7. Spatial/Time-lapse Data Management
Time-Lapse Seismic Crosswell Monitoring of CO2 Injected in an Onshore Sandstone Aquifer
Jesper Spetzler(1), Ziqiu Xue(2), Hideki Saito(3), Dai Nobuoka(3), Hiroyuki Azuma(3) and Osamu Nishizawa(4)
(1) Dept. of Geotechnology, Delft University of Technology, The Netherlands. (2) Research Institute of Innovative Technology for the Earth, Japan. (3) Oyo Corporation, Japan. (4)
Abstract
According to the Kyoto protocol, the amount of emitted carbon dioxide (CO2) into the atmosphere should be reduced in the next decade. Instead of emitting industrial CO2 to the atmosphere, it is possible to inject the CO2 back to where it originally came from. That is into the subsurface. In Japan, a pilot-scale CO2 sequestration project was carried out between 2003 and 2006 in Nagaoka, Niigata prefecture. In total, 10.400 tons of CO2 was injected into a porous reservoir sandstone at 1100 m depth. To monitor for 4D changes in the subsurface during CO2 injection, crosswell seismic data were measured before the injected started and after injection of 3.200, 6.200 and 10.400 tons CO2. Saito et al. (2006) used ray theory based travel time tomography to monitor this field for CO2 injection after 3.200 and 6.200 tons CO2 had been injected. We use time delay tomography on the time-lapse seismic data to compile 4D images of the injection area for all three injection states. Two approaches to describe the propagation of transmitted waves are used: 1) a linear finite-frequency wave theory and 2) the standard ray theory. The linear wave theory takes the finite-frequency effect of propagating wavefields into account. On the contrary, the ray theory is derived using a high-frequency approximation. Consequently, crosswell tomography based on linear finite-frequency wave theory than the ray theory is more accurate to image small-scale velocity anomalies. The seismic analysis reveals that most time-lapse changes in the sandstone reservoir occurred after 3.200 and 6.200 tons CO2 was injected, while between the production state of 6.200 and 10.400 tons injected CO2 the 4D differences are rather small. Additionally, the tomographic images compiled with the finite-frequency wave theory show that the time-lapse velocity anomaly is on the order of -10 % below the CO2 injection well head at 1100 m depth. The 4D velocity perturbation estimated from ray theory is about 25 % weaker.
Last modified: Sat May 20 18:18:38 2006