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

In the western Anadarko basin (Fig. 1), the lower Pennsylvanian Upper Morrow sands are both a prolific, and an elusive exploration target. Initial production from some of these sands can reach over 1000 BOPD and yet an offset well just 1000 feet away from a prolific producer can miss the sand entirely. This presentation explores the use of high resolution seismic data along with advanced geophysical modelling, seismic inversion processing, and seismic facies mapping, in detecting the sands before the drill bit.

Two Upper Morrow fields in the Texas Panhandle have been studied: the Lear and Darden Fields. The Morrow sands reach an isopach thickness of ten to fifteen metres at a depth below surface of 2500 to 3000 metres. The sands are within the ‘thin bed’ regime. They are also below the ‘tuning point where there is a linear relationship between the amplitude of the seismic reflection and the thickness of the sands. The sands have an interval velocity around 1000 metres per second faster than the encasing shales and are detectable on good signal to noise ratio seismic data. The comparison of geologic isopach mapping and geophysical seismic facies mapping shows a good correlation in the delineation of the Upper Morrow sands.

The use of these seismic stratigraphy methods should substantially increase exploration and development success when high resolution seismic data and advanced interpretation techniques are employed.

The objective of seismic exploration is to image the earth’s subsurface in order to improve the success rate of drilling wells. Since hydrocarbons are increasingly being found in subtle and complex traps, it is important to use all the information available for interpretation and to use it as accurately as possible. This implies combining all the surface seismic data with information from other sources, such as well logs and VSPs to provide an integrated interpretation. Moreover, this interpretation must usually be done within tight time constraints.

Interactive workstations organise and manipulate large quantities of data in a convenient and efficient manner, permitting either a faster interpretation, or more likely, a more detailed interpretation within a given time. The system described in this paper incorporates Landmark and Discovery workstations. It provides completely interactive structural interpretation, fault analysis, digitizing and final mapping. Well log, VSP processing and seismic attribute analysis may be performed on the workstations and the results incorporated in a detailed interpretation. Interactive modelling is also available. The complete system may be used for 2-D or 3-D datasets.

The North Island of New Zealand provides the setting for the majority of land based exploration. Most recent successes have occurred in overthrust structures on the Eastern margin of the Taranaki Basin. A combination of steep dips, deep targets and elevation induced static problems provide a challenge to quantitative seismic interpretation and mapping. Notable success in the area are the Mckee field (production 8000 bopd) and the offshore Maui gas field (360 million cuft/d).

The vertical VSP has been used in the area since 1982 for prediction ahead of the bit and qualitative dip analysis. The VSP data was typically processed on the wellsite and occasionally hands-on interpretation assisted in immediate sidetrack or plug and abandon decision making. In this role the VSP had only limited success mainly due to the early limitations of the wellsite processing along with the complexity of the problem.

In 1986 and 1987 several offset VSP (OVSP) surveys were recorded to further investigate the usefulness of the VSP technique.

This paper describes how the structural interpretations of the commercial gas production field ‘Tariki’ and the accompanying ‘Ahuroa’ field were aided by the addition of VSP/OVSP data. Particular reference is made to earlier data acquired with single axis well geophones in comparison with late 1986/1987 data acquired with triaxial well geophones.

In general, the OVSP/VSP results in a high quality seismic image of pre-estimated spatial extent. A new method of interpreting OVSP data via synthetic modelling is put forward as a practical solution in areas of exceptionally high dip.

The problem of interpreting long range crustal refraction profiles is usually addressed by a conventional estimation of the velocity-depth model from the travel time data followed by iterative forward modelling. In principle, the objective of this technique is to find the crustal model which gives the best agreement between the calculated and observed travel time data.

Recent advances in the design of ocean bottom seismometers (OBS) have facilitated the use of a large number of closely spaced OBS along such profiles. This has increased the resolution of the technique and the constraints put on the calculated models but at the same time increased considerably the complexity of the interpretation process and the computation time.

In order to reduce the tedious iterative interpretation process to manageable proportions, we propose to apply a development of the coherency inversion method, suggested for use in exploration reflection interpretation by Landa et al. (1987) and Koren et al. (1987), to the evaluation of long range crustal profiling data. This development is designed for the estimation of the velocity sequences and the geometry of the interfaces using selected record sections for a given profile. We are concerned with two-dimensional heterogeneous media and data which contain arrivals of reflected and refracted waves. The method is based on an iterative algorithm producing a model which maximises some measure of coherency; the latter is computed within a time gate for a record section along travel times generated by tracing rays through the model. This algorithm has the advantage that event picking on the record sections is avoided and it is not based on curve fitting of hyperbolic approximations of the arrival times.

The method is illustrated by a synthetic example as well as actual refraction record sections.

The finite-difference method is a well-established tool for the forward modelling and migration of seismic wavefields in complicated geological media. One problem, however, has been that one could not include the effects of intrinsic attenuation because of numerical limitations. It has been suggested that a Padé approximation to the viscoelastic moduli turns the equations into a numerically tractable form. To achieve this approximation we suggest an approach which is based partly on physical considerations and partly on numerical curve fitting, and is superior to earlier approaches both in accuracy and computational efficiency. Our finite-difference programs for acoustic and elastic wavefields only require about twice as much computer time and storage as for a calculation without absorption. We have carried out numerical simulations of channel waves in discontinuous coal seams and have investigated the reflection and transmission responses of different types of discontinuities for models with and without intrinsic attenuation. The results show that for the most likely range of attenuation (Qcoal~50−100) serious damping of the higher frequencies occurs, especially of the fundamental mode Airy phase. The differences in the reflectivities and transmissivities of different discontinuities are reduced considerably. This makes it much more difficult to detect faults and to discriminate between different fault types and throws than is expected from previous modelling of seam waves without attenuation.

Application of the stress/strain ellipsoid as defined by Harding (1974) to fractures observed in the Australian continent proved to be unsatisfactory. In a new modified ellipsoid, thrust faults are postulated to be bounded by both the synthetic fault, and an additional sub-parallel ‘K'-lineament which has an opposite sense of displacement to the synthetic. Thrusting occurs when these bounding faults are convergent, and where they are divergent there is extensional faulting. With this modification, a single E-W right-lateral stress/strain ellipsoid can account for the orientation of more than 90 per cent of the main fault systems in the Australian continent. Many other faults can also be accounted for when Tertiary E-W compression is superimposed on older fractures. Detailed fault analysis using the classical ellipsoid in producing Cooper/Eromanga Basin fields reveals en echelon sub-fractures as predicted by theory. However, some have opposite strike movements which are explained by the ‘K-lineament.

Recent oil discoveries in the Jurassic-Cretaceous Eromanga Basin of south-west Queensland have been in a variety of zones and many are believed to be controlled mainly by stratigraphic variations.

Talgeberry-1 was completed as a dual zone oil producer from the Cretaceous Wyandra Sandstone and mid-Jurassic Birkhead Formation. An appraisal well, Talgeberry-2, drilled at a structurally higher location, did not intersect either of these reservoirs but was completed as a producer from the Early Cretaceous Murta Member.

This paper presents a model for the Birkhead reservoir, which has been the most prolific producer to date.

At Talgeberry-1 the Birkhead reservoir is interpreted as a fining upwards channel fill sand overlain by a coarsening upward distributary mouth bar sand. The channel fill, with good reservoir parameters, is believed to be the main producing sand. Log analysis and core data support this interpretation. The 10 m sand has a recognisable seismic signature which has been confirmed by modelling. Mapping of this signature has identified a channel approximately 600 m wide meandering from east to west.

Structural mapping shows an updip location but it would have poorly developed reservoir sand and should be avoided.

Further drilling for development or pressure maintenance should be within the mapped channel.

A seismic inversion procedure has been developed that inverts data available from an unmigrated stacked section to produce an interval velocity model. The stacking velocity information is not used directly but incorporated in the determination of two-way reflection times for the offset rays believed to have greatest influence in the initial determination of the stacking velocities. This modification addresses the problem of non-hyperbolic moveout curves. The use of a generalized linear inversion technique addresses another difficulty inherent in many existing interval velocity modelling methods: velocity errors cumulative with depth. Lateral variations in the interval velocity are also permitted in our procedure.

The inversion procedure is applied to seismic data from the Gippsland Basin. At each location of interest, the input data consists of three stacking velocities and five horizon times per layer. The output from the inversion procedure consists of three depths, the interval velocity and the interval velocity gradient for each layer.

A significant reduction in the data mismatch is found relative to earlier results in which the moveout curves were assumed to be hyperbolic. This indicates that our treatment of stacking velocities is more appropriate when analysing the velocity structure of such regions.

Micro ground tremors were observed by an array of multiple 3-D geophones of 1 Hz with a short spacing along a traverse in a volcanotectonic zone of northeastern Japan. Instead of line spectral methods by single observation point for earthquake engineering purposes, a continuous spectrum was processed by cross correlation and 2-D filtering in time and then in the frequency domain. In addition to the power spectral profiling, 2-D representation of ground motion vectors assists in understanding a realistic structure. Frequency to depth conversion was made by using the law of a quarter wavelength in assumption of velocity distribution of S- or P-waves. Advanced discussion is proposed on this point considering dispersion of surface waves dominant in the tremors, especially of the longer period. As an example of interpretation of the results, geomagnetic intensity was correlated in detail, and the volcanotectonism was seen as double caldera walls (post Miocene).

We have developed a microcomputer-based unique seismic profiler which has a function of programmable time-variant gain (TVG) and time-variant filtering (TVF). An aim of the system is to improve the quality of single-channel analog seismic profiler records. The system consists of a voltage controlled amplifier for the TVG, low-pass and high-pass voltage tuneable filters for the TVF, D/A converters which convert digital control data to voltage for TVG and TVF, a sea bottom detector, and microcomputers. An operator can freely set the condition of TVG and TVF by programming of the control data on the microcomputer, and the TVG amplifier and TVF filters are controlled according to the programmed data. TVG and TVF control starts from near sea bottom using the sea bottom detector or external digital depth data. The control data can be changed at any time during survey. We named the system PRO-SEA II (Programmable Seismic Amplifier Two).

A field test of the system was successfully carried out. Resolution and penetration depth of the seismic records were improved remarkably compared with conventional singlechannel analog records.

Three dimensional (3-D) seismic surveys became commercially available in the late 1970s and their usefulness and value in field development, particularly offshore, has become widely accepted (Horvarth, 1985; Brown, 1986).

The disadvantage of 3-D surveying is that about ten times as much seismic is needed as is normally acquired in ‘conventional’ 2-D surveys.

In the last few years the concept of reconnaissance or exploration 3-D (Recon 3-D), using a wider line spacing to keep acquisition costs to a minimum, has been proposed as an exploration tool (Einarsson et at., 1987; Rasidi, 1987).

TCPL, as operator for three offshore prospecting licences in the Taranaki Basin (Fig. 1), has recently carried out three Recon 3-D surveys, each addressing a different set of problems and objectives. This paper discusses three topics: the background to each survey, the study carried out to decide on the critical parameter of line spacing, and the preliminary results.

Recon 3-D can be justified if it significantly reduces the risk involved in a drilling venture, and the cost is much less than the cost of a well.

Detailed geophysical mapping of the Canning Basin incorporating seismic, gravity, magnetic and well data was combined with published global tectonic studies to develop a model of the structural and depositional evolution of the basin. Data indicate that (1) the Cambro-Ordovician section in the Fitzroy Trough is similar to that penetrated on the platform areas, (2) Siluro-Devonian (?) Caribuddy evaporites were originally thin, and Devonian Carbonates are absent, within the northern Fitzroy Trough and (3) the Tandalgoo. Sandstone and Middle-Upper Devonian carbonate sequence are chronostratigraphic equivalents.

The evolution of the basin can be correlated to the orogenic history of Gondwanaland and is summarised as follows: (1) Cambro-Ordovician platform development, (2) Silurian- Devonian restricted basin development, uplift, followed by platform and trough initiation, (3) Devonian-Carboniferous extension, (4) Carboniferous convergent wrenching and uplift, (5) Carboniferous-Triassic downwarp, (6) Triassic convergent wrenching, (7) Jurassic uplift, erosion, followed by regional subsidence, (8) Post Jurassic uplift and extension.

Examination of the structural history of the basin has direct applications in: (1) extension of the Ordovician and Devonian carbonate plays to the south and along the basin margins, (2) further evaluation of the large anticlinal structures within the Fitzroy Trough, and (3) development of play concepts for the Kidson Sub-basin.

Thermal (warm) anomalies have been noted over a number of hydrocarbon fields overseas. To test our ability to use these anomalies as a possible low-cost exploration tool, we acquired NOAA weather satellite night-thermal imagery over two areas of ATP 354P in western Queensland to see if unusual thermal anomalies could be identified. The NOAA satellite provides temperature resolution to about 0.3 degrees Celsius.

Geological and cultural features created many anomalies which were discounted after matching the imagery with published topographic and geological maps. Several unexplained anomalies remain in both the Boulia and Springvale areas, and further exploration may shed light on their significance.

Further tests of the method in areas with known hydrocarbon production and simpler surface geology should provide a clearer picture of the applicability of the method. New, higher resolution satellite imagery will also provide a more detailed picture of surface temperatures.

Faults, thrust faults and intrusion boundaries as usually picked on seismic data are features quite different from normal reflection horizons.

Their definition is by terminations of reflecting interfaces rather than by discernible reflected energy off the fault or intrusion boundary.

Reflections off faults or intrusion boundaries are, in general, (except for cases of accurate depth migration) not located where the corresponding terminations of horizons line up. Frequently, they are not discernible.

Faults or intrusion boundaries that are not actual reflections may not be migrated or modelled. They have to be transformed from time to depth or depth to time domain. The transform parameters can only be derived from horizons available in both time and depth.

The azimuthal anisotropy due to a system of aligned vertical fractures, even when very weak, is sufficient to enable the strike of the fractures to be determined from shear VSP data. Synthetic VSP data were computed for several offsets using two perpendicular horizontal point forces into pairs of perpendicular horizontal receivers. The medium consisted of an anisotropic layer of orthorhombic symmetry (with the degree of azimuthal anisotropy due to the fractures an order of magnitude less than the degree of non-azimuthal transverse isotropy due to horizontal layering) over an isotropic halfspace. The line of sources from the well was at a fixed angle to the strike of the fractures. Source and receiver pairs were rotated numerically in small increments from the in-line to cross-line direction. When sources and receivers were aligned with the directions normal to, and parallel to, the fractures, the downgoing signal from a source to the receivers orthogonal to it decayed rapidly with depth. At zero offset this signal vanished identically at all depths.

The philosophy behind dip-moveout (Hale, 1984) is a slightly different approach to the problem of dip filtering, previously attacked by a number of methods such as a prestack process known generally as DEVILISH, which stands for dipping event velocity inequalities licked (Judson et al., 1978) and prestack partial migration (Yilmaz and Claerbout, 1980). The importance of the problem of dip filtering applied to common mid-point gathers is stated in great detail by the authors mentioned above.

On the one hand, the conventional normal-moveout and common-midpoint stacking process reinforces reflections having a particular slope in a given common mid-point gather; on the other hand it attenuates reflections having different slopes. Therefore this process behaves as a dip filter on a common mid-point gather and decreases lateral resolution of events on a seismic section.

A synthetic model and real field data have been employed to examine the algorithm of dip-moveout. These examples have shown that the algorithm may be included in the sequence of seismic data processing whenever it is needed. However, in terms of computer time and economy dip-moveout is not a cheap process. In order to overcome this difficulty, in this study some computational expense has been reduced by neglecting evanescent waves and only working in the seismic frequency spectrum.

The majority of the Bass Basin lies offshore between mainland Australia and Tasmania.

Detailed mapping of the top of Eastern View Coal Measures and four shallower horizons within the non-prospective Torquay Group led to the identification of regional north-south and east-west transfer/wrench zones. Mapping of deeper horizons has confirmed the existence of these trends and has resulted in a new tectonic model for the Bass Basin.

The basin was initiated during the Early Cretaceous as a northwest trending rift characterised by steeply dipping normal faults.

Australia separated from Antarctica at the end of the Early Cretaceous, causing crustal extension within the Bass Basin. This led to the rotation of existing normal faults and development of north-south transfer faults with wrench components.

A period of basin sag with limited basin extension followed. A second major phase of crustal extension was associated with the opening of the eastern margin of Australia. This event caused further rotation of earlier normal faults and development of east-west transfer faults with associated wrenching.

Basin sag resumed during the remainder of the Late Cretaceous and Palaeocene. Early Eocene re-activation of two north-south transfer/wrench zones controlled the development of the prospective Cormorant and Pelican troughs.

Late Miocene re-activation of north-south transfer/wrench zones is marked by ‘pop up’ structural features and igneous activity. Concurrently east-west transfer/wrench zones acted as hinge lines for uplift and erosion along the southern and northern parts of the basin.

The radio imaging method (RIM) of coal seam mapping relies on the natural waveguide effect exhibited by coal seams to electromagnetic wave propagation. Whenever coal seams are disrupted by a natural geological discontinuity, the waveguide disruption causes a measurable change in the wave attenuation. Attenuation measurements can be analysed directly or, if sufficient data are available by tomography, to indicate the location of the anomaly and its likely severity. Such information is of vital importance in modern underground coal mining.

In February, 1987 a series of trial surveys were conducted in three Australian mines by a survey team from Stolar Inc. of the USA. The results showed that the natural wave attenuation is 46dB/100 m in the Liddell Seam, 36 dB/100 m in the Wongawilli Seam and 13 dB/100 m in the Bulli Seam. In the Bulli Seam propagation over distances in excess of 600 m is possible.

The surveys were conducted over test sites where dykes and other geological features exist. Wave propagation was clearly disrupted by the dykes and it was possible to image their location by tomographic means. An application for RIM in Australian mines has been clearly indicated.

The generalized reciprocal method of seismic refraction interpretation is implemented on the IBM PC and compatible series of microcomputers. The implementation is heavily interactive, using spreadsheet style editors, templates and interactive graphics. Data input is in the form of raw traveltime curves, surface elevations, shot positions, uphole times and system delay times. No phantoming or other data processing needs to be done to the traveltime curves. Output is in the form of report ready graphics display of traveltime data, time depths, layer velocities and interface elevations in cross section form.

The program, called GREMIX, can accept spreads with up to 69 geophones and nine shots. The interpretation allows for up to four layers, and can be carried out in about one hour per spread.