ALBERT

Description: Determining thin layer thickness is very important for reservoir characterization and CO2 quantification. Given its high time-frequency resolution and robustness, the complex spectral decomposition method was applied on time-lapse 3D seismic data from the Ketzin pilot site for CO2 storage to evaluate the frequency-dependent characteristics of thin layers at the injection level. Higher temporal resolution and more stratigraphic details are seen in the all-frequency and monochromatic reflectivity amplitude sections obtained by complex spectral decomposition compared to the stacked sections. The mapped geologic discontinuities within the reservoir are consistent with the preferred orientation of CO2 propagation. Tuning frequency mapping shows the thicknesses of the reservoir sandstone and gaseous CO2 is consistent with the measured thickness of the sandstone unit from well logging. An attempt to discriminate between pressure effects and CO2 saturation using the extracted tuning frequency indicates that CO2 saturation is the main contributor to the amplitude anomaly at the Ketzin site. On the basis of determined thickness of gaseous CO2 in the reservoir, quantitative analysis of the amount of CO2 was performed and shows a discrepancy between the injected and calculated CO2 mass. This may be explained by several uncertainties, like structural reservoir heterogeneity, a limited understanding of the complex subsurface conditions, error of determined tuning frequency, the presence of ambient noise and ongoing CO2 dissolution.

Description: Full waveform inversion is an effective tool for velocity model building. Recently it has been introduced as a complementary method for interpreting time-lapse seismic data as it can be used to detect velocity changes in reservoirs. There are already some successful applications in the fields of oil/gas production and CO2 injection monitoring. We present a case study of data pre-processing of time-lapse data from the Ketzin CO2 geological storage site. Ketzin is a well-known onshore CO2 geological storage pilot site. Due to the restriction of the acquisi- tion geometry, the time-lapse seismic data sets here have limited maximum offset which makes direct inversion for reservoir velocity difficult. As shown by experiences from other case studies, the double difference full waveform inversion method is the best choice here. The success of double difference time-lapse full waveform inversion is highly dependent upon data pre-processing. This is because it only inverts the difference between the baseline and repeat shot gathers. In order to get the correct velocity change in the reservoir, it is important to apply some pre-processing steps to remove the time-lapse noise above the reservoir. In this study we apply cross equalization and time-lapse difference static corrections to remove the time-lapse noise in the shot gathers. We test our methods by using synthetic data sets. The results show that these methods can effectively remove the time-lapse noise in the shot gathers. We also apply these methods to the real time-lapse shot gathers from the Ketzin site. The time-lapse differences above the reservoir time sections are significantly reduced after pre-processing.