High-resolution tomographic inversion with image-guided preconditioning and offset- dependent picking

Summary A high-resolution approach to tomographic velocity model updating is described, combining advanced techniques to avoid simplifications which conventional tomography relies on, in order to yield geologically plausible models with greater accuracy and no significant increase in cost for time or effort. First, offset-dependent picking is implemented to better track gather events than traditional curvature-based picking, especially in situations demonstrating high anisotropy and complicated geology. Matching complex moveout can yield updates which more quickly flatten gathers and bring out fine detail in velocity models. To complement this, structure-oriented inversion preconditioning is applied, enforcing adherence to a stacked image’s geological features. This helps avoid the violation of faults and layering which may arise using conventional updates, as well as generating a higherresolution update which requires less postprocessing. Together, these refinements fit into a current tomographic flow to generate high-resolution velocity models without sacrifices regarding computation time or accuracy.