The cross-dip correction as a tool to improve imaging of crooked-line seismic data: a case study from the post-glacial Burträsk fault, Sweden

Abstract

Abstract. Understanding the development of post-glacial faults and their associated seismic activity is crucial for risk assessment in Scandinavia. However, imaging these features and their geological environment is complicated due to special challenges of their hardrock setting, such as weak impedance contrasts, often high noise levels and crooked acquisition lines. A crooked-line geometry can cause time shifts that seriously de-focus and deform reflections containing a cross-dip component. Advanced processing methods like swath 3-D processing and 3-D pre-stack migration can, in principle, handle the crooked-line geometry but may fail when the noise level is too high. For these cases, the effects of reflector cross-dip can be compensated for by introducing a linear correction term into the standard processing flow. However, existing implementations of the cross-dip correction rely on a slant stack approach which can, for some geometries, lead to a duplication of reflections. Here, we present a module for the cross-dip correction that avoids the reflection duplication problem by shifting the reflections prior to stacking. Based on tests with synthetic data, we developed an iterative processing scheme where a sequence consisting of cross-dip correction, velocity analysis and dip-moveout (DMO) correction is repeated until the stacked image converges. Using our new module to reprocess a reflection seismic profile over the post-glacial Burträsk fault in northern Sweden increased the image quality significantly. Strike and dip information extracted from the cross-dip analysis helped to interpret a set of southeast-dipping reflections as shear zones belonging to the regional-scale Burträsk Shear Zone (BSZ), implying that the BSZ itself is not a vertical but a southeast-dipping feature. Our results demonstrate that the cross-dip correction is a highly useful alternative to more sophisticated processing methods for noisy datasets. This highlights the often underestimated potential of rather simple but noise-tolerant methods in processing hardrock seismic data.