ASEG Extended Abstracts - ASEG2003 - 16th Geophysical Conference, 2003

Conventional prestack depth migration involves building a velocity model for the subsurface using velocities that are assumed to be isotropic. When the earth is anisotropic, it is impossible for this conventional earth model to correctly predict the seismic raypaths and hence to accurately migrate the recorded data.

Common problems with conventional prestack depth migration are overcorrected common-image point gathers and depth misties with available well data.

Applying anisotropic prestack depth migration involves making some assumptions about the symmetry of the anisotropy and then determining the vertical velocity and the anisotropic parameters, delta and epsilon. These parameters were determined using a combination of well information, higher moveout analysis and migration scans.

The results of a successful anisotropic prestack depth migration tie the available well information, have flat common-image point gathers and have higher resolution and improved focussing.

We present a system for inverting geological models in cases where there are no established numerical criteria to act as inversion targets. The method of interactive evolutionary computation provides for the inclusion of qualitative geological expertise within a rigorous mathematical inversion scheme, by simply asking an expert user to visually evaluate a sequence of model outputs. The traditional numerical misfit is replaced by a human appraisal of misfit. A genetic algorithm provides optimal convergence into the target parameter space, while optimising an ensemble of solutions, so that the non-uniqueness of the problem may be explored. In order to facilitate analysis of the results, we employ a visualisation technique known as self-organised mapping to represent the parameter space covered by the numerous model outputs. The result is a simple view of an otherwise complicated multi-dimensional problem. A user may infer much about the controlling parameters in the model through a few graphical displays of the data.

The potential of this interactive inversion and visualisation technique is demonstrated when we invert a geody-namic model for a conceptual pattern of fault spacing during crustal extension. We also present an example where the interactive scheme is linked to a numerical inversion of induced polarisation data. In this case, we are exploring for the numerical inversion parameters which lead to a particular geological output.

Airborne electromagnetic (AEM) data are presently inverted with one-dimensional (ID) models either as Conductivity Depth Images (CDI) or with full non-linear inversion to build model sections of concatenated ID models. If lateral conductivity changes are small, ID models are justified. However, AEM investigations are often carried out specifically to find localised conductors and in this case ID inversion is insufficient and will often produce artefacts in the model section.

We have developed an approximate inversion method that deals with laterally inhomogeneous sections. The method is based on the Adaptive Born Approximation previously applied by one of the authors (NBC) to the interpretation of central-loop ground EM profiles. The technique reproduces synthetic models of moderate conductivity contrasts without the artefacts typically seen in CDI sections. The computing speed is comparable to that of stitched ID inversions.

An example of processing field data over a massive Nickel Sulphide deposit shows promising results for routine application on large AEM data sets.

In this paper, we would like to examine the applicability of the time-lapse seismic reflection survey to the geothermal reservoir monitoring.

Akinomiya geothermal area, Akita pref. Japan is the study area where four-month steam production test was performed. The baseline and monitoring surveys were conducted just after and one year after the test. Because the maximum reservoir pressure difference and minimum seasonal difference are satisfied simultaneously.

We find the similarity and difference in the results of the baseline and monitoring surveys which suggests the large scale geology and the change occurred in the reservoir, respectively.

We also demonstrate the possibility of the simple 3-D survey named "handy 3-D. In the survey, vibrator shot points are located on the roads and the reservoirs are deployed as 3-D patch. If it is successful, number of the seismic reflection surveys carried out in the mountainous areas will be encouraged.

Sedimentary rocks can be regarded as transversely isotropic (TI) media. One difficulty in seismic processing is how to flatten the events for long offset data. P-wave reflections from horizontal reflectors in transversely isotropic media have non-hyperbolic moveout. For a multi-layered model, the reflection moveout formula is usually expressed as a Taylor series with higher order terms ignored. Alkhalifah and Tsvankin (1995) developed a three term Taylor series formula to calculate reflection travel times from a horizontal reflector in TI media with vertical symmetry axis (VTI). Using this formula, NMO correction works well for short spread lengths, but not so well for long spreads.

Zhang and Uren (2001 a, b) developed an approximate explicit analytical P-wave ray velocity function for transversely isotropic (TI) media. From this ray velocity function, a reflection travel time formula from a horizontal reflector in TI media was derived. This formula can be used for anisotropic NMO correction. It works well for both large offsets and small offsets. In order to obtain the unknown parameters required for seismic processing, a 3D-semblance analysis technique has been developed. We tested this method with numerical data from TI models with a single horizontal reflector and from multi-horizontal reflector models. The method was also tested with an isotropic model with multi-horizontal reflectors. The results show that the events can be completely flattened even for very large offsets and that both multi-layered TI and isotropic models may appear to be anisotropic.

In this paper we develop a new technique for 3-D cross-well electromagnetic imaging based on the localized quasi-linear (LQL) approximation introduced by Zhdanov and Tartaras, 2002. This approximation was specially designed for modelling the electromagnetic field generated with a moving transmitter. Using the LQL approximation, which is independent on the source position, one can model and invert the EM data for all transmitter and receiver positions at once. This remarkable property of the LQL approximation makes it a practical tool for 3-D imaging of the cross-well EM data. In the numerical inversion algorithm we implement the options for smooth or focusing regularized imaging, which helps generating a clear focused image of the geological target. The numerical examples demonstrate effectiveness of this technique in 3-D cross-well electromagnetic data interpretation for imaging both conductive and resistive targets.

In this paper, we develop a method to calculate the dispersion of seismic wave speed (phase velocity and group velocity) for a general anisotropic medium, which is defined by twenty-one elastic moduli. The solution includes, as special cases, the isotropic and transversely isotropic problems. We apply the plane-wave analysis to the general 3D anisotropic medium and obtain explicit expressions for three eigenvalues (phase velocities) and their corresponding group velocities, which are the propagation speeds of the wavefronts and the energy fluxes (ray-paths) of one qP wave and two qS waves. Basing on the solutions, we show that the phase and group-velocity vectors generally have different directions and they depend on twenty-one elastic moduli and the direction cosines of the incident wave. As examples of using the eigenvalue solutions, we numerically calculate the phase velocities and the group velocities for an isotopic medium, a VTI-medium and a gTI-medium. Two real models (clay shale and phenolic) were used for moduli selection. These results clearly show that the wave speeds vary with the azimuthal angle and the vertical angle of the incident wave, as well as the elastic moduli. This means that the solutions may be applied to investigation of kinematic features of real samples of rocks and the sensitivity of the wavespeed to each elastic modulus. We also show the application of the eigenvalue solutions to the 2D/3D ray tracing in the most general anisotropic media.

A time-lapse crosshole resistivity tomography trial was conducted at the Bolivar Aquifer Storage and Recovery (ASR) trial site, north of Adelaide between 1999 and 2001, to image aquifer properties and preferential flow paths. A bipole-bipole electrode configuration was used in order to reduce the effect of the top layers in the experiment. Five monitor wells were drilled on a circle of radius 75m for acquiring the crosshole resistivity data.

In total, seven time-lapse crosshole resistivity surveys were carried at different stages of fresh water injection. All survey data were processed, inverted and interpreted with the aid of numerical resistivity modelling and inversion code. It appears that the injected water flows in all directions, but mainly flows towards the south and the north; the injected water reached wells in the north and south but did not reach wells in the east and west by the end of the water injection. In general, the resistivity distribution in the region decreases with depth in the aquifer. A thin high resistivity layer at a depth of 130m was detected, which separates the T2 aquifer into two parts.

From this project, we learnt that resistivity inversion is a very necessary tool for crosshole resistivity data interpretation. There is no better way to analyse and interpret the data properly and accurately. However to obtain a good inversion result, adequate data and the right survey configuration are crucial.

Seismic imaging is a process of geophysical inversion and consists of data acquisition, processing and interpretation. Errors introduced in any of these stages can have a serious impact on later stages and the final results. The processing stage probably has the greatest scope for error. Modern processing packages allow considerable freedom in the selection of processing algorithms and parameters. Even if essentially the same processing/work flow is followed, different results may be obtained from different data processors. End-users of the data can be unwittingly placed in a very difficult position.

The seismic method itself is firmly based on the theory of elastic wave propagation. However the data processing and interpretation stages can be viewed as requiring the application of scientific art based on experience, a thorough understanding of the data in question, the field parameters, the options available for processing and the geological setting. This is not a new observation or theme, but in the light of recent experience in the application of seismic methods in mining, we revisit it in this paper. We use real data examples to illustrate the non-uniqueness of seismic data processing and the consequences of choosing different processing algorithms and parameters. The objective of this paper is to provide an awareness to some important issues in the contemporary use of seismic surveying in the hope that through this, better and more reliable results can be delivered and exploited by the end-users.