Quantitative Seismic Inversion
SAND has developed a family of seismic processing and inversion algorithms specifically tuned for application to high-frequency, marine near-surface seismic reflection data. Broadly these are grouped into two software suites, QSI and QSI-3D, which deal with 2D and 3D data, respectively.
Quantitive Seismic Imaging
The software includes bespoke imaging algorithms that have been developed to maximise the penetration and resolution of a variety of UHR marine seismic reflection data. These include algorithms capable of handling the irregular spatial sampling, high frequency content, and contamination by coherent noise (specifically source/receiver ghosts and seafloor multiples). Time and depth imaging versions are available.
High fidelity 3D imaging opens up the ability to apply a range of bespoke seismic inversion methods to derive quantitative information about the ground conditions.
The acquisition of multi-offset seismic reflection data offers a significant improvement over single-channel data, particularly for the derivation of a subsurface velocity model for pre-stack depth migration of the data.
High-resolution interval velocity model building for detailed pre-stack depth migration provides an excellent sediment discrimination. The figures below display depth migrated profile, alongside the same profile overlain by interval velocities.
Q-factor is a way of quantifying the seismic wavefield attenuation as it moves through the sub-surface and has been shown to be strongly linked to cohesive versus granular behaviour of sediments.
As a result, Q-factor analysis can provide useful information about the compressibility of near-surface sediments, differentiating between granular/cohesive sediments units and providing gas saturation estimates for shallow gas fronts.
By far the most novel component of QSI are a family of seismic inversion codes that use the amplitude, phase, and frequency content of seismic reflection data to derive quantitative information regarding the nature of the seabed and near-surface sediments. A range of different machine learning approaches can be applied, depending on the data/project requirements, permitting both bulk physical properties (e.g., P-wave velocity, bulk density, and porosity) as well as more advanced geotechnical properties (e.g., undrained shear strength, relative density, and soil classification) to be derived and mapped at a resolution controlled by the geophysical data.
The figure on the right shows a typical output from this process, including acoustic impedance, CPT tip resistance and sleeve friction.
Seismic imaging & inversion
High fidelity 3D imaging
These algorithms have been successfully applied to a broad spectrum of UHR marine seismic data, including Chirp, Boomer, Sparker, and Airgun sources. All inversion algorithms are cast within a stochastic framework, allowing property envelopes (i.e., 95% confidence estimates) to be provided as well as ‘best’ estimate solutions.
Please contact us with your project requirements for more details.