09:45 〜 10:00
[S13-02] The 2011 Tohoku-Oki Earthquake Near-Trench Structural Evolution by Time-lapse Seismic Depth Imaging
Previous comparative seafloor observations by the manned submersible Shinkai 6500 within the coseismic slip area of the 2011 Tohoku-Oki earthquake (M 9.0) showed several seafloor deformations, open fissures, and collapsed biological colonies associated with the earthquake. Differential bathymetry studies also showed that seafloor experienced a horizontal displacement of around 69 m near the Japan Trench, where the Pacific plate subducts beneath the Okhotsk plate. Despite these clear observations of the seafloor deformations, and many other researches using various types of geoscientific data, there are yet no detailed structural comparisons on sub-seafloor depth domain images of seismic reflection data before and after the earthquake. In this research, we did a time-lapse analysis between two seismic survey lines acquired on the same trajectory, but before and after the Tohoku-Oki earthquake, by applying Reverse Time Migration (RTM) on the multichannel seismic reflection data. For the baseline dataset we used 2D seismic line MY101, which had been acquired in 1999 using a streamer cable of 132 live channels at 25 m intervals with a maximum offset of 3400 m and shot intervals of 50 m. The monitor dataset is the 2D seismic line D13, which has been acquired in May 2011, just two months after the earthquake, using a similar shot spacing but a more advanced seismic acquisition system including a streamer cable of 444 channels at 12.5 m intervals with a maximum offset of 5700 m. Both surveys have been conducted by Japan Agency for Marine-Earth Science and Technology (JAMSTEC).
In order to maintain consistent survey geometry for both lines, we selected a subset of the monitor dataset which covers the same offset range as the baseline survey and skipped every other channel to have the same CMP (Common Mid-Point) fold coverage. Consequently, we used a similar pre-processing flow for both surveys to enhance the signal to noise ratio before imaging. For building a reliable RTM velocity model, we applied a layer-stripping migration velocity analysis approach followed by a reflection travel-time tomography using the full-offset monitor dataset, which provided us with a reasonable accuracy for constraining the velocity of the deeper reflections. We then used the velocity model and applied RTM with a maximum frequency of 20 Hz for both baseline and (offset-limited) monitor datasets.
The resulting seismic depth images show several distinct differences, which could be summarized as follows. (1) The plate boundary fault (i.e., decollement) reflection near the trench is continuously visible in the baseline but seems to be weak and partially broken in the monitor survey. Especially, the decollement reflectivity above a graben near the trench seems to be dimmed out in the monitor dataset, which implies that the seismic impedance contrast at this boundary has dramatically changed after the earthquake. (2) A number of thrust faults with strong reflectivity appear in the monitor line, which are not clearly visible in the baseline survey. Previously we have calculated pore-fluid pressures along the decollement and backstop interface, and observed that the backstop interface and several thrust faults in the accretionary wedge contribute to the fluid drainage from the underthrust sediments to the seafloor. The sharp reflections from the thrust faults observed in the monitor data may be caused by the impedance contrast of the fluid drainage paths along those faults. (3) The sedimentary units in the forearc of monitor line are partially deformed with different dips compared to the baseline survey. These structural deformations have not been reported before and we believe the complex rupture propagation pattern and coseismic slip of the Tohoku-Oki earthquake caused those unique features.
In order to maintain consistent survey geometry for both lines, we selected a subset of the monitor dataset which covers the same offset range as the baseline survey and skipped every other channel to have the same CMP (Common Mid-Point) fold coverage. Consequently, we used a similar pre-processing flow for both surveys to enhance the signal to noise ratio before imaging. For building a reliable RTM velocity model, we applied a layer-stripping migration velocity analysis approach followed by a reflection travel-time tomography using the full-offset monitor dataset, which provided us with a reasonable accuracy for constraining the velocity of the deeper reflections. We then used the velocity model and applied RTM with a maximum frequency of 20 Hz for both baseline and (offset-limited) monitor datasets.
The resulting seismic depth images show several distinct differences, which could be summarized as follows. (1) The plate boundary fault (i.e., decollement) reflection near the trench is continuously visible in the baseline but seems to be weak and partially broken in the monitor survey. Especially, the decollement reflectivity above a graben near the trench seems to be dimmed out in the monitor dataset, which implies that the seismic impedance contrast at this boundary has dramatically changed after the earthquake. (2) A number of thrust faults with strong reflectivity appear in the monitor line, which are not clearly visible in the baseline survey. Previously we have calculated pore-fluid pressures along the decollement and backstop interface, and observed that the backstop interface and several thrust faults in the accretionary wedge contribute to the fluid drainage from the underthrust sediments to the seafloor. The sharp reflections from the thrust faults observed in the monitor data may be caused by the impedance contrast of the fluid drainage paths along those faults. (3) The sedimentary units in the forearc of monitor line are partially deformed with different dips compared to the baseline survey. These structural deformations have not been reported before and we believe the complex rupture propagation pattern and coseismic slip of the Tohoku-Oki earthquake caused those unique features.