Prestack Q compensation with sparse tau-p operators

Abstract

The application of [Formula: see text] compensation to prestack marine data needs the proper removal of the water-layer time from the total traveltime, a process known as “time referencing.” To obtain the water-layer time, current industry practices use some form of normal moveout equation that requires subsurface velocities. We have derived a more straightforward and accurate formula for time referencing that does not require subsurface velocities and works under the same assumptions. The formula is based on a local angle decomposition via the tau-[Formula: see text] transform. Further complicating the [Formula: see text] compensation task in the prestack domain is the proper treatment of spatially aliased energy and high-frequency noise. We found out how time-slowness sparsity, used as a constraint for [Formula: see text] compensation, gives excellent immunity to incoherent noise and spatial aliasing, and we evaluate its role in accelerating the convergence rate for our iterative inversion algorithm.

Warped mapping–based blind deconvolution for resolution improvement

“…However, prestack seismic data that can be used for amplitude variation with offset inversion contains more information than poststack seismic data. Additionally, conventional prestack compensation methods directly apply analogue poststack compensation approaches to the CMP gather with different offsets under the assumption of a constant Q factor (Braga and Moraes, 2013; Aharchaou and Neumann, 2019). However, in practice, the Q model is time‐varying and prestack compensation is related to offset because the waves in a prestack gather travel along different raypaths.…”