Seismic full waveform is an emerging technique for determining the fine-scale velocity structure of the subsurface. Here, we present results of elastic full waveform inversion (FWI) along three multichannel seismic lines at the Lucky Strike volcano on the Mid-Atlantic ridge that provides a velocity image of the upper oceanic crust with unprecedented resolution (50–100 m). We have used a two-step process combining downward continuation with a time-domain, elastic FWI. The downward continuation procedure enhances the refracted arrivals and wide-angle reflections, and reduces the scattering noise due to rough seafloor. Since both sources and receivers are downward continued to the seafloor, the computational cost of FWI is reduced, as one does not need to model the thick water layer. Our results clearly demarcate two layers within seismic Layer 2A; a low-velocity, highly heterogeneous layer likely reflecting the complexity of accretion that is underlain by a more homogeneous high-velocity gradient layer. The base of Layer 2A is defined as a lithological boundary that can be offset by faulting. Thick (〉400 m) units of anomalously low-velocity material (〈2.5 km s –1 ) beneath different summital edifices on the central volcano indicate that a thick pile of high-porosity extrusive rocks can be supported without collapsing, suggesting that while in general there is pore closure with depth this is not the cause of high velocities we observe. Hydrothermal deposition sealing of small-scale porosity is shown to be a secondary process, which likely explains the upper crustal velocity increase with age, but is not responsible for the high-velocity gradient Layer 2A. Finally, the rapid thinning of the entire Layer 2A in the vicinity of the main normal faults suggests the tectonic thinning of a geologically defined layer, further confirming the lithological origin of the high-velocity gradient zone at the base of seismic Layer 2A.
Oxford University Press
on behalf of
The Deutsche Geophysikalische Gesellschaft (DGG) and the Royal Astronomical Society (RAS).