The Quake That Rocked Japan The 11 March 2011 magnitude 9.0 Tohoku-Oki megathrust earthquake just off the Eastern coast of Japan was one of the largest earthquakes in recorded history. Japan's considerable investment in seismic and geodetic networks allowed for the collection of rapid and reliable data on the mechanics of the earthquake and the devastating tsunami that followed (see the Perspective by Heki ). Sato et al. (p. 1395 , published online 19 May) describe the huge displacements from ocean bottom transponders—previously placed directly above the earthquake's hypocenter—communicating with Global Positioning System (GPS) receivers aboard a ship. Simons et al. (p. 1421 , published online 19 May) used land-based GPS receivers and tsunami gauge measurements to model the kinematics and extent of the earthquake, comparing it to past earthquakes in Japan and elsewhere. Finally, Ide et al. (p. 1426 , published online 19 May) used finite-source imaging to model the evolution of the earthquake's rupture that revealed a strong depth dependence in both slip and seismic energy. These initial results provide fundamental insights into the behavior of rare, very large earthquakes that may aid in preparation and early warning efforts for future tsunamis following subduction zone earthquakes.

The GPS/acoustic seafloor positioning has detected significant postseismic movements after the 2011 Tohoku-Oki earthquake (M9.0), just above the source region off the Pacific coast of eastern Japan. In contrast to the coastal Global Navigation Satellite Systems sites where trenchward-upward movements were reported, the offshore sites above the main rupture zone in the northern part of the source region exhibit landward displacements of tens of centimeters with significant subsidence from almost 3 years of repeated observations. At the sites above around the edge of the main rupture zone, smaller amount of trench-normal movements was found. Although the terrestrial movements were reasonably interpreted by afterslip beneath the coastal area, these offshore results are rather consistent with effects predicted from viscoelastic relaxation in the upper mantle, providing definitive evidence of its occurrence. On the other hand, the results in the southern part of the source region imply superposition of effects from viscoelastic relaxation and afterslip.

Kusatsu-Shirane volcano, Japan, is known for its active phreatic eruptions. We have investigated its hydrothermal system by conducting audio-magnetotelluric soundings at 22 stations along a profile that extends across the volcano. The final two-dimensional model is characterized by two conductors. One is a 300-to 1000-m-thick conductor of 1-10 m, which is located on the eastern slope and covered with 200-m-thick resistive layers of Kusatsu-Shirane lava and pyroclastics. This conductor indicates the presence of a Montmorillonite-rich layer of Pliocene volcanic rocks that may function both as an impermeable floor for the shallow fluid path from the peak to the hot springs to the east and as an impermeable cap for the deeper fluid path from the summit region to the foot of the volcano. The second conductor is found at a depth of 1-2 km from the surface, at the peak of the volcano, and its resistivity is as low as 1 m or less. This low resistivity can be explained by fluids containing high concentrations of chloride and sulfate which were supplied from the magmatic gases. Micro-earthquakes cluster above this conductor, and the cut-off of the earthquakes corresponds to the top of the conductor. This conductor infers the presence of the fluid reservoir, and the upward release of these fluids from the reservoir through the conduit presumably triggers the micro-earthquakes at the peak area of the volcano. Crustal deformation modeling using GPS and leveling data of the past 10 years revealed that the center of the deflation coincides with the top of the second conductor, indicating that the fluid reservoir itself can be hosting the deformation.

[1] Among electric field variations supposed to be associated with earthquakes, electric field variations coincident with the passage of seismic waves have been well documented and interpreted mostly in terms of the electrokinetic effect. Here we present two examples of electric field variations obtained in association with small artificial earthquakes caused by blasting and three examples for aftershocks of two large earthquakes of magnitude 6.9 and 7.2, respectively. The electric field turned out to be circularly polarized in some cases, whereas linearly polarized cases were also seen. Since it is unclear whether such a peculiar behavior is understood in terms of existing models, we propose another mechanism to explain circular polarization; here we call this mechanism as ''seismic dynamo effect,'' which would be regarded as an extended model of the so-called induction effect. In our model we consider ions motion in pores filled with water in the ground, which is driven by ground motion in the Earth's magnetic field. With this model we show that circular polarization of electric field is realized in association with resonance between the frequency of ground velocity due to seismic wave and the cyclotron frequency of ions, such as HCO 3 À or Cl À contained in pores, for the Earth's magnetic field at the observation site. Ions with positive charge, such as Na + , also seem to be responsible for circular polarization of electric field with rotation direction opposite to that for ions with negative charge. We also show that in this model the magnitude of electric field can be estimated in terms of the number density of ions.

Some examples of electric and magnetic field variations have recently been reported by Honkura and his colleagues in association with earthquakes, and these variations have been interpreted by them in terms of the seismic dynamo effect. In order to confirm that this effect is a universal phenomenon rather than a phenomenon appearing in a special local condition, we made magnetotelluric (MT) observations above the hypocentral area of the M7.1 earthquake which occurred off Miyagi Prefecture, northeastern Japan, on May 26, 2003. The MT site was selected at a location close to a seismic station belonging to the nation-wide seismic observation network called 'Hi-net', so that we can compare the MT signals with the seismic wave records. During the MT observation period after the mainshock, some moderate-size aftershocks of magnitudes between 2.8 and 4.1 occurred and MT signals appeared in association with all these aftershocks. In order to confirm that MT signals are not due to vibrations of MT equipment, we set up two sets of MT equipment at the same location; in the case of electric field measurements, we used independent electrodes and arranged cables connecting electrodes on the ground for one set and in the air for the other set, and in the case of magnetic field measurements, we buried the induction coils under the ground for one set and hang them in the air for the other set. As for the electric field, the two sets showed exactly the same records. On the other hand, the magnetic field was different from one set to another, but we conclude that the induction coils buried in the ground are more likely to represent the magnetic field due to electric currents flowing in the ground as a result of the seismic dynamo effect.