Exploration Geophysics - Volume 19, Issue 1-2, 1988

Microgravity and EM resistivity surveys were carried out on three areas for the purpose of site investigation, to detect any possible cavities in a karstic area where a water supply project is to be erected. The survey also included measurements near the project site, along two profiles crossing two previously drilled boreholes in which cavities were found. Gravity lows and resistivity highs coincided with the positions of known cavities. The survey on the three areas indicates possible near-surface cavities ranging in depth between 2.5–7 m and extended 3–4 m. These depths have been estimated by using inverse problem formulae and two-dimensional interpretation of the gravity anomalies. Meanwhile, comparison of the gravity and EM-resistivity maps help in delineating near surface cavities.

Future seismic activity forecasting, using the algorithm of McGuire based on Cornell’s probabilistic technique, has been used to evaluate ground acceleration and return period at the site. It is found that for 100 years return period, the value of peak ground acceleration is 0.05 g.

From these results it is recommended that detailed geophysical investigation prior to any drilling be carried out. This would locate the weak zones and, with drilling information, allow a better geophysical interpretation.

The down-hole magnetometric resistivity method (DHMMR) is a pseudo-DC grounded dipole geophysical survey technique which allows absolute direction to a conductor from a borehole to be established. Anomalies produced by elongated block conductors are similar in shape and can be modelled in a similar way to gravity anomalies produced by block-shaped bodies having a density contrast. The anomalies are therefore simpler in shape than TEM anomalies and are easier to recognize in areas of geological complexity. The DHMMR technique is faster and more directional than downhole TEM but has poorer resolution of conductor size and shape, and hence its ideal role is as a complement to downhole TEM surveys.

Case histories from two volcanogenic base metal prospects, near Charters Towers, Queensland, demonstrate the directivity of the method in the search for off-hole massive sulphides.

Wave test seismograms from an Aramco concession area show criss-cross patterns of noise which mask the reflections. The criss-cross pattern that follows the arrival of the Rayleigh wave is sometimes very strong and entirely masks any reflections that may be present. This zone is very often referred to as the ‘noise cone’. These noise patterns are also seen at larger offsets at times much earlier than the arrival of the direct Rayleigh wave. These patterns are believed to be due mainly to the scattering of Rayleigh waves into Rayleigh waves in the noise cone, and due to the conversion of refractions, multiply-reflected refractions, reflections and multiples into Rayleigh (or shear) waves outside the noise cone. A computer model was constructed using a random distribution of scatterers. Each scatterer was assumed to behave like the spherical obstacle in Lamb’s problem of acoustic scattering. A source was excited and the scattered impulses were recorded on an array of seismometers. The convolution of the recorded data with a typical wavelet resulted in a section that resembles the noise cone remarkably well. This suggests that the criss-cross pattern may be due to scattering by a large number of randomly distributed inhomogeneities or by a ‘perfectly’ random medium.

Data from three seismograph arrays in the Northern Territory are used to investigate the velocity structure of the upper mantle above the 400 km discontinuity. These arrays are the permanent 26 km aperture Warramunga (WRA) array southeast of Tennant Creek, a 90 km aperture temporary digital array northeast of Tennant Creek, which includes a closely spaced sub-array (RDA), and a 500 km aperture analog-tape array straddling the two medium aperture arrays. Together, these arrays have 66 stations. The WRA and RDA arrays have receiver spacing of 2 km and 5 km, respectively, and are separated by 100 km. Both arrays can be phased to enhance arrivals with a given slowness and azimuth. A composite record section is made with seismograms from 17 shallow Indonesian earthquake sources 1250 km to 2150 km distant. Receiver spacing of less than 5 km is achieved over most of the record section as a result of the medium aperture arrays, while phases from individual events can be tracked at 40 km intervals across the 500 km analog array. This record section overcomes shortcomings of some previous studies which relied on either sparse, very-large-aperture arrays where phase correlation across the array was difficult, or on WRA data alone where phase correlation among different earthquakes was uncertain. Beam forming is used to examine the character of the reflection from the 400 km discontinuity for two earthquakes 1600–1700 km away. Differences in beam-formed seismograms at the WRA and RDA arrays for one event suggest lateral heterogeneity near the 400 km discontinuity.

The offshore Otway Basin is a difficult area for seismic exploration, due to the presence of shallow, high-velocity carbonates. Both the shallow and deep-water parts of the basin were investigated in a single regional survey done by the Bureau of Mineral Resources vessel Rig Seismic in 1985. The survey and processing parameters used to obtain the best results for such a wide-ranging survey are discussed. A mixed group length cable was used to improve shallow resolution. By using an intensive approach to processing, good results were obtained on the shallow water shelf areas, and deep crustal features were well resolved in deeper areas. The processing sequence used F-K filtering in the shot domain routinely, with signature deconvolution and 2D filtering for multiple suppression in deep water.

A recursive operator based on evaluation of the sense of local curvature, named the ‘terracing operator’, transforms smoothly varying potential-field data into a step function, consisting of steeps and flats. The terraced function, rescaled to units of physical property, provides a first approximation for data inversion from which a satisfactory final solution is quickly derived. The method is directly applicable to gravity data and can be applied to aeromagnetic data that have been transformed to pseudo-gravity. The objectives of the method are somewhat akin to those of susceptibility mapping, but differ in that the terraced function, like a geologic map, favours uniform physical property domains with sharp domain boundaries. It suppresses both ringing at the boundaries due to Gibbs’ phenomenon (an advantage) and gradational physical property variation (possibly a disadvantage).

The Olympic Dam copper-uranium-gold deposit coincides with a major geothermal anomaly. Excess heat flow is expressed by elevated geothermal gradients in the flat-lying Cambrian and Late Proterozoic sediments which unconformably overlie the mid-Proterozoic basement (and the ore body). The Tregolana Shale within this sequence is assumed to act as an impermeable layer in which the lateral variation of thermal conductivity is negligible. Geothermal gradients in the Tregolana Shale range from 83.6°C/km at Roxby Downs to 52 ± 8°C/km elsewhere on the Stuart Shelf. Thermal conductivities were determined in representative core samples allowing estimates of heat flow ranging from 160 to 100 mW/m² respectively. Most of the anomalous heat flow can be attributed to the radiogenic decay of uranium concentrated in the orebody. However minor anomalies in heat flow are observed at depths of 750 to 1000 metres. Anomalies of 65 mW/m² at Roxby Downs can be explained assuming a one-dimensional model incorporating a 300 m thickness of 0.08% uranium oxide consistent with the published reserves.

Data structure and its management is of strategic importance to many organisations in the geosciences. As the complexity and quantity of data being handled grows, it is imperative that efficient and reliable means be provided for:

• (a)   data communication, between different organisations or on one computer system;
• (b)   data archiving and retrieval, so that the high cost of information may be protected; and
• (c)   data access, to enable efficient computation using the information.

Different data structure may be required for different purposes. The complexity of the data structure definition depends on the breadth of users and systems requiring access to the data. Computational data structures must be efficient, eg arrays in Fortran. Data structures for archival and retrieval purposes need to be sufficiently well defined that a change of computer environment is not hindered by the archived data. Data structures for communication are intermediate in complexity between the above.

Data Base Management Systems provide a means for storing data so that it persists between individual applications. The major benefits of such systems are that DBMS’s promote data sharing and impose standards so that the data can be easily converted into a form suitable for a particular requirement.

In shallow site investigations information is often required on both the thickness of overburden material and its physical properties prior to drilling the site. While conventional compressional wave methods may resolve the former, shear wave refraction shooting can also provide information on the sediment’s shear deformation characteristics. Furthermore, in some instances the shear wave is more suited to resolving depth to bedrock due to its characteristically large velocity contrast at a sediment/rock interface and thus shear wave methods should be given due consideration when planning any geophysical site survey.

Data from a research exercise on a potential opencast coal site clearly illustrate the case for shear wave refraction shooting. However, the accuracy of bedrock depth determination is shown to be dependent on interpretation procedures, the input of a shear wave velocity gradient function for the overburden being essential to the interpretation.

Given the availability of shear wave velocity data other engineering applications become apparent. In reconnaissance surveys such as the above they may be used in assessing the rippability of the overburden, while in more comprehensive site investigations, the geophysically-derived parameters from crosshole shooting may be used in the design of structural foundations.

Detailed gravity surveying and high resolution seismic profiling have been used in the Lake Vermont area in central Queensland in an attempt to establish structural control.

There are two coal seams of interest, the Leichhardt seam, 2-3 m thick, varying in depth from subcrop to 80 m, and the Vermont seam, 5-7 m thick, varying from subcrop to 130 m deep. They are cut by a series of faults trending N-NW with throws up to 15 m, and a less frequent pattern of faulting oriented E-NE.

A micro-gravity test survey along three lines in 1984 attempted to identify areas of faulting in these coal seams. A pattern of anomalies was correlated across the lines in the expected strike direction of the faults, but specific structures could not be modelled.

Following another phase of drilling a short seismic reflection survey was shot over the same test area in late 1986. The shallower Leichhardt seam was not able to be resolved but the deeper Vermont seam was well defined. The seismic verified structures interpreted from drilling, altered the interpretation of some structures detected by drilling and found additional features missed by the detailed drilling programme.

At the Centre for Geophysical Exploration Research at Macquarie University, computer programs have been developed to model the propagation of SH and P–SV waves in coal seams. These computer programs are based on the finite element method and include allowance for irregularities in the coal seam and for absorption in the coal and rock. In the longwall mining method, it is of the greatest economic importance to be able to detect faulting and other serious interruptions of the coal seam well ahead of the mining face. Scientists of the Broken Hill Proprietary Co. Ltd have conducted surveys of the transmission of SH and P–SV coal seam waves both by borehole observations and underground roadway observations. At frequencies below 200 Hz, interruptions of the coal seam are not readily detectable by SH and P–SV seam waves. At frequencies above 200 Hz, it is not easy to distinguish the effects on the transmitted SH and P–SV seam waves of interruptions of the seam from the effect of more than usual absorption in the seam. Thus, at a frequency of 400 Hz, for a coal seam offset one seam width by a vertical fault, finite element modelling shows that the effect of the fault is to reduce the amplitude of the SH fundamental mode by 82 percent. Allowance for a Q value of 25 in the coal seam and of 50 in the rock above and below the seam indicates a further decrease of 11 percent in only 24 m. Similar results have been obtained for the P–SV M23 mode in the same finite element model and for SH and P–SV modes in a model of a coal seam intersected by a vertical hard dyke. Because interruptions of the seam scatter SH and P–SV energy back and above and below the seam, it is desirable to make both reflection measurements and transmission measurements near the central plane of the seam to distinguish more readily the effect of a seam interruption from that of high absorption.

Conventional downhole EM surveys measure only one component of the EM field aligned with the drill hole axis. Consequently interpretations are often inconclusive. Three-component downhole TEM surveys can overcome the rotational ambiguities associated with current filament inversions. However, suitable probes are difficult to construct. New sensors have now been developed incorporating sensitive preamplifiers. Orthogonal components can be measured sequentially and the orientation of the probe can be determined using an absolute gravity reference. The system has been tested in a ground traverse over a major conductor and the results are consistent with the response obtained using coincident loop techniques.

Tomographic estimation of seismic velocity from directly transmitted arrivals in offset VSP or crosshole data requires the following processing operations: (a) statics estimation and corrections; (b) stacking to enhance arrivals of interest and to remove coherent noise; (c) deconvolution and time picking; (d) iteration between tomographic inversion of picked source-receiver delay times and ray tracing. Velocity tomography with reflection data may also require migration to estimate the position of reflectors for picking two-way travel times. A workstation based software system has been developed which permits interactive tomographic velocity estimation from shallow crosshole and offset VSP data. The flexibility gained by performing velocity tomography interactively has many advantages, such as easy comparisons of intermediate results and the ability to update filtering sequences to accommodate the results of interactive analyses of signal and noise characteristics. We illustrate this for shallow crosshole velocity tomography with examples of (a) interactive tau-p filtering for arrival enhancement and coherent noise removal; (b) the dependence of tomogram resolution on raypath coverage; (c) interactive smoothing to compensate for resolution artifacts related to incomplete tomographic raypath coverage. The use of interactive smoothing to cater for different types of geological prior information in tomography is discussed.

Vibrator sweep signals, transmitted into the subsurface, contain harmonic components due to nonlinearities in the vibrator system and due to the ground’s non-linear response to the vibratory force. The harmonic energy produces a distortion in the correlated records and appears as a noise train displaced from each primary reflection. The noise trails the primary when the pilot sweep begins at a high frequency and ends at a low frequency (downsweep). For upsweeps, the noise appears as a forerunner. This paper evaluates the characteristics of harmonic ghosts in terms of time of arrival, strength and frequency content. Adverse affects of the distortion, on adjacent reflections, are shown to be minimized by: (1) increasing the sweep length, (2) reducing the sweep bandwidth, (3) utilizing an upsweep, (4) stacking alternate polarity sweeps, (5) stacking variable phase sweeps and (6) stacking variable frequency sweeps. Advantages and disadvantages of each attenuation method are evaluated.

The audio range of high fidelity equipment is extremely wide, from 2 Hz to 15 kHz approximately. A single audio speaker cannot maintain efficient operation across the spectrum, so in order to compensate for the losses in efficiency, hi-fi enthusiasts use three speakers with over-lapping spectral operation. Thus, a broad-band signal is produced using a ‘woofer’ for low frequencies, a ‘squarker’ for medium range frequencies, and a ‘tweeter’ for the higher frequencies.

The useful seismic spectrum is limited to a range of 2 Hz to 400 Hz. The range of 2 Hz to 120 Hz is of particular interest to the oil explorer with present technology. In order to image reservoirs successfully, methods of signal enhancement and noise reduction have been designed to improve low signal-to-noise ratios.

In order to reduce the harmful signal-masking effects of surface waves, geophone strings spread over the surface waves’ length are used in an attempt to attenuate coherent ground roll. Where waves are dispersive, a higher valued geophone coil resonant frequency may provide assistance in attenuating such waves. Alternatively, in vibrator work, a change of sweep frequency character may assist not only surface wave reduction (by simply not sweeping over the surface wave frequencies) or a concentration of sweeps in the target zone spectrum may enhance reflected signal strength. One innovation — that of Varisweep* — allows a change of sweep frequency with vibrator position, thereby using the source array sweep frequency design to enhance the desired signal, and yet use the source array length to attenuate long wavelength coherent noise.

This paper proposes the combination of the Varisweep approach of source design to receiver design, such that the receiver array improves the spectral response in terms of both signal and noise.

An appreciation of errors is a prerequisite for scientific interpretation of any experimental result. In seismic deconvolution three main sources of error can be identified: ambient random noise, coherent noise, and uncertainty in the wavelet. In this paper a deconvolution method is described which (1) takes wavelet uncertainty into account during inverse filter design, and (2) determines the error in the deconvolved output. Wavelet uncertainty is characterised by the time domain covariance matrix for wavelet deviations and the vector of frequency domain wavelet variances. Thus a prerequisite for implementation of the method is a representative suite of source wavelets. This requirement is most readily satisfied for offshore data, where far-field signatures can be measured or calculated. Thus the method is introduced in the context of marine seismic, and illustrated with a synthetic example. Deconvolution with respect to a time-invariant far-field signature provides an estimate of the impulse response function. The wavelet-related errors in the impulse response model are model dependent and hence time dependent. Their distribution provides an immediate indication of the likelihood and size of errors in event amplitude, time, and shape.

It is known that a stack of traces of multiple coverage gather is usually accompanied by loss of resolution. Resolution could be essentially improved if sources and receivers are distributed on the free surface in such a manner that all traces in a gather correspond to one specular point. The corresponding method, called the Common Reflecting Element (CRE) method, is presented in this paper. The CRE method includes (1) a special binomial asymmetrical distribution of source-receiver pairs along a spread; (2) a special dynamic correction, called the Oblique Spherical Correction (OSC), depending on local characteristics of a medium and wavefront, emitted by a fictitious source located at the CRP on a reflector; (3) a special procedure of generating synthetic traces of so called corrected quasi-focusing gather, corresponding to binomial root distribution of the pairs and to the OCS, if sources and receivers had been distributed in the field arbitrarily and if the parameters of the binomial root distribution and of the OCS are known; (4) a special procedure of search of synthetic traces of an optimal corrected quasi- focusing gather, if sources and geophones had been distributed in the field arbitrarily and if only sets of possible values of the parameters of the binomial distributions and of the OSC are known.

The comparison of distribution of sources and receivers and of NMO, calculated by ray tracing program, with the binomial distribution and the OSC, shows high accuracy of the obtained formulas in CRE. Modelling testing of the principal blocks of the proposed procedures indicates that they should operate well.

The Trough Tank gold-copper prospect lies approximately 250 kilometres south-east of Mt Isa in Queensland, Australia. The prospect was initially located by airborne magnetics. Reconnaissance drilling located quartz-magnetite rock with anomalous gold beneath 30-40 metres of Mesozoic cover. There is no bedrock or anomalous geochemistry associated with the anomaly and so it represents a true ‘blind’ target.

Ground magnetics, induced polarisation, fixed and moving loop time domain EM and drilling have all been employed to detail the anomaly. Drill core laboratory measurements to determine physical properties revealed minimal remanence but strong demagnetisation. The structural complexity of the prospect plus the demagnetising effects made initial drill siting difficult with high magnetite content producing extremely high magnetic amplitudes. Important modifications to published algorithms and modelling programs were found necessary to calculate theoretical anomalies which correspond to measured anomalies

Dipole-dipole IP over the area defined a broadly anomalous chargeable zone coincident with the postulated sources determined by the ground magnetics. Fixed and moving loop EM using Zonge and SIROTEM equipment permitted greater resolution of the main conductive lithologies and a reasonable structural interpretation was developed. Drilling on EM and magnetic targets to date has intersected up to 41 metres of highly conductive, pyrite-chalcopyrite bearing, quartz- magnetite rocks.

Two of the distinct characteristics of the deep seismic reflection profiles from the Precambrian Arunta Block, Central Australia include the high frequency content (up to 60 Hz) of the reflections at two-way travel times of 6-10 s and the arrival of reflected energy to at least 14 s two-way time with the absence of a ‘non-reflective’ upper crustal zone.

The data quality is generally good, though individual reflections are short in horizontal extent. Typically reflector segments are continuous over 10-30 traces which makes correlation difficult. Different stacking techniques have been used to try to increase the distance range over which correlations can be made, but with limited success over a CMP stack.

The relatively high level of reflected energy from considerable depth suggested the development of an energy stack. The method is similar to ‘reflection strength displays’, but instead of modulating a stack with the envelope of the energy density, the energy within a specified time gate is displayed. The energy stack should be applied after equalizing the effective size of each shot to reduce contamination by near surface effects. This approach has a number of advantages:

(a) the interpreter can start to look at packets of energy rather than the normal amplitude display, and the display is not affected by signal polarity.

(b) the display enhances deeper events by making them stand out more clearly from a noisy background.

(c) the energy display enables an easier correlation of events over greater distances.

The events seen can still be discerned on conventional sections, but their clarity can be markedly improved.