Velocity analysis using shot‐indexed sensitivity kernels: application to a complex geological model

Abstract

We investigate the capability of inverting large scale velocity errors through wave equation migration velocity analysis using shot-indexed sensitivity kernels. We find that the global shift of the image (absolute residual moveout) is crucial for constraining the large-scale velocity errors in the model when the shift of the image is more than a number of wavelengths. We estimate the shift through fitting the curvature of RMOs on CIGs and add this information into the inversion system. Although the extraction of the shift information is based on a simple model, including this information into RMO data considerably improves the inversion result. We apply our method to Sigsbee benchmark model and demonstrate that the velocity analysis based on sensitivity kernels can handle both large-scale and small-scale velocity errors in a complex geological environment.