Modelling Long-Offset Seismic Wave Propagation and Scattering Within Heterogeneous Basalt Sequences

In this thesis, long-offset seismic wave propagation through heterogeneous layered sequences is modelled using ray- and wave-based methods. The primary consideration is for sub-basalt imaging, where long-offset waves are utilised due to their potential benefits over near-offset energy which is well known to be affected by scattering. The goal is to better understand how long-offset waves are influenced by geological heterogeneities and the potential limitations of using this part of the wavefield. Ray-tracing tomography is performed on a synthetic data set generated from a heterogeneous basalt model to develop a velocity model with a simple positive velocity gradient within the basalt layer, which when compared to the original model, often exhibits a different velocity structure. This exposes limitations in the method and may explain discrepancies between published velocity models from the Faroe- Shetland Basin. The thickness of the basalt layer may be over-or under-estimated, with implications for sub-basalt imaging. The influence of heterogeneity on long-offset refracted energy is assessed and found to affect the amplitude and apparent velocity of the wave. This means that layered sequences are unable to be replaced by a homogeneous equivalent medium (e.g. a simple positive velocity gradient) where scattering will be neglected. The heterogeneous basalt model generates a complex seismic response, with an en-echelon pattern of refracted arrivals that is difficult to model using ray-tracing tomography. The dominant cause of this is found to be the layering within the sequence, with each arrival coming from a higher-velocity package of lava flows separated by sedimentary layers. Refracted arrivals are best explained as a scattering loss from leaky guided waves from the high-velocity layers within a sequence. This response is also observed on field seismic data from the Rockall Trough. The key findings from this thesis are that the traditionally used ray-tracing method for velocity model building produces potentially unreliable results, and refracted arrivals within a basalt sequence may be a form of leaky guided waves.