Exploration Geophysics - Volume 28, Issue 1-2, 1997

Lode horizons north and south of the Broken Hill, New South Wales orebody contain a number of zones rich in sphalerite and poor in galena. One example is Pasminco Mining’s new opencut Potosi Mine, which averages 8.5% Zn and 2% Pb with little other sulphide.

The mineralisation at the Potosi Mine has been well defined by applied potential and induced polarisation (IP) surveys, but different techniques are required for deeper, downplunge exploration. Drillhole electromagnetics (DHEM) is routinely used but may miss those zones where sphalerite is almost the only sulphide. It was therefore decided to try a series of drillhole magnetometric resistivity (DHMMR) surveys, since this method responds to contrasting, rather than absolute, conductivity.

The results were very successful with the known sulphides (and some previously unsuspected zones) responding at downhole depths of more than 600 m. The method has detected mineralisation at distances greater than 150 m from the drillhole and has given clear responses to sulphides where DHEM has failed. The method has less resolving capability than DHEM, although it has proven to be more definitive where sulphide lenses occur both above and below the drillhole. The results also suggest that the magnetic field, which responds to changes in resistivity, has a much larger search radius than the phase angle, which measures the IP effect.

DHMMR is expected to play a wider role in the Broken Hill district and elsewhere, in the search for sphalerite-rich orebodies.

3-D acoustic wave propagation in a 2-D model is simulated in the frequency-wavenumber domain. Fourier transforms are used to reduce the problem of solving the 3-D equation to the solution of multiple 2-D boundary-value problems. Each of these 2-D boundary-value problems is associated with a frequency-wavenumber pair, and its discretisation via a finite-difference approximation leads to a sparse system of linear algebraic equations. The sparse linear system is repeatedly solved by an LU decomposition method for different pairs of wavenumber and frequency. The 3-D response for a specific frequency is obtained by an inverse Fourier transformation of these solutions with respect to the wavenumber. Time-domain synthetic seismograms are obtained by a further temporal inverse Fourier transformation of the frequency response.

A 2.5-D absorbing boundary condition is derived from the paraxial approximation of the 3-D wave equation. It is more efficient than the 2-D boundary condition in suppressing the artificial reflections from the edges of the computational domain. The numerical singularity is avoided by allowing the frequency to become complex. The Fourier transform with respect to the wavenumber can be accelerated by a non-uniform sampling in the wavenumber domain.

Conventional seismic processing is unable to properly image complex geological structures such as those found on the North West Shelf of Australia. Lateral variations resulting from faulting hinder complete imaging of the underlying structures. Prestack depth migration is the correct way to image such structures but requires very accurate velocity field information in order to do so. Prestack time migration and post-stack migration are quicker methods but they make assumptions that are quite often invalid. Rules are needed to identify the cases when simpler methods will work, and to decide the velocity models needed for each type of migration.

A synthetic seismic dataset was developed, based on the Oliver line from the North West Shelf of Australia, containing lateral velocity contrasts of 60%. Both time and depth algorithms were used to assess the amount of smoothing the velocity model was able to tolerate while still remaining suitable for interpretation.

Migrating with a smoothed velocity can cause two undesired effects. Events can be mispositioned vertically, and their amplitudes can be degraded. For different amounts of smoothing these two effects can be measured as a function of lateral position to evaluate the performance of the migration on the Oliver model.

Depth errors of 70 m and a 60% reduction in amplitude for a reflector at 2800 m resulted when the velocity was smoothed laterally over 200 m in prestack depth migration. The same smoothing resulted in a 40 ms time error with 50% amplitude reduction for prestack time migration. Rotating the smoothing to the vertical direction produced a 45 m depth error and an amplitude reduction of 40% for prestack depth migration. Prestack time migration produced results similar to those in the vertical direction.

Several developments aim to produce airborne gravity gradiometers with sufficient sensitivity to detect mineral deposits of several millions of tonnes at shallow depths. In a conventional ground gravity survey, a Bouguer slab correction is always necessary, with the terrain component of the full Bouguer correction often not needed unless the topography is severe. For airborne gravity gradiometers, however, the slab correction is exactly zero but the terrain corrections prove to be critical even with subdued or moderate topography.

This paper reports on research into airborne gravity gradiometer corrections, especially aspects that assessed the importance of regolith and its constraints of surface and bedrock topography, together with predicted target anomalies. Using an example of topography typical of the Western Australian shield, calculations show that densities of both regolith and bedrock need to be determined to better than 0.1 t/m³ for adequate correction to be made to raw gravity gradiometer data. Topography also needs to be defined to about ±3 m on 10 m pixels for adequate gravity corrections.

To take advantage of large quantities of new geophysical data and increased exploration interest due to the South Australian Exploration Initiative a compilation of the geophysical signatures of South Australian mineral deposits has been undertaken. This project has two stages which will be described in a two-volume publication. The first volume, scheduled for publication in 1996, summarises all publicly available data, and some proprietary data, on the geophysical signatures of the State’s mineral deposits. Separate chapters describe deposits according to the dominant commodity. These include copper, iron, uranium, lead-zinc, and other commodities such as gold, nickel, talc and graphite. The total number of deposits discussed is 44.

The second volume, scheduled for publication in early 1997, is primarily concerned with newly acquired airborne geophysical data, from the South Australian Exploration Initiative and also Australian Geological Survey Organisation and commercial surveys. The majority of the data described will be aeromagnetic, but other data will be included as required and available. For each deposit, data from the surrounding few tens of kilometres will be presented and interpreted with emphasis on lithological and structural control of the deposit. Regional-scale interpretations will also be included, again emphasising controls on known mineralisation.

A new gravity survey combined with ground-based magnetic surveys have enabled detailed modelling of basement structures in the Longford Basin of northern Tasmania. This basin began to form in the mid-Jurassic, but contains only Cainozoic sediments. The models suggest that the basin has been formed by a series of predominantly northwest-trending Jurassic normal faults which have produced a series of half-grabens in which blocks were downthrown to both east and west with dips of up to 10° W.

The basin blocks incorporate Palaeozoic to Mesozoic Parmeener Supergroup sedimentary rocks intruded by pre-to syn-fault Jurassic dolerites. These units overlie an older, geophysically distinct, Devonian fold–thrust terrane. Much of the tectonic activity from the Palaeozoic to the Cainozoic appears to have been controlled by the Tiers Fault system, a major gravity and magnetic lineament within the Tasmanian crust.

In contrast to many other Tasmanian Mesozoic–Cainozoic basins, the Longford Basin is exclusively continental, with up to 800 m of Paleocene to Quaternary fluvial sediments having been drilled in two wildcat oil wells.

The basin also contains intercalated Late Eocene to Pliocene basaltic volcanic rocks. Magnetic data indicate three eruptive centres localised in reactivated Jurassic faults. Massive flows, now lateritised, have infilled an Early Eocene to Pliocene drainage system.

A SIROTEM survey and experimental petrophysical data from drillcores have enabled a study of apparent conductance anomalies in and around the Longford Basin of northern Tasmania. Within the basin depocentre, conductance anomalies due to the thickness irregularity of highly conductive layers reflect stacking and migration of fluvial systems. Outside the basin margins, conductance anomalies reflect changes in the conductivity of drainage bedload which is a function of watershed lithology.

The anomalously conductive Tertiary sediments have three possible sources for stored salt: inheritance from erosion of conductive Triassic red bed sequences in the basement below the Tertiary sequence; relict meteoric salt from proximity to a drowned estuary; or ions released from extensive lateritisation of lavas after a period of Eocene to Miocene basaltic volcanism.

Interpretations of the gross geometry of the McArthur Basin from regional potential field data are presented as a 2.5-D component of a metallogenic geographic information system (GIS) developed for the region. The McArthur Group, host to the major HYC Pb-Zn deposit, is seen to extend well beyond its eastern limit of outcrop as defined by the Emu Fault. Units identified as prospective using lithological criteria encoded in the GIS contain all known stratiform base metal mineralisation. Such deposits are preferentially located on the periphery of the thickest accumulations of McArthur Group sedimentary rocks.

Volcanism in the upper and lower Tawallah Group is much more voluminous than its comparatively small strati-graphic thickness measured in outcrop would suggest. Over 15 km of basin fill (including volcanic rocks) is implied in some areas, but this may vary rapidly, implying considerable pre-McArthur Group structural development. A number of lineaments visible in the isochore images converge at the position of HYC, indicating bounding fault and strike-slip fault activity at this location during a large portion of basin evolution. These structures do not necessarily correspond to major regional faults interpreted from surface mapping.

Magnetotelluric impedances are estimated from electric and magnetic field data using a Kalman filtering method. Unlike most of the methods currently used in processing magnetotelluric data, this technique is a time-domain method which uses a linear, discrete-time filter to recursively obtain estimates of ground response. A three-step procedure is proposed, which allows the incorporation of remote reference data.

The Kalman filtering technique is applied to the "active interval" subset of the standard EMSLAB magnetotelluric dataset, and the results are found to be comparable with those from other commonly used analysis methods for magnetotelluric data.

The Generalised 3-D Normal Moveout (GMO) equation has been developed for a single, three-dimensionally dipping horizon, in which the velocity term has been separated from the dip term. The velocity term is the medium velocity above the chosen horizon whereas the dip term is partitioned into orthogonal dip directions. If a method could be found to establish the dip terms in the equation, then a single velocity term could be used to stack 3-D swath data, without the need for a stacking velocity analysis.

Firstly, a 3-D field method was devised in which the true dip of a reflecting horizon and its azimuth could be measured within a survey area. The method required a single horizon whose velocity was known, assuming a constant dip throughout the area. Two orthogonal swath surveys were then recorded, generating two orthogonal Common Midpoint (CMP) lines. The GMO equation then corrected data to a zero-offset coordinate origin point. An optimum value of dip and strike provided the best reflection line-up at the origin, producing the value of true dip and its azimuth for that horizon.

A physical model was used to demonstrate the method, which was able to predict dips and their azimuths to within half a degree. The two orthogonal CMP lines were then correctly stacked using the variable 3-D dip information only, without a need to perform a stacking velocity analysis, and the method shows potential for use in stacking low-fold data.

The geoid is the fundamental surface that defines the figure of the Earth. It is approximated by mean sea-level and undulates due to spatial variations in the Earth’s gravity field. The use of the geoid in regional geophysics is illustrated for the North West Shelf of Australia by removing long-wavelength geoid features, due predominantly to deep-Earth mass anomalies, in order to reveal near-surface structure. After this process, the residual geoid anomalies correlate well with known geological structures. Therefore, the geoid can provide information, complementary to other geophysical data, of the Earth’s internal structure.

Wide-angle seismic profiling along the Otway Continental Margin Transect defines the velocity structure from the continental crust onshore to the deep Southern Ocean across an area of extended and faulted basement to the Otway Basin. A significant section of Otway Basin sequences (up to 5 km) lies offshore on the continental slope overlying uppermost Palaeozoic basement (P-wave velocity 5.5 km/s to 5.7 km/s). A deeper basement velocity of 6.15 km/s to 6.35 km/s is prominent. The whole basement section thins from 16 km thickness onshore to 3.5 km at 120 km from the coast.

The lower crustal velocity of 6.4 km/s to 6.8 km/s extends down to a 1 km to 2 km thick upper mantle transition zone (6.9 km/s to 7.8 km/s) overlying the Moho at 30 km depth onshore, shallowing to 12 km in the deep ocean basin. The upper mantle velocity is 8.05 km/s, increasing to 8.3 km/s at 40 km depth.

There is no evidence of a high-velocity lower crust (>7 km/s) that would indicate that this is a ‘volcanic’ margin as defined on some North Atlantic margins. Analogues for the Otway margin include the ‘non-volcanic’ Galicia Bank off the Iberian Peninsula and the Nova Scotia margin off eastern Canada.

Artefacts in final processed 3-D seismic data volumes are becoming more frequently reported because such dataseis are used increasingly in reservoir management projects. These artefacts commonly take the appearance of amplitude variations which are not due to geology. This leads to confusion and some loss of detectability when using data in amplitude studies, as in stratigraphic plays. They may also demonstrate structural features, usually small-scale, which again do not depict geology. In this paper, a description of some causes of such artefacts are complemented with suggestions made on methodologies for minimising, if not eliminating, their effects. Practical examples of artefacts are shown to demonstrate these effects. Guidelines are given with respect to aspects of field survey design such that these designs will have minimal effect on data interpretability. The implications of recent techniques on survey design are also investigated. In particular, the introduction of DMO in the Radon domain has considerable impact on survey designs as it allows the use of wide azimuths and irregular offset sampling. Also, certain processing-induced artefacts are described, with suggestions given as to means of reducing or eliminating them.

Oil trapped in fluid inclusions in a sample of Plover Formation sandstone from the main oil zone in Jabiru-1A has been subjected to detailed geochemical comparison with production oil from the well. Fluorescence microscopy of the sandstone showed that predominantly blue-fluorescing oil inclusions occur in 69% of quartz grains, both within quartz overgrowths and on healed fractures within detrital quartz grains. There are some notable similarities and differences between the fluid inclusion oil and the production oil. Similarities include the n-alkane profiles (both maximising at n-Cis with slight odd over even predominance at high molecular weights), n-alkane to isoprenoid ratios and Pr/Ph ratios (2.7±0.04). Hopane and sterane biomarker maturity ratios are at or close to equilibrium values, typical of peak oil generation conditions. The fluid inclusion oil differs from the production oil in containing relatively lower amounts of rearranged hopanes, Ts, C29TS, C27 steranes and diasteranes. These differences are interpreted to be due to lower maturity rather than source rock facies variation. Aromatic hydrocarbon ratios confirm a small but consistently lower maturity for the fluid inclusion oil. For example, the calibrated methylphenanthrene index shows an equivalent reflectance of 0.84% for the fluid inclusion oil, compared to 0.92% for the production oil. These geochemical data suggest that the oil trapped in fluid inclusions in Jabiru-1A is from the same source rock but was generated at a lower maturity than the average of the oil now in the reservoir, and was trapped soon after initial charge mainly by thin quartz overgrowths. Further charge to the Jabiru structure was of progressively higher maturity oil.

It is commonly known that, for every magnetic field, whether observed on a plane surface or not, a distribution of magnetisation can be calculated on a surface below the surface of observation such that the magnetic field produced by the distribution duplicates the observed magnetic field. The distribution of magnetisation, which does not necessarily have any physical reality, is referred to as the "equivalent layer". The value of the equivalent layers is such that, once one has been calculated for an observed magnetic field, it can be used to recompute fields corresponding to the equivalent layer at other magnetic inclinations, on other surfaces of observation or on arbitrary spatial grid intervals. Equivalent layers can thus be used as a basis for reduction to the pole, elevation corrections and gridding algorithms. Equivalent layers can also be calculated for airborne radiometric data. The same properties apply to radiometric equivalent layers as apply to magnetic equivalent layers, with the added significance that a radiometric equivalent layer corresponding to the ground surface gives a mapping of the actual radioelement concentrations corresponding to the airborne observations. Accurate, stable, iteration-based algorithms can be used to calculate magnetic and radiometric equivalent layers corresponding to large grids of aeromagnetic and radiometric data.

In order to evaluate naturally fractured reservoirs, it is critical to assess whether natural fracture sets believed to exist at depth (eg, from surface mapping and/or seismic data) are likely to be open and productive or closed and nonproductive. The semi-log relation between stress and the closure of natural fractures is combined with the effective normal stress acting on fractures to yield:

which relates fracture closure (8), the constants in the semilog fracture closure/stress relation (k and s), and maximum effective horizontal stress (oH) magnitude with the effective horizontal stress ratio (n), and the angle between the normal to the fracture and the o"h direction (0).

This relation shows that: (i) for a given fracture, the sensitivity of fracture closure to the anisotropy of the in situ stress field can be constrained by the effective horizontal stress ratio (n); (ii) natural fracture closure is sensitive to fracture orientation with respect to the in situ stress field where n is high; (iii) the sensitivity of natural fracture closure to its orientation with respect to the in situ stress field decreases markedly as n drops; and (iv) the rate of change of closure with changing orientation is relatively low at very low and very high misalignment angles, and much greater at intermediate angles.

The forces that act on the Earth’s lithospheric plates are responsible for the stress regime within the plates at a regional scale, and thus influence issues pertinent to hydrocarbon exploration – such as the nature of fault reactivation, hydraulic seal integrity, natural and induced fracture orientation and wellbore stability. Four models of the intraplate stress field of the Australian continent have been produced by elastic finite element modelling of the forces acting on the Indo-Australian plate. All four models incorporate the push of mid-ocean ridges and of continental margins. In the four models the magnitude of the poorly constrained convergent boundary and basal drag forces are varied within reasonable limits. Despite being based on significantly different force magnitudes, regional stress orientations predicted by the three models that recognise the heterogeneity of forces acting along the convergent northeastern boundary of the Indo-Australian plate are considered reliable because they are consistent over much of the Australian continent and show broad agreement with the available in situ stress measurements.

In the absence of in situ stress measurements, for example from borehole breakouts, modelled stress orientations should be incorporated into the assessment of issues pertinent to hydrocarbon exploration that are influenced by the contemporary stress field. In the context of fault reactivation, pre-existing vertical faults striking at 30° to 45° to the maximum horizontal stress direction are the most prone to at least a component of strike-slip motion. Planes dipping in the minimum horizontal stress direction are the most suitably oriented to be reactivated in extension, and planes dipping in the maximum horizontal stress direction are the most suitably oriented to be reactivated in compression. Modelled stress orientations can also be used to predict open natural fracture orientation, with the preferred orientation being normal to the minimum horizontal stress, and to help assess the hydraulic integrity of reservoir seals, with faults and fractures normal to the minimum horizontal stress least likely to be sealing.

A new method has been developed that improves upon current methods of heterogeneous gravity data integration, with a much reduced computation time and memory requirement. Gridded datasets of satellite altimeter observations of the geoid height, and of shipboard measurements of the free-air gravity anomaly are combined through an iterated weighted superposition, using the fast Fourier transform for conversion between the datasets.

Although magnetic methods have long been used in basin studies the data are often underutilised since the formations of a relatively undeformed basin approximate sub-horizontal tabular sources except at basin edges. Horizontal sources are known to generate negligible anomalies. Very small angular deviations from tabularity, or horizontality, however, will generate very large responses at basin scale. This response is largely independent of source contrast, thickness, dip, or the orientation of either the source or the field but is very sensitive to small changes in dip differentials between upper and lower surfaces of the source.

Classical rift sequence forms may account for many of the anomalies observed in basins and no presumption of basement sources is universally justified. Anomalies must be treated with basin perspective and scale and not as isolated elements.