Previous   |   Next

doi:10.2204/iodp.sp.343.2012

Background

The Tohoku Mw 9.0 earthquake and tsunami devastated northern Japan and the catastrophe highlighted many of the gaping unknowns in earthquake science. The largest slip ever recorded in an earthquake occurred in a largely unexpected location. On the basis of historical records of seismicity and previous studies of the rheology and behavior of the shallow subduction megathrusts, earthquake scientists did not anticipate a Mw 9.0 earthquake near Japan that ruptured to the trench with slip magnitudes of ~50 m (Ozawa et al., 2011; Avouac, 2011; Ito et al., 2011; Ide et al., 2011; Sato et al., 2011).

The scientific community has an obligation to learn as much as possible from this extreme event about the forces that operate during and immediately after an earthquake. Specifically, we need to observe the effects of the earthquake on the fault zone so that we can recognize such an event elsewhere in the geological record. Furthermore, we need to determine the processes that resist slip so that we can predict the circumstances under which large slip is likely to happen at the seafloor where devastating tsunamis could be generated. The extreme slip of the Tohoku earthquake and the extraordinary monitoring data available in Japan together provide an unprecedented ability to address these questions.

Expedition 343 will drill into the Tohoku subduction zone in order to measure the fault zone physical properties, recover fault zone material, and directly record the temperature anomaly from coseismic frictional slip, which is the best method to determine the absolute resistive strength of a fault during dynamic slip (Brodsky et al., 2009). All of these observations have bearing on the most important observational gap inhibiting progress on physical models of faults: measures of stress. If the stresses are known, we can predict the behavior of a rupture and understand the conditions under which slip progresses over the complex geometry of the fault.

Stress both during and after the earthquake can be determined from a variety of observations accessible through a borehole (e.g., Lin et al., 2011). Stress measurements from breakouts and core relaxation measure the postseismic stress field. Geological textures and their relationship to laboratory experiments constrain the processes governing stress during the earthquake (e.g., Ujiie and Tsutsumi, 2010). Temperature records the absolute value of stress during the high-speed slip of the earthquake itself, which is perhaps the stage of stress least well known at this time. The frictional stress during the earthquake results in heating of the fault that persists after the earthquake.

Stress-related measurements need to be acquired soon after an earthquake to be useful. Expectations of fault stress vary widely (thus necessitating direct constraints), and several recent laboratory experiments suggest that during a large earthquake the frictional strength can be so low that the thermal signature would be small and decay to unobservable levels within 5 y. Since the drilling process itself temporarily disturbs the temperature field, a few months must pass after the completion of drilling for successful measurements. Furthermore, interpretation of the thermal anomaly can be greatly improved if the rate of temperature change is monitored for at least 1 y and initial measurements are taken within 1–2 y after the earthquake (Fulton et al., 2010). To obtain data within 2 y of the 11 March 2011 earthquake, drilling of the borehole must be completed by January 2013, allow thermal stabilization for 2 months, and then observe temperatures prior to March 2013 (continuous measurements will actually begin earlier, when the borehole is completed). Therefore, this project has been fast-tracked by the Science Planning Committee to begin as soon as possible to maximize the opportunity to capture this important and fleeting signal.

Geological setting

The 2011 Tohoku earthquake and tsunami originated from slip on the megathrust fault surface west of the Japan Trench where the Pacific plate of Cretaceous age subducts below Honshu Island (Fig. F1). The subduction zone is characterized by a relatively rapid convergence rate of ~8 cm/y (e.g., Apel et al., 2006), much seismic activity, and a deep trench. The convergent margin of the Japan Trench displays the features generally associated with subduction erosion (von Huene et al., 1994, 2004; Tsuru et al., 2000), specifically, evidence of subsidence in the Neogene with associated extensional faulting in the middle slope region, horst, and graben structure in the upper portion of the subducting plate and a relatively small frontal prism (5–15 km wide) containing landward-dipping reflectors and a backstop bounding the frontal prism on the landward side.

The structure and lithology of the forearc region of northern Japan was investigated during previous ocean drilling, specifically by Deep Sea Drilling Project (DSDP) Legs 56 and 57 and later drilling to establish observatories during Ocean Drilling Program Leg 186. Leg 186 Sites 1150 and 1151 are located above the slipped area of the Tohoku earthquake ~100 km north of the epicenter and the region of maximum slip at the trench, and Leg 56 and 57 Sites 434–441 are located approximately 50–100 km further north. In this region, the Japan Trench system consists of a deep-sea terrace, inner trench slope, midslope terrace, trench lower slope, the trench, and outer trench slope (Arthur and Adelseck, 1980). A forearc basin formed at the deep-sea terrace contains a sequence as thick as 5 km of Neogene sediments overlying a Cretaceous unconformity that correlates with the regional unconformity and geologic relations on land. The overlying sediments extend trenchward through the midslope terrace to the backstop boundary, with the frontal prism forming the trench lower slope (e.g., Tsuru et al., 2000). Seismic profiling indicates the structure in the northern Japan Trench is similar through the region to the south that ruptured during the Tohoku event (Tsuru et al., 2002).

The frontal prism is characterized by lower seismic velocity than just landward of the backstop, and the prism displays only disrupted-to-chaotic reflection patterns that likely indicate strong deformation (e.g., Tsuru et al., 2000). Coring of the toe region of the frontal prism at DSDP Site 434 revealed the prism is composed of a highly disrupted, very uniform hemipelagic deposit (Shipboard Scientific Party, 1980). The major constituents are terrigenous silty clay, biogenic silica, and vitric ash. Biostratigraphic observations indicate structural complexity in the prism, with repletion of assemblages that could record slumping, sliding, and faulting. Significant induration of the sediments occurs at depths below ~100 m, and the mudstones recovered are highly fractured with slickensided faces. The highly fractured and disrupted structure contributed to the difficulties of coring and poor core recovery at the site.

Seismic studies and site survey

The offshore Tohoku region is well characterized from decades of data collection, including some high-resolution surveys of bathymetric and seismic reflection data taken after the 2011 Tohoku earthquake (Fig. F2). A differential bathymetry analysis across the trench axis eastward of the hypocentral area, using bathymetric data collected along the same track before and after the earthquake, demonstrates considerable topographical changes on the landward side of the trench. From this analysis, Fujiwara et al. (2011) demonstrate movement of 50 m horizontal toward the southeast and 10 m vertically upward. In addition, a large submarine landslide scarp on the lower trench wall and associated slump deposits in the trench are indicated by the distinct negative and positive seafloor elevation changes at the trench. A critical result is that the coseismic displacement of the 2011 Tohoku earthquake extended all the way to the landslide scarp, if not to the trench axis proper (Fig. F2). The pervasive and similar magnitude of displacement eastward along the profile shows that the frontal prism was displaced as a unit along a fairly uniform dipping detachment surface. Analysis of other data sets also points to coseismic slip on the order of 50 m that extends to the trench (e.g., Ito et al., 2011; Ide et al., 2011; Simons et al., 2011).

Based on existing seismic data sets, three possible fault interfaces that may have experienced coseismic slip have been identified as potential targets for rapid response drilling:

• The interface near the trench between the subducting plate and the overriding frontal wedge (hereafter referred to as the plate interface or detachment surface),

• Backstop interface, including other reverse faults imaged in the wedge-shaped, low-velocity frontal prism between the plate interface and the Cretaceous unit to the west, and

• Normal faults cutting the unconformity at the top of Cretaceous unit.

All these interfaces occur at depths that can be reached with riserless drilling. Based on the differential bathymetry (Fig. F2) and the geometric continuity with the deeper main detachment surface, the plate interface is considered to be the most likely fault surface of large displacement during the earthquake. This fault is seen as the weak reflector located ~100 m above the strong reflector at ~900 meters below seafloor (mbsf) (Fig. F2). The strong reflector is interpreted as the sediment/basement interface of the subducting Pacific plate.

In general, the plate interface shallows toward the trench, but the water deepens. Based on technical information from CDEX engineers, the maximum water depth for drilling of the proposed borehole and observatory is 7000 m. Generally, the plate interface at a water depth of ~7000 m is shallower near 38°N and deeper further north. On the basis of existing seismic data, a primary site (JFAST-3) and an alternate site (JFAST-4) have been identified that minimize both water depth and depth below the seafloor.

The supporting site survey data for Expedition 343 are archived at the IODP Site Survey Data Bank.

Top of page   |   Previous   |   Next