Modeling 3D time-lapse seismic response induced by CO2 by integrating borehole and 3D seismic data – A case study at the Ketzin pilot site, Germany

• Detailed 3D physical property models of the reservoir at Ketzin are constructed.
• 4D seismic forward modeling is performed using the convolution method.
• Comparisons between synthetic seismic data and real seismic data are performed.
• Impact of source non-repeatability on CO2 monitoring is tested.

In order to simulate the 3D time-lapse seismic response from real data in a consistent manner detailed 3D property models at the Ketzin pilot site were constructed by robustly integrating borehole and 3D reflection seismic data. The spatial CO2 distribution and the detailed CO2 density in the reservoir were derived from dynamic flow simulations that had been history-matched to the site monitoring data. Changes in velocity and density after CO2 injection were estimated utilizing the CO2 saturation distributions at two repeat times and fluid substitution models. 4D seismic data were generated by convolving the property models at different times with an extracted wavelet.Time-lapse analysis was performed to qualitatively and quantitatively investigate the changes in reflection amplitude after CO2 injection. Comparison between the synthetic and real data at the corresponding time indicates that the 3D property models were built successfully and model the 3D time-lapse seismic response induced by CO2 injection. A synthetic experiment with two different source wavelets was implemented to investigate the impact of source non-repeatability on the seismic amplitude anomaly. Analysis shows that the same or similar sources should be used in time-lapse seismic monitoring to minimize the impact of source non-repeatability on the monitoring.