No abstract

No abstract

Basin analysis experts use distinctive seismic features like pockmarks or polygonal faulting to understand subsidence histories, migrations pathways, and sedimentation patterns in frontier basins. As geoscientists need to understand migrations paths and fluid accumulations, we rely on seismic methods to image features that aid us in gaining insights regarding fluid movements. Polygonal faulting and pockmarks are linked to multiple origins, including diagenetic, igneous, sedimentary, gravitational, or fluid-related processes. We present a brief review of the cause of polygonal faulting and pockmarks; then, we use seismic interpretation and seismic attributes to map these features and define a possible origin to the structures found in the Great South Basin region of New Zealand.

Fault identification is a critical component of seismic interpretation. During the past 25 years, coherence, curvature, and other seismic attributes sensitive to faults improved seismic interpretation, but human interaction is still required to generate a complete fault interpretation. Today, image enhancement of fault-sensitive attributes, multiattribute fault analysis using shallow learning, and deep learning algorithms based on extensive training and convolutional neural networks are promising fault interpretation workflows. We compare three workflows to test fault detection capabilities, these include image enhancement, Probabilistic Neural Networks (PNN), and Convolutional Neural Networks (CNN). We compared results to human-interpreted faults as our ground truth for a merged 3D seismic survey acquired in the Taranaki Basin, New Zealand. We extracted fault surfaces from the results of the workflows using them as seed points for an active contour method. Extracted faults are then compared to the human-interpreted surface using the Hausdorff distance. Data conditioning, including spectral balancing and structure-oriented filtering, improved the performance of all three workflows. Although all three approaches produce enhanced fault volumes, we find differences in fault location and different artifacts (mispredicted faults). While all three methods exhibit "false positive" predictions, in addition, the enhanced multispectral coherence method produces faults and stratigraphic edges in the final image, including residual stair-step artifacts. In our implementation, PNN produces many salt-and-pepper artifacts through the resulting image, suggesting that we might need to include better training data or reduce the volume size to reduce the number of relevant classes in order to obtain an improved classification. The CNN algorithm trained with synthetic data that provides rapid results, correctly identifying larger faults, but missing smaller faults and, in some cases, misclassifying mass transport deposits and angular unconformities as being faults.