A geomechanical correction for time-lapse amplitude variation with offset

Colin Macbeth, Samarth Bachkheti

Research output: Contribution to journalArticlepeer-review

4 Downloads (Pure)

Abstract

Accurate inversion of time-lapse amplitude variation with offset (AVO) to reservoir pressure change requires knowledge of strain/stress changes in the reservoir and overburden. These geomechanical effects have not been taken into account in previous studies. We have determined that there are three contributions of importance that must be included in the AVO equations: A density term in the reservoir, the seismic anisotropy generated by the reservoir deformations, and the impact of the changing overburden properties. To calculate the error in neglecting these contributions, we derived analytic expressions for the 4D intercept and gradient as functions of the strains developed in the reservoir and overburden during production. The 4D AVO behavior is found to be controlled by three R-factors that relate to time-lapse variations in P- A nd S-wave velocities. These factors can be estimated from rock-physics measurements in the laboratory or calibrated to field data measurements of time shifts via the well-known Hatchell-Bourne-Røste R-factor. Numerical computation indicates that the error in not accounting for the anisotropic contribution to the P-P reflectivity is 50%-300% for the 4D zero-offset intercept and 62%-155% for the 4D gradient. Thus, in most applications, the time-lapse anisotropy in the reservoir and the overburden cannot be neglected when estimating pressure changes using 4D AVO.

Original languageEnglish
Pages (from-to)M29-M40
Number of pages12
JournalGeophysics
Volume86
Issue number2
Early online date5 Feb 2021
DOIs
Publication statusPublished - 1 Mar 2021

Keywords

  • 4D
  • AVO/AVA
  • geomechanics

ASJC Scopus subject areas

  • Geochemistry and Petrology

Fingerprint

Dive into the research topics of 'A geomechanical correction for time-lapse amplitude variation with offset'. Together they form a unique fingerprint.

Cite this