Exploration Geophysics - Volume 29, Issue 3-4, 1998

Traditionally seismic exploration techniques have not been widely used in hard rock mining and exploration environments. To facilitate faster and more accurate shallow high-resolution seismic surveys a high-frequency electric discharge (sparker) seismic source was developed. The main design requirements of this source were that it (1) achieve rapid downhole firing, (2) fit down surface diamond drill holes (70mm diameter), (3) provide similar energy level and frequency content to a seismic detonator so as to produce useful sharp signals over a distance up to 200 meters, and (4) it should be highly repeatable and reliable.

The sparker system consists of a surface control and power source, winch cable and depth encoder, and a downhole probe. The 22kg downhole probe, which has a diameter of 60mm and length of approximately 3m, produces a discharge energy output of 480J. The innovation with this sparker is that the 60μF 5000V discharge capacitor is contained within the downhole probe, thus overcoming the problem of significant power loss through the high resistance of the cable and the effects of inductance experienced by other surface capacitor arc discharge sources. The sparker described does not include a discharge chamber, but instead relies on the saline conditions of a fluid filled borehole to enable explosive plasma bubble formation to occur.

Initial field tests in the hard rock environment of Kambalda Western Australia provided some very encouraging results. The ability of the sparker to operate in a borehole under hydrostatic heads in excess of 250m and to produce clearly received signal at distances in excess of 100m was demonstrated. The output signal of a single shot of the sparker compared very favourably with that produced by a seismic detonator under the same conditions. The firing rate in the field was found to be approximately 5 shots per minute. Repeatability of the signal was excellent.

New processing techniques for airborne radiometric data make use of the information contained in all 256 channels of a radiometric spectrum, improving the final quality obtained. However, visualisation and interpretation of the processed data require interpolation to a regular grid and current methods for doing this are generally unsatisfactory. We highlight alternative interpolation techniques (kriging, radial basis functions, tension splines, smoothing splines) that overcome some of the disadvantages of existing methods. These techniques are formulated in a common mathematical framework and can be used for exact or smooth interpolation of the processed data. The resulting grids can be made to inherit certain desirable characteristics, such as smoothness or minimum variance. Further, the framework generates a continuous model of the data that can be updated rapidly when the image is visualised at different scales. Until recently, the main impediments to the application of the technique to large geophysical surveys have been the computer memory and effort needed to solve the resulting matrix equations. We describe some recent advances that reduce the computational requirements to acceptable levels.

We describe an extension to the usual multi-channel technique that, during processing, preserves the original signal (as embodied in the 256-channel spectra) for as long as possible. We show that better images can result if the spectral components are gridded and the spectra reconstructed at the grid points. The reduction to standard 4-channel count-rates and conversion to ground concentration occur in the final processing step.

The Sub-Audio Magnetics (SAM) technique measures parameters related to electrical and magnetic characteristics of the sub-surface. A galvanic and/or electromagnetic source is used to excite current flow within the sub-surface. Measurements are taken with a rapid sampling, total field magnetometer while continuously traversing within the survey area.

By selecting the transmitted waveform appropriately, magnetic signals related to separate physical phenomena become distinguishable and can be observed within the measured magnetic field variations. Signal processing techniques are then used to obtain the following measurements:

• Total Magnetic Intensity (TMI)

• Total Field Magnetometric Resistivity (TFMMR)

• Total Field Magnetometric Induced Polarisation (TFMMIP)

• Total Field Electromagnetics (TFEM)

The measurements indicated are used to map variations in sub-surface resistivity, changeability and magnetic properties. While software and hardware limitations have previously restricted measurements to the TMI and TFMMR parameters, ongoing development of acquisition and processing software now allows all four parameters to be measured, potentially from a single survey.

Application of the SAM method to the Flying Doctor Lead/Zinc/Silver deposit showed that the mineralisation produced significant anomalous responses within the TFMMR. TFMMIP and TFEM parameters. Surface TFMMR and TFMMIP data are transformed to equivalent Magnetometric Resistivity (MMR) and Magnetometric Induced Polarisation (MIP) data, removing the data’s dependence on the direction of the Earth's ambient magnetic field direction.

A survey over an unexploded ordnance (UXO) test site showed that Total Magnetic Intensity (TMI) and transient electromagnetic (TFEM) data may be simultaneously acquired and used to locate these types of targets.

A number of electromagnetic and electrical methods have been applied jointly to map groundwater contamination near minesite landforms. The main methods being investigated are transient electromagnetic (TEM) and direct current (DC) resistivity soundings, DC profiling, self potential (SP), and induced polarisation (IP) measurements, combined with any existing hydrogeological and hydrogeochemical data.

At the tailings dam of the disused Brukunga pyrite mine in South Australia, reaction of groundwater with the tailings causes the formation and discharge of sulphuric acid. Geophysical methods have been investigated to determine whether they can be used to characterise variations in depth to watertable and map preferred groundwater flow paths through the tailings dam. The results of the geophysical surveys show that it is difficult to determine any preferred channels of groundwater flow from SP profiling data alone, but TEM and DC sounding measurements have enabled accurate determination of watertable levels and aquifer resistivity. The shallowest and most resistive part of the aquifer occurs in the southeast of the site, and we deduce that a possible source of fresh groundwater entering the site occurs here. It has been recommended that efforts to reduce acid formation in the tailings dam should concentrate on reducing this inflow of groundwater.

Regional compilations of gridded geophysical data from disparate individual surveys are playing an ever more important role in resource exploration. A key processing step in such compilations is the merging of overlapping grids to create a single grid. Traditional methods of connecting grids together can produce smooth final products but the process is time-consuming and has difficulty with differences that involve both long and short wavelength errors. A novel, completely automated method addresses several main challenges, such as determining how to select a path along which overlapping grids can be joined. The technique uses Fourier analysis to deconstruct the errors along a suture path, into a sum of functions with different spatial wavelengths, and applies corrections that propagate smoothly into the grids by a distance proportional to the individual wavelengths. The result is an almost seamless grid that minimises distortion from the correction process.

Based on the results of a large set of field data from Sirotem transient electromagnetic (TEM) soundings carried out in Taiwan, three case studies demonstrate the applications of TEM in Taiwan. The transmitter-receiver array used was the in-loop configuration with the dimensions of the square transmitter loops varying from 20 m to 400 m on a side for effective exploration depths of 20 m to 500 m. The Occam’s smooth model inversion was employed to reduce the temptation to over interpret the data and to eliminate arbitrary discontinuities in simple layered models. The inverted results were filtered and then organised into both depth slice and cross sections to study the 3D geoelectric structures of the survey area.

The first case study is located on the southwestern coast of Taiwan. The 3D resistivity images show fresh-water-saturated sand and/or gravel sediments in the northern part of the survey area, while a widespread sea-water intrusion is evident in the South. These findings are supported by more than 140 wells.

The second case study is located in eastern Taiwan. Based on the TEM sounding results, one predominant electrical discontinuity can be recognised in the Longitudinal Valley that agrees well with the known suture trace on the surface between the Philippine Sea and the Eurasian plates.

The third case study is in the urbanised Taipei Basin. The 3D resistivity images display basement topography and trace the probable faults of the Taipei Basin. These findings are supported by more than 180 wells. Therefore, TEM surveys can work around urbanised cities, where ambient electrical noise is high, and open spaces limited.

The induced magnetisation of a magnetic source is proportional to the ambient magnetic field and varies in response to natural geomagnetic variations, such as diurnal changes, storm fields and pulsations. In contrast, the remanent magnetisation is independent of changes in the ambient field. The local perturbation of the geomagnetic variations arising from a subsurface magnetic body can be determined by simultaneous monitoring of geomagnetic variations at two sites: one within the static magnetic anomaly associated with the body and another at a remote base station. Total field measurements can only provide a qualitative indication of the relative contributions of remanent and induced magnetisation to the anomaly. Monitoring of all three field components at the on-anomaly and base stations, however, allows the components of the second order gradient tensor of the anomalous pseudogravitational potential to be determined. This tensor depends only on the source geometry and the measurement location and is independent of the nature (remanent or induced), magnitude or direction of the source magnetisation.

Without making any assumptions about source geometry or location, the Koenigsberger ratio (Q), the direction of remanence and the direction of total magnetisation can be obtained from the components of this tensor. This information can constrain magnetic modelling prior to drilling and remove a major source of ambiguity in magnetic interpretation. The direction to the centre of a compact source can be determined directly from diagonalisation of the tensor. Values of Q constrain the magnetic mineralogy of the source and the remanence direction can discriminate sources of different ages or geological histories. Thus the method can also alleviate the geological ambiguity that afflicts magnetic interpretation.

Field trials of differential vector magnetometry (DVM) at several sites, including the Tallawang magnetite deposit, New South Wales, have demonstrated the validity of the proposed in situ method. However, a number of technical difficulties must be resolved before this method can be used routinely. Accurate characterisation of departures from mutual orthogonality of the components measured by each vector sensor, and the relative orientation of the anomaly and base station sensors, are crucial to the successful implementation of the method.

Very conductive ore bodies are difficult to detect using standard TEM techniques. Eddy currents are established at the surface of these bodies to oppose the flux associated with the primary field in accord with Faraday’s law of induction. These currents persist with negligible losses through resistivity when the primary field is removed. Consequently during the off-time dB/dt remains close to zero and signal levels at the surface are negligible. However an examination of the primary field, or ramp response, rather than the secondary field, will often show sudden changes in amplitude consistent with interference effects from source multiples. Ramp response data obtained for highly conductive targets have now been interpreted using models based on multiple current filaments. The results are consistent with conventional interpretations based on filament inversion theory and also with models based on a reflection of the primary field at the conductor surface.

Two methods of reducing noise in aerial survey data have been evaluated as to their performance on a set of spectra measured with a laboratory spectrometer and a section of an aerial survey. The two methods were Noise Adjusted Singular Value Decomposition (NASVD) and Maximum Noise Fraction (MNF). These are applied to the spectral data before subsequent processing to determine potassium, uranium and thorium concentrations. The MNF method was found most effective at removing noise from U and Th signals, reducing noise in U by a factor of 3.5 and Th by 1.5. No technique investigated gave an effective noise reduction for K. The MNF technique requires that a measure of the noise be obtained and this is best done using the differences of adjoining spectra along flight lines. Parameters such as the number of spectral channels to use in the analysis and the number of MNF components to retain in reconstructing the smoothed spectra were investigated. We recommend keeping around 40 MNF components to ensure minor spatially related signal is retained and using spectral data in the range 200 to 3000 keV.

Geophysical modelling combined with recent geological mapping indicates that the Koonenberry belt of far western NSW is a Mid Palaeozoic fold and thrust system.

This new interpretation relies upon the different petrophysical and structural attributes of four distinct tectonostratigraphic packages.

Package I comprises mixed, multiply deformed, Late Neoproterozoic-Late Cambrian rift and continental margin sequences. Package II comprises Late Cambrian-Early Ordovician mixed carbonate-siliciclastic facies, and has been subjected to two minor and one major deformation event. Package III contains fluvial-lacustrine red beds and volcanics with one major and one minor set of folds. Package IV is the Late Devonian Mulga Downs Group, a generally flat-lying fluvial cover sequence with restricted folding.

Fault kinematics are constrained by analysis and modelling of geophysical data, which indicate across-strike repetitions of various sequences. High spatial frequency magnetic linear features require steep surface dips, but listric character is demanded by lower frequency magnetic anomalies that are best fitted by sub-horizontal bodies at mid-crustal depths. Detailed analysis shows many anomalies are skewed to the east near the positions of suspected faults. However, the major Koonenberry Fault is a west-dipping back-thrust. These features strongly suggest that the Koonenberry belt is a west-vergent thrust package that detaches in the mid-crust. The principle fold-thrust deformation occurred during the Silurian, prior to deposition of package III, and probably corresponds to the Benambran event in the Lachlan Fold Belt. This deformation overprints a Late Cambrian event that has deformed package I.

These conclusions indicate that the Koonenberry belt is a zone of overlap between the Lachlan and Delamerian Orogens.

Seismic data quality in the deeper water regions of the Exmouth Sub-basin, Western Australia is generally excellent. Both the temporal and lateral resolution are very high, allowing the identification of fine geological detail. These data contrast with the data over much of Australia’s offshore exploration and producing regions, where multiples, velocity inversions, poor reflectivity contrasts, and sea floor reefs and channels seriously degrade data quality.

A quite different approach to seismic data processing is required for processing seismic data from the deeper waters of the Ex mouth Sub-basin than is normally used for the more noisy areas of offshore Australia. Rather than concentrating on processes that improve the signal-to-noise ratio of the data, it is more important to concentrate on processes that do not distort the seismic wavefield.

While processing seismic data recorded in 1994 in the Exmouth Sub-basin, I found that most processes typically applied to seismic data did not lead to any improvement in quality. In fact, it was quite difficult to find processes that actually improved the quality of the data. I found that the results of processing were very susceptible to the type of deconvolution applied to the data. Conventional trace-by-trace prestack deconvolution was found to introduce reverberations, degrading the quality of the final migrated section. Shot averaged deconvolution produced vastly superior results than trace-by-trace deconvolution. Other processes such as shot domain f-k filtering, trace summation and f-k demultiple had little effect on the quality of the final section. These processes, however, can introduce subtle distortions of the wavelet, and hence should only be used if it can be demonstrated that they produce noticeable improvements.

The final sequence used to process the data was extremely simple, and consisted of: (1) gain recovery; (2) large gate trace equalisation; (3) shot averaged deconvolution; (4) one pass velocity analysis; (5) NMO, mute and stack; and (6) migration followed by bandpass filtering. Probably many of the deeper water regions of Australia could benefit from the use of the simple processing sequence presented in this paper.

The SMARTem Electrical Methods Geophysical Receiver System has evolved during the last three years as a flexible new tool for TEM, IP and other electrical geophysical survey methods. This paper presents a brief description of that instrument and several examples of data collected recently in Australia.

First prototyped in mid-1995, the SMARTem receiver is now increasingly used in mineral exploration and ore delineation geophysics. Based on a rugged PC with a familiar operating system and programmed in a high level language, its aim is to increase the value of the data obtained in electrical geophysical campaigns. In addition to carrying out geophysical tasks in a graphics-rich environment it functions as a digital storage oscilloscope and spectrum analyser. SMARTem has been used in fixed-loop, moving-loop, borehole (axial and 3-component) and underground surveys in both direct-trigger and crystal-synchronised modes.

TEM data from Leinster and Kambalda in Western Australia illustrate the use being made of SMARTem in the exploration for nickel deposits in Western Australia. Data is typically collected in or around existing mine infrastructure where electrical interference from power grids and other sources is significant. Signal and data processing strategies have been developed and optimised to allow data of the desired quality to be collected at good rates of production. Software for automated acquisition and processing of 3-component borehole TEM data has been developed.

At Honeymoon Well, Western Australia, SMARTem work has been carried out over the Wedgetail Deposit – a popular site for tests of TEM instrumentation. Fixed-loop and moving-loop TEM data from the site is presented to illustrate this instrument’s performance in the mapping of this very difficult geophysical target.

Recently, SMARTem borehole TEM data has been collected underground at Mount Isa in the exploration for new deposits in the Deep Copper mine. This environment is especially difficult for TEM as a result of interference from underground equipment. Examples of data sets and processing results are presented.

Downhole electromagnetics (DHEM) is the principle geophysical tool used at Kambalda Nickel Operations for the detection and delineation of sulphidic ore zones. The case study presented here is from Miitel, a relatively new mine site, which expects to begin production by January 1999. Exploration at Miitel is a challenging proposition as mineralisation occurs in discrete blocks, which, due to limited drilling, are not yet well defined. For this reason, most holes drilled at Miitel are surveyed with DHEM to increase the investigation area and to assess the size of intersected conductors.

DHEM logging at Miitel faces a significant problem. Exploration targets are not only overlain by a conductive overburden but are also overlain by a layer of thick pyrrhotitic sediments, positioned 100m into the hanging wall. The consequence is a low amplitude response from the target and substantial overburden interference at the target area, below the sediments. These effects decrease the signal-to-noise ratios and increase the ambiguity in interpretation.

In DHEM logging, usually the optimum transmitter loop position is where coupling is maximal with the target and minimal with all other conductors. Typically this is accomplished by placing the transmitter loop in a normal coupled position where the field lines traverse a path from the centre of the loop through the hanging wall side of the target ore horizon. Reverse coupled loops couple poorly with the sediments because field lines couple through the footwall beneath them, while still coupling well with the targeted ore blocks.

Logging results from a surface drillhole at Miitel used both normal and reverse coupled transmitter loops to test the effectiveness of this non-conventional survey design. The results demonstrate how reverse coupling can successfully be used in this adverse environment to overcome the effects of amplitude reduction, current channelling, and drive delay thus improving data quality and interpretation reliability.

An understanding of the nature of possible velocity dispersion mechanisms is required for the quantitative interpretation of reflection seismic and acoustic well log data in terms of laboratory measurements of elastic wave velocities. Two mechanisms commonly used to explain dispersion in liquid-filled porous rocks are Biot dynamic poroelasticity and local flow. The results of Endres and Knight (1997) provide a basis for a consistent comparison of these two mechanisms using a single inclusion-based model to describe a porous medium. From the modeling study presented in this paper, it was found that both pore geometry and pore fluid properties significantly affect the absolute and relative magnitude of the dispersion produced by these two mechanisms.

Seismic methods are commonly used for the detection of faults in the exploration and production of coal. Surface seismic methods are not used in industry for the delineation of vertical structure, such as the imaging of dykes. This is because seismic waves transmit from the surface down to horizontal reflection surfaces, and reflect back up to the surface. Consequently, where sub-vertical structure such as dykes occurs, the surface seismic method fails.

Seismic methods can use different source-receiver geometries. Their ability to image dykes may therefore depend upon on the geometry used, the dyke thickness and the seismic wave propagation mode in relation to dyke composition and internal structure. Surface seismic methods have difficulty distinguishing between faults or fractures and very thin dykes (l–2m in thickness) when a dyke’s thickness is less than the seismic wavelength. Borehole seismic methods have to be used to detect the presence of dykes and avoid these problems.

This paper presents some results from a research project that is attempting to use both surface and downhole seismic methods to detect dykes. The paper shows how surface seismic methods imaged a sub-vertical dyke of some 40 m thickness and its associated faulting. The alternative approach of using downhole seismic sources and receivers (single borehole seismic profiling), showed that dyke sides can also be successfully imaged at depth. In the future it should be possible to produce an image of both sides of a dyke, in its correct orientation, using existing boreholes.

Recent years have seen a revival of interest in automatic interpretation of magnetic anomalies. This has resulted partly from a dramatic increase in the quantity and quality of aeromagnetic data, and partly from the development of the “improved Naudy” technique by Shi (1993) (referred to here by the term Naudy).

Application of a reliable automatic interpretation method to large aeromagnetic datasets as part of processing allows contractors to produce preliminary maps of structure and depth on a routine basis. This can be done in much the same way as maps of first vertical derivative.

In addition, the application of the Naudy technique on an individual line basis allows the interpreter to generate preliminary models rapidly for subsequent refinement by specialised modelling programs. This can reduce much of the setting up time traditionally associated with modelling long, detailed lines of data.

One critical parameter that is not determinable directly from the Naudy method is the strike of the body. In parallel with the development of an effective Naudy tool in Intrepid, we have introduced a trend detection method. We use this method to provide strike control during Naudy processing. Body strikes are inferred before the Naudy scanning takes place thereby increasing the reliability of the interpreted models by adding a third dimension. Trends corresponding to shallower and deeper structures are handled independently.

Simple two dimensional “dyke” models are constructed with attributes of strike, dip, width, depth and susceptibility. We show examples of the use of the Naudy method to provide a rapid, fully automatic, preliminary structural analysis and depth to basement map of a large petroleum exploration area. The map shows how secondary near-surface fracturing can be seen in the magnetic data. It also shows how pre-application of strike in an area where trends are well defined can significantly improve overall results.

The currently used gravity data reduction procedures were formulated in the petroleum industry and have not changed since the 1930’s. Simplifications needed to facilitate data reduction by hand have remained, in spite of the cheap and powerful computer processing that is now in common use. Corrections such as the Free Air Correction still use a linear relationship with height, even though the full latitude and height dependent formula is simple and available. More significantly, the Simple Bouguer Correction is still being used routinely in spite of the significant errors it introduces by assuming an infinite “Bouguer slab” and that the density of all rocks in the survey area can be approximated by one value. The result is that most gravity data is over or under corrected.

The Complete Bouguer Correction, incorporating a full terrain correction, is rarely applied, and only then when the topography is extreme. Yet it is this correction which promises to deliver an accuracy in data processing and presentation commensurate with the accuracies currently being achieved in data collection in the field with DGPS and modem gravimeters. The availability of high resolution Digital Terrain Models and faster terrain correction software certainly allows these corrections to be applied routinely. Even more rarely used is a variable density model. Obtaining such a model is the next major hurdle to be overcome.

Density models may be classified into three types: assumed, statistical, and inferred. Assumed densities are typically based on measurements of hand specimen samples and extended to the entire survey area. Statistical density models are based on minimising the correlation between topography and gravity whilst inferred densities are derived from a transformation of the actual gravity data.

The Hamersley Basin of Western Australia is a hostile environment for the gravity method. Precipitous cliffs of high density banded iron formations abut low lying plains of unconsolidated sediments. Exploration for iron is focussed on this topographically challenging range front. The strict application of the Complete Bouguer Correction coupled with a variable density model allows the gravity method to be used as a direct detector of iron deposits in spite of the obvious problems.

The ocean-bottom seismograph (OBS) data recorded in the Petrel sub-basin with a 100 m shot interval were not spatially aliased with respect to prevailing velocities and frequencies. This enabled the utilisation of digital seismic processing techniques not normally used to process refraction and wide-angle data. F-k filtering was used to enhance signal/noise ratio at large (tens of km) offsets. Depth migration of wide-angle reflections utilised the interpretation of signal dynamics to supplement the travel times-based interpretation conventionally used for refraction and wide-angle data. The project provided valuable velocity information to better constrain the depth conversion of AGSO’s regional deep reflection profiles in the region. Prominent reflectivity seen in the conventional reflection data at two-way times (TWT) greater than 4 s does not correspond to any velocity increase imaged by refraction/wide-angle techniques. On the other hand, the most significant velocity increase, which occurs at the Moho, does not produce high-amplitude near-vertical reflections. Interval velocities estimated from the conventional reflection data at TWT greater than 2 sec appear to be up to 1 km/s lower than those derived from the OBS data. If the former are used to depth convert reflection data, then depth to seismic boundaries in the centre of the basin at TWT 6-9 s would be underestimated by up to 2 km. Seismic reflection and refraction techniques are complementary to each other, and both are required to fully interpret the data.

In applied potential or mise à la masse surveys current is injected into an orebody in one drillhole and the resulting electric potential is measured as a function of depth in another drillhole, or as a function of horizontal position over the ground surface. The crosshole measurements are intended to establish the continuity or otherwise of the conducting ore. One can imagine the situation where metallic ore is encountered in both holes, but it is not known whether the holes intersect the same continuous orebody, or whether it is broken between the holes, or whether the intersections are of different ore surfaces. Alternatively, ore is intersected in only one hole and the question is what can one infer from DC electrical measurements about the extent of the conductor between the two holes.

In this paper we present a formulation for calculating the electrical potential distribution in an inhomogeneous 2-D or 3-D earth for any number of current electrode sources or sinks. Numerical modelling has been carried out for various classes of ore body structure, to understand the effects of conductor continuity, depth, thickness, dip and irregularity on the applied potential response. The effect of current electrode placement inside and outside the conductor, was also studied. It is possible from the shape and amplitude of the potential profiles to partially discriminate between continuous, terminating and faulted conductors. The modelling is especially useful as an aid to interpretation of field measurements and in the design of applied potential surveys.