10:45 〜 11:00
[SCG54-07] Seismic Imaging and Physical Properties of Underthrust Sediments at the Japan Trench Subduction Zone
キーワード：Subduction, Japan Trench, Seismic Imaging, Stress, Pore Pressure
Although 10 years has passed since the 2011 Tohoku earthquake (M 9.0) occurred in the Japan Trench subduction zone, still the physical properties of the underthrust sediments and plate boundary are poorly known. The stress pattern at the decollement level plays a key role in the seismic behavior of the plate boundary fault and rupture propagation of the megathrust earthquakes. In order to investigate the detailed structure of the Japan Trench underthrust sediments, we have conducted pre-stack depth migration (PSDM) using multichannel seismic reflection data which was acquired along line D13 by research vessel Kairei offshore Miyagi in 2011. After developing a reliable seismic velocity model which could plausibly image the subsurface geological features, we calculated vertical effective stress by using existing empirical relations between P-wave velocity and porosity, and data-fitting approaches. Then we calculated pore-fluid pressure and fluid overpressure ratio for the underthrust sediments. We also calculated shear stress by applying a friction coefficient provided by previous studies. The PSDM depth section revealed a backstop interface with a reverse-polarity reflection, implying on a low velocity zone which extends upward from the plate-boundary fault at a depth of 11 km and a distance of 40 km landward from the trench axis. We specially focused on this area, where the huge shallow slip occurred during the 2011 Tohoku event.
Our results showed that the effective stress at the decollement level is significantly lower than the expected effective stress (under normal consolidation). The very low effective stress might have enabled the 2011 Tohoku earthquake (M 9.0) rupture, which was nucleated at greater depth, to propagate along the shallow decollement up to the trench. JFAST temperature sensors across the fault zone have previously shown that the slip-averaged shear stress and the apparent coefficient of friction estimated from the temperature anomaly at the plate-boundary thrust are 0.54 MPa and 0.08, respectively. By applying a friction coefficient of 0.08 to our vertical effective stress we obtained a shear stress of 0.543 MPa at a location 6km far from the Trench axis, i.e. relatively at the same distance from the Trench axis as the JFAST site. In addition, we calculated fluid overpressure ratio (λ*), which corresponds to the ratio between fluid overpressure and overburden pressure. When λ* =1, it reflects undrained conditions or total fluid retention (the media is impermeable) and when λ* =0, it indicates optimal drainage conditions for fluid evacuation (the media is totally permeable). We noticed an area with a dramatically low λ* located between 7km to 15 km from the Trench axis. The PSDM section shows that there are several faults in this area which provide a drainage path for the sediments and the pore fluid pressure becomes nearly equal to hydrostatic pressure. On the other hand, we found bucket sediments trapped in the subducting grabens with a reflection polarity alteration, indicating that these sediments are fluid-rich and a considerable amount of pore fluid is trapped right below the decollement. The relatively low permeability results in a very low shear stress value and a mechanically weak behavior of the underthrust sediments. Our seismic velocity model, which is built by a layer stripping method and tuned by several iterations of grid-based tomography, has an uncertainty of less than 5% for a depth of 10 km at a location 30 km far from the trench axis. The uncertainty of the velocity model increases downdip to 10% at a location 40 km far from the trench axis. The uncertainty of Vp-porosity transformation is also around 3.5%. A curve fitting approach is used to derive the consolidation curve of the incoming sediments and a normalized RMS error of 15% is estimated based on the goodness of the fitted curve to the data points. Since the propagated uncertainty of Vp model and Vp-porosity transform are already included in the data points prior to data fitting, we believe the 15% uncertainty is the maximum total uncertainty in our calculated stress values. The maximum vertical effective stress is around 63 MPa and the maximum expected effective stress is 114 MPa, which is almost 80% larger than effective stress. Compared to 80% of difference in the stress values, 15% of uncertainty is negligible.
Acknowledgments: The authors would like to thank JAMSTEC for providing the seismic reflection data. We also thank Paradigm/Emerson (http://www.pdgm.com) for providing seismic data processing software.
Our results showed that the effective stress at the decollement level is significantly lower than the expected effective stress (under normal consolidation). The very low effective stress might have enabled the 2011 Tohoku earthquake (M 9.0) rupture, which was nucleated at greater depth, to propagate along the shallow decollement up to the trench. JFAST temperature sensors across the fault zone have previously shown that the slip-averaged shear stress and the apparent coefficient of friction estimated from the temperature anomaly at the plate-boundary thrust are 0.54 MPa and 0.08, respectively. By applying a friction coefficient of 0.08 to our vertical effective stress we obtained a shear stress of 0.543 MPa at a location 6km far from the Trench axis, i.e. relatively at the same distance from the Trench axis as the JFAST site. In addition, we calculated fluid overpressure ratio (λ*), which corresponds to the ratio between fluid overpressure and overburden pressure. When λ* =1, it reflects undrained conditions or total fluid retention (the media is impermeable) and when λ* =0, it indicates optimal drainage conditions for fluid evacuation (the media is totally permeable). We noticed an area with a dramatically low λ* located between 7km to 15 km from the Trench axis. The PSDM section shows that there are several faults in this area which provide a drainage path for the sediments and the pore fluid pressure becomes nearly equal to hydrostatic pressure. On the other hand, we found bucket sediments trapped in the subducting grabens with a reflection polarity alteration, indicating that these sediments are fluid-rich and a considerable amount of pore fluid is trapped right below the decollement. The relatively low permeability results in a very low shear stress value and a mechanically weak behavior of the underthrust sediments. Our seismic velocity model, which is built by a layer stripping method and tuned by several iterations of grid-based tomography, has an uncertainty of less than 5% for a depth of 10 km at a location 30 km far from the trench axis. The uncertainty of the velocity model increases downdip to 10% at a location 40 km far from the trench axis. The uncertainty of Vp-porosity transformation is also around 3.5%. A curve fitting approach is used to derive the consolidation curve of the incoming sediments and a normalized RMS error of 15% is estimated based on the goodness of the fitted curve to the data points. Since the propagated uncertainty of Vp model and Vp-porosity transform are already included in the data points prior to data fitting, we believe the 15% uncertainty is the maximum total uncertainty in our calculated stress values. The maximum vertical effective stress is around 63 MPa and the maximum expected effective stress is 114 MPa, which is almost 80% larger than effective stress. Compared to 80% of difference in the stress values, 15% of uncertainty is negligible.
Acknowledgments: The authors would like to thank JAMSTEC for providing the seismic reflection data. We also thank Paradigm/Emerson (http://www.pdgm.com) for providing seismic data processing software.