The 2011 off the Pacific coast of Tohoku Earthquake, hereafter referred to as Tohoku-Oki earthquake, occurring off northeastern Japan’s Pacific coast on March 11, 2011 had a moment magnitude of 9.0 and generated a tsunami responsible for most of the deaths of the event’s 19,000 victims. Identifying scientifically what happened before, on, and after March 11 is one starting point for a discussion on how to reduce casualties and mitigate the impact of such natural disasters. The 14 papers in this special issue cover incidents related to pre-, co- and post-seismic phenomena, including volcanoes. Three papers discuss why and how such a large quake occurred. Three more papers go into the implications of short- and long-term crustal deformations seen in northeastern Japan. Four papers detail short- and long-term phenomena leading to the Tohoku-Oki quake. Two papers discuss real-time tsunami forecasting based on off-shore and on-shore geodetic, seismic and tsunami observation data. The last two papers explore the effects of the 2011 temblor on volcanic phenomena.

The magnitude 9.0 produced in the 2011 event is the largest historically recorded in Japan and may not necessarily have been anticipated beforehand, and the generation mechanism behind such a gigantic occurrence is not yet completely understood. Even so, preparations should be made for such earthquakes in other parts of Japan and in other countries. The Nankai trough is an example of areas that require our attention.

A national project for observation and study for earthquake prediction is now being integrated into a new program, Earthquake and Volcano Hazards Observation and Research Program (2014-2019). Studies presented in this special issue are also being supported in part by this program.

We are certain that readers will find that this special issue will contribute much to our understanding of gigantic earthquakes and at least some of the measure to be taken in preparation for such natural phenomena. Finally, we extend our sincere thanks to all of the contributors and reviewers involved with these articles.

We should estimate the size of the largest earthquakes carefully to minimize future earthquake losses. Taking the earth’s size and the thickness of the lithosphere into account, the largest earthquakes are estimated to have a magnitude (M) of 11. The occurrence of such a gigantic earthquake, however, could wipe out the human race because energy from an M11 earthquake would be as large as the asteroid impact that is thought to have caused the extinction of the dinosaurs. This would make it meaningless to prepare for M11 earthquakes. It would be more realistic to calculate the seismic moment expected from the longest plate boundary, i.e., from the northern to northwestern boundary of the Pacific plate. The plate boundary is about 10,100 km long and the magnitude of an earthquake occurring there is estimated at approximately 10, assuming that the fault is 200 km wide and average slip is 20 m. This means that M10 events would be the largest earthquakes we could estimate as occurring. It will not be pragmatic, however, to attempt to devise measures against such extremely rare events. It is nonetheless important to numerically simulate beforehand what might happen during such events to quickly and accurately determine the initial response to quake shaking and tsunamis.

Some observational studies have suggested that the 2011 great Tohoku-Oki earthquake (M_w9.0) released a large portion of the accumulated elastic strain on the plate interface owing to considerable weakening of the fault. Recent experimental and theoretical studies have shown that considerable dynamic weakening can occur at high slip velocities because of thermal pressurization or thermal weakening processes. This paper reviews several models of the generation of megathrust earthquakes along the Japan Trench subduction zone, that considers thermal pressurization or a friction law that exhibits velocity weakening at high slip velocities, and it discusses the causes of megathrust earthquakes. To reproduce megathrust earthquakes with recurrence intervals of several hundreds of years, it will be necessary to consider the existence of a region at the shallow subduction plate boundary where significant dynamic weakening occurs due to thermal pressurization or other thermal weakening processes.

Earthquakes occur in a complex hierarchical fault system, meaning that a realistic mechanically-consistent model is required to describe heterogeneity simply and over a wide scale. We developed a simple conceptual mechanical model using fractal circular patches associated with fracture energy on a fault plane. This model explains the complexity and scaling relation in the dynamic rupture process. We also show that such a fractal patch model is useful in simulating long-term seismicity in a hierarchal fault system by using external loading. In these studies, an earthquake of any magnitude appears as a completely random cascade growing from a small patch to larger patches. This model is thus potentially useful as a benchmarking scenario for evaluating probabilistic gain in probabilistic earthquake forecasts. The model is applied to the real case of the 2011 Tohoku-Oki earthquake based on prior information from a seismicity catalog to reproduce the complex rupture process of this very large earthquake and its resulting ground motion. Provided that a high-quality seismicity catalog is available for other regions, similar approach using this conceptual model may provide scenarios for other potential large earthquakes.

Numerous source models of the 2011 Tohoku earthquake have been proposed based on seismic, geodetic and tsunami data. Common features include a seismic moment of ∼4×10²² Nm, a duration of up to ∼160 s, and the largest slip of about 50 m east of the epicenter. Exact locations of this largest slip differ with the model, but all show considerable slip near the trench axis where plate coupling was considered to be weak and also at deeper part where M∼7 earthquakes repeatedly occurred at average 37-year intervals. The long-term forecast of large earthquakes made by the Earthquake Research Committee was based on earthquakes occurring in the last few centuries and did not consider such a giant earthquake. Among the several issues remaining unsolved is the tsunami source model. Coastal tsunami height distribution requires a tsunami source delayed by a few minutes and extending north of the epicenter, but seismic data do not indicate such a delayed rupture and there is no clear evidence of additional sources such as submarine landslides along the trench axis. Long-term forecast of giant earthquakes must incorporate non-characteristic models such as earthquake occurrence supercycles, assessments of maximum earthquake size independent of past data, and plate coupling based on marine geodetic data. To assess ground shaking and tsunami in presumed M∼9 earthquakes, characterization and scaling relation from global earthquakes must be used.

Ground motion from the M_w9.0 March 11, 2011, Off-Tohoku earthquake recorded by dense seismic networks in Japan, K-NET and KiK-net, clearly demonstrated the high-frequency seismic wavefield radiating from the earthquake source and developing long-period ground motion in sedimentary basins. The photographic sequence of the visualized wavefield demonstrated the process in which the high-frequency seismic waves radiated from large slips at the top of the subducting Pacific Plate at relatively deeper depth of 25-50 km, which caused multiple large shocks of large (>1000-2000 cm/s²) ground acceleration and several minutes lasting ground motions over a wide area along the Pacific Ocean side of northern Japan. An efficient seismic wave propagation along the subducting Pacific slab and ground motion amplification in a superficial thin low-velocity layer overlying rigid bedrock also enhanced high-frequency (>5 Hz) ground motions very drastically. However, the dominant frequency of the strong ground motion recorded in near-field station was too high such as to cause serious damage to wooden-frame residences having relatively longer-period resonance period (T=1-2 s); The velocity response in this frequency band was only about one third to one half of those observed in severely damaged area during the destructive M_w6.9 1995 Kobe earthquake. The 2011 Off-Tohoku earthquake also produced long-period ground motion in sedimentary basins such those at Tokyo’s population center but observation of the long-period ground motion within T=6-8 s was rather weak and of a level comparable to that of an M7.9 Tonankai earthquake occurring along the Nankai Trough in 1944. This was because the surface wave in this period band was not generated efficiently by the relatively deeper slip over the source fault of the Off-Tohoku earthquake.

The 2011 Tohoku-oki earthquake caused large eastward displacement and subsidence along the Pacific coast of northeastern Japan. This earthquake partly solved a well-known paradox holding that sense and rate of deformation differ greatly between geologic and geodetic estimates. A paradox remains, however, in explaining long-term uplift along the Pacific coast on a geologic time-scale. Geodetic data show that coastal subsidence continued at a nearly constant rate of ∼5 mm/yr with small fluctuations associated with M7-8 interplate earthquakes for ∼120 years before the Tohoku-oki earthquake. In an area near the Oshika Peninsula where coseismic subsidence is largest, extrapolation of a logarithmic function fitting observed postseismic deformation suggests that coseismic subsidence may be compensated for by the postseismic uplift for several decades but it is difficult to expect the postseismic uplift exceeding 2 meters, so it is implausible that the observed rapid subsidence continued throughout an entire interseismic period in a great megathrust earthquake cycle. We propose a hypothetical model in which the sense of vertical deformation changes from uplift to subsidence during the interseismic period. Using simple elastic dislocation theory, this model is explained by the shallow coupled part of a plate interface in an early interseismic period and the deep coupled part of a late interseismic period.

In this article, we review papers on precursors for the 2011 off the Pacific coast of Tohoku Earthquake. We discuss phenomena such as seismic, geodetic, electromagnetic, ionospheric, and macroscopic anomalies for which time scales range from a few decades to a few hours, from long-, mid- to short-term precursors. Of these, we treat ionospheric anomalies in the greatest detail. Through our review, we found that many “signals” preceded the 2011 off the Pacific coast of Tohoku Earthquake.

I review a spatiotemporal evolution of slow-slip transients on the plate interface of the subducting Pacific plate that happened in and around the mainshock rupture area prior to the 2011 Tohoku-Oki earthquake. Based on foreshock activity before the mainshock, two sequences of slow-slip transients were identified by earthquake migrations toward the initiation point of the mainshock rupture. These two slow-slip transients were also detected by geodetic measurement. The second sequence of slow-slip transients, which involved large slip rates, may have caused significant stress loading onto the hypocenter of the mainshock and prompted the initiation of unstable dynamic rupture. In addition, decadal slip-behavior on the plate interface revealed by geodetic measurement and small repeating earthquakes show that slow-slip transients occurred in the down-dip and up-dip portions of the mainshock rupture area. These slow-slip transients imply the reduction of coupling between the subducting and overlying plates, that could be interpreted as the late stage of mega-thrust earthquake cycle, although this notion remains conjectural.

Studies of slow earthquakes during the last decade have suggested a relationship between various types of earthquakes occurring at the interface between subducting oceanic plates and overlying continental plates. Such a relationship has been postulated for slow earthquakes, which are distributed between the stable sliding zone and the locked zone, and megathrust earthquakes, which are located in the locked zone. The adjacency of the respective sources of slow and megathrust earthquakes suggests expected interactions between these two types of earthquakes. Observed interactions between different types of slow earthquakes located at neighbor area suggest a common triggering mechanism in the seismogenic zone. Also, it is expected that stress accumulations in the locked zone should influence stress regimes in surrounding regions; thus, slow earthquake activity in the stable sliding zone may change in response to stress build-up in the locked zone. Numerical simulations reproducing both megathrust and slow earthquakes show a shortening of the recurrence interval between slow earthquake episodes leading up to the occurrence of a megathrust earthquake. Similarities between the activities of slow and megathrust earthquakes, such as those related to periodicity and patterns of multisegment ruptures, are useful for understanding megathrust earthquakes, particularly given the higher frequency of occurrence of slow earthquakes. From this perspective, the continuous and accurate monitoring of slow earthquake activity is important for evaluating the occurrence potential of megathrust earthquakes.

Because the 2011 great Tohoku earthquake was accompanied by phenomena similar to those associated with the 869 Jogan earthquake, as reconstructed on the basis of historical and geological evidence, paleoseismology is recognized for its potential effectiveness in earthquake forecasting. In attempts to avoid such unexpected situations as the 2011 Tohoku event when taking disaster prevention measures, the Japanese government and local administrations announced a maximum class model for earthquakes and tsunamis that is not based on paleoseismological evidence. Thus, paleoseismologists must both inductively study the reconstruction of evidence from the past and deductively evaluate the maximum class earthquake and tsunami.

This paper reviews recent studies on methods of real-time forecasting for near-field tsunamis that use either offshore tsunami data or onshore global navigation satellite system (GNSS) data. Tsunami early warning systems for near-field coastal communities are vital because evacuation time before tsunami arrival is usually very short. We focus on forecasting between the occurrence of a tsunamigenic earthquake and the arrival of the first tsunami at a near-field coast – typically a few tens of minutes or less after the earthquake. Offshore tsunami measurement that provides coastal communities with direct information on impending tsunamis is very effective in providing reliable tsunami predictions. Crustal deformation due to coseismic slips at an earthquake fault detected by real-time GNSS analysis is quite useful in estimating fault expansion and the amount of slip, which in turn contributes to timely tsunami warnings, e.g., within 10 minutes, even for huge interplate earthquakes. Our review encompasses methods on the leading edge of research and those already in the process of being applied practically. We also discuss an effective combination of methods developed for mitigating tsunami disasters.

Paleotsunami studies have shown that several large tsunamis hit the Pacific coast. Many tsunami deposit data were available for the 17^th century tsunami. The most recent tsunami deposit study in 2013 indicated that the large slip of about 25 m along the plate interface near the Kurile trench would be necessary and the seismic moment of this 17^th century earthquake was 1.7 × 10²² Nm. If a great earthquake like the 17^th century earthquake occurs off the Pacific coast of Hokkaido, the devastating disaster along the coast is expected. To minimize the tsunami disaster, a development of the real-time forecast of a tsunami inundation area is necessary. Estimating a tsunami inundation area requires tsunami numerical simulation with a very fine grid system of less than 1 arc-second. There is not enough time to compute the tsunami inundation area after a large earthquake occurs. In this study, we develop a real-time tsunami inundation forecast method using a database including many tsunami inundation areas previously computed using various fault models. After great earthquakes, tsunamis are computed using linear long-wave equations for fault models estimated in real time. Simulating such tsunamis takes only 1-3 minutes on a typical PC, so it is potentially useful for forecasting tsunamis. Tsunami inundation areas computed numerically using various fault models and tsunami waveforms at several locations near the inundation area are stored in a database. Those computed tsunami waveforms are used to choose the most appropriate tsunami inundation area by comparing them to the tsunami waveforms computed in real time. This method is tested at Kushiro, a city in Hokkaido. We found that the method worked well enough to forecast the Kushiro’s tsunami inundation area.

The 2011 Tohoku mega-thrust earthquake caused huge crustal deformation over a wide are of Mainland Japan. Many mega-thrust earthquakes worldwide have triggered volcanic eruptions nearby, and it is assumed that stress changes due to the Tohoku earthquake resulted in a perturbation to the magma system. The objectives of our study is to evaluate this perturbation quantitatively and to analyze the mechanism of the interaction between mega-thrust earthquakes and volcanic eruptions. This paper focuses on quasi-static stress change due to viscous relaxation of a source region and the surrounding area.

Studies using spaceborne interferometric synthetic aperture radar (InSAR) analysis showed that two megathrust earthquakes – the 2011 M_w9.0 Tohoku-oki earthquake in Japan and the 2010 M_w8.8 Maule earthquake in Chile – triggered unprecedented subsidence in multiple volcanoes. There are strong similarities in the characteristics of the surface deformation in Japan and Chile: (1) Maximum subsidence is about 15 cm. (2) Areas of subsidence are elliptically elongated in a north-south direction perpendicular to the principal axis of the extensional stress change. (3) Most of this subsidence is coseismic. These similarities imply that volcanic subsidence triggered by the megathrust earthquakes is a ubiquitous phenomenon. Nonetheless, the mechanism of subsidence is yet to be investigated. Two main hypotheses have been proposed thus far: 1) The localized deformation of hot and weak plutonic bodies. 2) Water release from large hydrothermal reservoirs beneath the volcanoes.

Problem: Strategic action planning and scheduling (SAP) in the coordination of a disaster response team involves selecting and decomposing an objective into sub-goals, grouping available units into coalitions and assigning them to the sub-goals, allocating units to tasks, and adjusting the decisions that have been made. The primary responsibility of a team’s incident commander (IC) in SAP is to coordinate the actions of operational units in disaster crisis/emergency response management by making macro/strategic decisions. Objective: In this paper, we completely model a real-world problem and present data related to the SAP problem. This data model is used to support the design and development of an appropriate approach to SAP. Method: The employed methodology is to analyze and study the SAP problem, which is composed of six essential dimensions: the problem domain, geographic information, geospatial-temporal macro tasks, strategic action planning, strategic action scheduling, and team structure. Result: The contribution of this paper is the SAP problem data model. It is designed as a unified modeling language (UML) class diagram consisting of entity types, attributes, and relationships associated with SAP problem data modeling. Conclusion: To evaluate the quality of SAP data modeling, the SAP problem data model is used to propose and develop an intelligent assistant software system to assist and collaborate with incident commanders in SAP. The study makes five novel contributions: 1) a complete data model for SAP problem modeling, 2) a presentation and aggregation of task information in geographic objects, 3) the expression and encoding of human intuition as human high-level strategy guidance for SAP, 4) the formulation of a strategic action plan, and 5) the integration of strategic action schedule information with other entities.