Decimetric-resolution elastic characterisation of the shallow subseabed using pre-stack waveform inversion

• Main Author: Provenzano, Giuseppe
• Other Authors: Vardy, Mark
• Published: University of Southampton 2018
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.759271

Currently, engineering-scale shallow sediment characterisation relies heavily on core analysis and cone-penetrometer tests. Unlike reservoir-scale exploration, little quantitative information is derived from seismic reflection data, leaving largely unexploited their intrinsic value as a remote characterisation tool. In this work, we develop seismic modelling and inversion techniques custom-built for limited-offset, limited-bandwidth, ultra-high-frequency (UHF, 0.2-4.0 kHz) seismic reflection data, to obtain a robust decimetric-resolution elastic model of the top 50 m below the seabed. Dedicated signal processing procedures are devised in order to account for the specific bandwidth and acquisition characteristics of UHF data, with the accuracy and the efficiency necessary for pre-stack waveform inversion. A deterministic pre-stack waveform inversion strategy is presented and tested on both complex synthetic and real data. Robustness to noise and sensitivity to the multi-parameter elastic model are tested in relation to the acquisition geometry and the range of physical properties. An original stochastic inversion strategy based on a genetic algorithm is developed in order to improve the robustness to inaccurate starting models. Complex synthetic tests show that this outperforms a conventionally parametrised genetic algorithm, and a real case study demonstrates that this is capable of characterising remotely decimetre-thick shallow weak layers. By using limited a priori information and minimal data pre-processing, an excellent agreement with the geotechnical ground truth is attained. The deposition of submarine slopes with realistic permeability layerings is simulated, and time-lapse multi-channel UHF seismic data are computed to detect the development of localised excess pore pressure anomalies. The results show that seismic data are sensitive to destabilising excess pore pressure levels with a decimetric resolution. Provided that adequate signal-to-noise ratio data are available, these can be quantitatively interpreted to constrain the in-situ effective stress conditions, and therefore better characterise the stability of a slope. This work demonstrates that the inversion of UHF seismic data has the potential to become an important practical tool for submarine ground model building in spatially heterogeneous areas, reducing the reliance on expensive and time-consuming coring campaigns.