Exploration Geophysics - Volume 24, Issue 3-4, 1993

Alcoa of Australia’s bauxite treatment plant at Kwinana discharges alkaline liquor to impoundments constructed on Quaternary formations of the Perth Basin. Contaminated groundwater has been detected outside some of these impoundments. The impoundments are constructed with an impermeable basal clay layer overlain by a permeable, thin sand blanket. Wet tailings slurry is poured in, the fine residue fraction is dried and stacked higher than the dyke walls allowing access of tracked vehicles and geophysical equipment to the impoundments. Leakage is initially detected from monitoring bores placed outside the impoundments. Approximate leak locations can then be detected from water level lows in piezometers placed within the semi-confined aquifer of the sand blanket. The mise-a-la-masse technique uses current electrodes placed within the impoundment and within conductive, leaking water at some distance away. A current pathway is provided between the impoundment and the conductive groundwater via the leakage zone and this can be mapped as voltage highs using potential electrodes within piezometer holes near the base of the impoundment. In some cases the detection of these leakage points is complicated, but not impossible, by multiple leaks, PVC liners placed above the basal clay layer and the strongly varying conductivity within the impoundment. The mise-a-la-masse anomalies only give leak location with differences in intensity between anomalies not being diagnostic of leak properties.

Recent work using radio wave frequencies to define ore shape between boreholes shows promise for changing the way mine geologists evaluate a deposit.

Traditionally, orebody evaluation at an advanced exploration site or a mine involves three distinct tiers of information, exploration drilling (>200 m borehole separation), follow-up drilling (40 m to 100 m separation), and in-fill drilling (<20 m). the inexact nature of ore definition at the follow-up and in-fill stages inevitably results in poor mine design or dilution. this has been shown to have a significant cost to any operation. the clear incentive is to improve evaluation techniques by using high resolution sensing methods between physical intersections.

A new geophysical technique using comparative radio wave attentuation values from cross-borehole measurements and advanced tomographic imaging procedures has been applied at a number of mine sites in Australia. Radio wave attentuation is a function of the host medium conductivity. Conductive ore will attenuate the signal more than resistive host rock. These variations in signal attenuation may be expressed in the form of a tomographic image for analysis.

This paper evaluates results from controlled experimentation at the Broken Hill and Osborne deposits and reviews the likely future role for radio imaging at metalliferous mine sites.

It is suggested that the clear role for radio scanning is particularly in the follow-up stage of drilling (40 m to 100 m). The integration of the images into mine planning packages, as an input to geostatistics and as a visualisation aid for mine planners and geologists is likely to improve the accuracy of ore definition and overall metal recovery at individual mine sites.

A detailed gravity traverse is described along the Western Mining Corporation — Australian Geological Survey Organisation (WMC-AGSO) Widgiemooltha-Tramways crustal seismic profile. Positive Bouguer anomalies of up to 20 mGal over the greenstone sequence imply that up to 3.5 km of mafic stratigraphy occur along the profile. Comparisons to the seismic data imply the presence of a felsic stratigraphic unit forming the base of the supracrustal greenstone succession. The 3-D structure of local successor basins (Black Flags and Merougil Beds) above the mafic stratigraphy are also constrained by the gravity data to a maximum vertical thickness in the order of 3 km. The Widgiemooltha granodiorite is modelled by the seismic and gravity data as a steep-sided intrusion underlain at depth (~ 3 km) by felsic greenstone stratigraphy.

The geological history of the Norseman-Wiluna belt inferred from the Widgiemooltha-Tramways seismic and gravity data analysis indicates that:

2.9 Ga: Intracratonic basin of predominantly felsic composition (Penneshaw Formation and lateral equivalents) forms over preexisting continental basement.

2.7 Ga: Main Kambalda mafic-felsic stratigraphy develops above antecedent felsic basin.

2.7 Ga to 2.65 Ga: Polyphase basin inversion produces the presently observed upper crustal section.

Ground penetrating radar (GPR), of all the commonly practiced geophysical techniques, has the greatest ability to provide clear high resolution images of shallow sub-surface structure. To date, however, the perceived unpredictability of its performance at different sites has limited its use.

Factors which control the performance of GPR can be summarised by the radar range equation. The site dependence of GPR is a result of the wide variation between the wave attenuation rates of different geological materials and the variation in the reflectivities of the different targets. The attenuation rate of a material depends on its conductivity and dielectric constant while the reflectivity depends on the contrast of these properties between the target and host materials. Unfortunately, since conductivity varies with frequency, conductivities obtained by resistivity and other low frequency electromagnetic geophysical measurements are different to those at GPR frequencies.

One way to obtain these properties is to make measurements of the radio-frequency electrical properties of rock samples from prospective GPR sites. For this purpose the open-ended coaxial line technique has been found to be the most practical method. Above 100 MHz measurements can be made on small samples with a single flat surface. Below 100 MHz the accuracy of this technique deteriorates and the capacitance bridge technique, which requires small disk-shaped samples, is more suitable.

In homogeneous media, nomograms can be used to convert the attenuation rates determined from the sample measurements, to a maximum depth that can be imaged by a GPR system. The maximum depth in layered media can be determined by a simple graphical method involving summing the attenuation and spreading in each layer. Case histories from 4 different sites show that despite the uncertainties involved in making any measurement on samples which have been removed from their in-situ conditions, ranges calculated in this way provide a valuable guide as to how well GPR will perform at a particular site.

In the quest for improved resolution, the application of good statics routines is often ignored. Using coalfield seismic data, a model independent statics routine which uses the refraction events to align reflections is discussed with examples of its application. Good statics allows improved imaging of 2-D seismic reflection data, which then allows a novel work station method to produce pseudo-3-D volumes from 2-D pre-stack data. Hence subtle features, such as faults with small throws, can be well defined in the pre-stack domain, where they could not be readily observed on the conventional 2-D sections after stack.

Geophysical surveys often result in the inference of Earth properties based on the data collected from organised experiments. Such problems are known as inverse problems. The mathematical theory relating the data and the Earth properties is generally non-linear which, combined with ill-posed and underdetermined attributes, leads to ambiguous solutions. A single solution needs to be chosen for the geophysical experiment to be of benefit. Of the range of possible solutions it appears that some are better than others — simply by imposing common sense.

The entropy principle consists of selecting the solution that has the maximum entropy value. Such a solution will be consistent with all experimental data and maximally non-committal with regard to unavailable data. This solution contains structure only if it is needed to satisfy the given data. The entropy function is a simple convex function of logarithmic form. The convex nature of the function is enough to provide most of the benefits associated with maximum entropy solutions. However, a simple demonstrative example shows that the logarithmic form is essential to ensure that the solution is maximally non-committal to unavailable data.

The entropy principle also provides a natural mechanism for using prior information. Prior information is that obtained from experiences prior to the execution of the geophysical experiment. Such information is generally not of the form of a hard constraint and is frequently difficult to use. The entropy values can be measured relative to a prior estimate of the model parameters. This allows prior information to be used as default values and as a guide without restricting the inversion. The addition of useful prior information models can be beneficial to the outcomes of geophysical inversions.

A synthetic cross-hole tomography example shows the benefits of maximum entropy solutions when compared to the results of the popular SIRT inversion algorithm. The most obvious of these are the lack of spurious background structure and increased resolution of the anomalies.

Studies of Amplitude Variations with [Versus] Offset (AVO) have been used in Australia and worldwide to evaluate possible hydrocarbon prospects. This study concentrates on modelling the effects of thin beds (carbonates, shales) above and within sandstones on AVO behaviour.

Forward modelling of synthetic gathers from the Campbell-2 well has confirmed that the strongly anomalous AVO response observable over the gas and oil field is due to the hydrocarbon column in the Flag Sandstone of the Barrow Group. Modelling presented here shows that a 5 m thick shale with its top 15 m below the top of the reservoir modified the AVO response of the top of the reservoir, but this response is still strongly anomalous. However, modelling a reservoir with two 5 m thick shale beds located 5 m and 15 m below the top of the reservoir modified the AVO response of the top of the reservoir to the extent that the AVO response is not significantly anomalous.

The ground UTEM electromagnetic system was used to discover a large extention to the known massive sulphides at the Cafelandia Prospect in Goias state, Brazil. Large in-loop surveys were used to delineate a blind mineralized horizon with a 3.5 km strike length, shallow dip and high conductivity. Subsequent drilling has shown that the conductive horizon consists of bands of massive sulphides within an large layer of disseminated sulphides. The capacity of the UTEM time domain electromagnetic system in separating multiple conductors at depth, and the quantitative interpretation possible from the data collected make this transient EM system a highly effective tool in searching for deep massive sulphides in the Cafelandia Project area.

Direct waves generated as a result of field VSP surveys are used to obtain tomographically reconstructed images in anisotropic and inhomogeneous media.

A method is described to determine the anisotropic characteristic of velocity which assumes that velocity changes with depth and arrival angle. There is no need to assume any form of anisotropy. The form of velocity anisotropy can be determined by the results of the velocity inversion. Then a method is outlined for velocity tomography with an anisotropic correction based on the inverse theory of Newton’s non-iterative solution. Instead of using 2-D or 3-D ray tracing to deal with strongly vertical inhomogeneous media, the method uses vertical changes and anisotropic velocity determined from the velocity analysis as an a priori model of reference velocity. The method circumvents image distortion caused by straight line rays and the ray divergence caused by ray tracing. The method was designed for 3-D Vertical Seismic Profiling (VSP) application and to give a 3-D volume tomographic image of velocity disturbance by displaying 2-D cross-section slices in 3-D space.

Two field VSP data sets have been analysed, one which is the result of flat layers with obvious anisotropy and the other with no obvious anisotropy but obvious horizontal velocity variation. Strata logs from the two sites confirmed the tomographic results at each well position.

The method described in this paper can be conveniently used in 2-D and 3-D VSP and cross hole seismic surveys for the petroleum and mining industry.

Extrapolating recorded elastic displacement data from the surface downwards is the first step of a two-step elastic migration scheme. The second step is the image formation condition. Although an elastic (vector) reverse time migration can extrapolate elastic displacement data downwards, the subsurface reflections and the loss of energy due to the use of the two-way wave equations are unavoidable. The elastic Kirchhoff-Helmholtz type integrals can not be used to extrapolate elastic displacement data downwards in media in which the velocity changes in both the x-direction and the z-direction. The present method attempts to first obtain the P-and S-wave displacement potentials near the surface. Then the P- and S-wave displacement potentials are each extrapolated (in a scalar fashion) with the split-step Fourier method, subject to modest velocity gradients. Finally, the elastic displacement fields within the subsurface can be obtained from the extrapolated potential fields.

In order to obtain the P- and S-wave displacement potentials near the surface, the finite difference elastic reverse time migration method is used to extrapolate the recorded elastic surface displacement data downwards only a few depth intervals, assuming a constant velocity over this small depth region, so that the P- and S-displacement potentials can be computed from the calculated elastic displacement fields. Therefore the method can be used to extrapolate recorded elastic displacement fields from the surface downwards with almost no velocity limitation except near the surface (only a few depth intervals). It was tested on the numerical model data produced by a fourth order finite difference elastic forward modelling program with good results. The finite difference elastic reverse time migration method was used because a finite difference elastic forward modelling program had already been developed and could be run inverse. In addition, the use of the Kirchhoff-Helmholtz type integrals is preferable for the initial extrapolation of elastic surface displacement data downwards a few depth intervals to calculate potentials because of its high accuracy, no velocity limitation in the z-direction and no spurious boundary reflections. However such a scheme is computationally demanding and was not used in the present study.

The interpretation of a magnetic anomaly profile is achieved by solving a Fredholm integral equation of the first kind. Numerical solutions to such a problem require the conversion from the original form to a system of linear equations. The commonly used mid-point quadratic scheme for converting the integral equation into a system of linear equations is impractical because the obtained system is highly ill-conditioned and a poor approximation to the integral equation. In contrast, a strip integral quadrature method can give a very good approximation; further it enables the application of techniques of Computerised Tomography (CT) to potential field inverse problems. The new method significantly improves the accuracy and stability of the solution without any a priori knowledge and is very likely to have applications beyond the interpretation of magnetic field data.