Time-lapse Velocity Change Tomography

Jonathan Edgar and Nicolas Mastio. ( 2017 )

Abstract

Many prevalent methods for estimating time-lapse velocity changes from the shifts between vintages of time-lapse seismic data make use of a 1D approximation: that seismic energy propagates vertically through a horizontally layered Earth of homogeneous and isotropic layers. While this approximation has produced operational benefits, it is in conflict with real seismic data propagating through real Earth structures. Here we present a method of estimating production related time-lapse velocity changes from pre-stack seismic data without assuming vertical seismic propagation through a horizontally layered Earth. Instead seismic energy is modeled by ray tracing through an existing interval velocity model, previously estimated for seismic imaging. A tomographic system of equations can be formed from these ray paths and the shifts between time-lapse vintages measured in either the pre-stack data (before migration) or image (after migration) domains. Assuming isotropic velocity changes and negligible changes in reflector depths, the solution of this system yields the causal time-lapse velocity changes. We demonstrate that these estimated velocity changes are correctly positioned and are of the right magnitude, whereas those estimated using the aforementioned 1D approximation are not.

Download / Links

BibTeX Reference

@PROCEEDINGS{,
    author = { Edgar, Jonathan and Mastio, Nicolas },
     title = { Time-lapse Velocity Change Tomography },
 booktitle = { 79th EAGE Conference and Exhibition 2017 },
      year = { 2017 },
       doi = { 10.3997/2214-4609.201700810 },
  abstract = { Many prevalent methods for estimating time-lapse velocity changes from the shifts between vintages of time-lapse seismic data make use of a 1D approximation: that seismic energy propagates vertically through a horizontally layered Earth of homogeneous and isotropic layers. While this approximation has produced operational benefits, it is in conflict with real seismic data propagating through real Earth structures. Here we present a method of estimating production related time-lapse velocity changes from pre-stack seismic data without assuming vertical seismic propagation through a horizontally layered Earth. Instead seismic energy is modeled by ray tracing through an existing interval velocity model, previously estimated for seismic imaging. A tomographic system of equations can be formed from these ray paths and the shifts between time-lapse vintages measured in either the pre-stack data (before migration) or image (after migration) domains. Assuming isotropic velocity changes and negligible changes in reflector depths, the solution of this system yields the causal time-lapse velocity changes. We demonstrate that these estimated velocity changes are correctly positioned and are of the right magnitude, whereas those estimated using the aforementioned 1D approximation are not. }
}