2-D velocity inversion/imaging of large offset seismic data via the tau‐p domain

Abstract

We describe a method for determining a two‐dimensional (2-D) velocity field from refraction data that has been decomposed into some function of slowness. The most common decomposition, intercept time‐slowness or [Formula: see text], is used as an intermediate step in an iterative wave field continuation procedure previously applied to one‐dimensional (1-D) velocity inversions. We extend the 1-D approach to 2-D by performing the downward continuation along numerically computed raypaths. This allows a correction to be made for the change in ray parameter induced by 2-D velocity fields. A best fitting velocity model is chosen as a surface defined by critically reflected and refracted energy that has been downward continued into a three dimensional (3-D) space of velocity, offset, and depth. Synthetic data are used to demonstrate how this approach can compensate for the effects of known lateral inhomogeneities while determining an underlying 1-D velocity field. We also use synthetic data to show how multiple refraction lines may be used to determine a general 2-D velocity model. Large offset field data collected with an Ocean Bottom Hydrophone are used to illustrate this technique in an area of significant lateral heterogeneity caused by a sloping seafloor.