Long-period Seismic Waveforms Research Articles

Dynamic movement history of the Iiyama slope failure revealed from seismic data

The Iiyama slope failure occurred in the early morning of May 19, 2017, in Nagano Prefecture, central Japan. It caused a debris flow in the Idegawa River, resulting in the evacuation of the residents of the neighboring areas for half a year. Since the source was located in a mountainous area, the cause of the debris flow was not discovered immediately.To investigate the mechanism of this slope failure, we performed a field survey and obtained aerial photos of the slope failure using an unmanned aerial vehicle. Additionally, we analyzed seismic data recorded by the surrounding stations, to identify the ground motion generated by the mass movements.We found that the Iiyama slope failure involved two major mass movements: a large landslide and a subsequent debris flow triggered by the landslide. The movement of the landslide was recorded with long-period seismic waveforms, and the estimated force history showed a unidirectional movement from north to south for 50 s, reaching a maximum velocity of about 6.6 m/s. Following, debris flew for about 5 km from the top of the landslide source area. The initial collapse produced a large shaking with a short-period component, and the signal gradually decayed as the debris flow traveled along the river.Our study shows that the two mass movements in the Iiyama slope failure generated ground motions with completely different frequencies. This suggests that the ground motions recorded by the seismometer are significant for understanding the movement mechanism of slope failure. The seismic signal, combined with the aerial photos, helps us to understand the dynamic movement history of the landslide.

Variation patterns of landslide basal friction revealed from long-period seismic waveform inversion

A catastrophic landslide struck the Xiaoba village in Fuquan, Guizhou, southwestern China at about 8:30 p.m. (Beijing Time, UTC + 8) on August 27, 2014. The landslide and induced impulse water waves destroyed two villages and killed 23 persons. By reprocessing seismic signals from a seismic network deployed in the surrounding area of the landslide, we recognized the event from low-frequency seismic signals and subsequently performed a long-period seismic waveform inversion to obtain its force–time history. The inversion results reveal that the maximum force for the landslide is 5 × 109 N, and the duration of the landslide is 38.4 s. The landslide reached its maximum velocity of 12.4 m/s at 13.2 s after its initiation, and the mass center plugged into the quarry at 24.2 s. Based on the inversion results, we estimated basal friction of the landslide. We found the friction coefficient rapidly reduces to a relatively steady-state value of ~ 0.4 at a steady-state distance of 35 m and subsequently reduces in a near-linear manner that satisfies the empirical formula mu = - 1.4d + 0.44 , where d is sliding distance in km. The reduction in friction revealed by the formula is compatible with the finding of previous studies for landslides of similar volume in landslide acceleration stage. However, our result does not make it possible for the friction coefficient to increase again in landslide deceleration stage that a velocity-dependent friction law would allow. The friction variation patterns can be used to constrain input parameters in numerical landslide simulation, which can predicate runout distance and deposit areas for massive landslides to carry out landslide hazard assessment.

Broadband-seismic analysis of a massive landslide in southwestern China: Dynamics and fragmentation implications

We analyse broadband seismic records generated by the massive Xinmo landslide that occurred in southwestern China at 5:39 on 24 June 2017 (Beijing Time, UTC + 8). Characteristics of landslide seismic signals, including propagation velocity, duration and frequency contents, are presented and discussed by comparing them with seismic signals generated by a local earthquake of similar magnitude. We perform a long-period seismic waveform inversion in the frequency band of 0.01–0.1 Hz to acquire the landslide force history using 46 seismic traces from 21 seismic stations with distances from the epicentre extending to 360 km. Based on the inversion, we calculate the dynamic parameters of the landslide and recognize its stages and movement characteristics. Landslide basal friction coefficients are also acquired using both a constant dip and varying dips along the slope. A combined analysis of near-field short-period seismic records and dynamic inversion results indicates that the high-frequency energy of landslide signals is most likely generated by rock fragmentations during collision and scraping.

Focal Mechanisms of the 2016 Central Italy Earthquake Sequence Inferred from High-Rate GPS and Broadband Seismic Waveforms

Numerous shallow earthquakes, including a multitude of small shocks and three moderate mainshocks, i.e., the Amatrice earthquake on 24 August, the Visso earthquake on 26 October and the Norcia earthquake on 30 October, occurred throughout central Italy in late 2016 and resulted in many casualties and property losses. The three mainshocks were successfully recorded by high-rate Global Positioning System (GPS) receivers located near the epicenters, while the broadband seismograms in this area were mostly clipped due to the strong shaking. We retrieved the dynamic displacements from these high-rate GPS records using kinematic precise point positioning analysis. The focal mechanisms of the three mainshocks were estimated both individually and jointly using high-rate GPS waveforms in a very small epicentral distance range (<100 km) and unclipped regional broadband waveforms (100~600 km). The results show that the moment magnitudes of the Amatrice, Visso, and Norcia events are Mw 6.1, Mw 5.9, and Mw 6.5, respectively. Their focal mechanisms are dominated by normal faulting, which is consistent with the local tectonic environment. The moment tensor solution for the Norcia earthquake demonstrates a significant non-double-couple component, which suggests that the faulting interface is complicated. Sparse network tests were conducted to retrieve stable focal mechanisms using a limited number of GPS records. Our results confirm that high-rate GPS waveforms can act as a complement to clipped near-field long-period seismic waveform signals caused by the strong motion and can effectively constrain the focal mechanisms of moderate- to large-magnitude earthquakes. Thus, high-rate GPS observations extremely close to the epicenter can be utilized to rapidly obtain focal mechanisms, which is critical for earthquake emergency response operations.

The Puzzle of the 1996 Bardarbunga, Iceland, Earthquake: No Volumetric Component in the Source Mechanism

A volcanic earthquake with M w 5.6 occurred beneath the Bardarbunga caldera in Iceland on 29 September 1996. This earthquake is one of a decade-long sequence of ![Graphic][1] events at Bardarbunga with non-double-couple mechanisms in the Global Centroid Moment Tensor catalog. Fortunately, it was recorded well by the regional-scale Iceland Hotspot Project seismic experiment. We investigated the event with a complete moment tensor inversion method using regional long-period seismic waveforms and a composite structural model. The moment tensor inversion using data from stations of the Iceland Hotspot Project yields a non-double-couple solution with a 67% vertically oriented compensated linear vector dipole component, a 32% double-couple component, and a statistically insignificant (2%) volumetric (isotropic) contraction. This indicates the absence of a net volumetric component, which is puzzling in the case of a large volcanic earthquake that apparently is not explained by shear slip on a planar fault. A possible volcanic mechanism that can produce an earthquake without a volumetric component involves two offset sources with similar but opposite volume changes. We show that although such a model cannot be ruled out, the circumstances under which it could happen are rare. [1]: /embed/inline-graphic-1.gif

A new technique for moment tensor inversion with applications to the 2008 Wenchuan MS8.0 earthquake sequence

A MS8.0 earthquake occurred in Wenchuan County, Sichuan Province, China, on May 12, 2008, and subsequently, numerous aftershocks followed. We obtained the moment tensor solutions and source time functions (STFs) for the Wenchuan earthquake and its seven larger aftershocks (MS5.0∼6.0) by a new technique of moment tensor inversion using the broadband and long-period seismic waveform data from the Global Seismic Network (GSN). Firstly, the theoretical background and technical flow of the new technique was briefly introduced, and an aftershock of the Wenchuan earthquake sequence was employed to illustrate the real procedure for inverting the moment tensor; secondly, the moment tensor solutions and STFs of the eight events, including the main shock, were presented, and finally, the interpretation of the results was made. The agreement of our results with the GCMT results indicates the new approach is efficient and feasible. By using this approach, not only the moment tensor solution can be obtained but also the STF can be retrieved; the inverted STFs indicate that the source rupture process may be complicated even for the moderate earthquakes. The inverted focal mechanisms of the Wenchuan earthquake sequence show that the most of the aftershocks occurred in the main faults of the Longmenshan fault zone with predominantly thrustingwith minor right-lateral strike-slip component, but some of them may have occurred in the subfaults with strike-slip faulting in the vicinity of the main faults.

Insights into the nature of the transition zone from physically constrained inversion of long-period seismic data

Imposing a thermal and compositional significance to the outcome of the inversion of seismic data facilitates their interpretation. Using long-period seismic waveforms and an inversion approach that includes constraints from mineral physics, we find that lateral variations of temperature can explain a large part of the data in the upper mantle. The additional compositional signature of cratons emerges in the global model as well. Above 300 km, we obtain seismic geotherms that span the range of expected temperatures in various tectonic regions. Absolute velocities and gradients with depth are well constrained by the seismic data throughout the upper mantle, except near discontinuities. The seismic data are consistent with a slower transition zone and an overall faster shallow upper mantle, which is not compatible with a homogenous dry pyrolite composition. A gradual enrichment with depth in a garnet-rich component helps to reduce the observed discrepancies. A hydrated transition zone would help to lower the velocities in the transition zone, but it does not explain the seismic structure above it.

Rupture Kinematics of the 2005Mw 8.6 Nias–Simeulue Earthquake from the Joint Inversion of Seismic and Geodetic Data

AbstractThe 2005 Mw 8.6 Nias–Simeulue earthquake was caused by rupture of a portion of the Sunda megathrust offshore northern Sumatra. This event occurred within an array of continuous Global Positioning System (gps) stations and produced measurable vertical displacement of the fringing coral reefs above the fault rupture. Thus, this earthquake provides a unique opportunity to assess the source characteristics of a megathrust event from the joint analysis of seismic data and near-field static co-seismic displacements. Based on the excitation of the normal mode data and geodetic data we put relatively tight constraints on the seismic moment and the fault dip, where the dip is determined to be 8° to 10° with corresponding moments of 1.24 × 1022 to 1.00 × 1022 N m, respectively. The geodetic constraints on slip distribution help to eliminate the trade-off between rupture velocity and slip kinematics. Source models obtained from the inversion of various combinations of the teleseismic body waves and geodetic data are evaluated by comparing predicted and observed long-period seismic waveforms (100–500 sec). Our results indicate a relatively slow average rupture velocity of 1.5 to 2.5 km/sec and long average rise time of up to 20 sec. The earthquake nucleated between two separate slip patches, one beneath Nias and the other beneath Simeulue Island. The gap between the two patches and the hypocentral location appears to be coincident with a local geological disruption of the forearc. Coseismic slip clearly tapers to zero before it reaches the trench probably because the rupture propagation was inhibited when it reached the accretionary prism. Using the models from joint inversions, we estimate the peak ground velocity on Nias Island to be about 30 cm/sec, an order of magnitude slower than for thrust events in continental areas. This study emphasizes the importance of utilizing multiple datasets in imaging seismic ruptures.

Glacial earthquakes.

We have detected dozens of previously unknown, moderate earthquakes beneath large glaciers. The seismic radiation from these earthquakes is depleted at high frequencies, explaining their nondetection by traditional methods. Inverse modeling of the long-period seismic waveforms from the best-recorded earthquake, in southern Alaska, shows that the seismic source is well represented by stick-slip, downhill sliding of a glacial ice mass. The duration of sliding in the Alaska earthquake is 30 to 60 seconds, about 15 to 30 times longer than for a regular tectonic earthquake of similar magnitude.

Simultaneous waveform inversion for three-dimensional Earth structure and earthquake source parameters considering a wide range of modal coupling

Summary We perform iterative inversion of long-period seismic waveform data to determine simultaneously the 3-D structure of the upper and uppermost lower mantle and earthquake source parameters. We perform direct inversion of the observed data using the methods of Geller & Hara (1993) and Hara (1997), thus intermediate parameters such as ‘splitting functions’ are not used. We analyse data in the period bands 178-185, 282-299 and 347-376 s using the Direct Solution Method (DSM) to compute synthetic seismograms and their partial derivatives. A wide range of modal coupling is considered in these computations. We conduct numerical experiments that show that neglecting toroidal-spheroidal coupling can significantly bias estimates of 3-D Earth structure for the period band 282-299 s; note that previous studies for this band generally did not consider such coupling. A numerical experiment for the period band 347-376 s suggests that our inversion approach can extract more information on odd angular order structure for the period band 347-376 s than previous techniques.

Three-dimensional velocity structure and resolution of the core-mantle boundary region from whole-mantle inversions of body waves

The nature of the core-mantle boundary (CMB) has important implications for deep-Earth processes, particularly those which have their origin in the lower mantle or outer core. The possibility of the existence of lateral P-wave velocity variations in the outermost core is explored here by obtaining models of three-dimensional, CMB region velocity structure using an inversion of waveform data. The numerical inversions involve determination of the three-dimensional structure of the Earth's outermost core and lower mantle, using over 5000 long-period seismic waveforms collected from global digital seismic networks. Spheroidal modes with periods between 33 and 100 s were selected to model the body-wave portion of seismograms recorded from global earthquakes. First-order perturbations in S-wave velocities in 11 mantle layers 200–350 km thick and P-wave velocities in one outermost core layer comprise the least-squares solutions to the inverse problem. Although the statistical F test supports the additional layer in the inversions, this layer is not required by the F ratio to occur in the outermost core. Pattern-retrieval resolution tests, in which synthetic data constructed from an initial synthetic Earth model are inverted for model parameters, are used to compare parameters for one region to those of another in which some coefficient power smearing is suspected. Depth resolution tests indicate that the best radial resolution at the CMB is 500 km. Thus, this method does not produce the resolution required for outermost core lateral velocity variation analysis. An examination of the limitations of this approach demonstrates the need for waveform inversion techniques which more accurately predict raypaths in the deep mantle. Improvements in radial resolution are needed before whole-mantle velocity models can begin to distinguish whether significant heterogeneity related to core-mantle interactions occurs in the CMB region.

Comparison of global waveform inversions with and without considering cross-branch modal coupling

SUMMARY In this study we present a new approach to inverting long-period seismic waveforms. The effect of lateral heterogeneity is partitioned into two. The first part represents the effect of horizontally averaged structure along the great circle between the source and receiver, and is allowed to remain in non-linear form in the formulation. The second part incorporates any further correction due to cross-branch modal coupling, which has been neglected in the more conventional path average approximation (PAVA). This term is linearized and then treated asymptotically so that the seismogram depends only upon the structure within the great-circle section determined by the source and receiver (Li & Tanimoto 1993a). We refer to this new method as the non-linear asymptotic coupling theory (NACT). The sensitivity kernels predicted by the PAVA and NACT are compared. While the sensitivity kernels are similar for surface waves and shallow-turning body waves, they are very different for body waves that sample the deep mantle. By examining the inversion algorithms for the PAVA and NACT, we demonstrate that the computation time required by the NACT tends to be of the same order of magnitude as that required by the PAVA, as the number of model parameters increases. Based upon a realistically large data set (5041 body-wave seismograms and 1531 mantle-wave seismograms), formal resolution analyses are performed using both PAVA and NACT. We find that the NACT is significantly more powerful in resolving 3-D structure in the deep mantle. We compare the models obtained for the same observed data set using the two approaches. As expected, they differ more in the lower mantle than in the upper mantle. The difference in their amplitude spectra increases with spherical harmonic degree. The model developed using the NACT predicts the observed surface geoid better than that developed using the PAVA, based upon geodynamic flow modelling.

Moment tensor inversion for near earthquakes using long-period digital seismograms.

The possibility of rapid determination of source parameters (seismic moment and fault parameters) for near earthquakes is examined by means of the moment tensor inversion method. Numerical experiments of linear moment tensor inversion are made for a hypothetical source-station geometry appropriate for the Japanese region and suggest that it may be possible to determine the source parameters for the near earthquakes within about 10 min after the earthquake occurrence, provided that high-quality long-period seismic waveform data can be used. Source parameters of eight earthquakes (MW > 6.0) which occurred near the Japanese region during the period from November 1986 to April 1987 are determined by applying the inversion method to three-component long-period digital data recorded at a single station in Japan. The obtained source parameters are in good agreement with the solutions derived from global networks. The centroid-moment tensor (CMT) inversion method, solution of which is a result of simultaneous inversion of waveform data for the hypocentral parameters (epicentral coordinates, depth, origin time, and moment tensor) of the best point source, is applied to both synthesized and observed data. The inversion of synthesized data suggests that the CMT inversion to determine the hypocentral and source parameters simultaneously for near earthquakes may be possible in the Japanese region. We applied this method to the eight events, but we could obtain the CMT solution for only one event. The short propagation time and inadequate station coverage in the case of the single-station observation of the near earthquakes might have caused the non-uniqueness and instability in the determination of the CMT solutions. The numerical experiments suggest that this problem may be significantly remedied by the improvement of the station coverage.

Surface waves and free oscillations in a regionalized earth model

The linearized equation is derived which relates observed long-period seismic waveforms to the aspherical perturbations of a spherically symmetric earth model. This is accomplished by formulating the theory of spectral splitting in the time domain. It is shown to be possible greatly to simplify the resulting equations in a way which makes it apparent that for each modal multiplet the ‘scattered’ field depends only upon three local functionals of earth structure. The effect of regional structural variations may then be quantified in a manner analogous to that assumed in the ‘pure path technique’, but without making the usual asymptotic approximations. These results are used to investigate the validity of the asymptotic result for the locations of the centroids of spectral peaks in individual recordings, for a regionalized model of the Earth. A technique is suggested for retrieving information about geographical structural variations from low-frequency waveform data.