Full text loading...
Guided waves contain significant information about the near surface in seismic data, but prove difficult to remove through conventional velocity discrimination methods. Here, we analyse guided waves jointly with ground roll to characterise the near-surface properties and then we remove them by modelling methods. We begin by retrieving guided wave and ground-roll phase velocities. Then, we jointly invert them to build S- and P-wave velocity models using a robust multimodal inversion algorithm. Finally, we use the inferred guided wave and ground-roll properties to model the dispersive coherent noise and subtract it while preserving the signal.