comparison of near-surface seismic velocity estimation methods with application to on-shore Peru and offshore Malaysia
The near-surface of the Earth often has rapidly varying velocity changes that distort the seismic images to a greater or lesser degree. Depending on the complexity of the velocity variations, a variety of methods have been developed to remove these distortions so that the deeper structures can be more properly imaged. If the near-surface is characterized by layers, where the layers below the top layer have velocity variations that slowly vary laterally, an efficient and effective approach is the Refraction Delay-Time method (RDT). In this technique first-arrival picks are used to infer layer information in the shallow sub-surface. If the layer assumption fails and the velocity still generally increases with depth, then the First-Arrival Traveltime Tomography (FAT) technique may produce a better solution using the same or more extensively picked traveltime data. However, if one or more of the layers decreases in velocity with depth (the “hidden layer” problem), the interpretation of the traveltimes can be difficult. In such a situation Early-Arrival Waveform (EAT) Tomography may be able to handle the complexity at a greater computation cost. These three techniques were applied to a 2D onshore example from Peru and a 2D Ocean Bottom Cable (OBC) dataset from Malaysia. In both cases the more sophisticated techniques of FAT and EAT produced superior results than the RDT technique. In addition, the EAT method gave further improvement to the land example.