Optimisation of Near-Surface Seismic Reflection Data

Abstract

Geological ground models of the shallow subsurface, incorporating stratigraphic interpretations and physical properties are used for applications such as offshore engineering foundation design, geohazard assessment and carbon capture and storage monitoring. In the past they have relied on integrating qualitative interpretations of seismic imagery with 1D in situ geotechnical measurements, limiting the ability to capture lateral variations in physical properties. Ultra-High Frequency Multi-Channel Seismic (UHFMCS; 0.2 - 4.0 kHz) reflection data potentially enables quantitative characterisation of shallow sediments at resolutions < 1 m vertically and 5 m horizontally. However, this potential is limited by conventional processing workflows and imaging strategy. In this thesis I focus on optimising the entire seismic processing workflow to improve both the seismic images used in a ground model and the ability to estimate physical parameters from the data. Firstly, I develop a new method to define source-receiver geometries. I determine relative source-receiver positions using first break travel-times in a hierarchical Genetic Algorithm (GA) inversion. I test the method against synthetic examples, allowing  assessment of noise contamination effects, before a real data case study demonstrates the ability of the GA to correctly position traces. Secondly, I use a UHF-MCS dataset acquired in the East Solent, UK, to understand how imaging technique and velocity modelling methodology affect physical property estimations. I compare interval velocities from conventional NMO-Semblance time-domain velocity picking against a layer stripping depth migration and use empirical relationships to derive porosity from the velocity models. Uncertainty analysis demonstrates improved reliability of physical property estimations , as well as higher quality seismic images from pre-stack depth domain velocity modelling and imaging strategies. Finally, I apply these approaches to a commercial dataset from the Dutch Sector of the North Sea, using UHF-MCS data to quantitatively evaluate a strongly variable stratigraphic boundary. Implementing seismic quality factor and post-stack acoustic impedance inversions alongside lithostratigraphic interpretations, the amplitude anomalies at this stratigraphic boundary are interpreted to over-pressures generated by the trapping of pore fluids as a result of dewatering during the Saalian Glacial Complex. The implication of these findings on offshore engineering design are then discussed.

More information

Identifiers

ORCID for Samuel Callum Clay: orcid.org/0000-0001-9819-0428

ORCID for Tim Henstock: orcid.org/0000-0002-2132-2514

Catalogue record

Export record