Patterns of significant seismic quiescence on the Mexican Pacific coast

Many authors have proposed that the study of seismicity rates is an appropriate technique for evaluating how close a seismic gap may be to rupture. We designed an algorithm for identification of patterns of significant seismic quiescence by using the definition of seismic quiescence proposed by Schreider (1990). This algorithm shows the area of quiescence where an earthquake of great magnitude may probably occur. We have applied our algorithm to the earthquake catalog on the Mexican Pacific coast located between 14 and 21 degrees of North latitude and 94 and 106 degrees West longitude; with depths less than or equal to 60 km and magnitude greater than or equal to 4.3, which occurred from January, 1965 until December, 2014. We have found significant patterns of seismic quietude before the earthquakes of Oaxaca (November 1978, M_w = 7.8), Petatlán (March 1979, M_w = 7.6), Michoacán (September 1985, M_w = 8.0, and M_w = 7.6) and Colima (October 1995, M_w = 8.0). Fortunately, in this century earthquakes of great magnitude have not occurred in Mexico. However, we have identified well-defined seismic quiescences in the Guerrero seismic-gap, which are apparently correlated with the occurrence of silent earthquakes in 2002, 2006 and 2010 recently discovered by GPS technology.     

Earthquake ground motion predictive equations for Garhwal Himalaya,India

Predictive equations based on the stochastic approach are developed for earthquake ground motions from Garhwal Himalayan earthquakes of 3.5≤M_w≤6.8 at a distance of 10≤R≤250 km. The predicted ground motion parameters are response spectral values at frequencies from 0.25 to 20 Hz, and peak ground acceleration (PGA). The ground motion prediction equations (GMPEs) are derived from an empirically based stochastic ground motion model. The GMPEs show a fair agreement with the empirically developed ground motion equations from Himalaya as well as the NGA equation. The proposed relations also reasonably predict the observed ground motion of two major Himalayan earthquakes from Garhwal Himalayan region. For high magnitudes, there is insufficient data to satisfactorily judge the relationship; however it reasonably predicts the 1991 Uttarkashi earthquake (M_w=6.8) and 1999 Chamoli earthquake (M_w=6.4) from Garhwal Himalaya region.     

Ground response analysis of Kanpur soil along Indo-Gangetic Plains

The present work deals with 1D and 2D ground response analysis and liquefaction analysis of alluvial soil deposits from Kanpur region along Indo-Gangetic plains. Standard penetration tests and seismic down hole tests have been conducted at four locations namely IITK, Nankari village, Mandhana and Bithoor at 1.5 m interval up to a depth of 30 m below the ground surface to find the variation of penetration blows and the shear wave velocity along the depth. From the selected sites undisturbed as well as representative soil samples have been collected for detailed soil classification. The soil profiles from four sites have been considered for 1D and 2D ground response analysis by applying the free field motions of three Himalayan earthquakes namely Chamba earthquake (M_w—5.1), Chamoli earthquake (M_w—6.4) and Uttarkashi earthquake (M_w—6.5). An average value of Peak Ground Acceleration (PGA) obtained from 1D and 2D analysis is considered for liquefaction analysis and post-liquefaction settlement. The excess pore water pressure ratio is greater than 0.8 at a depth of 24 m from ground surface for IITK, Nankari village, Bithoor sites. More than 50% of post liquefaction settlement is contributed by layers from 21–30 m for all sites. In general, the soil deposits in Kanpur region have silty sand and sand deposits and are prone to liquefaction hazards due to drastic decrease of cyclic resistance ratio (CRR) at four chosen sites in Kanpur.     

Analysis of the relations between slope failure distribution and seismic ground motion during the 2008 Wenchuan earthquake

The 2008 Wenchuan earthquake with M_w7.9 occurred at Yingxiu County in Longmenshan thrust fault belt, southwest China, having triggered a huge amount of slope failures. This paper applied a detailed inventory with more than 190,000 slope failures and strong ground motion records of 187 seismic stations to analyze the qualitative and quantitative relations between slope failure distribution and seismic parameters. The results revealed that slope failure distribution was exponentially decreased with the increment of epicentral distance and distance from surface fault rupture; peak ground acceleration (PGA) on the hanging wall side was apparently larger than that on footwall side. Linear correlation between concentration of slope failures (LNC) and the percentage of the area affected by slope failure (LAP) and PGA was demonstrated by statistical analysis, which revealed that 0.18~0.21 g horizontal PGA was the threshold value of the occurrence of slope failure. Furthermore, this paper presented an empirical model for the attenuation relation of slope failure distribution.     

Temporal and spatial variations of seismicity scaling behavior in Southern México

R/S analysis is used in this work to investigate the fractal correlations in terms of the Hurst exponent for the 1998–2011 seismicity data in Southern Mexico. This region is the most seismically active area in Mexico, where epicenters for severe earthquakes (e.g., September 19, 1985, M_w = 8.1) causing extensive damage in highly populated areas have been located. By only considering the seismic events that meet the Gutenberg–Ritcher law completeness requirement (b = 0.97, M_GR = 3.6), we found time clustering for scales of about 100 and 135 events. In both cases, a cyclic behavior with dominant spectral components at about one cycle per year is revealed. It is argued that such a one-year cycle could be related to tidal effects in the Pacific coast. Interestingly, it is also found that high-magnitude events (M_w ≥ 6.0) are more likely to occur under increased interevent correlations with Hurst exponent values H > 0.65. This suggests that major earthquakes can occur when the tectonic stress accumulates in preferential directions. In contrast, the high-magnitude seismic risk is reduced when stresses are uniformly distributed in the tectonic shell. Such cointegration between correlations (i.e., Hurst exponent) and macroseismicity is confirmed for spatial variations of the Hurst exponent. In this way, we found that, using the Hurst exponent standpoint, the former presumed Michoacan and the Guerrero seismic gaps are the riskiest seismic zones. To test this empirical finding, two Southern Mexico local regions with large earthquakes were considered. These are the Atoyac de Alvarez, Guerrero (M_w = 6.3), and Union Hidalgo, Oaxaca (M_w = 6.6), events. In addition, we used the Loma Prieta, California, earthquake (October 17, 1989, M_w = 6.9) to show that the high-magnitude earthquakes in the San Andreas Fault region can also be linked to the increments of determinism (quantified in terms of the Hurst exponent) displayed by the stochastic dynamics of the interevent period time series. The results revealed that the analysis of seismic activity by means of R/S analysis could provide further insights in the advent of major earthquakes.     

GIS-based soil liquefaction susceptibility map of Mumbai city for earthquake events

The problem of liquefaction of soil during seismic event is one of the important topics in the field of Geotechnical Earthquake Engineering. Liquefaction of soil is generally occurs in loose cohesionless saturated soil when pore water pressure increases suddenly due to induced ground motion and shear strength of soil decreases to zero and leading the structure situated above to undergo a large settlement, or failure. The failures took place due to liquefaction induced soil movement spread over few square km area continuously. Hence this is a problem where spatial variation involves and to represent this spatial variation Geographic Information System (GIS) is very useful in decision making about the area subjected to liquefaction. In this paper, GIS software GRAM++ is used to prepare soil liquefaction susceptibility map for entire Mumbai city in India by marking three zones viz. critically liquefiable soil, moderately liquefiable soil and non liquefiable soil. Extensive field borehole test data for groundwater depth, standard penetration test (SPT) blow counts, dry density, wet density and specific gravity, etc. have been collected from different parts of Mumbai. Simplified procedure of Youd et al. (2001) is used for calculation of factor of safety against soil liquefaction potential. Mumbai city and suburban area are formed by reclaiming lands around seven islands since 1865 till current date and still it is progressing in the area such as Navi Mumbai and beyond Borivali to Mira road suburban area. The factors of safety against soil liquefaction were determined for earthquake moment magnitude ranging from M_w = 5.0 to 7.5. It is found that the areas like Borivali, Malad, Dahisar, Bhandup may prone to liquefaction for earthquake moment magnitude ranging from M_w = 5.0 to 7.5. The liquefaction susceptibility maps were created by using GRAM++ by showing the areas where the factor of safety against the soil liquefaction is less than one. Proposed liquefaction susceptibility map of Mumbai city can be used by researchers for earthquake hazard analysis, for the preventive measures in disaster management, for urban planning and further development of Mumbai city and suburban area.     

Seismic microzonation studies for the city of Ragusa (Italy)

The seismic history of the city of Ragusa (Italy), the geotechnical characterisation of the subsoil and the site response analysis should be correctly evaluated for the definition of the Seismic Geotechnical Hazard of the city of Ragusa, through geo-settled seismic microzoning maps. Basing on the seismic history of the city of Ragusa, the following earthquake scenarios have been considered: the “Val di Noto” earthquake of January 11, 1693 (with intensity X–XI on MCS scale, magnitude M_W=7.41 and epicentral distance of about 53 km); the “Etna” earthquake of February 20, 1818 (with intensity IX on MCS scale, magnitude M_W=6.23 and epicentral distance of about 64 km); the Vizzini earthquake of April 13, 1895 (with intensity I=VII–VIII on MCS scale, magnitude M_W=5.86 and epicentral distance of about 26 km); the “Modica” earthquake of January 23, 1980 (with intensity I=V–VI on MCS scale, magnitude M_W=4.58 and epicentral distance of about 10 km); the “Sicilian” earthquake of December 13, 1990 (with intensity I=VII on MCS scale, magnitude M_W=5.64 and epicentral distance of about 50 km). Geotechnical characterisation has been performed by in situ and laboratory tests, with the definition of shear wave velocity profiles in the upper 30 m of soil. Soil response analyses have been evaluated for about 120 borings location by some non-linear 1-D models. Finally the seismic microzonation of the city of Ragusa has been obtained in terms of maps with different peak ground acceleration at the surface; shaking maps for the central area of the city of Ragusa were generated via GIS for the earthquake scenarios.     

Site response effects on an urban overpass

Strong ground motion variability due to rapid changes in subsoil conditions may lead to different site responses, which in turn yields to beneficial or detrimental soil–foundation–structure interaction. This technical note presents the results of a seismic soil–structure interaction analysis conducted using a 2D finite difference model, developed with the program FLAC, of a critical section of a 60 km long strategic urban overpass, which is under construction in Mexico City, for a M_w 8.7 earthquake. Initially, the response of the free field was calibrated comparing the values obtained with FLAC, with those gathered using the computer code QUAD4M. Good agreement was observed between the results generated with these programs. Accelerations and displacements were determined at the upper deck and foundation of the urban overpass. Important seismic soil–structure interaction was observed along the overpass at the supports analyzed. This numerical study helps to gain insight regarding the site response ground motion incoherence effects that influence the dynamic behavior of this kind of structures during extreme events.     

Seismic response of deep Quaternary sediments in historical center of L’Aquila City (central Italy)

On 6 April 2009 a M_w=6.1 earthquake produced severe destruction and damage over the historic center of L’Aquila City (central Italy), in which the accelerometer stations AQK and AQU recorded a large amount of near-fault ground motion data. This paper analyzes the recorded ground motions and compares the observed peak accelerations and the horizontal to vertical response spectral ratios with those revealed from numerical simulations. The finite element method is considered herein to perform dynamic modeling on the soil profile underlying the seismic station AQU. The subsurface model, which is based on the reviewed surveys that were carried out in previous studies, consists of 200–400 m of Quaternary sediments overlying a Meso-Cenozoic carbonate bedrock. The Martin-Finn-Seed's pore-water pressure model is used in the simulations. The horizontal to vertical response spectral ratio that is observed during the weak seismic events shows three predominant frequencies at about 14 Hz, 3 Hz and 0.6 Hz, which may be related to the computed seismic motion amplification occurring at the shallow colluvium, at the top and base of the fluvial-lacustrine sequence, respectively. During the 2009 L’Aquila main shock the predominant frequency of 14 Hz shifts to lower values probably due to a peculiar wave-field incidence angle. The predominant frequency of 3 Hz shifts to lower values when the earthquake magnitude increases, which may be associated to the progressive softening of soil due to the excess pore-water pressure generation that reaches a maximum value of about 350 kPa in the top of fluvial-lacustrine sequence. The computed vertical peak acceleration underestimates the experimental value and the horizontal to vertical peak acceleration ratio that is observed at station AQU decreases when the earthquake magnitude increases, which reveals amplification of the vertical component of ground motion probably due to near-source effects.     

Bayesian analysis on earthquake magnitude related to an active fault in Taiwan

It is understood that sample size could be an issue in earthquake statistical studies, causing the best estimate being too deterministic or less representative derived from limited statistics from observation. Like many Bayesian analyses and estimates, this study shows another novel application of the Bayesian approach to earthquake engineering, using prior data to help compensate the limited observation for the target problem to estimate the magnitude of the recurring Meishan earthquake in central Taiwan. With the Bayesian algorithms developed, the Bayesian analysis suggests that the next major event induced by the Meishan fault in central Taiwan should be in M_w 6.44±0.33, based on one magnitude observation of M_w 6.4 from the last event, along with the prior data including fault length of 14 km, rupture width of 15 km, rupture area of 216 km², average displacement of 0.7 m, slip rate of 6 mm/yr, and five earthquake empirical models.     

A new fault rupture scenario for the 2003 Mw 6.6 Bam earthquake,SE Iran: Insights from the high-resolution QuickBird imagery and field observations

The December 26, 2003 M_w 6.6 Bam earthquake is one of the most disastrous earthquakes in Iran. QuickBird panchromatic and multispectral satellite imagery with 61 cm and 2.4 m ground resolution, respectively provide new insights into the surface rupturing process associated with this earthquake. The results indicate that this earthquake produced a 2–5 km-wide surface rupture zone with a complex geometric pattern. A 10-km-long surface rupture zone developed along the pre-existing Bam fault trace. Two additional surface rupture zones, each 2–5 km long, are oblique to the pre-existing Bam fault in angles of 20–35°. An analysis of geometric and geomorphic features also shows that movement on the Bam fault is mainly right-lateral motion with some compressional component. This interpretation is consistent with field investigations, analysis of aftershocks as well as teleseismic inversion. Therefore, we suggest that the 2003 Bam earthquake occurred on the Bam fault, and that the surface ruptures oblique to the Bam fault are caused by secondary faulting such as synthetic shears (Reidel shears). Our fault model for the Bam earthquake provides a new tectonic scenario for explaining complex surface deformations associated with the Bam earthquake.     

Time-Domain Moment Tensors for shallow (h ≤ 40 km) earthquakes in the broader Aegean Sea for the years 2006 and 2007: The database of the Aristotle University of Thessaloniki

We present a catalog of moment tensor (MT) solutions and moment magnitudes, M_w, for 119 shallow (h ≤ 40 km) earthquakes in Greece and its surrounding lands (34°N–42°N, 19°E–30°E) for the years 2006 and 2007, computed with the 1D Time-Domain Moment Tensor inversion method (TDMT_INV code of Dreger, 2003). Magnitudes range from 3.2 ≤ M_w ≤ 5.7. Green's functions (GF) have been pre-computed to build a library, for a number of velocity profiles applicable to the broader Aegean Sea region, to be used in the inversion of observed broad band waveforms (10–50 s). All MT solutions are the outcome of a long series of tests of different reported source locations and hypocenter depths. Quality factors have been assigned to each MT solution based on the number of stations used in the inversion and the goodness of fit between observed and synthetic waveforms. In general, the focal mechanisms are compatible with previous knowledge on the seismotectonics of the Aegean area. The new data provide evidence for strike-slip faulting along NW–SE trending structures at the lower part of Axios basin, close to the heavily industrialized, and presently subsiding, region of the city of Thessaloniki. Normal faulting along E–W trending planes is observed at the Strimon basin, and in Orfanou Gulf in northern Greece. A sequence of events in the east Aegean Sea close to the coastline with western Anatolia sheds light on an active structure bounding the north coastline of Psara–Chios Islands about 20–25 km in length exhibiting right lateral strike-slip faulting.     

Systematic deflection and offset of the Yangtze River drainage system along the strike-slip Ganzi-Yushu-Xianshuihe Fault Zone,Tibetan Plateau

The Ganzi-Yushu-Xianshuihe Fault Zone (GYXFZ) is a typical active strike-slip fault that has triggered many large historic earthquakes, including the 2010 M_w 6.9 Yushu earthquake in the central Tibetan Plateau. This fault zone extends for ca. 800 km from the central Tibetan Plateau to its southeastern margin and varies in trend from WNW-ESE in the northwestern segment of the fault zone to NNW-SSE in the southeastern segment, having the geometry of an arc projecting northeastwards. In this study, we present evidence for the systematical sinistral deflection and/or offset of the Yangtze River and its branch stream channels and valleys along the GYXFZ. Topographic analysis of three-dimensional (3D) perspective images constructed using digital elevation model (DEM) data, 0.5 m-resolution WorldView and GeoEye images, and 15 m-resolution Landsat-Enhanced Thematic Mapper (ETM+) images, together with analysis of geological structures, reveals the following: (i) the main river channels and valleys of the Yangtze River drainage system show systematic sinistral deflections and/or offsets along the GYXFZ; (ii) various amounts of sinistral offset have accumulated on the tributary stream channels, valleys, and gullies of the Yangtze River along the fault, with a linear relation, D = aL, between the upstream length L from the deflected point and the offset amount D with a certain coefficient a; (iii) the maximum amount of sinistral offset is up to ca. 60 km, which was accumulated in the past 13–5 Ma; and (iv) the long-term average strike-slip rate is ca. 4.6–12 mm/year. Geological and geomorphic evidence, combined with geophysical data, demonstrates that the GYXFZ is currently active as one of the major seismogenic faults in the Tibetan Plateau, dominated by left-lateral strike-slip motion. Our findings supply important evidence for the tectonic evolution of strike-slip faults in the Tibetan Plateau since the Eurasia-India continental collision.     

Prediction of ground motion due to the collapse of a large-scale cooling tower under strong earthquakes

The ground motion owing to the collapse of a large-scale cooling tower under strong earthquakes was appropriately predicted using a comprehensive approach. The predicted results can be used for the safety evaluation of nuclear-related facilities adjacent to the cooling tower as well as in the planning of nuclear power plant construction in China. In this study, a cooling tower–soil model was first developed based on a falling weight–soil model, which the authors verified by falling weight tests. Then the collapse process of a cooling tower was simulated, and the collapse-induced ground vibrations were assessed by using the proposed model. Finally, the ground motion, which was a combination of the earthquake-induced ground motion and the collapse-induced ground vibrations, was estimated based on the superposition principle of waves. It was found that the cooling tower may collapse under strong earthquakes with the peak ground accelerations (PGAs) in the range of 0.35–0.45 g in x (EW) and y (NS) directions, respectively. These PGAs are far beyond the PGA range of major earthquakes in the common seismic design in China. The types of the site geologies of towers can significantly affect the collapse-induced ground vibrations. For a typical hard soil consisting of strongly weathered sandy slate, moderate ground vibrations may occur in the considered region. The collapse-induced PGAs were in the range of 0.017–0.046 g for the observed points at distances of 350 m in radial direction. For a rock-like foundation, the collapse-induced radial PGAs may be as high as 0.08 g at distances of 350 m, indicating that the effect of the collapse-induced ground vibrations on the nuclear-related facilities should be seriously assessed in certain scenarios.     

Seismic microzoning from synthetic ground motion earthquake scenarios parameters: The case study of the city of Catania (Italy)

The city of Catania (Italy) in the South-Eastern Sicily has been affected in past times by several destroying earthquakes with high values of estimated magnitude. The seismogenic area to the south of Volcano Etna, known as Iblean Area, is placed between the African and the Euro-Asiatic plates on the west of the Ibleo-Maltese escarpment, to the south of the Graben of the Sicilian channel and on the east of the overlapping front of Gela. Basing on the seismic history of Catania, the following earthquake scenarios have been considered: the “Val di Noto” earthquake of January 11, 1693 (with intensity X-XI on MCS scale, magnitude M_W=7.41 and epicentral distance of about 13 km); the “Etna” earthquake of February 20, 1818 (with intensity IX on MCS scale, magnitude M_W=6.23 and epicentral distance of about 10 km). The soil response analysis at the surface, in terms of time history and response spectra, has been obtained by 1-D equivalent linear models for about 1200 borings location available in the data-bank of the central area of Catania of about 50 km², using deterministic design scenario earthquakes as input at the conventional bedrock.Seismic microzoning maps of the city of Catania have been obtained in terms of different peak ground acceleration at the surface and in terms of amplification ratios for given values of frequency.     

Nonlinear geodynamics of the Baikal rift system: An evolution scenario with triple equilibrium bifurcation

This is an attempt to analyze the current lithospheric stress pattern in the Baikal rift in terms of nonlinear dynamics as an open self-organizing system in order to gain more insights into the general laws of regional seismicity. According to the suggested approach, the stress pattern inferred from seismic moments of 70,000 M_LH ≥ 2.0 events that occurred in the region between 1968 and 1994 is presented as a phase portrait in the phase spaces of the seismic moments. The obtained phase portrait of the system evolution fits well a scenario with triple equilibrium bifurcation where stress bifurcations account for the frequency of M > 5.5 earthquakes. Extrapolation of the results into the nearest future indicates probability of such a bifurcation (a catastrophe of stress), i.e., there is growing risk that M ≈ 7 events may happen in the region within a few years.     

Modification of fault slip models of the Mw 9.0 Tohoku Earthquake by far field GPS observations

GPS data from Crustal Movement Observation Network of China (CMONOC) are used to derive far-field co-seismic displacements induced by the M_w 9.0 Tohoku Earthquake. Significant horizontal displacements about 30 mm, 10 mm, and 20 mm were caused by this large event in northeast China, north China, and on the Korean peninsula respectively. Vectors of relatively large horizontal displacements with dominant east components pointed to the epicenter of this earthquake. The east components show an exponential decay with the longitude, which is characteristic of the decay of the co-seismic horizontal displacements associated with earthquakes of thrust rupture. The exponential fit of the east components shows that the influence of the co-seismic displacements can be detected by GPS at a distance of about 3200 km from the epicenter of the earthquake. By considering the capability of the far field displacements for constraining the inversion of the fault slip model of the earthquake, we use spherically stratified Earth models to simulate the co-seismic displacements induced by this event. Using computations and comparisons, we discuss the effects of parameters of layered Earth models on the results of dislocation modeling. Comparisons of the modeled and observed displacements show that far field GPS observations are effective for constraining the fault slip model. The far field horizontal displacements observed by GPS are used to modify the slips and seismic moments of fault slip models. The result of this work is applicable as a reference for other researchers to study seismic source rupture and crustal deformation.     

Crustal velocity structure in the southern edge of the Central Alborz (Iran)

Inversion of local earthquake travel times and joint inversion of receiver functions and Rayleigh wave group velocity measurements were used to derive a simple model for the velocity crustal structure beneath the southern edge of the Central Alborz (Iran), including the seismically active area around the megacity of Tehran. The P and S travel times from 115 well-located earthquakes recorded by a dense local seismic network, operated from June to November 2006, were inverted to determine a 1D velocity model of the upper crust. The limited range of earthquake depths (between 2 km and 26 km) prevents us determining any velocity interfaces deeper than 25 km. The velocity of the lower crust and the depth of the Moho were found by joint inversion of receiver functions and Rayleigh wave group velocity data. The resulting P-wave velocity model comprises an upper crust with 3 km and 4 km thick sedimentary layers with P wave velocities (V_p) of ～5.4 and ～5.8 km s^?1, respectively, above 9 km and 8 km thick layers of upper crystalline crust (V_p ～6.1 and ～6.25 km s^?1 respectively). The lower crystalline crust is ～34 km thick (V_p ～ 6.40 km s^?1). The total crustal thickness beneath this part of the Central Alborz is 58 ± 2 km.     

Elasticity of antigorite,seismic detection of serpentinites,and anisotropy in subduction zones

Serpentinization of the mantle wedge is an important process that influences the seismic and mechanical properties in subduction zones. Seismic detection of serpentines relies on the knowledge of elastic properties of serpentinites, which thus far has not been possible in the absence of single-crystal elastic properties of antigorite. The elastic constants of antigorite, the dominant serpentine at high-pressure in subduction zones, were measured using Brillouin spectroscopy under ambient conditions. In addition, antigorite lattice preferred orientations (LPO) were determined using an electron back-scattering diffraction (EBSD) technique. Isotropic aggregate velocities are significantly lower than those of peridotites to allow seismic detection of serpentinites from tomography. The isotropic V_P/V_S ratio is 1.76 in the Voigt–Reuss–Hill average, not very different from that of 1.73 in peridotite, but may vary between 1.70 and 1.86 between the Voigt and Reuss bonds. Antigorite and deformed serpentinites have a very high seismic anisotropy and remarkably low velocities along particular directions. V_P varies between 8.9 km s^? 1 and 5.6 km s^? 1 (46% anisotropy), and 8.3 km s^? 1 and 5.8 km s^? 1 (37%), and V_S between 5.1 km s^? 1 and 2.5 km s^? 1 (66%), and 4.7 km s^? 1 and 2.9 km s^? 1 (50%) for the single-crystal and aggregate, respectively. The V_P/V_S ratio and shear wave splitting also vary with orientation between 1.2 and 3.4, and 1.3 and 2.8 for the single-crystal and aggregate, respectively. Thus deformed serpentinites can present seismic velocities similar to peridotites for wave propagation parallel to the foliation or lower than crustal rocks for wave propagation perpendicular to the foliation. These properties can be used to detect serpentinite, quantify the amount of serpentinization, and to discuss relationships between seismic anisotropy and deformation in the mantle wedge. Regions of high V_P/V_S ratios and extremely low velocities in the mantle wedge of subduction zones (down to about 6 and 3 km.s^?1 for V_P and V_S, respectively) are difficult to explain without strong preferred orientation of serpentine. Local variations of anisotropy may result from kilometer-scale folding of serpentinites. Shear wave splittings up to 1–1.5 s can be explained with moderately thick (10–20 km) serpentinite bodies.