Lawrence Berkeley National Laboratory

A continuous active source seismic monitoring data set was collected with crosswell geometry during CO2 injection at the Frio-II brine pilot, near Liberty, TX. Previous studies have shown that spatiotemporal changes in the P wave first arrival time reveal the movement of the injected CO2 plume in the storage zone. To further constrain the CO2 saturation, particularly at higher saturation levels, we investigate spatial-temporal changes in the seismic attenuation of the first arrivals. The attenuation changes over the injection period are estimated by the amount of the centroid frequency shift computed by local time-frequency analysis. We observe that (1) at receivers above the injection zone seismic attenuation does not change in a physical trend; (2) at receivers in the injection zone attenuation sharply increases following injection and peaks at specific points varying with distributed receivers, which is consistent with observations from time delays of first arrivals; then, (3) attenuation decreases over the injection time. The attenuation change exhibits a bell-shaped pattern during CO2 injection. Under Frio-II field reservoir conditions, White's patchy saturation model can quantitatively explain both the P wave velocity and attenuation response observed. We have combined the velocity and attenuation change data in a crossplot format that is useful for model-data comparison and determining patch size. Our analysis suggests that spatial-temporal attenuation change is not only an indicator of the movement and saturation of CO2 plumes, even at large saturations, but also can quantitatively constrain CO2 plume saturation when used jointly with seismic velocity.