Search

Earthquakes are observed to occur in subduction zones to depths of approximately 680 km, even though unassisted brittle failure is inhibited at depths greater than about 50 km, owing to the high pressures and temperatures. It is thought that such earthquakes (particularly those at intermediate depths of 50-300 km) may instead be triggered by embrittlement accompanying dehydration of hydrous minerals, principally serpentine. A problem with failure by serpentine dehydration is that the volume change accompanying dehydration becomes negative at pressures of 2-4 GPa (60-120 km depth), above which brittle fracture mechanics predicts that the instability should be quenched. Here we show that dehydration of antigorite serpentinite under stress results in faults delineated by ultrafine-grained solid reaction products formed during dehydration. This phenomenon was observed under all conditions tested (pressures of 1-6 GPa; temperatures of 650-820 degrees C), independent of the sign of the volume change of reaction. Although this result contradicts expectations from fracture mechanics, it can be explained by separation of fluid from solid residue before and during faulting, a hypothesis supported by our observations. These observations confirm that dehydration embrittlement is a viable mechanism for nucleating earthquakes independent of depth, as long as there are hydrous minerals breaking down under a differential stress.

The 2003 Tokachi-oki earthquake of M 8.0, which occurred off the south-east coast of Hokkaido, Japan, was preceded by noticeable deep seismicity in the subducting slab, including a deep-focus earthquake of M 7.1. The 1952 Tokachi-oki earthquake of M 8.2 and the 1993 Kushiro-oki earthquake of M 7.5, which occurred also off the south-east of Hokkaido along the Kurile Trench, were also preceded by high deep seismicity including large deep-focus earthquakes of the M 7 or over (focal depths 200–600 km). The relations between the deep seismicity and the ‘shallow’ large earthquake (focal depth ≤ 100 km) in these three cases are quite similar. This result strongly supports the author’s view (Mogi, 1973, 1988) that large ‘shallow’ earthquakes along the subduction zones are sometimes preceded by the occurrence of large deep-focus earthquakes.

On September 26, 2003, a large earthquake with a magnitude of 8.0 occurred along the Kuril trench off Tokachi, Hokkaido, Japan. We investigated the source process by using teleseismic P- and SH-wave data. The main source parameters are as follows: the seismic moment 1.0 × 1021 Nm (Mw= 8.0); (strike, dip, rake) = (230°, 20°, 109°); the depth of initial break point 25 km; source duration 40 sec; and the maximum slip 5.8 m. We estimated the fault area to be 90 × 70 km2, the average slip 2.6 m, and the stress drop 5.0 MPa. This earthquake was an interplate earthquake associated with the subduction of the Pacific plate. The rupture propagated northward from a shallow to a deep region. In this area, a great earthquake (magnitude 8.2) occurred in 1952. We also made limited inversion of nearfield records of the 1952 event and found that the 2003 asperity was also ruptured in 1952. Our result suggests that the 2003 Tokachi-oki earthquake was a recurrent event of the 1952 Tokachi-oki earthquake.

It is well established that an earthquake in the Earth's crust can trigger subsequent earthquakes, but such triggering has not been documented for deeper earthquakes. Models for shallow fault interactions suggest that static (permanent) stress changes can trigger nearby earthquakes, within a few fault lengths from the causative earthquake1,2,3, whereas dynamic (transient) stresses carried by seismic waves may trigger earthquakes both nearby and at remote distances4,5,6,7,8. Here we present a detailed analysis of the 19 August 2002 Tonga deep earthquake sequences and show evidence for both static and dynamic triggering. Seven minutes after a magnitude 7.6 earthquake occurred at a depth of 598 km, a magnitude 7.7 earthquake (664 km depth) occurred 300 km away, in a previously aseismic region. We found that nearby aftershocks of the first mainshock are preferentially located in regions where static stresses are predicted to have been enhanced by the mainshock. But the second mainshock and other triggered events are located at larger distances where static stress increases should be negligible, thus suggesting dynamic triggering. The origin times of the triggered events do not correspond to arrival times of the main seismic waves from the mainshocks and the dynamically triggered earthquakes frequently occur in aseismic regions below or adjacent to the seismic zone. We propose that these events are triggered by transient effects in regions near criticality, but where earthquakes have difficulty nucleating without external influences.

Quantitative estimates of the crack density (ϵ), saturation rate (ξ) and porosity (ψ) parameters from seismic velocities (Vp, Vs) and Poisson’s ratio (σ) in the 2001 Bhuj earthquake area in western India indicate that the 2001 Bhuj earthquake hypocenter is associated with high-ϵ, high-ξ and high-ψ in the depth range of about 23–28 km, extending 15–30 km laterally. These anomalies may be due to a fluid-filled, fractured rock matrix, which might have contributed to trigger the 2001 Bhuj earthquake in the intraplate, stable continental region of the Indian Peninsula. This feature is similar to that of the 1995 Kobe earthquake [D. Zhao et al., Science 274 (1996) 1891–1894]. High-ψ areas are generally consistent with high-ϵ areas, but high-ξ areas have a wider distribution, indicating that micro-cracks exist in more localized areas within the crust, and that permeation of fluids in the hypocenter zone might have occurred extensively through the intergranular and fractured pores due to hidden intersecting fault geometry. Here we suggest the possibility that earthquake occurrence is closely related to in situ material heterogeneities, rather than stress conditions alone.

Researchers use computer simulations to find the date and earthquake source of an ancient tsunami that deposited sediment in a Hawaii sinkhole.

Several destructive earthquakes have occurred in Tien-Shan region at the beginning of 20th century. However, the detailed seismological characteristics, especially source parameters of those earthquakes are still poorly investigated. The Chon-Kemin earthquake is the strongest instrumentally recorded earthquake in the Tien-Shan region. This earthquake has produced an approximately 200 km long system of surface ruptures along Kemin-Chilik fault zone and killed about similar to 400 people. Several studies presented the different information on the earthquake epicentre location and magnitude, and two different focal mechanisms were also published. The reason for the limited knowledge of the source parameters for the Chon-Kemin earthquake is the complexity of old analogue records processing, digitization and analysis. In this study the data from 23 seismic stations worldwide were collected and digitized. The earthquake epicentre was relocated to 42.996NA degrees and 77.367EA degrees, the hypocentre depth is estimated between 10 and 20 km. The magnitude was recalculated to m(B) 8.05, M-s 7.94 and M-w 8.02. The focal mechanism, determined from amplitude ratios comparison of the observed and synthetic seismograms, was: str = 264A degrees, dip = 52A degrees, rake = 98A degrees. The apparent source time duration was between similar to 45 and similar to 70 s, the maximum slip occurred 25 s after the beginning of the rupture. Two subevents were clearly detected from the waveforms with the scalar moment ratio between them of about 1/3, the third subevent was also detected with less certainty. Taking into account surface rupture information, the fault geometry model with three patches was proposed. Based on scaling relations we conclude that the total rupture length was between similar to 260 and 300 km and a maximum rupture width could reach similar to 70 km.

The precise hypocenter was determined location using double difference method around subduction zone in Moluccas area eastern part of Indonesia. The initial hypocenter location from MCGA data catalogue of 1,945 earthquake events. Basically the principle of double-difference algorithm assumes if the distance between two earthquake hypocenter distribution is very small compared to the distance between the station to the earthquake source, the ray path can be considered close to both earthquakes. The results show the initial earthquakes with a certain depth (fix depth 10 km) relocated and can be interpreted more reliable in term of seismicity and geological setting. The relocation of the intra slab earthquakes beneath Banda Arc are also clearly observed down to depth of about 400 km. The precise relocated hypocenter will give invaluable seismicity information for other seismological and tectonic studies especially for seismic hazard analysis in this region.

We measured acoustic emission energy during antigorite dehydration in a multianvil press from 1.5 to 8.5 gigapascals and 300° to 900°C. There was a strong acoustic emission signal on dehydration, and analysis of recovered samples revealed brittle deformation features associated with high pore-fluid pressures. These results demonstrate that intermediate depth (50 to 200 kilometers) seismicity can be generated by dehydration reactions in the subducting slab.

A great earthquake is expected to occur along the Suruga Trough where the Philippine Sea plate subducts beneath the Eurasian plate. We estimate crustal deformation prior to this hypothetical earthquake, called the Tokai earthquake, based on a simulation model for plate subduction, a three-dimensional extension of the models by Stuart (1988) and Kato and Hirasawa (1996). In this model, the time evolution of a dislocation field in an elastic half-space is governed by a quasi-static interaction induced by the dislocation field itself as well as by a rate- and state-dependent friction law acting on the plate interface. Observed geophysical constraints taken into account are the shapes of the curved subducting plate interface, the recurrence interval of past great earthquakes (≃ 90–150 years), the average seismic coupling coefficient (≃0.5), the annual rate of level change (≃5 mm/year) between the two bench marks at Hamaoka and Kakegawa, and the backslip distribution on the plate interface estimated from the result of GPS observation. The simulation shows the following. The occurrence of great earthquakes is approximately periodic. During interseismic times, a strongly coupled region is formed in a seismic zone, and it shrinks as the next great earthquake approaches. The strongly coupled region is confined to an area where the plate inclination is gentle. The rupture of the earthquake starts at a point close to the maximum shear stress acting on the plate boundary. Subsidence of Hamaoka relative to Kakegawa changes to uplift several years before the earthquake. Volumetric strain undergoes a change from compression to extension, or vice versa, at certain observation sites several years before the earthquake. Changes in volumetric strain accelerate as the occurrence of the earthquake impends. The maximum strain change for 1 day observed on the ground surface just before the earthquake is 10 −8 to 10 −7 , a value about one-tenth of the value predicted by a related two-dimensional model. The pattern of shear stress accumulation and the location of the preslip are affected by the shape of the plate interface.

A localized shear stress concentration, provoked by the tectonic escape process, in the source region of the Chi-Chi (Taiwan) Earthquake has been used to explain the mechanism of the ML=7.3 devastating earthquake. This study, by employing the geophysical magnetotellurics (MT) soundings technique, mapped the deep electrical structures beneath the seismogenic zone of the Chi-Chi Earthquake and shows prominent effects. About fifty MT soundings were conducted and analyzed to recover 3D resistivity distributions. Results show a distinct low resistivity anomaly, less than 10-ohm-m, with depth ranging to greater than 10 km, which extends laterally no less than 10 km beneath the hypocenter of the earthquake. Correlations of the resistivity model with the seismogenic zone of the Chi-Chi Earthquake, raises the important possibility that the fluids released though crust dehydration may have been one of the most important causes of the Chi-Chi Earthquake. The MT model also suggests that the potential sources of these fluids may be: (1) fluids trapped in pore spaces at shallow depths, less than 8 km, and (2) the accumulation of fluids in the crust, greater than 12 km, contributed by the subducted ocean crust on the top of the slab.

A change of Coulomb Failure Stress (CFS) is commonly used to explain a mechanism of earthquake triggering due to both earthquake faulting and volcanic sources. First, we introduced crustal deformation and source model for five episodes of volcanic activities off Ito, central Japan and around the Mt. Iwate volcano, northern Japan based on dense geodetic measurements. And then, we calculated spatial distribution of a CFS change due to the estimated source models and compared the observed seismic activities.In the case of the 1998 earthquake swarm off Ito, the region where opening of dike and the largest earthquake (M5.7) increase CFS was concordant with the hypocenter distribution observed after the largest earthquake. However, the largest earthquake occurred in the area where CFS decreased.In the case of the Iwate volcano, a volcanic inflation source increased CFS by 0.33 MPa at the hypocenter of the M6.1 earthquake that occurred on September 3, 1998. The calculated CFS change is 7 % of the coseismic stress drop. The M6.1 earthquake occurred at the northern edge of the Nishine Fault found by geographical and geological studies. It suggests that the inflation promoted the rupture of a known Quaternary fault. Stress shadow caused by the M6.1 earthquake explains the observed quiescence of the seismic swarm activity west of E140.95° after the occurrence of the earthquake.

Two great earthquakes have occurred recently along the Sunda Trench, the 2004 M9.2 Sumatra-Andaman earthquake and the 2005 M8.7 Nias earthquake. These earthquakes ruptured over 1600 km of adjacent crust within 3 mo of each other. We quantitatively present poroelastic deformation analyses suggesting that postseismic fluid flow and recovery induced by the Sumatra-Andaman earthquake advanced the timing of the Nias earthquake. Simple back-slip simulations indicate that the megapascal (MPa)–scale pore-pressure recovery is equivalent to 7 yr of interseismic Coulomb stress accumulation near the Nias earthquake hypocenter, implying that pore-pressure recovery of the Sumatra-Andaman earthquake advanced the timing of the Nias earthquake by ∼7 yr. That is, in the absence of postseismic pore-pressure recovery, we predict that the Nias earthquake would have occurred in 2011 instead of 2005.

The 2008 Iwate-Miyagi Nairiku earthquake (M 7.2) was a shallow inland earthquake that occurred in the volcanic front of the northeastern Japan arc. To understand why the earthquake occurred beneath an active volcanic area, in which ductile crust generally impedes fault rupture, we conducted magnetotelluric surveys at 14 stations around the epicentral area 2 months after the earthquake. Based on 56 sets of magnetotelluric impedances measured by the present and previous surveys, we estimated the three-dimensional (3-D) electrical resistivity distribution. The inverted 3-D resistivity model showed a shallow conductive zone beneath the Kitakami Lowland and a few conductive patches beneath active volcanic areas. The shallow conductive zone is interpreted as Tertiary sedimentary rocks. The deeper conductive patches probably relate to volcanic activities and possibly indicate high-temperature anomalies. Aftershocks were distributed mainly in the resistive zone, interpreted as a brittle zone, and not in these conductive areas, interpreted as ductile zones. The size of the brittle zone seems large enough for a fault rupture area capable of generating an M 7-class earthquake, despite the areas distributed among the ductile zones. This interpretation implies that 3-D elastic heterogeneity, due to regional geology and volcanic activities, controls the size of the fault rupture zone. Additionally, the elastic heterogeneities could result in local stress concentration around the earthquake area and cause faulting.

The August 23, 1995 (Mw=7.0, depth 586 km) Mariana event produced one of the strongest deep earthquake aftershock sequences known. We combined teleseismic arrival times with arrival times picked from a local network operating directly above the event in the Mariana Islands, and relocated the main shock and aftershocks. We derived the source characteristics from teleseismic body waves by inverting waveform data of the Global Seismograph Network. The main shock ruptured downward across the entire width of the active slab along a plane dipping southwestward. Most of the 17 well-located aftershocks occurred along or very near the rupture plane. The different focal mechanisms of the main shock and aftershocks indicate that the stress field varies across the width of the Mariana slab at the area of the August 1995 sequence.

We have identified three groups of deep earthquakes showing nearly identical waveforms in the Tonga slab. Relocation with a cross-correlation method shows that each cluster is composed of 10 to 30 earthquakes along a plane 10 to 30 kilometers in length. Some of the earthquakes are colocated, demonstrating repeated rupture of the same fault, and one pair of events shows identical rupture complexity, suggesting that the temporal and spatial rupture pattern was repeated. Recurrence intervals show an inverse time distribution, indicating a strong temporal control over fault reactivation. Runaway thermal shear instabilities may explain temporally clustered earthquakes with similar waveforms located along slip zones weakened by shear heating. Earthquake doublets that occur within a few hours are consistent with events recurring before the thermal energy of the initial rupture can diffuse away.

This article reviews the development of the study of relationship between the earth's rotation and the spatio-temporal distribution of global seismicity by stationary model of seismicity rates and annual seismic energy released. Observed variations of seismic energy release in shallow, intermediate and deep focus earthquakes and their frequency distribution confirms this correlation. The peaks of these parameters are controlled by the earth rotation. There exists a phase relationship among earth's rotation rate minima and maxima with the maxima (peaks) of the above parameters as well as Thrust (T) and the strike slip (S) dominating periods of global seismicity. In order to compare our results with observations, the space-time dependence of the frequency of earthquakes and annual energy release has been established. The results are in very good agreement with previous studies and further enhance our knowledge for global seismicity distribution.

Stress changes in the crust due to an earthquake can hasten the failure of neighbouring faults and induce earthquake sequences in some cases1,2,3,4,5. The 1999 Hector Mine earthquake in southern California (magnitude 7.1) occurred only 20 km from, and 7 years after, the 1992 Landers earthquake (magnitude 7.3). This suggests that the Hector Mine earthquake was triggered in some fashion by the earlier event. But uncertainties in the slip distribution and rock friction properties associated with the Landers earthquake have led to widely varying estimates of both the magnitude and sign of the resulting stress change that would be induced at the location of the Hector Mine hypocentre—with estimates varying from -1.4 bar (ref. 6) to +0.5 bar (ref. 7). More importantly, coseismic stress changes alone cannot satisfactorily explain the delay of 7 years between the two events. Here we present the results of a three-dimensional viscoelastic model that simulates stress transfer from the ductile lower crust and upper mantle to the brittle upper crust in the 7 years following the Landers earthquake. Using viscoelastic parameters that can reproduce the observed horizontal surface deformation following the Landers earthquake, our calculations suggest that lower-crustal or upper-mantle flow can lead to postseismic stress increases of up to 1–2 bar at the location of the Hector Mine hypocentre during this time period, contributing to the eventual occurrence of the 1999 Hector Mine earthquake. These results attest to the importance of considering viscoelastic processes in the assessment of seismic hazard8,9,10,11.

The study addresses the depth distribution of stresses in the Kamchatka Wadati-Benioff zone, based on detailed geometry of the deep seismically active structures, homogeneous data of earthquake focal mechanisms and the inverse technique by Gephart and Forsyth (Gephart, J., Forsyth, D., 1984. An improved method for determining the regional stress tensor using earthquake focal mechanism data: application to the San Fernando earthquake sequence. J. Geophys. Res. 89, 9305–9320). The used data set includes 205 CMT Harvard solutions for earthquakes which belong to the Wadati–Benioff zone and 17 for a deep-seated subduction segment. The inverse technique used allows for determining the best fit principal stress directions σ1, σ2, σ3 and the ratio R=σ2−σ1/σ3−σ1 in several depth intervals in the Wadati–Benioff zone and for the deep-seated slab segment considered as a single body. The depth ranges 0–40 and 41–60 km are characterized by maximum compressive stress σ1 which is essentially horizontal and trends SE, indicating the NW–SE convergence between the Pacific and Eurasian lithospheric plates. The minimum compression σ3 in the above depth ranges is down-dipping due to the slab pull. The stress tensors obtained for the depth layer 61–80 km and K2, the lower part of the subduction zone, indicate a heterogeneous stress field. Further analysis of the above sub-volumes showed that the stress heterogeneity is attributed to different stress regimes in the frontal and in-slab parts of the Wadati–Benioff zone: the frontal part is under slab-parallel compression and extension which dips at approximately 40° to SE; the in-slab part is subjected to down-dipping to slab-parallel extension and shallow dipping to SE compression. This stress distribution suggests unbending of the subducting plate at depths greater than 61 km. The stresses in the depth range 81–90 km are of similar orientations to those in the in-slab parts of 61–80 and 91–230 km but of different dips. The deep-seated slab segment is characterized by compression which is parallel to the dip direction of this volume and by nearly horizontal extension of SE trend. These results show that the main geodynamic forces which drive the presently active processes at different depths in the Kamchatka and Northern Kurille region are the NW-SE convergence between the Eurasian and Pacific plates (ridge push) causing the observed almost horizontal compression of SE trend at 0–60 km depth, the slab pull, unbending forces at depths 61–230 km causing the observed slab-parallel compression in the frontal part of the Wadati–Benioff zone and down-dip to slab-parallel extension in the in-slab part.

The kinematics and dynamics of plate tectonics are frontal subjects in geosciences and the strong earthquake occurred along the plate boundary result directly from plate movement. By analyzing Ibaraki earthquake sequence, it has been found that the focal fault plane shows a special image of grading expansion along the direction of strike and adjustment along the dip direction respectively. With the consideration of strike, dip and slip directions of focal mechanism, we have confirmed that Ibaraki earthquake belongs to a thrust fault earthquake occurred under the Japan Trench. The cause of the earthquake sequence is discussed in the paper. The study on the temporal-spatial distribution of the earthquake sequence with a time-scale between the year-scale spatial geodetic data and the second-scale moment tensor of the strong earthquake has indicated the dynamic process of Pacific Plate subduction under the Eurasia Plate. According to the average slip distance of earthquake and the velocity of plate movement, it is predicted that a strong earthquake might occur in recent years.

The differential attenuation beneath the Gauribidanur seismic array has been estimated using the spectral ratio method. Deep focus earthquakes originating in the Hindukus region have been considered for this study. In fact, the estimated Q is the combined effect of the intrinsic attenuation and the attenuation due to scattering. The estimated standard deviation and mean of the differential attenuation have been found as 0.18 s and -0.06 s for the array, respectively. The Q GBA is estimated to be 25. It is significant enough and may be due to the scattering by the inhomogeneities beneath the array. __

The test and application of the time and magnitude predictable model in the subduction zones around the circum-Pacific belt is performed. Each subduction zone was separated in several seismogenic regions on the basis of tectonic and geometrical features of the subducting lithosphere and for three different depth ranges. The data derived from these seismogenic regions were used in order to check the suitability of the regional time and magnitude predictable model, as this model was defined for shallow events. The results obtained in the present study confirm that this model appropriately describes the behavior of deep seismicity in these regions, since a positive ( c =0.31±0.003) and a negative ( C =−0.13±0.047) dependence were found for the logarithm of the interevent times and the magnitude of following mainshock, respectively, on the magnitude of the preceding mainshock. Moreover, the slip-predictable model was rejected, since a negative dependence was found ( E =−0.19) between the interevent times and the magnitudes of the following mainshock. Probabilities of occurrence of the future strong (M≥7.0) events in each seismogenic region have been determined for the next decade conditioned on the previous occurrence of such events. The magnitudes of the expected events have also been estimated. Statistical tests for the estimated probabilities exhibit a good correlation with the occurrence rate of such events. Furthermore, the statistical significance of the model has been studied through additional tests. The results of this study demonstrate the validity of the regional time and magnitude predictable model in the areas examined.

It has been proposed that intermediate- and deep-focus earthquakes occur on preexisting planes of weakness that were created at shallow depth. The purpose of this study is to test this hypothesis. We examined fault plane solutions for 360 events that occurred in the Tonga subduction zone from January 1976 to February 1999. For events in the slab, each fault plane solution is rotated by the slab dip angle so that the slab is in its original horizontal position. We then compare fault plane solutions for events in the slab with those for events at the outer rise. We find that an asymmetric fault system, corresponding to that found for outer rise events, persists down to about 450 km depth. This suggests that earthquakes down to this depth are caused by the reactivation of preexisting faults created in the oceanic plate before subduction. A similar pattern, though less well constrained, is found in the Kurile subduction zone down to 200 km depth. For earthquakes deeper than 450 km the fault plane solutions are much more scattered. In some places the fault plane pattern suggests the creation of new fault planes along orientations of maximum shear with zero internal friction. The pattern of fault plane solutions in particular localized volumes cannot be easily explained by the creation of new faults in the direction of maximum shear under a coherent regional tectonic stress field. Thus, either a highly heterogeneous stress field exists or preexisting planes of weakness also account for some deep earthquakes.

We investigate the nature of laterally varying P wave speeds (Vp) and their relationship to a group of unusual deep earthquakes beneath the backarc of Tonga. This subhorizontal swath of “outboard” earthquakes occurs within the transition zone of the mantle but is several hundred kilometers away from the inclined Wadati-Benioff zone. Surprisingly, the seismogenic region is not associated with fast Vp. It follows that the effect of cold temperature, as implied by the presence of large earthquakes, must have been counteracted by other factors such as compositional or mineralogical effects. Our results are based on a sequence of seismic profiles in a fan-shot geometry. For a given azimuthal sector, we seek the simplest model of radially varying Vp that explains observed broadband waveforms and travel times. The lateral extent of anomalous Vp is then constrained by comparing the results from a wide range of azimuths. This approach fully utilizes information from triplicate P arrivals that are sensitive to seismic wave speeds at the depths of interest. Our results identify prominent anomalies of fast Vp (∼3% over average Earth models) both in the transition zone and in the lower mantle beneath the Tonga backarc. However, the region of fast Vp in the transition zone is offset by ∼300 km to the south of that in the lower mantle, and the region of offset coincides with the outboard earthquakes. A straightforward interpretation is that a large flux of subducted Pacific lithosphere lingers in the transition zone before descending into the lower mantle. Within this broad remnant of subducted lithosphere, the outboard earthquakes occur in a region where impounding of subducted material is most significant, a condition favorable for the accumulation of met astable olivine or volatiles. Such materials reduce Vp and are candidates for triggering earthquakes, thus accounting for the absence of fast Vp in the source zone of deep earthquakes.

Teleseismic recordings (Δ > 87°) of a deep earthquake beneath the Banda Sea at stations in Tanzania show a difference in the arrival time of the radial (SSV) and transverse component (SSH) S wave ranging from 1–3 s. Shear velocity anisotropy in the lowermost mantle beneath the Indian Ocean is the likely cause of this signal because recordings at the same stations of closer-in events (Δ < 80°) in the same source region do not present a comparable differential travel time. For the Banda Sea event, the SSH signals are broader than SSV signals, suggesting that a discontinuity (or strong vertical gradient) in primarily VSH marks the sudden onset of transverse isotropy in D” (with a magnitude of 1.4–2.7%) about 350 km above the core-mantle boundary. SKKS coda, S-to-p converted phases at the Moho, and upper mantle heterogeneity beneath the stations obscure the onset of SSV and complicate wave shapes. It is therefore difficult to evaluate whether general anisotropy needs to be invoked into a model of shear velocity anisotropy.

We use a finite difference algorithm to compute P-SV synthetic seismograms for deep earthquakes recorded at regional seismic stations located above a subducting slab. We calculate synthetics for 2-D slab models in which the 410 km-discontinuity is uplifted, owing to the exothermic nature of the α → β transition (the equilibrium model), and models in which the 410 km-discontinuity is depressed, owing to inhibited kinetics of the α → β olivine transition (the metastable model). In the latter case the existence of the low-velocity wedge of metastable olivine has a profound influence on the waveforms of deep-focus earthquakes. The wedge acts as a geometrical waveguide, with the energy-focusing effect apparent even after the wavefront has traveled the several hundred kilometers from the wedge boundary to the Earth's surface. The guided energy is most pronounced for receivers located near the surface projection of the deep slab but is observable over a range of 300 km and so may provide a valuable diagnostic for discriminating between equilibrium and metastable models of subduction. A recent search for such guided energy, using regionally recorded P and S waves from deep events in Tonga, yielded negative results.

We consider different formation models of deep focus earthquakes at the boundary between the Pamir and Hindu Kush mountains including the Hindu Kush megasource. Reconstruction of the Late Cretaceous‐ Paleogene basin between the Northern and Central Pamirs allows us to connect anomalously high seismicity with the deformation of the subduction plate (slab) of the oceanic lithosphere in this basin during its submersion into the mantle. In this sense the Vrancea zone in the Carpathians is an analog of the Hindu Kush megasource. The concentration of mantle earthquakes of the intermediate type with hypocenter depths up to 300 km and magnitudes up to 8 and greater [1] within a small megasource region at the boundary of tectonic zones in the Hindu Kush, Central, and Southwestern Pamirs has long attracted the interest of researchers (Fig. 1). Here, 90% of strong earthquakes in the Pamir‐Hindu Kush zone and 95% of released energy are concentrated on a territory of 200 × 120 km [2]. In the western region, mantle earthquakes are sharply interrupted at 69 ° 30 � E. The zone of mantle earthquakes approximately 350 km long with hypocenter depths up to 200 km is extended to the northeast of the megasource already on the territory of the Pamirs. Thus, the belt of mantle seismicity with a total length up to 500 km is clearly pronounced. The Hindu Kush source is confined to its southwestern part. L.I. Lobkovskii suggested a model based on the concept of two-level plate tectonics to explain the nature of the Hindu Kush earthquakes [3‐5]. Calculations demonstrated that dissipative heating of the upper parts of the mantle up to 700‐750 ° C occurs during the underthrust of the continental lithosphere in the collision zones. At such temperatures, the mechanism of brittle deformation (earthquake) becomes impossible. At the same time, the mantle part of the lithosphere separated by the asteno-layer from the crust can submerge into the upper mantle. This process is generally aseismic. Deep-focus mantle earthquakes appear only in the anomalous regions either as a result of insufficient heating, or due to the penetration of water or other fluids. In the latter case, the rocks become brittle and capable of seismogenic deformations. According to the other version [2], mantle seismicity in the Hindu Kush is related to the submersion of rocks of the ancient oceanic crust (Paleozoic or Early Mesozoic ophiolites) to the depths of 40‐70 km under the weight of the tectonic nappes. In this case, the basalt-eclogite phase transition occurs and their submersion to the upper mantle of lower density. This is an earthquake-related process. We eventually admit that the local source of mantle seismicity can be related to submersion, deformation, and separation of a slab of oceanic lithosphere in the basin that had closed recently. For example, seismicity in the Vrancea zone (East Carpathians) is related to the residual phenomena in the subduction zone of a deep basin that completely closed in the Middle‐Late Miocene (see Fig. 4.11 in [5] and references in this publication). Such an explanation was not applied to the Hindu Kush earthquakes because the existence of a young basin and related subduction were not assumed in this region.

We use the slip model of the Bolivian earthquake obtained by Ihmle [1998] to investigate the spatial distribution of stress drop and calculate the frictional heat generated locally on the fault. We find that the stress drop associated with the earthquake averages between 5 and 10 M Pa over the fault. The total frictional energy is estimated at between 3 × 1017 and 2 × 1018J. If the fault width is as thin as the one produced by transformational faulting in laboratory experiments, where it is about two orders of magnitude smaller than the local shear displacement, melting likely occurred on a significant portion of the fault. In this event, as proposed by Kanamori et al. [1998], once rupture is triggered, shear melting can quickly promote extensive slip.

Following a quasi-dynamic scheme proposed by Minear and Toksoz (1970), thermal structures of subduction zones for different models by finite element method (FEM) were calculated. Density distribution and P-wave anomaly of subduction zones were calculated at the same time. Comparing with seismological evidences and results of laboratories, it is proposed that earthquakes occurred below 400 km depth are probably controlled by anti-crack mechanism.

—We analyze the seismic moment-frequency relation in various depth ranges and for different seismic regions, using Flinn-Engdahl's regionalization of global seismicity. Three earthquake lists of centroid-moment tensor data have been used the Harvard catalog, the USGS catalog, and the Huang et al. (1997) catalog of deep earthquakes. The results confirm the universality of the β-values and the maximum moment for shallow earthquakes in continental regions, as well as at and near continental boundaries. Moreover, we show that although fluctuations in earthquake size distribution increase with depth, the β-values for earthquakes in the depth range of 0–500 km exhibit no statistically significant regional variations. The regional variations are significant only for deep events near the 660 km boundary. For declustered shallow earthquake catalogs and deeper events, we show that the worldwide β-values have the same value of 0.60 ± 0.02. This finding suggests that the β-value is a universal constant. We investigate the statistical correlations between the numbers of seismic events in different depth ranges and the correlation of the tectonic deformation rate and seismic activity (the number of earthquakes above a certain threshold level per year). The high level of these correlations suggests that seismic activity indicates tectonic deformation rate in subduction zones. Combined with the universality of the β-value, this finding implies little if any variation in maximum earthquake seismic moment among various subduction zones. If we assume that earthquakes of maximum size are similar in different depth ranges and the seismic efficiency coefficient, χ, is close to 100% for shallow seismicity, then we can estimate χ for deeper earthquakes for intermediate earthquakes χ≈ 5%, and χ≈ 1% for deep events. These results may lead to new theoretical understanding of the earthquake process and better estimates of seismic hazard.

Seismic moment of aftershocks during the August 23, 1995 deep earthquake sequence in the Mariana amounts up to 17% of that of the main shock (Mw 7.0). Such a high aftershock activity is extremely unusual. Based on the inversion of broadband P and SH waveforms, however, we found nothing anomalous in the source mechanisms. Among the 33 largest deep earthquakes (Mw≥7.0) since 1958, only events in the Mariana have high aftershock activity. We associate this observation with the unique configuration of the Mariana slab which seems to plunge sub-vertically into the lower mantle with little distortion.

The occurrence of the March 9, 1994 (Mw 7.6) deep Tonga earthquake during a temporary deployment of broadband seismographs provides an excellent opportunity to study the seismic discontinuity structure in the vicinity of a subducting slab. Near vertical ScS reverberations from the 410, 520, and 660 discontinuities are visible at 10 stations in the South Pacific at frequencies between 0.003–0.033 Hz without stacking. We use a regional velocity model and ray theory to model the interval from ScS to ScS3. A grid search is then performed on the depth and impedance contrast at each of the seismic discontinuities to optimize the fit to both the underside and topside reflections. Stacked records indicate prominent discontinuities at 400 km and 665 km depth beneath and to the west of the slab, and 410 km, 500 km, and 640 km to the east of the slab. Analysis of individual records suggests a local depression of 30 km in the 660 and a broadening of the transition zone directly beneath and to the west of the Tonga slab, in a region of possible near horizontal slab. These results are consistent with a large cool region just to the west of the intersection of the Tonga slab with the 660.

According to the feature of Hindukush-Pamirs intermediate focus earthquake belt with a S-shaped pattern of dip direction reversion by using the data of earthquake catalogues obtained by seismic networks in Xinjiang and Mid-Asia area of the former Soviet Union, by means of focal mechanism solution and tectonic stress analysis, it is considered that the intermediate focus earthquake belt is possibly formed by the compression rupture which is caused by the collision between Indian and European Plates in the lithosphere of the upper mantle. Under the action of torsion moment, the continuous torsional break of reverse part of the earthquake belt might be the reason why the intermediate focus strong earthquakes occur repeatedly in the same place. In this paper, the boundary line between the intermediate focus earthquake belt and the shallow focus earthquake region of the western part of south Tianshan is also defined from the angle of seismicity division.

Characteristics of two recent moderate earthquakes from India (1993 Killari, Mw 6.1 and 1997 Jabalpur, Mw 5.8) are presented in this paper to illustrate the influence of geologic and tectonic setting on the nature of seismicity in stable continental regions (SCRs). The Killari event, which occurred in the middle of a cratonic region of low background seismicity, long recurrence interval and poorly developed neotectonic features, may be qualified as a common type of SCR event. This earthquake is also characterized by a shallow focal depth and a long sequence of aftershocks. In contrast, the Jabalpur earthquake shows a spatial association with a well-defined tectonic structure with significant background seismicity. A remarkable feature of this earthquake is its deep focus, not commonly observed in SCR seismicity. This event was associated with fewer aftershocks unlike most other SCR earthquakes. Analogous events are found in other shield regions as well. Based on their general characteristics, a broad classification of moderate SCR earthquakes is attempted here. Two broad groups have been identified in this study: (1) a set of deep focus earthquakes that are related to intracratonic paleorifts, and (2) shallow-focus earthquakes associated with discrete faults in the less deformed Precambrian terrains. A synthesis of earthquake data from different geologic environments will improve our understanding of the seismogenesis in the shield regions. © 1999 Elsevier Science B.V. All rights reserved.

We analyze the coda of teleseismic P waves to deterministically map seismic scatterers in a 1000 km on-a-side cube beneath the Izu-Bonin trench by migrating and stacking waveforms. This method was applied to several hundred short-period vertical-component western United States seismometer recordings of 17 deep-focus Izu-Bonin earthquakes. Except for isolated arrivals, the midmantle appears devoid of sharp discontinuities. S-to-P conversions at the 660-km discontinuity generated the largest signals; the 410-km discontinuity occasionally generated signals. Horizontal discontinuities poorly explain other signals; however, S-to-P conversions generated at a nearly north-south trending, steeply dipping discontinuity at 30°N, 145°E, and 1000 km beneath and parallel to the Izu-Bonin trench explain most additional signals. Observed coherent S-to-P conversions of 2-s period waves limit the width of a gradient zone to

We analyze residual spheres for teleseismic S wave arrival times from 10 intermediate and deep focus earthquakes in the Kurile subduction zone. Long-period WWSSN and CSN recordings of transverse component S and ScS arrivals provide good azimuthal coverage for each event, with arrival time estimates that compare well with short-period observations. In order to isolate the near-source structure, substantial lower mantle and receiver path effects are suppressed both by using empirical station/path statics derived from S, sS, ScS, and sScS arrivals for a total of 28 northwestern Pacific events and by constructing differential residual spheres. Ray tracing through three-dimensional aspherical shear velocity models supports the inferred large magnitude of the deep mantle and near receiver effects, although existing models do not appear to provide accurate corrections. The events are relocated in the PREM structure, and a modest spatial smoothing of the residuals is applied. The resulting residual spheres are modeled using a finite difference travel time method that accounts for three-dimensional ray path perturbations. Our starting models are derived from slab structures determined by previous P wave residual sphere analyses. We explore the compatibility of the P and S wave observations to constrain the Vs/Vp velocity ratio of the slab and to test whether the S wave arrival times place additional constraints on the slab geometry. The modeling supports the previous results favoring penetration of the Kurile slab to at least 800 km depth for most of the length of the arc, with some evidence that the northern portion of the slab broadens by as much as a factor of 3 upon entering the lower mantle. The preferred slab shear velocity models have relatively low (2–3%) shear velocity contrasts below 670 km depth, which can be accounted for by slab temperature reductions of 300°–400°.

It is usually assumed that earthquakes in intraplate regions occur in the upper crust, and northwestern Italy is generally assigned to this kind of ‘normal’ seismicity. In this work, the depth distribution of the events localized in this area by the Istituto Geofisico Geodetico (IGG) seismic network in the period 1991–1997 is analyzed in detail. In particular, the location capability of the network is discussed, adopting as reference quarry blasts (for the epicentral position) and the locations obtained from a dense temporary network (for the depth estimate). Within the so-obtained error limits, the depth distribution of events show a characteristic pattern: while for most of the area covered by the network the well-located seismicity lies within the first 20 km of depth, in a band following the inner arc of the Western Alps, numerous events have anomalously large focal depths, reaching a maximum of 114 km. These depth determinations cannot be attributed to instabilities of the location procedure: different choices of the propagation models used for the hypocentral determination led to very similar depth values, always significantly larger than the standard values for the surrounding areas. A strong correlation has been found between the 3-dimensional distribution of these foci and the P-wave propagation anomalies obtained from tomographic studies, suggesting a direct link between elastic and rheological properties of lower crust and upper mantle in this area.

Both seismological observations and mineral physics experiments suggest that old and cold subducting slabs might contain a wedge of metastable olivine (`MO'). Dynamically consistent models of subducting slabs, in which the kinetics of the olivine to spinel (wadsleyite or ringwoodite) transition is treated in a simplified way, are carried out to evaluate the buoyancy effect of MO on the subduction velocity. Assuming slab thicknesses according to the cooling half-space model, we find that MO forms in slabs with ages greater than approximately 70 Myr, and that a significant slowing down of subduction velocity occurs for ages greater than 100 Myr. Because a decrease in the subduction velocity for old lithospheric ages is not observed, our models suggest that the amount of MO in slabs is less than approximately 5000 km2 in cross section, and that for old lithosphere the cooling half-space model is not applicable.

Double Benioff zones provide opportunities for insight into seismogenesis because the underlying mechanism must explain two layers of deep earthquakes and the separation between them. We characterize layer separation inside subducting plates with a coordinate rotation to calculate the slab-normal distribution of earthquakes. Benchmark tests on well-established examples confirm that layer separation is accurately quantified with global seismicity catalogs alone. Global analysis reveals double Benioff zones in 30 segments, including all 16 subduction zones investigated, with varying subducting plate ages and stress orientations, which implies that they are inherent in subducting plates. Layer separation increases with age and is more consistent with dehydration of antigorite than chlorite.

To constrain dynamic source properties of deep earthquakes, we have systematically constructed broadband time functions of deep earthquakes by stacking and scaling teleseismic P waves from U.S. National Seismic Network, TERRAscope, and Berkeley Digital Seismic Network broadband stations. We examined 42 earthquakes with depths from 100 to 660 km that occurred between July 1, 1992 and July 31, 1995. To directly compare time functions, or to group them by size, depth, or region, it is essential to scale them to remove the effect of moment, which varies by more than 3 orders of magnitude for these events. For each event we also computed short-period stacks of P waves recorded by west coast regional arrays. The comparison of broadband with short-period stacks yields a considerable advantage, enabling more reliable measurement of event duration. A more accurate estimate of the duration better constrains the scaling procedure to remove the effect of moment, producing scaled time functions with both correct timing and amplitude. We find only subtle differences in the broadband time-function shape with moment, indicating successful scaling and minimal effects of attenuation at the periods considered here. The average shape of the envelopes of the short-period stacks is very similar to the average broadband time function. The main variations seen with depth are (1) a mild decrease in duration with increasing depth, (2) greater asymmetry in the time functions of intermediate events compared to deep ones, and (3) unexpected complexity and late moment release for events between 350 and 550 km, with seven of the eight events in that depth interval displaying markedly more complicated time functions with more moment release late in the rupture than most events above or below. The first two results are broadly consistent with our previous studies, while the third is reported here for the first time. The greater complexity between 350 and 550 km suggests greater heterogeneity in the failure process in that depth range.

Latent heat release by equilibrium mineralogical transformations in an adiabatically subducting slab reversibly perturbs temperatures and pressures so as to conserve entropy. However, latent heats of metastable transformations in such a slab yield irreversible isobaric temperature changes which increase entropy despite adiabatic constraints. As a result, latent heat release by metastable exothermic transformations can yield local superheating above the background adiabat, with the degree of potential superheating increasing with extent of metastable overstep. In real slabs, however, regions of metastably persisting low-pressure phases should undergo conductive warming from surrounding transformed material. Such warming should proceed more rapidly than warming of the bulk slab from the surrounding mantle, and the resulting decrease in metastable transition pressures will slightly decrease the degree of local superheating. Nonetheless, such local temperature increases may trigger seismic release of accumulated strain energy via a number of proposed mechanisms of shear instability. Adiabatic instability, in the form of shear localization in material with temperature-dependent rheology, is one mechanism which may be triggered by such latent heat release in metastable regions yet produce rupture that extends beyond the boundaries of such regions.

We stack the teleseismic depth phases sS, sP, and pP produced by deep focus earthquakes to image precursory arrivals that result from near-source, underside reflections off the 410-km seismic velocity discontinuity (hereinafter referred to as the 410) and use differential time measurements between these phases and their precursors to compute discontinuity depths near seven subduction zones around the Pacific Ocean margin. We begin by selecting seismograms with high-quality depth phase arrivals recorded by several long-period networks between the years 1976 and 1996. Filtering the waveforms and stacking them along theoretical travel-time curves reveals clear precursors which vary in shape and timing. We compute confidence levels to evaluate the reliability of the observed precursory features using a bootstrap method that randomly resamples the seismograms prior to stacking. We measure the differential travel time between the reference pulse and the precursor using a cross-correlation technique and convert this time to an apparent discontinuity depth using the isotropic Preliminary Reference Earth Model (PREM) at 25-s period, corrected to an oceanic crustal thickness. The lateral resolution of our long-period stacks for 410 topography is limited compared to that sometimes achieved in short-period analyses but is much higher than that obtained from global SS precursor studies. For most subduction zones the results indicate little change in the average depth to the 410-km discontinuity in the local areas sampled by the precursor bounce points compared to broad regional depths inferred from SS precursor results. This implies that any large variations in depth to the 410-km discontinuity near subduction zones are limited to a narrow zone within the slab itself where they may be difficult to resolve with long-period data. Coverage for the Tonga and Peru-Chile subduction zones is sufficiently dense that we can observe lateral variations in 410 depths. In Tonga the results suggest depth variations perpendicular to the slab of up to 33 km, after correcting for probable lateral heterogeneity in velocity above 400 km depth, and variations parallel to the slab orientation as large as 13 km. The cross-slab variation is consistent with the elevation of olivine phase transformations in cold regions; the variation along strike suggests a more complex thermal heterogeneity that may be related to the subduction history of this region. We see evidence for additional reflectors above the 410 in some of the waveform stacks, but the inconsistency and weak amplitude of these features preclude definitive interpretations.

The duration of each subevent of 48 earthquakes with magnitude larger than 5.5 and depth greater than 100 km was determined from stacked traces of broadband records of Global Seismograph Network stations. We fitted the source time function by one or more triangles convolved with attenuation. We found that global stacks of displacement seismograms yield reliable estimates of the rupture duration. The durations, scaled to a moment of 1019 N m, of both the subevents and the entire earthquake show a slight decrease with depth from 9 s for events at 100 km depth to about 7 s for events at 600 km depth. Assuming that the rupture velocity is a constant fraction of the shear wave speed, this decrease can be completely explained by the increase in shear velocity of 20%. In this sense, deep earthquakes are comparable to intermediate ones. For some intermediate-depth events, Vidale and Houston [1993] found durations up to twice as long. We find that almost all of their slow events have been recorded at large epicentral distances. At these distances, we conjecture that the end of the P wave train may be extended by the arrival of reflections from the D″ layer.

We evaluate the methodologies that aim at resolving the space-time slip distribution during large earthquakes. In one common approach, the source process of an event is assumed to be composed of a series of subevents. Critical points in teleseismic body waves, such as peaks and coherent arrivals, are viewed as discrete points in the spatiotemporal history of an earthquake. We show that these features can be interpreted as the instantaneous centroid locations of the moment release history. We demonstrate this result with an analysis of the Mw=8.2 Bolivia 1994 deep event. We reconstruct the slip distribution in two contrasting ways. First, broadband P waves are interpreted as being composed of subevents. Five peaks in the moment release rate are located relative to the onset of the rupture using the azimuthal variations of onset-to-peak times. Second, we invert broadband P waves for the slip distribution on a subhorizontal fault. We use a nonlinear tomographic imaging technique called simulated annealing and an L1 misfit norm. Rupture velocities appear to be

One of the largest (Mw = 7.8) deep earthquakes (595 km) in recorded history occurred beneath the Flores Sea on June 17, 1996. We analyze broadband body waveforms to determine the seismic source characteristics of this important earthquake. The source time function has a small initial subevent (5 s duration) followed by three major subevents with a total source duration of 29 s. Directivity analysis indicates an unilateral rupture to the east-southeast with a subhorizontal rupture length of 75 km. A time-independent moment tensor (TIMT) inversion gives a total seismic moment of 5.4×1020 N m ( Mw = 7.8). A time-dependent moment tensor (TDMT) inversion indicates rupture on a near-vertical fault (strike of 130°, dip of 80°SW, rake of −166°) during the first 5 s, releasing 8% of the seismic moment. The major part of the moment was released along a shallower dipping plane (strike of ∼100°, dip of ∼55°, rake of ∼−45°). This change of faulting geometry may imply that the rupture initiated at a structural, thermal, or phase boundary of high stress concentration and then triggered the major moment release along a regional weak zone. The mainshock produced at least 23 (mb>2.9) aftershocks within the first 3 days. Seventeen of the largest aftershocks were located using the Joint Hypocenter Determination algorithm. The aftershocks define a plane dipping antithetically to the dip of the subducting slab with lateral and depth dimensions of 145 km and 76 km, respectively. The aftershock locations correlate with the southeast striking nodal plane of the moment tensor mechanism.

The June 17, 1996 (Mw 7.8) Flores Sea event is one of the largest deep earthquakes, and also produced an energetic aftershock sequence, allowing study of the spatial relationships between moment release, aftershock occurrence, and slab geometry. The locations of the mainshock, rupture sub-events, aftershocks, and background slab seismicity are determined using a relative location algorithm. The event occurred at the edge of a 150 km wide aseismic region that represents a tear in the lowermost part of the Indonesian slab. The aftershocks define a 90 km long region parallel to slab strike. Two other aftershocks locate 10–25 km outside the previously active slab. The mainshock propagated about 75 km eastward along an E-W striking, southward dipping plane, with a rupture velocity of 3.5 km/s. There was a substantial change in focal mechanism during the rupture, with the fault plane strike changing by about 25°. Overall, the Flores Sea deep earthquake shows numerous large aftershocks, and a moderate rupture velocity (3.5 km/s) and stress drop (∼ 7-20 MPa.) These characteristics are similar to the 1994 Tonga deep earthquake, but different from the 1994 Bolivia event, suggesting they may be typical of large deep earthquakes in colder slabs. Systematic differences in rupture parameters for deep earthquakes in warm and cold slabs suggest that the rupture mechanism is highly sensitive to temperature.

Information concerning a total number of 13700 instrumentally recorded earthquakes is used to study the geographical and the vertical distribution of the Earth's seismicity. From these earthquakes, which form four complete samples of data (M ≥ 7.0, 1894–1992; M ≥ 6.5, 1930–1992; M ≥ 6.0, 1953–1992; M ≥ 5.5, 1966–1992), 11511 are shallow (h ≤ 60 km), 2085 are of intermediate focal depth (61 ≤ h ≤ 300 km) and 564 are deep focus earthquakes (301 ≤ h ≤ 720 km). The parameters a and b of the frequency-magnitude relationship were calculated in a grid of equally spaced points at 1° by using the data of earthquakes located inside circles centered at each point. The radius of the circles increased from 30 km with a step of 10 km until the information for the earthquakes located inside the circle fulfil three criteria which concern the size of the sample used to compute these parameters at each point of the grid. The results are given in a qualitative way (epicenter maps) as well as in a quantitative way (mean return periods).

The 16 April 2013 southeast Iran earthquake ( M w 7.8) is the second major earthquake in the region, including the Makran subduction zone, since 1945. This intraslab earthquake, within the subducting Arabian plate, occurred as a result of normal faulting at an intermediate depth. We report coseismic offsets from the two nearby continuously operating Global Positioning System sites that are located within 300 km of the epicenter. The coseismic offsets are consistent with the slip on the northward steeply dipping fault plane. This earthquake probably occurred in response to the slab‐detachment process, and its occurrence has implications for the plate interface coupling of the Makran subduction zone.

Thermal perturbation of mantle phase relations in subduction zones gives rise to significant buoyancy anomalies. Finite element modeling of stresses arising from these anomalies reveals transition from principal tension to compression near ∼400 km depth, down-dip compression over ∼400–690 km (peaking at ∼550 km), and transition to rapidly fading tension below ∼690 km. Such features, even when complicated by olivine metastability, are consistent with observed patterns of deep seismicity. That such a simple model, neglecting all effects other than buoyancy anomalies due to temperature and to thermal perturbation of olivine phase relations, successfully generates so many observed features of deep seismicity suggests that these buoyancy anomalies are significant contributors to the stress field in subducting slabs. It also suggests that the depth distribution of deep seismicity may largely reflect the state of stress in the slab rather than simply a particular mechanism of stress release.

We report the observation of T phases recorded at a number of Pacific island sites, following the great Bolivian deep earthquake of June 9, 1994. By studying their arrival times at the various stations, we unravel the process of conversion from seismic to acoustic energy at the South American coastline. We find that the maxima of amplitude of the phase are incompatible with propagation along the great circle from source to receiver but rather require the generation of the acoustic wave at a common point at the Arica Bight for all stations in the South Pacific, and even farther south for stations in Hawaii and the Bonin Islands. The timing of these conversions corresponds to the arrival of regional S waves at the conversion points. The high-frequency nature of the T phase requires the channeling of the S wave from the seismic source to conversion point through low-attenuation material, which in turns adds supportive evidence for the mechanical continuity of the downgoing slab under that section of South America, despite the observed gap in seismicity between 300 and 600 km.