Time-lapse seismic and electrical resistivity tomography combined for monitoring of the CO2 storage site, Ketzin, Germany
Imaging of subsurface ﬂuid migration is a longstanding geophysical research task, in which CO2 storage monitoring is a relatively recent application. It is well known that such monitoring can beneﬁt from a combination of diﬀerent geophysical methods, which complement each other with regard to imaging characteristics and sensitivity to variable saturation of CO2. At the Ketzin site, Germany, pilot-scale CO2 injection is performed with then aim to improve the understanding of in-situ physical, chemical and biological processes, and to provide practical and operational experience for future geological storage of CO2. This thesis addresses the geophysical monitoring of the CO2 that was injected at the Ketzin site. Therefore, time-lapse analyses are carried out on the basis of repeated seismic experiments and electrical resistivity tomographies (ERT), both of which are presented separately in the ﬁrst place. Finally, a methodical combination of these methods is proposed and applied to image the injected CO2. The thesis gathers these investigations as follows: First, a synthetic modelling of crosshole ERT and surface seismic experiments is performed. This study is carried out in order to identify key sensitivities of the time-lapse images from the Ketzin site and to provide assistance for the subsequent processing of real data. Therefore, a static reservoir model is derived from well logs and the structural interpretation of the baseline 3D seismic data. Further, multiphase ﬂow simulations are used to establish dynamic CO2 distributions in this model. The study indicates that a CO2 quantiﬁcation may not be possible on the basis of the seismic amplitude information only, since the thicknesses of the CO2 distributions do not exceed the seismic resolution limit with regard to typical wavelengths that are provided by surface seismic data. Secondly, the time-lapse processing of the baseline and ﬁrst repeat 2D seismic surveys from the Ketzin site is presented. The ﬁrst survey was performed before the start of the injection in 2005 and the second survey in 2009, after approximately 22 kilotons of CO2 had been injected. Although the datasets show a good repeatability, near-surface velocity changes are found to have a degrading impact on the quality of the time-lapse images. As these changes could only be imperfectly resolved by refraction static corrections, a pre-stack static correction is proposed, which decomposes the timing delays of baseline and repeat traces in a surface-consistent manner. Together with a test of a post-stack static correction, this pre-stack static correction is shown to improve the quality of the time-lapse images considerably. In the subsequent interpretation of the time-lapse images, this provided evidence that in 2009 no CO2 related amplitude change is observable where the 2D lines allow for monitoring of the reservoir. This ﬁnding is further conﬁrmed by an AVO analysis and a comparison with the corresponding 3D surveys. Thirdly, large-scale ERT surveys are presented, which have been carried out repeatedly during the site startup and the ﬁrst year of CO2 injection. The experimental setup of these surveys combines surface measurements with downhole measurements by utilizing a permanent electrode array which has been deployed in the three Ketzin wells. One baseline and three repeat experiments are presented, which show a resistivity increase over time at the CO2 injector and indicate a preferential CO2 migration towards the northwest. Using an experimental resistivity-saturation relationship, CO2 saturations of up to 70% are mapped near the injection well, which is consistent with CO2 saturations obtained from pulsed neutron-gamma logging. Finally, a combination of seismic and geoelectric investigations is presented using a structurally constrained inversion approach. For this purpose, lithostructural constraints are interpreted from the seismic reﬂection data and implemented in the geoelectric inversion by means of a local regularization technique. Consequently, seismics and constrained resistivity inversion are arranged in a sequential workﬂow which is based on a structural similarity of seismic parameters and electric resistivity. In other words, a change in elastic impedance is expected to occur together with a change in resistivity and vice versa. Prior to an application to the Ketzin datasets, this approach was tested through a synthetic data example. In consistency, both the synthetic and the real data example demonstrated that the constrained inversion allows for an enhanced resistivity imaging along the caprock-reservoir boundary than a conventional ERT inversion. The practical demonstration for the Ketzin datasets shows, that the approach has a potential for an integrated geophysical monitoring of CO2 storages and is also signiﬁcant for imaging of other subsurface processes, which trigger changes in elastic parameters and electric resistivity.