Search

In this paper, we explore the use of seismicity data on a single-station basis in site response characterization. We train a supervised deep-learning model, SeismAmp, to recognize and separate seismic site response with reference to seismological bedrock (VS  = 3.45 km s−1) in a broad frequency range (0.2–20 Hz) directly from single-station earthquake recordings (features) in Japan. Ground-truth data are homogeneously created using a classical multistation approach—generalized spectral inversion at a total number of 1725 sites. We demonstrate that site response can be reliably separated from single-station seismograms in an end-to-end approach. When SeismAmp is tested at new sites in both Japan (in-domain) and Europe (cross-domain), it achieves the lowest standard deviation among all tested single-station techniques. We also find that horizontal-to-vertical spectral ratio (HVSR) is not the optimal use of single-station recordings. The individual components of each record carry salient information on site response, especially at high frequencies. However, part of the information is lost in HVSR. SeismAmp could lead to improved site-specific earthquake hazard prediction in cases where recordings are available or can be collected at target sites. It is also a convenient tool to remove repeatable site effects from ground motions, which may benefit other applications, for example, improving the retrieval of seismic source parameters. Finally, SeismAmp is trained on data from Japan, future studies could explore transfer learning for practical applications in other regions. [ABSTRACT FROM AUTHOR]

Peak ground acceleration (PGA) is a key parameter used in earthquake early warning systems to measure the ground motion strength and initiate emergency protocols at major projects. The traditional P -wave peak displacement-dependent PGA prediction model (Pd-PGA model) tends to underestimate the PGA for large earthquakes because it cannot make full use of the fault continuity rupture information hidden in the time-varying process of ground motion. In this paper, a continuous PGA prediction long short-term memory (LSTM) neural network model is proposed. The model takes eight sequential features of stations that are proxies of the energy and other physical parameters as input and provides the recorded PGA at the station as the target output. A total of 5961 records from 119 earthquakes recorded by the Japanese Strong-Motion Earthquake Network (K-NET) in Japan are used to train the neural network and 3433 records from 73 earthquakes are used as the test set to verify the model's generalization ability. The results show that within the same data set, the residuals of the predicted PGA for the proposed model are smaller than those of the Pd-PGA model and that the problem of PGA underestimation is resolved. The prediction accuracy also improves with increasing sequence length, which indicates that the LSTM neural network learns the rules hidden in the time series. To further verify the model's generalization ability, the model performance is analyzed for an M 7.3 earthquake that was not included in the training or test data sets. The results show that the residuals of the predicted PGA for the event are consistent with those for the test data set, indicating that the model has good generalization ability. [ABSTRACT FROM AUTHOR]

Recovery of ground displacements from strong motion accelerographs alone is still performed in an empirical way. This paper starts with a theoretical description of strong motion instruments. Then, based on observations from the non-tilt 2-D shake table tests, this paper proposes a baseline correction procedure with a cumulative correction pattern. The procedure is validated by recorded displacements of the shake tables. The paper also extends the baseline correction procedure to deal with 3-D strong motions by introducing a tilt correction pattern. The results from the proposed baseline correction procedure are compared with those from an existing empirical baseline correction procedure as well as nearby GPS measurements. The accuracy of the proposed procedure is studied considering 41 accelerographs and the corresponding GPS stations that recorded the 2011 M w 9.0 Tohoku earthquake in Japan. We find that the corrected displacements from recorded accelerations alone are not always consistent with the GPS measurements. In this specific Tohoku event of which the primary direction of motion was eastward, the ground displacements derived from strong motion stations using the proposed baseline correction procedure achieve the best agreement in the east–west direction. Meanwhile, the proposed procedure is computationally efficient. [ABSTRACT FROM AUTHOR]

This paper describes an efficient approach to high-resolution time-reversal source imaging simulation. Dense seismograph network records are backpropagated from stations to the hypocentre through a 3-D subsurface structure model to estimate the initial source wavefield at the earthquake initiation time. By assimilating high-density observational data into the time-reversal wave propagation, a clearer source image can be determined, even for deep and distant earthquakes, than is achievable with conventional source imaging. The effectiveness of data-assimilation-based source imaging by a time-reversal wave propagation simulation with a 3-D heterogeneous structural model was demonstrated using recordings from a nationwide strong-motion seismograph network during the 2007 Off Niigata, Japan, M w 6.6 earthquake, and the 2007 Off Ibaraki, Japan, M w 6.8 earthquake. Such data-assimilated-based simulations are also effective for early forecasting of strong ground motions caused by large earthquakes through fast time-advancing simulations based on the current assimilated wavefield. We will discuss the feasibility of a disaster prevention system for the early forecasting of strong motion disasters due to large earthquakes, based on repeatedly estimating source parameters and forecasting strong motions in future time based on the current assimilated wavefields. [ABSTRACT FROM AUTHOR]

Slip on major faults primarily occurs during ‘characteristic’ earthquakes. The recurrence statistics of characteristic earthquakes play an important role in seismic hazard assessment. A major problem in determining applicable statistics is the short sequences of characteristic earthquakes that are available worldwide. In this paper, we introduce a rescaling technique in which sequences can be superimposed to establish larger numbers of data points. We consider the Weibull and log-normal distributions, in both cases we rescale the data using means and standard deviations. We test our approach utilizing sequences of microrepeaters, micro-earthquakes which recur in the same location on a fault. It seems plausible to regard these earthquakes as a miniature version of the classic characteristic earthquakes. Microrepeaters are much more frequent than major earthquakes, leading to longer sequences for analysis. In this paper, we present results for the analysis of recurrence times for several microrepeater sequences from Parkfield, CA as well as NE Japan. We find that, once the respective sequence can be considered to be of sufficient stationarity, the statistics can be well fitted by either a Weibull or a log-normal distribution. We clearly demonstrate this fact by our technique of rescaled combination. We conclude that the recurrence statistics of the microrepeater sequences we consider are similar to the recurrence statistics of characteristic earthquakes on major faults. [ABSTRACT FROM AUTHOR]

The concept of determining the dislocation Love numbers using GNSS (Global Navigation Satellite System) data and calculating the corresponding Green's functions is presented in this paper. As a case study, we derive the dislocation Love number h of vertical displacement by combining 1232 onshore GPS data and 7 GPS-Acoustic data with the 2011 Tohoku-Oki earthquake (M w 9.0). Three fault-slip distributions are used to compare and verify the theory and results. As the GPS stations are only located in Japan Island and along the Japan trench, we use the theoretical vertical displacements of a spherically layered Earth structure to constrain the low-order signal. The L-curve and an a priori preliminary reference skill are applied in the inversion method. Then, the GPS-observed vertical displacement changes are used to invert for the vertical displacement dislocation Love numbers h based on three different fault-slip models. Our results indicate that the estimated dislocation Love numbers |$h$| fluctuate significantly from the earth model (i.e. the preliminary reference earth model), especially for the |$h_{n1}^{32}$| component, and these changes in |$h_{n2}^{12}$| and |$h_{n0}^{33} - h_{n0}^{22}$| are relatively small. The vertical displacements derived from the inversion results agree well with the GPS vertical observations. The inverted dislocation Love numbers are considered to profile the regional structure which differs from the mean 1-D heterogeneous structure of the Earth, and the corresponding Green's functions of four independent dislocation sources describe a more reasonable seismic deformation field. [ABSTRACT FROM AUTHOR]

Earthquake swarms in the area east of the Izu Peninsula, Central Japan have been active and have been repeated intermittently since 1978 after 40 years of quiescence. The activities were always accompanied by crustal deformations, which were well modelled by dyke intrusions. To study the process ora dyke intrusion, precise hypocentres of the latest activity occurring in 1998 are obtained in this paper using waveform similarity, by which new images of volcanic processes have been successfully acquired at several volcanoes. The relocated hypocentres are mainly aligned on a thin vertical plane with a circular shape at a depth of 3-7 km, and there is an aseismic area at the centre. The normal direction to the plane coincides well with the direction of tectonic extensional stress around the hypocentral area and matches theoretical models well. At the beginning of the activity, a small fraction of the events occurred at greater depth, where hypocentres align on a vertical line and migrate upward at a rate of 1 km h[sup -1]. The migration rate in this stage agrees well with a theoretical model of buoyancy-driven dyke propagation. After 1 day from the beginning, earthquakes began to occur around the aseismic area, and hypocentres spread on the thin circular plane described above. During this activity, the hypocentres seemed to migrate downwards and upwards from the centre. The earthquakes are caused by shear fracture at the tip of the dyke and the migration of hypocentres reflects that the dyke expands downward and upward from the neutrally buoyant layer that is located at the centre of the swarm. In this paper, we also propose a process of magma intrusion based on the precise migration of hypocentres, and show that the 1998 activity is composed ora few simple volcanic processes: magma rising by buoyancy, it staying at a neutral buoyancy point and spreading outward from the neutral point with elastic fracture by an inside excess magma pressure. [ABSTRACT FROM AUTHOR]

SUMMARY: To rapidly and accurately provide alerts at target sites near the epicentre, we develop an on-site alert-level earthquake early warning (EEW) strategy involving P -wave signals and machine-learning-based prediction equations. These prediction equations are established for magnitude estimation and peak ground velocity (PGV) prediction accounting for multiple feature inputs and the support vector machine (SVM). These prediction equations are called SVM-M model for estimating magnitude and SVM-PGV model for predicting PGV, respectively. According to comparison between the predicted magnitude and PGV values with the predicted threshold values (M = 5.7 and PGV = 9.12 cm s–1, respectively), different alert level (0, 1, 2, 3) is issued at the different recording site when the predicted magnitude or PGV values exceed the given threshold values. Alert level 3 means that both the predicted magnitude and the predicted PGV exceed a given threshold, and there may be serious damage in this recording site. We apply the method to three destructive earthquake events (M ≥ 6.5) occurred in Japan, and our results indicate that with regard to the performance of SVM-PGV model for predicting PGV, at 3 s after P -wave arrival, the percentage of successful alarms (SAs) for these three events is higher than 95, 73 and 94 per cent, respectively, and the percentage of false alarms approaches 0. Additionally, with regard to the performance of SVM-M model for estimating magnitude, at 3 s after P -wave arrival, the percentage of SAs for these three events exceeds 95 per cent, and the percentage of missed alarms approaches 0. Moreover, almost all stations in the areas PGV ≥ 16 cm s–1 (IMM ≥ VII) near the epicentre issue alert level 3. The proposed method provides potential applications in EEW system. [ABSTRACT FROM AUTHOR]

Data from dense seismic arrays allow nowadays to map directly the radiation of the seismic source in space. However, the main features observed on these maps are usually linked to the propagation, its attenuation and some site effects. Only few parameters like the magnitude or fault extension may be inferred through the shape and the size of the epicentral area. In this work, we propose a new strategy able to produce fast qualitative maps, 'Sourcemaps', that highlight directly some basic characteristics of the seismic source. They are based on a parameter which is robust and fast to compute with a minimal processing: the displacement response spectrum (DRS) computed at low frequency. In order to highlight the source effects, we correct the DRS by the geometrical spreading. We have applied this strategy to some earthquakes occurred in Japan in the last 25 yr using K-NET/KiK-net data. In these examples, we show that the source parameters such as the focal mechanism (nodal planes) and the directivity may be clearly seen on the Sourcemaps. Some site effects at large scale are still visible. The strategy proposed in this work is completely automatic without the need of a visual inspection of the input data. Thus it provides a robust assessment of key parameters of the source. We suggest a systematic computation of these 'Sourcemaps' after each event inside a dense network. [ABSTRACT FROM AUTHOR]

In this study, numerical investigation is performed on a realistic source-to-site earthquake scenario, with the aim to assess the role of complex 3-D geological structures on the predicted wavefield.With this respect, the paper pointedly targets the seismic response of nuclear power plants in near-field conditions and the verification of some simplified assumptions commonly adopted for earthquake ground motion prediction and site effects analysis. To this purpose, the Kashiwazaki-Kariwa Nuclear Power Plant (Japan) is assumed as reference case-study. In 2007, the nuclear site and its surroundings were struck by the Niigata-Ken Chūetsu-Oki seismic sequence, which caused some of the peak ground motion design limits to be largely overpassed. The dense observation network deployed at the site recorded a highly incoherent and impulsive earthquake ground motion. Many studies argued that the intricate synclineanticline geology lying underneath the nuclear facility was highly responsible of the observed seismic response. Therefore, a physics-based numerical model of the epicentral area is builtup (≈60 km wide) and tested for small aftershocks, so to discount the effect of extended source on the synthetic site-response. The numerical model (based on the Spectral Element Method) reproduces the source-to-sitewave propagation by embracing the effects of the surface topography along with the presence of the Japan Sea (i.e. the bathymetry, the coastline and the fluid--solid interaction). Broad-band (0-5 Hz) synthetic waveforms are obtained for two different aftershocks, located at the two opposite sides of the nuclear facility, aiming to assess the influence of the incidence angle the radiated wave field impinges the foldings beneath it. The effect of the folding presence is assessed by comparing it to a subhorizontally layered geology, in terms of numerical outcome, and by highlighting the differences with respect to the observations. The presence of an intricate geology effectively unveils the reason behind the observed ground motion spatial variability within a relatively small area, stressing its crucial role to properly reproduce the modification the wavefield undergoes during its propagation path towards the surface. The accuracy of the numerical exercise is discussed along with its results, to show the high-fidelity of these deterministic earthquake ground motion predictions. [ABSTRACT FROM AUTHOR]

The Global Navigation Satellite System—Acoustic ranging combination technique (GNSS-A) is a seafloor geodetic observation technique that achieves an accuracy of centimetres by combining high-rate GNSS data with acoustic ranging. The technique determines the seafloor position by acoustic ranging between the sea surface and multiple seafloor stations, using GNSS data from the sea surface station. Here, the gradient state of the underwater sound speed structure (SSS) is a significant source of error. The open-source software GARPOS can reduce the effect from underwater gradient structures by retrieving the underwater disturbance as a parameter projected onto the sea surface and seafloor. To evaluate the effects of underwater disturbances, a quantitative comparison of the model parameters is necessary. In this study, we developed a representation method to evaluate features of the ocean field. Here, the expression method was described in the order of a formulation and an interpretation in the case of a 2-D cross-section and extension to the case of an actual 3-D field. This method makes it possible to evaluate SSS states in the GNSS-A observations. As an example, we showed the correlation between the anomaly of the expressed ocean state and the anomaly of the seafloor station position, showing that this expression method is an effective index for correcting bias errors. Additionally, we used the data from sites located in the Kuroshio area, a strong current near Japan, to show that the ocean state can be quantitatively interpreted using this expression method. [ABSTRACT FROM AUTHOR]

The physical mechanisms controlling deep-focus earthquakes, or those observed at depths greater than 300 km, remain enigmatic. The leading processes by which deep-focus earthquakes are thought to occur include transformational faulting, thermal runaway and dehydration embrittlement, but distinguishing observations in support of one or more prevailing mechanisms are needed. In this study, we use a modified back-projection method, data recorded by the Hi-net array in Japan and a 3-D velocity model to produce source images of 19 deep-focus earthquakes within the Izu-Bonin subduction zone. We find that the rupture properties and fault plane orientations of imaged events separate according to reported moment magnitude, indicating the distinct operation of two moment-dependent causal mechanisms of deep-focus earthquakes in this region. We discuss these results in the context of previous observational, laboratory and numerical studies and emphasize the importance of continued research to validate the dual-mechanism hypothesis both in and outside Izu-Bonin. Such work may not only improve our understanding of the nucleation and propagation of deep-focus earthquakes, but also help clarify slab structure and subduction zone dynamics. [ABSTRACT FROM AUTHOR]

Site amplification factors (SAFs) of seismic ground motions are essential in evaluating and estimating seismic hazards. In our previous study, the authors proposed a simple and cost-effective method to estimate a SAF based on a deep neural network (DNN) model and microtremor horizontal-to-vertical spectral ratio (MHVR). Since the previous DNN model was based on the observed SAFs and MHVRs within a limited district in Japan, the applicability of the previous model to non-source regions with different site conditions was limited. This study explored the application of a transfer learning (TL) technique to adapt an existing (pre-trained) DNN model to new regions and a different database. The SAFs obtained through generalized spectral inversion technique (GIT) at the seismic observation stations (K-NET and KiK-net) in Japan were collated as the ground truth for site effects. MHVRs recorded at the stations in several districts of Japan were collected to construct a data set for the development of the TL model. Subsequently, a TL model was constructed, leveraging the neural network layers and their weights from the pre-trained model while incorporating additional neural network layers to enhance the performance. During the training process, a total data set of 112 sites was divided into training set, validation set, and external test set by 1:1:5. Utilizing a cross-validation strategy, the residuals between pSAFs (pseudo-SAFs) estimated by the TL model and the observed SAFs were analysed for the external test set containing 80 sites. The results showed that the TL model outperformed the pre-trained DNN model. The cross-validation results demonstrated that almost consistent prediction results were obtained regardless of any combination of 16 sites selected as the training set. Furthermore, by contrasting the influence of varying training set sizes on the performance of the TL model and comparing the TL model to a DNN model with an extended training set, the effectiveness of constructing the model with the limited number of data (16 sites) was ascertained. Finally, the effectiveness and limitations of the TL model were evaluated using MHVRs with peak frequencies falling outside the training set's range. [ABSTRACT FROM AUTHOR]

Hyuga-nada, off the Pacific coast of Kyushu along the Nankai Trough in southwest Japan, is one of the most active slow earthquake regions around Japan. We estimated the energies of shallow tremors and moments of shallow very low frequency earthquakes (VLFEs) in Hyuga-nada using data from a permanent onshore broadband network and temporary ocean–bottom seismometer observations. The energies and moments of these slow earthquakes have a similar along-strike variation and are generally higher south of the subducted Kyushu–Palau Ridge than near the top of the ridge. This spatial variation is also related to the characteristics of slow earthquake migration. The along-strike migration speed was faster at initiation in the south, where the moments of slow earthquakes are higher. After migration entered the subducted Kyushu–Palau Ridge, its speed was decelerated with a parabolic pattern and their moments became smaller. Assuming a constant patch size of slow earthquakes, we estimated that the stress drop of VLFEs in the south of the subducted ridge was approximately three times higher than that near the top of the subducted ridge. According to our observations and a physical model, this stress drop difference between adjacent regions may cause parabolic migration. We also estimated the scaled energy of slow earthquakes from the ratio of the seismic energy rates of tremors to the seismic moment rates of accompanying VLFEs. The spatial variation in scaled energy is not identified inside the Hyuga-nada. Since the range of scaled energy is similar between the south and near the top of the subducted ridge, the apparent stress may be similar if the rigidity is the same. The dominant range of scaled energy of slow earthquakes in Hyuga-nada is 10−11.5–10−8.5. In addition to having similar or one order smaller values compared to other slow earthquake regions, the range of scaled energy in Hyuga-nada is broader. This broader range suggests wide range of characteristic time and various spectral features of slow earthquakes in Hyuga-nada. Based on a Brownian slow earthquake model, the wide range of characteristic time in this area suggests width variations of slow earthquake source area. [ABSTRACT FROM AUTHOR]

The south-eastern offshore of the Boso Peninsula in Japan periodically experiences short-term slow slip events (SSEs) every few years. On 2018 June, an SSE occurred with the maximum surface horizontal displacement reaching up to 4.7 cm by according to the operational global navigation satellite system (GNSS) network. This study performed a time-series analysis of interferometric synthetic aperture radar (InSAR) with Sentinel-1 SAR images to investigate detailed spatial pattern of surface displacements caused by the SSE. With the assistance of an atmospheric delay correction with a regional numerical weather model output, the InSAR time-series analysis successfully captured displacement signals in three paths, whose maximum amplitudes in line-of-sight directions were 1.46, 1.86 and −0.80 cm. A checkerboard test revealed that the resolution of the slip inversion was higher when InSAR was used than that using GNSS, especially in and around the inland. The slip inversion with the actual displacement data derived from the InSAR time-series analysis was performed with the L-curve optimization, showing that the estimated slip area was concentrated offshore south-eastward from the Boso Peninsula with the maximum slip of 5 cm and the estimated moment magnitude of 6.4. As similar to previous SSEs in the Boso Peninsula, a seismic swarm simultaneously occurred in the downdip area adjacent to the estimated slip with the SSE occurrence, suggesting a different friction characteristics between them. This study demonstrates usefulnesses of the InSAR observation for capturing detailed spatial characteristics of small-displacement events like SSEs and of the hybrid use of the externally derived delay correction with the time-series analysis to improve the displacement detection accuracy. [ABSTRACT FROM AUTHOR]

The Shinmoe-dake volcano started with three sub-Plinian eruptions from 26 to 27 January 2011, followed by a magma effusive stage from 28 to 31 January 2011, and Vulcanian eruptions occurred frequently during 1 to 10 February 2011. During the magma effusive and Vulcanian stages, multiple episodes of harmonic tremors were observed at stations near the summit crater. Although harmonic tremors have been observed at various volcanoes worldwide, the source mechanism remains poorly understood. This paper proposes a source model for harmonic tremors, which is composed of a non-linear viscous fluid flow in a flexible channel. A simple lumped parameter model is used to consider the process. The dynamics are described by a third-order system of ordinary differential equations using model variables for a cross-sectional area of the constricted segment and the fluid velocities in the upstream and downstream tubes. This model produces various kinds of trajectories for self-sustained oscillations that change the reservoir pressure connected on the upstream channel of the model. Linearization analysis around the stationary point and global analysis using nullcline planes reveal the mechanism of self-sustained oscillations of the system qualitatively. To consider both the frequency peaks of the harmonic tremor and the characteristics of observed phase spectra, the qualitative characteristics of an observed phase portrait are compared to those of a simulated one. This tremor model simulates the frequency peaks and the phase portraits of typical harmonic tremors observed during the 2011 Shinmoe-dake eruption. Because this model involves several geometrical configuration parameters, it has the potential to reveal the source mechanism of various kinds of harmonic tremors. [ABSTRACT FROM AUTHOR]

Current earthquake early warning (EEW) systems lack the ability to appropriately handle multiple concurrent earthquakes, which led to many false alarms during the 2011 Tohoku earthquake sequence in Japan. This paper uses a Bayesian probabilistic approach to handle multiple concurrent events for EEW. We implement the theory using a two-step algorithm. First, an efficient approximate Bayesian model class selection scheme is used to estimate the number of concurrent events. Then, the Rao-Blackwellized Importance Sampling method with a sequential proposal probability density function is used to estimate the earthquake parameters, that is hypocentre location, origin time, magnitude and local seismic intensity. A real data example based on 2 months data (2011 March 9-April 30) around the time of the 2011M9 Tohoku earthquake is studied to verify the proposed algorithm. Our algorithm results in over 90 per cent reduction in the number of incorrect warnings compared to the existing EEW system operating in Japan. [ABSTRACT FROM AUTHOR]

Real-time seismology has made significant improvements in recent years, with source parameters now available within a few tens of minutes after an earthquake. It is likely that this time will be further reduced, in the near future, by means of increased efficiency in real-time transmission, increasing data coverage and improvement of the methodologies. In this context, together with the development of new ground motion predictive equations (GMPEs) that are able to account for source complexity, the generation of strong ground motion shaking maps in quasi-real time has become ever more feasible after the occurrence of a damaging earthquake. However, GMPEs may not reproduce reliably the ground motion in the near-source region where the finite fault parameters have a strong influence on the shaking.In this paper we test whether accounting for source-related effects is effective in better characterizing the ground motion. We introduce a modification of the GMPEs within the ShakeMap software package, and subsequently test the accuracy of the newly generated shakemaps in predicting the ground motion. The test is conducted by controlling the performance of ShakeMap as we decrease the amount of the available information. We then update ShakeMap with the GMPE modified with a corrective factor accounting for source effects, in order to better constrain these effects that likely influence the level of (near-source) ground shaking.We investigate two well-recorded earthquakes from Japan (the 2000 Tottori, Mw 6.6, and the 2008 Iwate-Miyagi, Mw7.0, events) where the instrumental coverage is as dense as needed to ensure an objective appraisal of the results. The results demonstrate that the corrected GMPE can capture only some aspects of the ground shaking in the near-source area, neglecting other multidimensional effects, such as propagation effects and local site amplification. [ABSTRACT FROM AUTHOR]

Interferometric synthetic aperture radar (InSAR) and global positioning system (GPS) have obvious deficiencies for monitoring surface deformations, for example, 1-D line-of-sight (LOS) measurements for InSAR and spatially very sparse observations for GPS. In this paper, InSAR and GPS measurements are integrated to derive spatially high-resolution 3-D coseismic surface displacement fields of the 2011 Mw 9.0 Tohoku-Oki, Japan earthquake. A unified simultaneous least squares (USLS) approach is developed to minimize the inconsistence between the InSAR results from adjacent paths. 3-D ground displacements are then derived by integrating the InSAR and the GPS measurements with the method of weighted least squares (WLS). Comparisons with independent GPS measurements show that the root mean square errors (RMSEs) of the derived 3-D displacements are 6.30, 4.57 and 1.29 cm for the vertical, east and north components, respectively. The 3-D coseismic displacement map shows that the Honshu Island moved eastwards towards the epicentre and subsided in the eastern part. The maximal displacements in the vertical, east and north directions are −1.5, 5.0 and −2.0 m, respectively. The effects of the density of GPS sites on the InSAR/GPS integration are also investigated. The experimental results reveal that lower to 70 km spatial resolution's GPS observations are adequate to guarantee the accuracies of the 3-D displacements for the Tohoku-Oki earthquake. This demonstrates the applicability of the developed WLS-based InSAR/GPS integration method, as in general the GPS observations are not as dense as those in this study area. Based on the spatially high-resolution 3-D surface displacement fields, we estimate the high-resolution and 3-D strain of the Tohoku-Oki earthquake. The preliminary results show that the Honshu Island suffers from an evident dilatation and shear during the seismic event. [ABSTRACT FROM AUTHOR]

SUMMARY This paper evaluates a forecasting procedure, which was based on the data for the period 1926-1993, in and around Japan. The period of the experiment is from 1994 to April 2011. According to the procedure, the probability that the first earthquake will be a foreshock varies in a range between 1 and 10-odd per cent, depending on its location. This location-dependent forecasting performed significantly better than the unconditional foreshock probability (3.7 per cent average) throughout the Japan region. Furthermore, when multiple earthquakes were observed as a cluster, the foreshock forecast probabilities varied in a much wider range, depending on the space-time distances and magnitude increments, between the earthquakes. Such forecasts performed better than the average foreshock probability, and the forecast probabilities were basically consistent with the outcomes. The improvements of the forecasts were objectively evaluated by using the log likelihood score. It is also shown that the forecasting procedure was robust enough and can be adopted for automatically determined earthquakes in real-time. [ABSTRACT FROM AUTHOR]

SUMMARY Earthquake forecasts are usually underinformed, and can be plagued by uncertainty in terms of the most appropriate model, and parameter values used in that model. In this paper, we explore the application of two different models to the same seismogenic area. The first is a renewal model based on the characteristic earthquake hypothesis that uses historical/palaeoseismic recurrence times, and fixed rupture geometries. The hazard rate is modified by the Coulomb static stress change caused by nearby earthquakes that occurred since the latest characteristic earthquake. The second model is a very simple earthquake simulator based on plate-motion, or fault-slip rates and adoption of a Gutenberg-Richter magnitude-frequency distribution. This information is commonly available even if historical and palaeoseismic recurrence data are lacking. The intention is to develop and assess a simulator that has a very limited parameter set that could be used to calculate earthquake rates in settings that are not as rich with observations of large-earthquake recurrence behaviour as the Nankai trough. We find that the use of convergence rate as a primary constraint allows the simulator to replicate much of the spatial distribution of observed segmented rupture rates along the Nankai, Tonankai and Tokai subduction zones. Although we note rate differences between the two forecast methods in the Tokai zone, we also see enough similarities between simulations and observations to suggest that very simple earthquake rupture simulations based on empirical data and fundamental earthquake laws could be useful forecast tools in information-poor settings. [ABSTRACT FROM AUTHOR]

SUMMARY This paper presents a scheme to compute coseismic deflection change, and presents a set of Green's functions for four independent dislocation sources. To compare the theoretical deflection changes with the GRACE-observed ones, the dislocation Love numbers are truncated and the Green's functions are computed with application of a Gaussian filter. Using this computing scheme, the modelled coseismic geoid and deflection changes can be compared directly with the GRACE-observed ones. Numerical computation and comparison show that the computing scheme is valid and efficient, and that it can obviate much unnecessary computing time for high-degree Love numbers. This study further examines the problem of sea water correction to modelled geoid and deflection changes because it is an important and necessary step to compare the modelled results with GRACE-observed deformations. As an application of the dislocation theory and the computing scheme described as a result of this study, we consider the 2011 Tohoku-Oki earthquake ( Mw 9.0) using three different fault-slip models. Using the fault models, we compute the coseismic geoid and deflection changes for an area around Japan, considering sea water corrections. Results indicate that the coseismic geoid and deflection changes can be detected clearly by GRACE observation, and the coseismic geoid change is not sensitive to the fault-slip models: the three slip models yield identical coseismic geoid changes; whereas the coseismic deflection changes are very sensitive to the fault-slip models because the modelled deflection changes indicate pretty large difference, especially for the E-W component. These behaviours provide us a new and useful approach to invert seismic faults using GRACE-observed deflection changes as constraints. [ABSTRACT FROM AUTHOR]

If aseismic slip occured on a fault or its deeper extension, both seismicity and crustal deformation around the source would be affected. Such anomalies are determined from earthquake occurrence data and geodetic records during the periods leading up to the 2007 March Noto Hanto earthquake of M6.9 and the 2007 July Chuetsu-oki earthquake of M6.8, which successively occurred after the 2004 October Chuetsu earthquake of M6.8 on the Japan Sea coast. Assuming such aseismic slips, seismic zones corresponding to negative and positive increments of the Coulomb failure stress show seismic quiescence and activation, respectively, relative to the rate predicted by the epidemic type aftershock sequence model. These interpretations are further supported by transient crustal movements around the sources preceding the ruptures, for example, the time-series of baseline distances between permanent Global Positioning System stations deviated from their long-term linear trends in a manner that is consistent with the assumed slip. In particular, this paper emphazes on the investigation of the interacting tectonic processes within such a narrow area and a short period, considering the precursory slow slips on the focal faults or their downdip extensions based on the observed anomalies in seismicity and crustal movement. [ABSTRACT FROM AUTHOR]

Since the discovery of deep low-frequency tremors in southwestern Japan, various types of low-frequency oscillation phenomena have been reported and analysed. Among these phenomena, non-volcanic, deep, low-frequency tremors in southwestern Japan are outstanding low-frequency oscillation phenomena occurring over a wide region along the Philippine Sea Plate. However, the weakness of these signals makes it difficult to apply usual analyses to derive source characteristics. In this paper, we apply a new algorithm to obtain a frequency structure from highly noisy data and reveal the characteristic frequency structure of deep low-frequency tremors with peaks lined up from 1 to 5 Hz at intervals of 0.5 Hz. In addition, we present one piece of evidence that the deep low-frequency tremors and low-frequency earthquakes have a common source mechanism. [ABSTRACT FROM AUTHOR]

Propagation of short period S waves through the crust is strongly controlled by scattering loss and intrinsic absorption. Scattering loss controls the shape of the seismic envelope and the spatial distribution of the energy while intrinsic absorption produces an exponential decay as a function of time which is independent of the coordinates. Since both parameters have a different effect, it is possible to compute them by means of the Multiple Lapse Time Window Analysis (MLTWA). In this paper, high resolution maps of Japan for both seismic S-wave attenuation parameters are obtained by using the MLTWA with data provided by the Hi-net seismic network. The maps show strong regional variation of the parameters. The variations depend mainly on the tectonic setting of each region and volcanism mechanisms. In the scattering loss maps, in the 1–2 Hz map, there is a very different behaviour between the northeast (stronger scattering loss) and the southwest (much weaker scattering loss). In the northeast of Japan, areas with strong scattering loss correspond mainly to the location of the volcanic arcs. In Southwest Japan, Chugoku region and the west side of Kii peninsula show strong scattering levels for the 4–8 Hz and higher frequency bands. In Hokkaido, strong scattering loss and intrinsic absorption can be observed for the lower frequency bands. In the Sendai plain, we find strong scattering loss for all the frequency bands. Under all these areas low velocity anomalies have been detected in previous works. In Tohoku area the east side shows lower intrinsic attenuation than the west side. This characteristic seems to extend to the southern part of Hokkaido and to central Japan. In Kanto area, the values of intrinsic absorption are lower than in the surrounding areas; but in the region around the Itoigawa-Shizuoka tectonic line and in its western side strong absorption areas are observed. The behaviour of each area for intrinsic absorption is similar to the behaviour observed in the maps. The MLTWA is based on the hypothesis of multiple isotropic scattering in a medium with the homogeneous distribution of scattering mean free path and intrinsic absorption. Although these hypotheses are very simple, the results are very informative on the characteristics of each region, and show the usefulness of studying the properties of scattered seismic waves in the understanding of the properties of the crust. [ABSTRACT FROM AUTHOR]

The 2003 Tokachi-oki earthquake ruptured a large area of approximately 100 km2. The location of the largest asperity was estimated to be several dozen kilometres offshore of Hokkaido, Japan. The magnitude measured 8.0 on the Japan Meteorological Agency scale, and several studies used waveform inversion analysis to estimate the moment-magnitude as Mw 8.0–8.2. Several studies reported that there was a minor asperity at the northeastern edge of the fault plane, and that the rupture velocity towards the minor asperity was less than that towards the main asperity. One of them illustrated that the location and timing of the minor asperity were poorly constrained. In this paper, we introduce a procedure based on semblance analysis to image the location of the minor northeastern asperity with improved resolution. We group 15 strong-motion seismographs into three arrays, and we perform semblance analysis on impulsive waves that were possibly generated from the minor asperity and were conspicuously observed at stations in eastern Hokkaido. By projecting the semblance values onto the fault plane, we estimate the location of the minor asperity. We find it to be shallower and farther from the coast than the previous results indicated. The average rupture velocity towards the asperity is estimated to be 2.5 km s−1, which is slower than the 3.6 km s−1 obtained by waveform inversion analysis. [ABSTRACT FROM AUTHOR]

In this paper, we present a new approach to estimate high-resolution teleseismic receiver functions using a simultaneous iterative time-domain sparse deconvolution. This technique improves the deconvolution by using reweighting strategies based on a Cauchy criterion. The resulting sparse receiver functions enhance the primary converted phases and its multiples. To test its functionality and reliability, we applied this approach to synthetic experiments and to seismic data recorded at station ABU, in Japan. Our results show Ps conversions at approximately 4.0 s after the primary P onset, which are consistent with other seismological studies in this area. We demonstrate that the sparse deconvolution is a simple, efficient technique in computing receiver functions with significantly greater resolution than conventional approaches. [ABSTRACT FROM AUTHOR]

An important element of seismic tomography is the inversion process. In this work, we use P -wave arrival times of local earthquakes recorded at onshore and offshore seismic stations in East Japan to investigate the influence of two well-known inversion algorithms (LSQR and L-BFGS-B) on anisotropic tomography. Our synthetic tests show that a large damping parameter in the LSQR algorithm can lead to a stable and fast convergence, but it can result in many small value disturbances. The L-BFGS-B algorithm, which has second-order convergence, could converge fast to the optimal solution without damping regularization, but an inappropriate bound on the unknown parameters makes them hard to be recovered fully and causes strong trade-off between isotropic velocity and azimuthal anisotropy. If appropriate control parameters are adopted, the two inversion algorithms lead to almost the same results, though the L-BFGS-B provides a more efficient convergence and leads to a slightly better fit to the data than LSQR does. The two algorithms are applied to investigate the 3-D P -wave velocity (V p) structure and azimuthal anisotropy of the East Japan subduction zone. Our results show that high- V p anomalies and trench-normal fast-velocity directions (FVDs) exist in the forearc crust beneath the Pacific Ocean off South Hokkaido, which may reflect a cold and hydrated forearc crust with aligned microcracks or fractures. Significant low- V p anomalies and trench-parallel FVDs exist at 40–80 km depths beneath Hokkaido, reflecting a water-rich mantle wedge with aligned B-type olivine. In the subducting Pacific slab, strong anisotropy with trench-parallel FVDs is revealed, reflecting localized horizontal bending of the slab. [ABSTRACT FROM AUTHOR]

Time-domain analyses of seismic waveforms have revealed diverse source complexity in large earthquakes (M w > 7). However, source characteristics of small earthquakes have been studied by assuming a simple rupture pattern on the frequency domain. This study utilizes high-quality seismic network data from Japan to systematically address the source complexities and radiated energies of M w 3–7 earthquakes on the time domain. We first determine the apparent moment-rate functions (AMRFs) of the earthquakes using the empirical Green's functions. Some of the AMRFs show multiple peaks, suggesting complex ruptures at multiple patches. We then estimate the radiated energies (ER) of 1736 events having more than ten reliable AMRFs. The scaled energy (eR = ER / M 0) does not strongly depend on the seismic moment (M 0), focal mechanisms, or depth. The median value of eR is 3.7 × 10−5, which is comparable to those of previous studies; however, eR varies by approximately one order of magnitude among earthquakes. We measure the source complexity based on the radiated energy enhancement factor (REEF). The values of REEF differ among earthquakes, implying diverse source complexity. The values of REEF do not show strong scale dependence for M w 3–7 earthquakes, suggesting that the source diversity of smaller earthquakes is similar to that of larger earthquakes at their representative spatial scales. Applying a simple spectral model (e.g. the ω 2-source model) to complex ruptures may produce substantial estimation errors in source parameters. [ABSTRACT FROM AUTHOR]

This study presents the shear wave velocity (VS) structures of sedimentary sequences and a section of the upper crustal layer in the Fukushima forearc region of the Japan Trench subduction zone, which were obtained by analysing the horizontal-to-vertical (H/V) spectral ratios of ambient vibration records. The H/V curves were derived using 31 d of continuous seismic data from 3 broad-band and 16 short-period ocean bottom seismometer (OBS) stations. Using the broad-band data, H/V ratios from 0.01 to 10 Hz were derived, but the ratios below 0.1 Hz frequencies were unusually large and temporally unstable. Characterization of seismic noise energy from ∼1 yr of seismic data of three broad-band OBSs revealed variable and elevated energy conditions below 0.1 Hz due to typical long-period oceanic noise; we link these observations with the unstable H/V ratios below this frequency. Therefore, H/V analysis was performed in the frequency range of 0.1–10 Hz for both broad-band and short-period OBSs to obtain subsurface VS profiles. For the forward calculation of the H/V ratios in the inversion process, we used the recently developed 'hvgeneralized' method, which is based on the diffuse field assumption, and accounts for the water layer on top of stratified media. Moreover, available prior geological and geophysical information was utilized during the inversion of the H/V curves. We found that subsurface VS ranged from approximately 30 m s−1 at the seabed to approximately 4900 m s−1 at 7000 m below the sea floor (mbsf). Starting with the best model candidate at each OBS location, the effect of the water layer on the H/V curve in the deep ocean was investigated by comparing synthetic H/V curves with and without the water layer. The synthetic H/V analysis revealed that the water layer had a significant effect on H/V amplitudes at higher frequencies (>1 Hz), whereas comparatively little effect was observed at lower frequencies (<1 Hz). This study provides an empirical basis for H/V analysis using OBS data to determine VS down to several kilometres of sedimentary sequences to the upper crust with high-resolution. [ABSTRACT FROM AUTHOR]

The average S-wave velocity (VS) in the upper 30 m (V S 30) is a proxy for seismic wave amplification. Microtremor array exploration is one of the available methods for site characterization, but the recorded data require complicated processing that can lead to different estimations of V S 30 depending on the analyst and processing software. We propose a method of estimating V S 30 by using derivatives obtained in the early stages of microtremor array data processing. Statistical analysis with 2376 virtually generated subsurface VS structure models revealed that the frequencies at which the spatial-autocorrelation (SPAC) coefficients and centreless circular array (CCA) coefficients take specific values strongly correlate with V S 30, which we used to develop formulas for estimating V S 30. Numerical validations using actual VS profiles at 616 sites in Japan showed that the proposed method could estimate V S 30 with a root-mean-square deviation (RMSD) of 57–80 m/s with SPAC coefficients and 56m/s with CCA coefficients. Our proposed methods were applicable for 98–100 per cent of the VS profiles when we limited our estimation to sites with V S 30 < 760 m/s. The results indicated that SPAC coefficients from arrays with radii of 8–20 m can be used for V S 30 estimation and are less affected by incoherent noise. In contrast, CCA coefficients are much more sensitive to incoherent noise, which resulted in the overestimation of V S 30. The estimated V S 30 values from the recorded microtremor array data were in good agreement with the reference values from the actual VS profiles. The proposed method allows for robust and efficient V S 30 estimation without relying on the analyst's skills or software. [ABSTRACT FROM AUTHOR]

How are the sizes of the earthquakes determined? To solve this important problem, we analysed the data from a dense temporary seismic observation network installed in the aftershock area of the 2016 Mw 6.2 Central Tottori earthquake, which occurred in an intraplate region in Japan. We compared the stress field estimated from approximately 10   000 accurate focal mechanisms of aftershocks with the calculated post-earthquake stress field and found that the differential stress before the earthquake was very small near both horizontal edges. These results did not depend significantly on the modeled slip distribution and the orientation of the principal stress before the earthquake. Similar results were obtained for the 2000 Mw 6.7 Western Tottori earthquake, which also occurred in the same intraplate region in Japan. These results suggest that the fault size of large intraplate earthquakes can be determined by the region of small differential stress surrounding future earthquake faults. [ABSTRACT FROM AUTHOR]

During moderate to deep (35–260 km) earthquakes within the Philippine-sea slab along the Ryukyu arc, distinctive later phases after S are observed across the Japanese archipelago for epicentral distances from 1500 to 2200 km, producing anomalous amplification of ground motion in central and northern Japan. Broad-band observations show that these later phases have a faster apparent wave speed (7–9.5 km s−1) than S , and are dominant at low frequencies (0.05–1 Hz), indicating S -wave reflections returned from the upper mantle with strong attenuation for high frequencies. Numerical simulation of seismic wave propagation in 3-D models including subduction of both the Pacific (PAC) and Philippine-sea (PHS) plates reveal the origin of these reflections. The triplicated S wave front from the 410 km discontinuity undergoes wide-angle reflection at the top of the Pacific slab bringing strong amplitudes in a narrow band of epicentral distance near 1500 km. Also, wide-angle S reflections at the 660 km discontinuity are reinforced by refraction when travelling through the high-wave speed Pacific slab; these arrive at the surface beyond 2000 km epicentral distance. The characteristics of the deep mantle reflections and the distribution of large ground motion across Japan due to the reflections are strongly dependent on the source depth and distance to the Ryukyu earthquakes. Tomographic imagery for western Japan indicates loss of the high-wave speed signal of the PAC slab below 200 km depth; the character of the large S reflections and pattern of enhanced ground motions favours a model in which the PAC slab is thinned rather than fully broken. The peculiar pattern of ground motion from Ryukyu earthquakes can be useful for constraining deep slab structures that are difficult to identify based on tomography. [ABSTRACT FROM AUTHOR]

We present new absolute archaeointensity data from six archaeological sites situated in the Okayama Prefecture, Japan. The materials studied are well-dated fragments from pottery, ceramic coffins and haniwa artefacts. Their ages range from 160 AD to 675 AD, covering the Late Yayoi and Kofun periods. Rock magnetic experiments suggest the presence of magnetite and/or Ti-magnetite as the main carrier of the remanence, with a possible minor contribution of higher coercivity minerals. After thermal demagnetization experiments, the most magnetically stable samples were selected for archaeointensity analysis performed following the double-heating method proposed by Thellier and modified by Coe. Partial thermoremanent magnetization (pTRM) checks and pTRM tail-checks were performed for monitoring possible chemical alterations during heating. All measurements were corrected for both anisotropy and cooling-rate effects. Successful archaeointensity determinations, following rigorous selection criteria, were obtained for samples from all the investigated archaeological sites. Compared with literature data from Japan, the new high-quality data show significantly lower intensity values. They also reveal possible fast secular variation changes during the Late Yayoi period and very weak geomagnetic intensity field around 630 AD. Such values offer evidence of a possible recurrence of weak intensity field in East Asia, suggesting an ancient recurrence of the West Pacific Anomaly. The new data might change the archaeomagnetic field models interpretations in the area, even though more data are still necessary to better understand the secular variation in Japan and the temporal evolution of the geomagnetic field's behaviour in East Asia. [ABSTRACT FROM AUTHOR]

The ability to accurately and reliably obtain images of shallow subsurface anomalies within the Earth is important for hazard monitoring and a fundamental understanding of many geologic structures, such as volcanic edifices. In recent years, machine learning (ML) has gained increasing attention as a novel approach for addressing complex problems in the geosciences. Here we present an ML-based inversion method to integrate cosmic-ray muon and gravity data sets for shallow subsurface density imaging at a volcano. Starting with an ensemble of random density anomalies, we use physics-based forward calculations to find the corresponding set of expected gravity and muon attenuation observations. Given a large enough ensemble of synthetic density patterns and observations, the ML algorithm is trained to recognize the expected spatial relations within the synthetic input–output pairs, learning the inherent physical relationships between them. Once trained, the ML algorithm can then interpolate the best-fitting anomalous pattern given data that were not used in training, such as those obtained from field measurements. We test the validity of our ML algorithm using field data from the Showa-Shinzan lava dome (Mt Usu, Japan) and show that our model produces results consistent with those obtained using a more traditional Bayesian joint inversion. Our results are similar to the previously published inversion, and suggest that the Showa-Shinzan lava dome consists of a relatively high-density (2200–2400 km m–3) cylindrical anomaly, about 300 m in diameter. Adding noise to synthetic training and testing data sets shows that, as expected, the ML algorithm is most robust in areas of high sensitivity, as determined by the forward kernels. Overall, we discover that ML offers a viable alternate method to a Bayesian joint inversion when used with gravity and muon data sets for subsurface density imaging. [ABSTRACT FROM AUTHOR]

Enigmatically strong tube waves continue to exist long after the direct wave during repeat crosswell-monitoring surveys at the Nagaoka CO2 injection site in Japan. The tube waves, which have linear moveouts with velocities of 1.29 and 1.41 km s−1 at plastic and steel casings, respectively, are generated by double scattering on the shallow side of the wells. We characterize wavefields to confirm that these late coda waves are tube waves that are excited at the source well, propagate upward, then travel to a receiver well as a body wave, and finally propagate downward along the receiver well. Even though these tube waves result from second-order scattering, they have large amplitudes because of the relatively short distance between source and receiver wells. Because these waves propagate along both wells many times, we can extract detailed information of wave propagation by averaging them to increase the signal-to-noise ratio. We first use the tube waves to relocate sources and receivers between multiple monitoring surveys, demonstrating our ability to correct the severe noise caused by location errors that frequently degrades the repeatability of time-lapse surveys. Then, we apply an inversion based on the physics of tube waves to estimate the location and strength of the scatterers and find that they are predominantly located in the shallow segment of the wells, above the sources, and infer that they are related to local fractures and/or wellbore conditions because their locations do not correspond to the well geometry. Lastly, we use the tube waves to accurately estimate subsurface velocities along the wells. The estimation is stable and robust, and the velocities follow the general trend of subsurface structure seen in the well log data. Due to the receiver spacing, tube-wave analysis cannot resolve a thin, high-velocity layer at the CO2 reservoir. By combining tube waves observed during different stages of the monitoring survey, we can estimate the time-lapse changes of the subsurface velocities. [ABSTRACT FROM AUTHOR]

Ocean bottom distributed acoustic sensing (OBDAS) is emerging as a new measurement method providing dense, high-fidelity and broad-band seismic observations from fibre-optic cables deployed offshore. In this study, we focus on 35.7 km of a linear telecommunication cable located offshore the Sanriku region, Japan, and apply seismic interferometry to obtain a high-resolution 2-D shear wave velocity (VS) model below the cable. We first show that the processing steps applied to 13 d of continuous data prior to computing cross-correlation functions (CCFs) impact the modal content of surface waves. Continuous data pre-processed with 1-bit normalization allow us to retrieve dispersion images with high Scholte-wave energy between 0.5 and 5 Hz, whereas spatial aliasing dominates dispersion images above 3 Hz for non-1-bit CCFs. Moreover, the number of receiver channels considered to compute dispersion images also greatly affects the resolution of extracted surface-wave modes. To better understand the remarkably rich modal nature of OBDAS data (i.e. up to 30 higher modes in some regions), we simulate Scholte-wave dispersion curves for stepwise linear VS gradient media. For soft marine sediments, simulations confirm that a large number of modes can be generated in gradient media. Based on pre-processing and theoretical considerations, we extract surface wave dispersion curves from 1-bit CCFs spanning over 400 channels (i.e. ∼2 km) along the array and invert them to image the subsurface. The 2-D velocity profile generally exhibits slow shear wave velocities near the ocean floor that gradually increase with depth. Lateral variations are also observed. Flat bathymetry regions, where sediments tend to accumulate, reveal a larger number of Scholte-wave modes and lower shallow velocity layers than regions with steeper bathymetry. We also compare and discuss the velocity model with that from a previous study and finally discuss the combined effect of bathymetry and shallow VS layers on earthquake wavefields. Our results provide new constraints on the shallow submarine structure in the area and further demonstrate the potential of OBDAS for high-resolution offshore geophysical prospecting. [ABSTRACT FROM AUTHOR]

The aftershock distribution of the 1933 Sanriku-oki outer trench earthquake is estimated by using modern relocation methods and a newly developed velocity structure to examine the spatial extent of the source-fault and the possibility of a triggered interplate seismicity. In this study, we first examined the regional data quality of the 1933 earthquake based on smokedpaper records and then relocated the earthquakes by using the 3-D velocity structure and double-difference method. The improvements of hypocentre locations using these methods were confirmed by the examination of recent earthquakes that are accurately located based on ocean bottom seismometer data. The results show that the 1933 aftershocks occurred under both the outer- and inner-trench-slope regions. In the outer-trench-slope region, aftershocks are distributed in aĀ280-km-long area and their depths are shallower than 50 km. Although we could not constrain the fault geometry from the hypocentre distribution, the depth distribution suggests the whole lithosphere is probably not under deviatoric tension at the time of the 1933 earthquake. The occurrence of aftershocks under the inner trench slope was also confirmed by an investigation of waveform frequency difference between outer and inner trench earthquakes as recorded at Mizusawa. The earthquakes under the inner trench slope were shallow (depth ≦30 km) and the waveforms show a low-frequency character similar to the waveforms of recent, precisely located earthquakes in the same area. They are also located where recent activity of interplate thrust earthquakes is high. These suggest that the 1933 outer-trenchslope main shock triggered interplate earthquakes, which is an unusual case in the order of occurrence in contrast with the more common pairing of a large initial interplate shock with subsequent outer-slope earthquakes. The off-trench earthquakes are distributed about 80 km width in the trench perpendicular direction. This wide width cannot be explained from a single high-angle fault confined at a shallow depth (depth ≦50 km). The upward motion of the 1933 tsunami waveform records observed at Sanriku coast also cannot be explained from a single high-angle west-dipping normal fault. If we consider additional fault, involvement of highangle, east-dipping normal faults can better explain the tsunami first motion and triggering of the aftershock in a wide area under the outer trench slope. Therefore multiple off-trench normal faults may have activated during the 1933 earthquake. We also relocated recent (2001- 2012) seismicity by the same method. The results show that the present seismicity in the outer-trench-slope region can be divided into several groups along the trench. Comparison of the 1933 rupture dimensions based on our aftershock relocations with the morphologies of fault scarps in the outer trench slope suggest that the rupture was limited to the region where fault scarps are largely trench parallel and cross cut the seafloor spreading fabric. These findings imply that bending geometry and structural segmentation of the incoming plate largely controls the spatial extent of the 1933 seismogenic faulting. In this shallow rupture model for this largest outer trench earthquake, triggered seismicity in the forearc and structural control of faulting represent an important deformation styles for off-trench and shallow megathrust zones. [ABSTRACT FROM AUTHOR]

We present a systematic study of seismic velocity changes associated with a megathrust and five strong crustal earthquakes in Japan. We perform both cross-correlation and singlestation cross-correlation analysis for station pairs and stations, respectively. The correlation of ambient seismic noise allows us to reconstruct the Green's functions of the wave propagation. By relating the coda parts of the daily Green's functions with the long-term reference Green's functions, shear wave velocity changes are determined. We analyse data from four areas in Japan where large earthquakes occurred: Iwate-Miyagi (2008 MW 6.9 Iwate-Miyagi Nairiku earthquake), Niigata (2004 MW 6.6 Chūetsu, 2007 MW 6.6 Chūetsu-oki and 2011 MW 6.2 Nagano/Niigata earthquakes), Noto Peninsula (2007 MW 6.7 Noto Hantō earthquake) and Fukuoka (2005 MW 6.6 Fukuoka earthquake). In all areas, we analyse time-series which start before the respective earthquakes and last until after the 2011 MW 9.0 Tōhoku-oki earthquake. The analysis in five different frequency ranges between 0.125 and 4.0 Hz yields time-series of the velocity changes for the different station pairs or stations. At the time of the respective earthquakes, we observe coseismic velocity drops in all areas which are followed by a partial post-seismic recovery process. For the Tōhoku-oki earthquake, coseismic velocity drops can also be observed in all regions. There is a general trend of increasing coseismic velocity drops with frequency in all four areas. The largest coseismic drops are observed close to the fault zones. Over the observed time range, the post-seismic recovery is only partial and around half of the coseismic velocity drops do not recover. The characteristic recovery times for the recovering part are similar in all areas and frequency ranges, with an average value of 0.55 yr. We model the volumetric strain changes for the different earthquakes and find that the observed pattern of the coseismic velocity drops cannot be explained by these models. The coseismic velocity drops at the different stations are better related with the peak ground velocities and the associated dynamic strain than with the peak ground accelerations, but the correlation is still poor. This suggests that non-linear effects caused by the strong ground motion during the earthquake can explain at least part of the coseismic velocity drops. [ABSTRACT FROM AUTHOR]

To obtain an internal S -wave velocity structure, we conducted a passive seismic campaign with 21 1-Hz seismometers on and around the Showa-Shinzan lava dome, which emerged during the 1943–1945 eruption of Usu Volcano, Japan. Before the campaign, we calibrated seismometers and found slight phase-response differences between seismometers of less than 1–2°. After the campaign, we extracted seismic wavefield by taking cross-correlations of vertical-component ambient noise records between seismic sites. We developed a new method to measure phase velocities of the Rayleigh wave automatically by assuming layered structure and finally obtained 1-D S -wave velocity models in summit, roof and base regions. The obtained S -wave velocity right beneath the intruded lava dome is higher than that in surrounding areas by a few tens of per cent down to a few hundred metres below sea level, indicating narrow but deep existence of the root of the lava dome. The obtained S -wave velocity at depths shallower than ∼50 m inside the lava dome in the summit area was ∼1 km s−1, significantly lower than that predicted from the density of  2.3 × 103 kg m−3 estimated in previous muon-radiography studies and a conventional scaling, indicating the effect of cracking in the lava dome. [ABSTRACT FROM AUTHOR]

Sedimentary basins can strongly amplify seismic waves from earthquakes. To better predict future ground motions, detailed knowledge of the sediment thickness and internal structure of basins is required. We image the sediment-to-bedrock interface of the Kanto Basin in Japan using the P -wave reflectivity response from earthquake and ambient seismic noise autocorrelation functions (ACFs) at 286 shallow borehole stations. Earthquake ACFs are computed using P -wave records from 50  M w 6+ teleseismic events. Noise ACFs are obtained using 1 month of continuous data. Both methods are used to retrieve P -wave traveltimes between the surface and the bedrock interface and map the basin basement geometry. Our prediction of the basement depth agrees generally well with that from a reference velocity model, except for smoother variations in the central part of the basin. Using full-wavefield simulations, we show that the nature of the autocorrelated wavefield has a significant impact on the shape of the ACF waveforms and that earthquake ACFs yield more accurate results in the Kanto Basin. [ABSTRACT FROM AUTHOR]

We propose a Bayesian approach for jointly determining the locations of earthquake hypocentres and velocity structures in a region divided by a structural boundary, such as a fault zone. To accommodate the Markov chain Monte Carlo (MCMC) method in a region containing lateral velocity contrast, we incorporated a data-driven station clustering scheme when estimating earthquake hypocentre locations, a 1-D velocity model assigned to each cluster and station corrections for traveltimes. The proposed method realizes the clustering of stations reflecting velocity structures beneath the station network in the framework of the MCMC method. We verified the performances of the proposed method using numerical experiments with different distributions of earthquake hypocentres relative to the structural boundary. The results demonstrated that the proposed method reconstructed earthquake hypocentre locations and velocity structures involving lateral contrast robustly to earthquake distributions. Additionally, the number of 1-D velocity models (the number of station clusters) was optimized by comparing resultant posterior probability proportional to distances between the observed and theoretical traveltimes. We also applied the proposed method to estimate the earthquake hypocentre locations and velocity structures around the source area of the 2004 Mid-Niigata prefecture earthquake, Japan, to examine the applicability of the proposed method to real observation data. The stations were divided into east and west clusters, coinciding with the lateral velocity variation related to the geological setting. Therefore, the proposed method accurately determines earthquake hypocentres in a region divided by a structural boundary without any knowledge regarding the geological structure. [ABSTRACT FROM AUTHOR]

We present a fully Bayesian inversion of kinematic rupture parameters for the 2011 Mw9 Tohoku-oki, Japan earthquake. Albeit computationally expensive, this approach to kinematic source modelling has the advantage of producing an ensemble of slip models that are consistent with physical a priori constraints, realistic data uncertainties, and realistic but simplistic uncertainties in the physics of the kinematic forward model, all without being biased by non-physical regularization constraints. Combining 1 Hz kinematic GPS, static GPS offsets, seafloor geodesy and near-field and far-field tsunami data into a massively parallel Monte Carlo simulation, we construct an ensemble of samples of the posterior probability density function describing the evolution of fault rupture. We find that most of the slip is concentrated in a depth range of 10–20 km from the trench, and that slip decreases towards the trench with significant displacements at the toe of wedge occurring in just a small region. Estimates of static stress drop and rupture velocity are ambiguous. Due to the spatial compactness of the fault rupture, the duration of the entire rupture was less than approximately 150 s. [ABSTRACT FROM PUBLISHER]

Classification of volcanic tremors and earthquakes is an important part of volcano monitoring. Conventional classification technique relies on visual characterization of the amplitudes, frequency, or duration of seismic signals. In some cases, such classification is difficult to be maintained continuously, especially during high eruptive activities. In the present study, we develop a classification technique based on analyses of seismic wave correlation. The technique is based on a measure of spatial coherence using the distribution width of eigenvalues from the seismic covariance matrix that is referred to as 'spectral width'. We use the spectral width characteristics to extract the frequencies where the seismic signals are the most spatially coherent. We analyse 6-month continuous seismic data at six seismic stations deployed on Sakurajima volcano, Japan. The classification is performed every 10 min to recognize volcanic tremor and B-type earthquakes, both of which show unclear onsets and are dominant at around 1–4 Hz. Their frequencies of high spatial coherence are different from each other, providing a basis for automatic classification of both types of seismic events. Our classification results show that an increase in seismic activity, particularly volcanic tremors, occurred during high eruptive activity and is well matched with the events that are routinely determined by the Japan Meteorological Agency (JMA) based on visual examination of the data. We discuss the volcanic activities by combining our classification results with those from detections and location methods based on seismic correlation. Most of the source locations of volcanic tremor and B-type earthquakes are distributed at the same region with depths of less than 4 km beneath the active craters, which are almost consistent with the source locations of explosion earthquakes and B-type earthquakes determined from P -wave onsets in the previous studies. Also, the distribution of horizontal locations is spatially elongated at shallow depths above a dyke that is inferred to extend from Aira caldera at a depth of 7–9 km. We obtain deep and shallow sources for the tremor that occurred before and after an explosive eruption, respectively. We also obtain shallow sources beneath the active crater during a continuous tremor that accompanies increasing eruptive activity. The source locations of the classified volcanic tremor and B-type earthquakes may represent the movement of magma and gas from the shallow magma chamber at 5 km depth to the active craters. [ABSTRACT FROM AUTHOR]

Iterative least-squares reverse time migration (LSRTM) is the state-of-the-art linearized waveform inversion method to obtain quantitative subsurface parameters. The main drawback of such an iterative imaging scheme is the significant computational expense of many modelling/adjoint cycles through iterations. In the context of the extended domain, an interesting alternative to LSRTM is the asymptotic linearized direct waveform inversion, providing quantitative results with only a single iteration. This approach was first proposed for constant-density acoustics and recently extended to the variable-density case. The former is based on the application of the asymptotic inverse Born operator, whereas the latter has two more extra steps: building an angle-dependent response of the asymptotic inverse Born operator and then solving a weighted least-squares approach for simultaneous inversion of two acoustic parameters. To examine the importance of accounting for density variations, we compare constant- and variable-density linearized direct waveform inversion techniques applied to a marine real data set from the Eastern Nankai Trough, offshore Japan. The inversion results confirm the efficiency of the asymptotic linearized direct waveform inversion in estimating quantitative parameters within a single iteration. The variable-density direct inversion yields subsurface images that (1) exhibit a superior resolution and (2) better reconstruct the field data than does the constant-density approach, even if the data set does not contain large enough surface offset to fully decompose velocity and density perturbations. [ABSTRACT FROM AUTHOR]

Magnetotelluric (MT) observations have revealed subvertical electrical conductors that extend from shallow depths into the mid-crust at various geothermal zones, active volcanoes and active faults worldwide. These deeply rooted subvertical conductors have typically been interpreted to represent entire zones of dedicated fluid transport through the crust. We estimate the high-resolution 3-D crustal resistivity structure below the Kuju Volcanoes, Japan, using dense observations from 153 broad-band MT measurement sites and 40 telluric measurement sites. The resistivity structure highlights subvertical conductors that merge into a deep conductor to the north of the volcanoes, with deep low-frequency earthquakes occurring near the southeastern edge of this subvertical conductor at 10–30 km depth. This deep conductor branches into several subvertical conductors at 2–10 km depth, coinciding with a shallow zone where tectonic earthquakes rarely occur. The surface expressions of active geothermal areas and past volcanic eruptions are all located above the edges of the conductors at 2–6 km depth. Widespread conductive layers exist around the volcanoes above 2 km depth, and their distribution approximately corresponds to a low-gravity-anomaly zone. We discuss the nature of these subvertical conductors, the potential causes of their complex structure and their relationship to local magmatic fluid transport. These subvertical conductors, a shallow clay-rich layer, developed fracture systems and high-strength solidified magma may all contribute to magmatic fluid transport to the surface at the Kuju Volcanoes. In this study, we add the possibility that the edges of these subvertical conductors act as important magmatic fluid pathways. [ABSTRACT FROM AUTHOR]

Palaeomagnetic records from geological archives provide significant information about the nature of geomagnetic polarity reversals; however, there are few detailed palaeomagnetic records of pre-Pleistocene reversals. The lower Mammoth Subchron boundary (late Pliocene) is recorded in a 10-m interval of a marine succession deposited at high accumulation rates (9–66 cm kyr–1) in the Boso Peninsula, central Japan. Here, we report a continuous palaeomagnetic record of the lower, normal to reverse boundary interval of the Mammoth Subchron, including the geomagnetic field direction and relative palaeointensity, with an average temporal resolution of ca. 800 yr. A hybrid method of thermal demagnetization at 200 °C and progressive alternating field demagnetization were used to effectively extract the primary palaeomagnetic component, which is carried by magnetite. The lower Mammoth transition is characterized by palaeomagnetic direction of instability and decay of the relative palaeointensity, and occurred from late Marine Isotope Stage MG3 (3351 ka) to MG2 (3336 ka) or MG1 (3331 ka), spanning 15–20 kyr. Virtual geomagnetic poles (VGPs), calculated from primary palaeomagnetic directions, rapidly rebounded twice from southern latitudes to northern latitudes within the transition. In contrast to the complex lower Mammoth reversal behaviour recorded in the Boso Peninsula succession, records from a lava sequence in O'ahu (Hawai'i) reveal a rebound following a 180° directional change, and those from a marl succession in Sicily (Italy) indicate a single rapid directional change. Diverse geomagnetic field evolution among these three sections is reflected resolution difference among the records likely in combination with an influence of non-axial dipole field. [ABSTRACT FROM AUTHOR]

Gravity monitoring might require observation of temporal changes of gravity below the nanometre per second squared (nm s−2) level, which can be achieved by precise isolation of the signal of interest from all other disturbing effects. One method of signal isolation is elimination of disturbing effects by taking the difference between gravity changes measured simultaneously using two gravimeters installed close together. Herein, we describe differences in temporal gravity changes below the nm s−2 level in the tidal frequency bands as observed through simultaneous measurements taken with three superconducting gravimeters (SGs) located 80, 93 and 94 m from the coastline in Tomakomai, Hokkaido, northern Japan. Those changes are consistent with differences in ocean tide loading effects on gravity at the SG locations computed using the software package GOTIC2, which uses a highly accurate land–sea boundary and ocean tide model near our site. The observed ocean tide loading differences were found to result from Newtonian attraction of the ocean tide mass within an angular distance of 0.003° from the SG locations. This result suggests that coastal observations of differential tidal gravity variations at different distances from the coast help to validate ocean tide loading computation models in the immediate vicinity of the SG stations. This method enables observation of non-periodic gravity changes occurring below the nm s−2 level over a few hours. Its salient benefit is that rapid and simple observation can be achieved without long-term continuous measurements, which is necessary for observing that level of gravity change with only one SG. [ABSTRACT FROM AUTHOR]

We propose a new array covariance matrix analysis method, named 'maximum covariance direction method' (MCD method), to detect and locate unconventional seismic sources of weak signals without clear onsets. The MCD method builds a normalized-covariance matrix of time-shifted seismic waveforms recorded in a seismic array and determines the existence of source based on the consistency of the maximum covariance direction with the theoretical prediction. Synthetic tests demonstrate effectiveness of the MCD method in detecting one and multiple isolated sources with low signal-to-noise ratios. As a data application, we study 1-hr long-period tremors (LPTs) around Aso Volcano of Japan in 2014 November 24. A total of 26 LPTs are detected near the Naka-dake first crater of Aso Volcano, with the uncertainties of source location of about 7 km. Using the recorded background noise at the seismic stations, we show that the MCD method can detect LPTs even when the LPT signals are buried in the background noise and become indiscernible in the seismic data. Unlike traditional methods that employ the coherent features of seismic signals for source detection, the MCD method places emphases on both the coherence of seismic signals and consistency of the direction of the coherent signals from a potential source location. The synthetic tests and data application indicate that the MCD method provides a good alternative to other traditional methods for detecting and locating unconventional seismic sources, with a major improvement of avoiding source misidentification in the presence of strong incoherent signals. [ABSTRACT FROM AUTHOR]