Mapping geological features such as faults, lineaments and channels is important in interpreting 3D seismic data. Detecting subtle features of interest and identifying the fold related fractures and deformations are also of paramount importance in the course of effective seismic interpretation. Seismic discontinuity attributes such as coherence and curvature attributes are excellent tools to extract these features from seismic data.

Our study area is in Sable sub-basin in offshore Nova Scotia. In this study, we have devised a robust work-flow for mapping the geological features of interest from the seismic data in an effective way. The workflow includes application of the cutting-edge techniques for seismic preconditioning and generation of coherence and curvature attributes.

Seismic data is usually contaminated with random noise and also coherent noise in the form of acquisition footprint especially in the shallower sections, even after proper migration and de-multiple of the seismic data. To bring out the subtle features effectively from seismic data, preconditioning of the seismic data is necessary. For footprint removal, we have applied attribute assisted footprint suppression using 2D continuous wavelet transform (Alali et al., 2018) which is an advanced and latest technique to suppress footprints effectively from seismic data. Also, for eliminating random noise from seismic data, edge preserving dip steered median filter and Principal Component – filtered amplitude (Chopra et al., 2011) have been used. These are effective means of eliminating random noise from the seismic data while preserving the dip information as well as the edges intact in the data. On application of the preconditioning steps, we observe a much cleaner seismic data, while the discontinuities and edges are preserved effectively.

We applied advanced coherence algorithms, such as Sobel-filter based similarity and Energy ratio coherence (Marfurt, 2006; Chopra et al., 2014) to extract the discontinuities and subtle features from seismic data. While Sobel-filter based similarity algorithm, like semblance-based coherence, is sensitive to changes in both waveform and amplitude, energy ratio coherence is only sensitive to lateral changes in waveform. Application of these coherence algorithms on the preconditioned seismic data shows precise delineation of fault-fractures and subtle stratigraphic features. We observe fault zones more clearly on the stratal slice of the energy ratio coherence, while channels and other subtle stratigraphic features are more evident on the Sobel-filter based similarity slices.

We also computed volumetric multispectral curvature attributes (Al-Dossary and Marfurt, 2006; Chopra and Marfurt, 2007) in our study. Curvature is a second order derivative of the structure and first order derivative of the reflector dip and thus depicts enhanced sensitivity to the structural deformations, flexes, folds and faults. For the same reason, it also helps in mapping subsurface channels and other sag features caused due to differential compaction (Chopra, 2008). Besides the conventional structural curvatures, more sophisticated amplitude curvature attributes (Chopra and Marfurt, 2013) were also computed on our dataset. Application of amplitude curvature attributes in this study was found to be useful in delineating fault-fractures, channel and other subtle stratigraphic features in greater detail compared to the structural curvature attribute.

Therefore, application of the advanced algorithms of coherence and curvature attributes on the properly preconditioned seismic dataset helped in extracting the fault-fracture network, channel and other subtle stratigraphic features effectively. Judicious choice and application of the workflow, integrating data preconditioning and various advanced discontinuity attributes, provide valuable geological information and work as a powerful aid to seismic interpretation.