all surveys can be complicated by the presence of external sources of interference such as power transmission lines. We argue that significant improvements can be made by optimising the frequency content of the transmitter waveform. Additionally, in the case of ground surveys, the duration of an individual reading can be controlled in order to achieve rapid production and desired data quality. Variables are the EM noise spectrum (a function of the sensor and environment) and the conductivity of terrain. These change along the TEM profile and best results are obtained by regular re- evaluation in light of the target sought.

There are several options available to modify the effective frequency content of a TEM transmitter waveform. For a conventional square wave transmitter, an irregular pattern of polarity reversals can be transmitted. .Another method is to use two or more base frequencies sequentially. The survey can be automated and data can be combined automatically into a single decay with optimal signal-to-noise ratio over the entire decay.

Theoretical examples and field data are presented to illustrate improvements in performance.

PRESENTER PROFILE

Andrew Duncan is the Managing Director of ElectroMagnetic Imaging Technology Pty Ltd (EMIT) and Absolute Geophysics Pty Ltd, based in Perth, WA. EMIT’s products include the SMARTem electrical methods receiver system, Maxwell EM software and the Atlantis borehole magnetometer tool for EM. Absolute Geophysics operates a unique total field EM system with particular utility in nickel exploration. Andrew has a background in the development of technology for electrical geophysics, EM in particular. His experience includes the development of airborne EM systems and distributed systems for geophysical measurements. One of his interests is EM techniques for highly conductive targets.

Email: [email protected]

Electromagnetic (EM) methods provide vital constraints on fundamental flow and storage processes in the earth, at different spatial scales. Recent advances in multidimensional electromagnetic modelling and inversion have brought EM methods close to their theoretical resolving power, but there are still limitations in the way that data from multi-component and multi-scale field experiments in heterogeneous geological media are currently interpreted, especially because simulating the responses of ‘field-realistic’ earth structures is computationally demanding and the presence of measurement uncertainties limits model resolution to a large extent. Also, the fact that EM methods require large amounts of data to accurately image geological heterogeneity presents further computational challenges.

In this seminar, I will present my contributions in recently developed 2.5D and 3D EM inversion techniques for subsurface characterisation, methods developed for quantifying the impact of uncertainty in the interpretation of electromagnetic data (extreme bounds analysis), a recently developed method for simultaneous interpretation of electromagnetic and seismic travel-time data for improved characterisation of geological heterogeneity, and innovative multi-scale joint electromagnetic and seismic practical experiments whose realisations are presently hindered by lack of appropriate computational platform. Strategies to overcome some of the computational difficulties will be discussed and I will draw on examples from near-surface to mantle depths.

PRESENTER PROFILE

MAX MEJU joined Petronas Research in September 2008 as a principal researcher in EM and team leader. Before then, he was a Reader at Lancaster University (2002-2008) where his team developed a novel cross-gradient joint EM and seismic multidimensional inversion method. He was a lecturer in Geophysics at Leicester University (1988-2002) where he developed EM inversion and uncertainty analysis methods. He studied Geophysics at Imperial College London and Edinburgh University, published over 40 research papers in international journals and 6 textbooks, won national and international (including the G.W. Hohmann) awards for research excellence in EM. Current emphasis is on marine EM inversion and quantitative integration of multi-physics data to reduce interpretational ambiguity.

[email protected]

As a check on frequency domain systems for calibration and drift, we aim to predict in-phase from quadrature data and vice versa. It is known that, for causal systems, the real and imaginary components of the response of the system under excitation are related through dispersion relations. The formal relationships between the components of any complex response function are the Rramers-Kronig relations, which are satisfied by any complex function which is analytic in the upper half plane, representing causality. Geophysical signals are inherently causal and as such it is in principle always possible to relate in-phase to quadrature measurements. In the frequency domain these dispersion relations take the form of the Hilbert transformation relationship between scaled in-phase and quadrature measurements which for specific ideal Earth structures are able to be solved analytically. However real Earth responses are rarely captured by these idealised models and numerical approaches are required. The major limitation on the successful implementation of the numerical Hilbert transform is limited bandwidth measurements which introduce significant noise constraints. Frequency domain measurements rely on the existence of frequency dependent responses of materials for the detection of anomalies. The response function is nominally the resistivity (conductivity) and experimental determination of this response function involves two measurements per frequency, the in-phase and quadrature components.

There are several possible applications of the prediction of in-phase response from quadrature data including 1) base level, calibration and phase checks on electromagnetic and induced polarisation systems 2) prediction and validation of noise levels in in-phase from quadrature measurements and vice versa and 3) interpolation and extrapolation of sparsely sampled data enforcing causality and better frequency-domain - time-domain transformations. This could be used for example to speed up forward and inverse modeling without loss of accuracy. In practice, using recalibrated Resolve HFEM data, in-phase data points can only be predicted using a scaled Hilbert transform with a standard deviation between 20 ppm (central frequencies) and 100 ppm (outer frequencies. This error is too large to be used reliably in any of the potential applications listed above.

Transient electromagnetic (TEM) soundings have become one of die standard methods of environmental geophysics (Fittermann and Stewart, 1986; Buselli et al., 1990; Hoekstra and Blohm, 1990; Christensen and Sørensen, 1998; Auken et ah, 2006; Lane and Pracilio, 2000). Over die past decades, airborne TEM methods have found widespread use in hydrogeophysical investigations, making it possible to cover large areas in a cost-effective way. Several helicopterbome TEM systems have been developed and they now represent die most up-to-date method of airborne hydrogeophysical investigations, collecting 100,000s of data sets.

We present a fast approximate method for one-dimensional inversion of time domain electromagnetic data and apply it to Sky TEM helicopterbome data from the Toolibin Lake area of West Australia. The method is based on fast approximate forward computation of transient electromagnetic step responses and their derivatives with respect to die model parameters of a ID model. The inversion is carried out with multi-layer models in a state-of-the-art formulation of a least-squares iterative inversion scheme including explicit formulation of the model regularization through a model covariance matrix. The method is 50 times faster than conventional inversion for a layered earth model and produces model sections of concatenated ID models and contoured maps of mean conductivity in elevation intervals almost indistinguishable from those of conventional inversion.

To ensure lateral smoothness of the model sections and to avoid spurious artifacts in the mean conductivity maps, the Lateral Parameter Correlation method is applied to the results of the individual inversion models. In this way, well determined parameters are allowed to influence the more poorly determined parameters in the survey area.

Applied to the Sky TEM Toolibin data set, the inversion produces model sections and conductivity maps that reveal the distribution of conductivity in the area and thereby the distribution of salinity. This information is crucial for any remediation effort aimed at alleviating the salinity problems. The results are in accordance with previous results from other investigators.

Despite advances in computer speed and code efficiency, the most often employed interpretation method for frequency domain helicopterbome electromagnetic (FDHEM) data is still inversion with one-dimensional (ID) models. In areas where die lateral rate of change of conductivity is small, inversion with ID models is justified and can be sufficient. These conditions are often found in hydrogeophysical investigations in a sedimentary environment.

Even for ID inversion, the task of inverting a survey volume of 100,000s of data points can be time-consuming, and more or less sophisticated approximate methods have been developed. These are often included in software packages used for data and model display and handling of geographical information, e.g. EMflow (Macnae et al., 1998), EMax (Fullagar and Reid, 2001), etc. A fairly comprehensive comparison between different inversion and transformation techniques used for FDHEM data is published in Sattel (2005).

We present a fast approximate method for one-dimensional (ID) inversion of frequency domain data and apply it to frequency domain helicopterbome data from the Bookpumong area of die Murray River, South Australia. The method is based on fast approximate forward computation of transient electromagnetic step responses and their derivatives with respect to die model parameters of a ID model, and die frequency domain responses and derivatives are found through Fourier transformation of the time domain counterparts using Fast Hankel Transform filters. The inversion is carried out with multi-layer models in a state-of-the-art formulation of a least-squares iterative inversion scheme including explicit formulation of the model regularisation through a model covariance matrix. The approximate and ordinary ID inversion approaches thus share die inversion formulation, die difference lying only in die forward mapping. The method is 30 times faster than ordinary inversion for a layered earth model and produces model sections of concatenated ID models and contoured maps of mean conductivity in elevation intervals almost indistinguishable from those of ordinary inversion.

Applied to the Bookpurnong data set, the inversion produces model sections and conductivity maps that reveal the distribution of conductivity in the area and thereby the distribution of salinity. This information is crucial for any remediation effort aimed at alleviating the salinisation problems of the river and modifying the irrigation practises in the area.

The past decade has witnessed increasing policy demands to demonstrate cost-effective outcomes in environmental management. This has brought focus on the need to understand and map the complexity of biophysical systems in order to maximise the effectiveness of targeted management investments. A rise in demand for more accurate hydrogeological predictions for salinity and groundwater management, has seen increasing emphasis on multidisciplinary geoscience systems approaches and the use of new or improved geospatial mapping technologies that enable specific gaps in the biophysical knowledge framework to be mapped and characterised with greater certainty.