S44A-01 16:00h
Physical properties of the top of the subducting Philippine Sea plate beneath the SW Japan arc, derived from onshore - offshore integrated seismic survey
The Nankai trough region, where the Philippine Sea Plate is subducting beneath the southwestern Japan arc, is a well-known seismogenic zone of interplate earthquakes (e.g. the 1944 Tonankai Earthquake (M=7.9) and the 1946 Nankai Earthquake (M=8.0)). A detailed crustal and upper mantle structure of the subducted Philippine Sea Plate and the overlying SW Japan arc is inevitably important to constrain the physical process of earthquake occurrence. In the summer of 1999, we conducted a highly dense onshore-offshore integrated seismic experiment in the eastern part of Shikoku Island and the adjacent Nankai trough, SW Japan. The most remarkable feature of the record sections is that extremely high amplitude reflections (bright reflections) can be recognized. This phase was interpreted as a reflected wave from the top of the subducting Philippine Sea plate at a depth of 18-30km (Kurashimo et al., 2002). Physical properties across the reflecting interface control amplitude versus offset (AVO) response. To obtain physical properties of the material between the subducting Philippine Sea plate and island arc crust, we investigated AVO response on this bright reflection. Analyzing this bright reflection, we could obtain the reflection coefficient (Rpp) as a function of the incident angle. Rpp tends to increase beyond about 50 degrees. To discuss about this characteristic, we calculated reflection coefficient for different velocity models. The single interface models (positive velocity contrast exists between the interface. negative velocity gradient exists upper side of the reflecter) can not explain the characteristic of the Rpp. Thin layer model (about 200 m) with a P-wave velocity of 4.0 km/s (a thin layer with a negative reflection coefficient at its upper boundary and a much larger, but positive reflection coefficient at its base exists) explains the characteristic of the Rpp. The P-wave velocity of the sediments shows 2.0-4.2km/s off Shikoku Island (Kodaira et al., 2002). These results suggest that sedimentary material subducts beneath the southern part of Sikoku Island with the oceanic crust.
S44A-02 16:15h
The Kanto Subduction Zone: Seismogenic Layers, Slab Deformations and Volcanism Associated With the two Subducting Philippine Sea Plate and Pacific Plate
Understanding the details of complex seismicity and tectonic processes in the Kanto subduction zone, which is located at the junction of northeastern Japan arc and Izu-Ogasawara arc, is important in various fields. The most important factor characterizing plate convergence in the Kanto area is that the young, hot and laterally heterogeneous forearc portion of the Philippine Sea plate (PHS) descends into the mantle wedge and collides with the old and cold Pacific plate (PAC). The interaction among overriding, subducting and mutually colliding plates, each having different plate movement, is responsible for complex seismic activity and various subduction modes as have been observed by recent seismic and geodetic networks. Among several noticeable features of seismicity, we focus on the intermediate-depth seismicity of the double Wadati-Benioff seismic zones (WBZ) of the Pacific slab subducting beneath the Izu-Ogasawara arc, and its relation with the overlapping boundary of two slabs, i.e., the contact interface between the two slabs (PHS-PAC interface). Along the upper seismic plane of WBZ, we recognize a remarkable band of intermediate-depth seismicity, which distributes in southeast-northwest direction with a systematic increase of depth northwest from about 90 to 150 km. Focal mechanism solutions of this band are characterized by normal faultings. Associated with this intense seismicity band, we examine the intersection of the bottom of PHS and the upper boundary of PAC (the loci of the deeper limit of PHS-PAC interface), by taking cross-sections of hypocenter distributions. We find the deeper limit of interface also locates southeast-northwest in the depth range between 70 and 100 km and almost parallel with the intense seismicity band with distance 30-40 km apart. To explain reasonably the northwest deepening of the intermediate-depth seismic band, we use the model of dehydration embrittlement of serpentine and temperature structure in the slab proposed by Yamasaki and Seno (2003, JGR, 108). Because PHS-PAC interface deepen northwest, the colder temperature structure of the Pacific slab underneath may remain deeper northwest. As a result the dehydration loci and the corresponding seismogenic layer in the slab may deepen northwest. Further in the overlying asthenospheric mantle the ascending path of aqueous fluid and magma formation point may also be deepened northwest. This inference accords with the location of the volcanic front in western Kanto, which runs northwest above the upper seismic plane of WBZ across isodepth lines with depth range bewteen about 110 and 160 km along the Izu-Ogasawara arc.
S44A-03 16:30h
Seismogenic Mega-splay Fault in Subduction Zone-Modern and Ancient Examples from Southwest Japan
In the Nankai Trough, SW Japan, 1944 Tonankai earthquake took place along a mega splay fault branching from the plate boundary between the upper Eurasian and lower Phillipine Sea Plate (Park et al., 2002). The splay fault is recognized as a strong reflector showing partly negative polarity with a wavelength of ca. 200 m (Park et al., 2002). The fault appears to make a boundary between the underthrusted sediments in the deep portion ($>$ 6km), cut the Nankai accretionary prism in a shallow part, and appears a cold seep spot on the sea bottom surface (Ashi, person. Com.). A major out-of-sequence thrust (the Nobeoka Thrust) cutting the ancient accretionary prism of the Shimanto Belt into two subbelts is observed in the Kyushu Island, SW Japan (Murata, 1986; Kondo et al., in press). The fault was located under the thermal condition ca. 250-OC in terms of vitrinite reflectance and fluid inclusion geo-thermo-barometry, that is enough within the seismogenic zone by thermal model (Hyndman et al., 1997) and is a good analog for the modern splay fault in the Nankai Trough mentioned above. Structural analysis of the thrust shows that the damage zone around the thrust is a few hundred meters thick (Kondo, et al., in press) and is characterized by strong brittle deformation with ubiquitous development of crack-seal and shear-parallel veins. Hanging-wall side of the Nobeoka Thrust is composed of shale/sandstone dominated phyllites, of which deformation is characterized primarily by plastic flow and pressure solution overprinted by brittle cataclastic shear and cracking. Several thin cataclasitc shear zones represent extremely fine fragmental components or re-crystalized clays. Their chemical aspect indicates preferential Fe and Mg concentration, which is suggestive of preferential melting of micaceous and chrolite of host rocks due to frictional heating. Asymmetric development of cracks surrounding the thin shear zones is also characteristic. The asymmetric cracks are quite resemble to that developed in the process zone of fault (Vermilye and Sholz, 1998), that indicates upward propagation of rupture. The footwall side of the Nobeoka Thrust is completely brittle, and crack seal or shear zone filling quartz or calcite veins are developed. Geometric relationship of these veins with the fault is quite systematic, therefore indicates quite active fluid deposite and flow in the damage zone. Fluid inclusion analysis indicates that fluid passing through strongly deformed part of damage zone was hotter than that in less deformed part. Such a passage may be possible only during dynamic dilatant rupture associated with seismic event.
S44A-04 16:45h
Permeability of underthrust sediments at the Costa Rican subduction zone: Scale dependence and implications for dewatering and fault strength
Sediment permeability is a key parameter controlling fluid overpressures and fluid egress at subduction zones. Data compilations for argillaceous sediments illustrate that scale effects on permeability are generally minor. However, parallel studies of permeability at multiple scales for the same sediments are rare. We report on laboratory permeability measurements on sediments incoming to the Costa Rican subduction zone, at porosities from 26% to 80%. We then apply a method for directly estimating average permeability at the scale of sediment layer thickness, constrained by published estimates of excess pore pressure. Permeability values from both methods are in excellent agreement, and exhibit a trend consistent with permeability-porosity relationships used in numerical models at scales of several to 10's of km. Our results indicate no scale dependence of permeability. The permeability-porosity relationship we derive is comparable to those inferred at the Nankai and Barbados subduction zones, indicating that the high porosity, and hence high permeability, of incoming sediments at Costa Rica is the most likely explanation for comparatively rapid dewatering there. One key implication is that at non-accretionary subduction zones, especially those with thin incoming sediment sections, the high porosity (and thus high permeability) of subducted sediments should contribute to rapid compaction, high volumes of fluid expulsion, better drainage, and higher strength at the wedge base when compared with accretionary systems characterized by thick incoming sediments.
S44A-05 17:00h
Modeling Short-Interval Silent Slip Events in Deeper Subduction Interfaces Caused by Pore-Fluid Pressure Changes due to Frictional Dilatancy and Compaction
Recent high-resolution observations of crustal movements reveal the occurrence of two types of silent slip events in deeper subduction interfaces; long- and short-interval silent slip events. Short-interval silent slip events occur in the deeper Cascadia subduction zone (Dragert et al., 2001) and in the southwest Japan subduction zone (Obara and Hirose, 2004). These silent slip events occur simultaneously with the activity of distinct, low-frequency, non-earthquake tremors. Recurrence interval of these silent events is from 3 to 14 months. Short-interval silent slip events are thought to be caused by high pore-fluid pressure associated with dehydration in deeper subduction interfaces since these events are accompanied by low-frequency tremors. We developed a 2D model of short-interval silent slip events considering frictional dilatancy due to slip and compaction in the deeper frictional stable region. Pore-fluid pressure increases at very low slip velocity due to compaction. When pore-fluid pressure increases and shear stress reaches a certain level, slip starts to accelerate. With an increase of slip velocity, porosity increases due to frictional dilatancy; as a result, pore-fluid pressure decreases. This decrease in pore-fluid pressure results in strengthening the fault and deceleration of slip. We found that short-interval silent slip events can occur only at the condition in which pore fluid pressure is very close to lithostatic pressure. The recurrence interval of these silent slip events is from several months to a few years. The maximum slip velocity of these silent slip events reaches 10$^{-8}$ m/s. Important parameters are the critical displacement for the porosity equation and the dilatancy coefficient which determines the amount of frictional dilatancy. To reproduce short interval silent slip events, we need to take the value of the critical displacement for the porosity equation to be around 1 mm. Approaching the main event, there are some changes in intervals and the maximum slip velocity of silent slip events. By the slow precursory slip of the main event, occurrence of silent slip events becomes irregular. Our numerical results suggest that monitoring silent slip events will be very important for forecasting the main earthquake. We also investigated rupture velocity of these silent slip events by a 3D modeling. We can reproduce silent slip events which propagate horizontally with a velocity of a few km per day below the bottom of the seismogenic region. Observations of silent slip events have shown that the velocity of the migration ranges from 5 to 15 km per day (Dragert et al., 2004). Our numerical results are consistent with the observations.
S44A-06 17:15h
Contrasts in Seismicity Along the 1964 Great Alaska Earthquake Rupture Zone
We have examined seismicity occurring over 35 years prior to and following the 1964 great Alaska earthquake. These studies indicate that the regions associated with the Prince William Sound (PWS) and Kodiak asperities have behaved very different seismically over time. Prior to 1964, the PWS region experienced an increase in crustal seismicity, especially in the region located down-dip of maximum slip in 1964. Since 1964 this same region has been aseismic at the M$>$ 3 level. PWS intraslab seismicity has remained relatively constant, although down-dip migration of M $>$ 4.5 events was observed following the 1964 mainshock. Low-level seismicity has occurred at the northeastern end of the Kodiak asperity throughout the past 70 years, but much of the central Kodiak asperity has been aseismic at the M $>$ 5 level. In contrast, the southwestern edge of the Kodiak asperity has been associated with moderate to large (Mw 5.5 to 7) crustal, interplate, and intraslab events throughout the past 70 years. These changes in seismic behavior along the 1964 rupture zone are consistent with GPS/geodesy estimates of seismic coupling across the interface and with known changes in plate geometry.
S44A-07 17:30h
Seismic imaging of the 1923 Kanto Earthquake Source Area on the subducting Philippine Sea plate, in Japan
Large devastating earthquakes sometimes occur on a mega-thrust source fault which underlies the Tokyo metropolitan region. The source faults are located on the upper boundary of the subducting the Philippine Sea plate (PSP), which was so far estimated mainly from seismicity distribution. To get better prediction of the strong ground motion, we need to precisely characterize the source fault of those large earthquakes. A depth and geometry of the PSP is a key for characterization. We deployed four controlled source seismic lines in the Kanto area from 2002 to 2003: The 150-km long Boso line in 2002, the 80-km Sagami line in 2003, the 80-km Tokyo Bay line in 2003, and the 165-km Eastern part of the Kanto Mountain line in 2003. We, for the first time, directly identify the source fault using deep seismic reflection profiling images to be the upper surface of the PSP, which is found to be much shallower than previous indirect estimation based on seismicity distribution by 10 to 20 km. This geometry serves as a new constraint for studies of Kanto seismotectonics and seismic imaging using earthquakes such as high resolution 3D tomography and receiver function analysis. In the seismic profiles, we also found that highly reflective zones correspond to regions of silent slip plus areas outside the asperity zones of the 1923 Kanto earthquake. We propose that an asperity on a plate boundary can be detected by an evaluation of its reflectivity and represents possible regions of rupture for future earthquakes. Together with the obtained velocity structure, our result provides the essential control for the research on crustal deformation, modeling of source fault and strong ground motion estimation.
S44A-08 17:45h
Spatio-temporal stress states estimated from seismicity rate changes in the Tokai region, central Japan
Since unprecedented large-scale silent slip was detected by GPS in 2001 in the Tokai region, evaluating whether such movement is uniquely connected to the expected Tokai earthquake or repeatedly occurs in this area becomes vitally important. Because of short history of GPS observations and the limited areal coverage surrounding the Suruga trough, we take advantage of continuously recorded seismicity that is presumed to be sensitive to the deformation of the seismogenic depth. Together with the well-maintained NIED earthquake data, we employ the seismicity-to-stress inversion approach of rate/state friction to infer the spatio-temporal stress changes in and around the presumed hypocentral zone of the future Tokai earthquake. Mapping stress changes inverted from microseismicity year by year, we find that the stress under Lake Hamana, the western future Tokai source, has been decreasing since 1999, during which the GPS data showed a normal trend. In contrast, stresses in the surrounding regions are calculated to have increased by transfer from Lake Hamana region. We interpret that this continuous process is associated with the 2000-2004 Tokai slow slip event. The characteristic patterns related to aseismic stress-release are also identified in the early 1980s and during 1987-1989, when slow events are inferred to have occurred on the basis of conventional geodetic measurements. Revisiting the seismotectonics and taking into account the mechanical implications of the inversion results, we argue that the transition zone situated between a deep stable creeping zone and a locked zone undergoes episodic creep and plays an important role in the transfer of stress to the locked zone. Consequently, even though we speculate that the current (2000-to-present-day) silent slip event might be one of the repeating events, the inferred enlargement of the stress releasing area is significant and possibly raises the likelihood of the next Tokai earthquake.