Exploration Geophysics - Volume 44, Issue 2, 2013

Reflection coefficients and arrival times, together with seismic velocities, are significantly important for possible evaluation of reservoir properties in exploration seismology. Reflectivity inversion is one of the robust inverse techniques used to estimate layer properties by minimising misfit error between seismic data and model. On the other hand, the layer-stripping method produces subsurface images via a top-down procedure so that a given layer is modelled after all the upper layers have been inverted. In this paper, we have combined these two methods to develop a new random layer-stripping scheme which first determines the reflectivity series via a random-search algorithm and then estimates P-wave velocities. The first step can be viewed as a variant of sparse spiking deconvolution, and the second step is accomplished by considering empirical relations between density and P-wave velocity. The method has been successfully applied to Marmousi synthetic data to examine dipping reflectors and velocity gradients, and it has been found to be quite reliable for analysing complex structures. A comparison with minimum entropy deconvolution showed that our inversion algorithm gives better results in detecting the amplitudes and arrival times of seismic reflection events.

With advanced computational power, prestack reverse-time migration (RTM) is being used increasingly in seismic imaging. The accuracy and efficiency of RTM strongly depends on the algorithms used for numerical solutions of wave equations. Hence, how to solve the wave equation accurately and rapidly is very important in the process of RTM. In this paper, in order to improve the accuracy of the numerical solution, we use a time-space domain staggered-grid finite-difference (SFD) method to solve the acoustic wave equation, and develop a new acoustic prestack RTM scheme based on this time-space domain high-order SFD. Synthetic and real data tests demonstrate that the RTM scheme improves the imaging quality significantly compared with the conventional SFD RTM. Meanwhile, in the process of wavefield extrapolation, we apply adaptive variable-length spatial operators to compute spatial derivatives to decrease computational costs effectively with little reduction of the accuracy of the numerical solutions.

Applications of seismic inversions strongly depend on inversion methods, data quality, and reservoir complexity. An advanced inversion scheme to integrate seismic data, well data, and geological knowledge is employed by combining statistical Caianiello convolutional networks with a hierarchical seismic convolutional model for impedance estimation and with nonlinear petrophysical models for porosity and clay-content inversions. The method used to measure the reliability of seismic inversions for different geological complexities is important for reservoir characterisation. The widely used cross-validation may not be the best for the evaluation of the reliability of seismic inversions because of different geological conditions away from wells. As a supplementary means and also to understand failed cross-validations, we propose a systematic methodology to measure the reliability of seismic inversions through prior seismic-to-well correlation analyses for the fidelity of seismic data. The resulting correlation coefficients at the main frequencies of seismic data may express what degree the seismic data reflect the subsurface reliably in both amplitude and phase. First, the low-cut filtered borehole impedance logs are correlated with the seismic relative impedance traces computed by trace integration of seismic traces at wells. The resulting correlation coefficients within the seismic frequency band could be an index with which to evaluate the reliability of seismic inversions for impedance estimation. Second, the correlation between borehole impedance and porosity/clay-content is analysed by measuring the overall trend across the cloud of data points in the logging-databased cross-plot. The resulting correlation coefficients could be used to evaluate the reliability of mapping seismic impedance to porosity/clay content. Case studies from several oilfields across China show that the prior seismic-to-well correlation analyses are an excellent way to test the reliability of seismic inversions before the implementation of inversions.

Lineament analysis is typically applied to geoscientific data to identify lithological contacts, faults, fractures and dyke swarms. We implement lineament analysis as a method for quantifying the adequacy of pre-processing of airborne magnetic datasets. This is accomplished through the identification of noise due to inappropriate levelling. Lineament analysis involves the extraction of linear features from a dataset using visual and/or automatic interpretation techniques and the statistical and directional analyses of these extracted lineaments. We apply lineament analysis to a levelled high resolution aeromagnetic dataset from the Northwest Territories, Canada, to assess the levelling quality. A priori knowledge will include geology defining regional tectonic trends such as fault systems and dyke swarms. Analysis of a lineament’s azimuth separates assumed geologic sources and noise associated with data acquisition or processing artefacts. The lineament azimuths are assessed as rose diagrams. This is an alternative method to standard computation of 2D radially averaged power spectrums in the frequency domain and sunshading orthogonal to flight path. The rose diagrams are compared with the 2D power spectrums which both provide quantitative directional information; however, the power spectrum provides spectral frequency content and rose diagrams provide frequency of occurrence. Calculation of the number of lineaments along a particular azimuth before and after pre-processing quantifies the degree to which flight-line variations have been suppressed and geological signal more apparent.

Acquisition of magnetic gradient tensor data is anticipated to become routine in the near future. In the meantime, modern ultrahigh resolution conventional magnetic data can be used, with certain important caveats, to calculate magnetic vector components and gradient tensor elements from total magnetic intensity (TMI) or TMI gradient surveys. An accompanying paper presented new methods for inverting gradient tensor data to obtain source parameters for several elementary, but useful, models. These include point dipole (sphere), vertical line of dipoles (narrow vertical pipe), line of dipoles (horizontal cylinder), thin dipping sheet, and contact models. A key simplification is the use of eigenvalues and associated eigenvectors of the tensor. The normalised source strength (NSS), calculated from the eigenvalues, is a particularly useful rotational invariant that peaks directly over 3D compact sources, 2D compact sources, thin sheets, and contacts, independent of magnetisation direction. Source locations can be inverted directly from the NSS and its vector gradient.

Some of these new methods have been applied to analysis of the magnetic signature of the Early Permian Mount Leyshon gold-mineralised system, Queensland. The Mount Leyshon magnetic anomaly is a prominent TMI low that is produced by rock units with strong reversed remanence acquired during the Late Palaeozoic Reverse Superchron. The inferred magnetic moment for the source zone of the Mount Leyshon magnetic anomaly is ~10¹⁰ Am². Its direction is consistent with petrophysical measurements. Given estimated magnetisation from samples and geological information, this suggests a volume of ~1.5 km × 1.5 km × 2 km (vertical). The inferred depth of the centre of magnetisation is ~900 m below surface, suggesting that the depth extent of the magnetic zone is ~1800 m. Some of the deeper, undrilled portion of the magnetic zone could be a mafic intrusion similar to the nearby coeval Fenian Diorite, representing part of the parent magma chamber beneath the Mount Leyshon Intrusive Complex.

In this paper, some strategies to improve previous edge detection methods are given, and two different forms of improved theta map filters are presented. These filters use the power transform and the exponential transform of the theta map to recognise the edges of the sources. The new filters are demonstrated on synthetic and measured gravity and magnetic anomalies, and the resolving power of the new filters is tested by comparing the results with those obtained by the conventional theta map filter. The advantage of the new filters lies in their ability to delineate the source edges more clearly.

This paper presents the results of 34 new heat flow estimates taken in 2004 from 16 water bores and 18 petroleum exploration wells in the western Otway Basin. The average estimated heat flow measured across the study area is 65.6 ± 9.4 mW/m², with a range of 42–90 mW/m². There are three recognisable sectors within the study area where heat flow is slightly elevated relative to the background levels. These sectors can be broadly classified as Mount Schank (73.5 ± 0.5 mW/m²), Mount Burr (71.2 ± 7.6 mW/m²) and Beachport (78.3 ± 10.4 mW/m²). Thermal conductivity values for each unit involved in heat flow estimation were determined from laboratory measurements on representative core using a divided bar apparatus. Borehole thermal conductivity profiles were then developed in this study by assigning a constant value of conductivity to each geological formation. The process of collecting temperature data involved measuring temperature profiles for 16 water bores using a cable, winch and thermistor, and compiling well completion temperature data from 18 petroleum wells. The precision of temperature data was higher in the water bores (continuous logs) than in the petroleum wells (largely bottom-of-hole temperature estimates). Inversion heat flow modelling suggests heterogeneous heat flow at 6000 m depth, with two zones where vertical heat flow might exceed 90 mW/m², and several zones where vertical heat flow might be as low as 40 mW/m². Therefore, while slightly higher surface heat flow does coincide with some of the volcanic centres, heterogeneous basement heat production is a more likely explanation, as there are no heat flow anomalies greater than 5–10 mW/m² associated with the Pleistocene–Recent Newer Volcanics Province. The distribution of heat flow in south-east South Australia is most simply explained by non-volcanic phenomena.