Seismic Diffraction Imaging for Improved Structural Detection in Complex Geological Environments

Rather than ignoring or suppressing seismic diffraction data, they can be retained as they are valuable for identifying small geological discontinuities such as faults, dykes and other features that cause small scatters in the subsurface. Through previous projects C22016 Enhancing Fault Detection by Seismic Diffraction Imaging and Stage I of this project (see annexure), a moving average error filter (MAEF) has been adopted to extract diffraction imaging from reflection seismic data. This MAEF can be applied to seismic data from any geological environment without the assumption of relatively flat or subparallel coal seam strata and without the need for a seismic event flattening process. This is achieved by first estimating the local seismic dip and then applying the MAEF filter in the estimated reflection dip direction. The key to this approach is in the estimation of local seismic dips, realised by a non-iterative method using the gradients or the derivatives of the seismic wavefield based on the plane wave propagation.

The diffraction extraction algorithms have been implemented into a standalone Windows program DiffractionWin for both 2D and 3D seismic data. This program is freely available to ACARP members. Numerical results have confirmed that diffraction imaging for small fault identification is feasible for both unmigrated1 , post-stack-time and prestack-depth migrated2 seismic data. The application of diffraction imaging algorithms to unmigrated post-stack seismic data illustrates that diffractions can be extracted from existing final stacked seismic time sections without the need for extensive reprocessing.

The assessment of the new 3D diffraction algorithm developed for this project has highlighted two major findings:

• The new 3D algorithm successfully eliminates directional bias; and
• The new 3D algorithm has maintained the resolution needed for fault analysis.

As found in 2017, the diffraction images improved the precision of interpreting larger faults and added confidence to the interpretation of smaller faults close to the detection limit.

Using additional case study data, an analysis of suitable trace filters, and underlying seismic processes, it was shown that:

• The 21-trace filter generally achieved the best results for fault interpretation, but this may vary with different surveys;
• Compared with both post-stack time- and pre-stack depth migrated seismic volumes, unfiltered final stacked 3D seismic volumes produced the best diffraction results (plausibly due to inaccurate migration velocity model used); and
• The most suitable display method for diffraction interpretation includes horizon and bracket images, but this may be different for other surveys. Several approaches should be tested at the beginning of each new interpretation study.

A test using the new 3D diffraction algorithm with the Mine A data from the 2017 case study showed little difference between the 2017 and 2019 interpretations. The new process eliminated the directional bias but did not improve the resolution of the diffraction images. The case study showed that diffraction anomalies on their own may not be a reliable indicator of faulting but can improve the reliability of seismic data interpretation if they are used along with the reflection seismic data.

The purpose of diffraction imaging is to support the interpretation of reflection data by adding confidence and potentially detail to fault interpretations. As found in 2017, the diffraction images improved the precision with which larger faults can be mapped and added confidence to the interpretation of smaller faults that are close to the detection limit. Diffraction images do not lower the fault detection limit, nor do they attempt to improve the inherent location accuracy of seismic data.

The case study at Mine B followed the same workflow producing similar results. Diffraction analysis significantly improved the interpretation of subtle structures but did not lower the fault detection limit. Mine B also included several shallow-dipping thrust faults with 4-7m throw that were interpreted from reflection data but occur outside the mining footprint. These faults presented weak diffraction anomalies that could be used to trace the faults more accurately.

This project has demonstrated that even in complex geological environments, if diffraction analysis is undertaken as part of the interpretation of 2D and 3D seismic survey data, the diffractions add confidence to the interpretation of small-scale faults near the seismic detection limit. Informed diffraction analysis can therefore significantly improve the usefulness of seismic reflection surveying to coal mine exploration and planning.