Datapages, Inc.Print this page

Click to view presentation in PDF format.

 

Reducing the Structural Uncertainty in Poor 2-D Seismic Data, Gambier Embayment, Otway Basin, Australia - a Minimum Strain Approach*

Peter Boult1, Brett Freeman2, and Graham Yielding2

 

Search and Discovery Article #40371 (2008)

Posted December 2, 2008

 

*Adapted from oral presentation at AAPG Annual Convention, San Antonio, TX, April 20-23, 2008.
 

 

1 Petroleum & Geothermal Group, PIRSA, Adelaide, SA, Australia.

2 Badley Geoscience Ltd, Hundleby, Spilsby, United Kingdom.([email protected])

 

Abstract

In faulted reservoirs, structural uncertainty arises from two major sources: systematic error of the seismic method and human error of the interpreter. When the distance between faults along a sample line is greater than the distance between lines, the pattern of faulting is usually clear. The error in lateral positioning of structures is then similar to the error in the vertical dimension, and both are dominantly systematic. This is typical of good 3D seismic data. For 2D data, poor 3D data, and situations when faults are more closely-spaced than lines of the interpretation sub-grid, the pattern of faulting becomes much more interpretive. The effects of systematic errors become second-order relative to those inherent in the interpreter’s “model”. The Otway Basin is a passive margin displaying multiple rift events and a complex history of faulting. The Gambier Embayment, within the basin, contains what we now recognise as two vintages of faults with a small angular offset of trends. The older, deep-seated, faults have displacement increasing with depth, and a younger set have their displacement maxima at a higher level in the overburden. In the original structural model (work by a 3rd party) the distinction between these two sets was not made, and consequently prospect maps were largely erroneous. In this new work we quantify the relative uncertainty by assuming that the fault configuration should represent a minimum volume strain given the interpreted, local, offsets. By mapping the displacement patterns on the fault planes from the original structural model we show that the displacement gradients, locally, are unrealistically high. Using the regions of implied high displacement gradient as an indicator for higher uncertainty we were able to break out the faults into two self-consistent sets each of lower implied volume strain.

 

uAbstract

uFigures

uConclusions

uReferences

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigures

uConclusions

uReferences

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigures

uConclusions

uReferences

 

Conclusions

  • Mapping displacement over the entire fault surface highlights interpretation busts.
  • Estimates of wall rock strains may provide a suitable OBJECTIVE metric for discriminating between correct and incorrect interpretations.
  • Interpretations that minimize the wall rock strains appear to be geologically more plausible.
  • 3D quality interpretation from 2D data?
  • Summary of obvious problems:
    • The mapped fault polygons overestimate the heave by a factor of two or three (reduces size of target).
    • Not all fault traces are picked.
    • Fault planes are interpreted as linear trends taking the largest fault from seismic line to line (could be correct …).
    • SOME of the structural interpretation is made by “inventing” polygons at the mapping stage.
    • There has been no notion of structural integrity.
    • Fault displacement patterns are erratic, ungeological.
    • Strains implied by the displacements are, locally, geologically unrealistic.
  • Towards a better solution:
    • Iterative interpretation.
    • Pick ALL obvious fault traces before horizons.
    • Pick horizons in 3D; i.e., work from a well constrained part of the structure by interpreting all horizons locally.
    • Correlate faults from line to line as part of the interpretation process.
    • Use displacement analysis to (a) assess the correlations AND (b) suggest alternative horizon interpretations.
    • Ensure that each piece of completed interpretation makes sense in terms of displacements.
    • Use on-the-fly gridding to interpolate between lines of good data quality on to lines of poor data quality.
    • Generate polygons by modeling horizon fault intersections.
    • A 2D work flow? Not really, it’s also good for reconnaissance mapping of 3D data, good and poor quality.

References

Bailey, W.R., J.J. Walsh and T. Manzocchi, 2005, Fault populations, strain distribution and basement fault reactivation in the East Pennines Coalfield, UK: Journal of Structural Geology, v. 27/5, p. 913-928.

Freeman, B., G. Yielding, and M.E. Badley, 1990, Fault correlation during seismic interpretation: First Break, v. 8/3, p. 87-95.

 

Return to top.