Lateral Velocity Heterogeneity Research Articles

Regional earthquake location using empirical traveltimes in a region of strong lateral velocity heterogeneity

SUMMARY The unmodelled effects of lateral heterogeneity are a primary cause of earthquake mislocation using routine methods. Approaches that account for the traveltime effects of lateral heterogeneity are a common feature of specialized studies, but some earthquake location methods have been developed that are also suitable for routine location. Here we investigate the use of one such method (3-D empirical traveltimes, ETTs) for local and regional earthquake location in an area of pronounced lateral velocity heterogeneity. We focus on New Zealand and the surrounding region due to the presence of strongly heterogeneous subduction zone environments. The New Zealand example poses significant challenges since most of the seismicity occurs outside the national seismograph network, and many of the observed ray paths traverse the high-velocity subducting slab. We show that in this environment, >60 per cent of the P and S residuals observed when a 1-D model velocity model is used stem from the resolvable lateral heterogeneity signal. For stations that have been in operation for more than five years, 85 per cent of observed residuals can be ascribed to this resolvable signal. We produce consistent 3-D traveltime corrections for all of New Zealand and the surrounding region and demonstrate their application to clusters of earthquakes occurring off the east coast of the North Island.

Improved GPR interpretation through resolution of lateral velocity heterogeneity: Example from an archaeological site investigation

In a typical common-offset ground-penetrating radar (GPR) survey, lateral velocity contrasts may go undetected leading to misinterpretation. Resolution of lateral velocity heterogeneity requires multi-fold acquisition and analysis. Further, pre-stack depth migration (PSDM) is required to produce accurate images in the presence of large lateral velocity gradients. In an archaeological investigation conducted near Boise, Idaho, we delineated a portion of what we interpret to be an abandoned dump site. Using multi-fold acquisition with reflection tomography, we identified an abrupt lateral velocity increase of ∼ 40% resulting in a substantial velocity pull-up in the time domain. PSDM corrected for the velocity pull-up enabling a more accurate interpretation and identification of additional structures of potential historical significance. The migration velocity model provided additional constraints on materials which enhanced our understanding of the subsurface.

Crustal‐scale prestack depth imaging of the 1994 and 1999 LARSE surveys

ABSTRACTWhile seismic imaging for crustal and mantle structures has traditionally relied on surface wave and refraction data, the use of reflection data for crustal‐scale targets has been largely limited to the common midpoint (CMP) stack techniques. The rapid increase in the number of seismograph array deployments in recent years in crustal and mantle seismology has reached a level such that a re‐examination of the imaging techniques is becoming necessary. In this paper we show the advantage of prestack depth imaging for crustal reflection studies, based on data from two reflection surveys of the Los Angeles Regional Seismic Experiment (LARSE) to map faults and crustal‐scale structures. Our analysis indicates that the quality of the previous images of these surveys is limited by the CMP stack technique. For comparison, we present here depth images of the same LARSE data using wave equation prestack depth imaging and a tomographic velocity model based on first arrivals of the LARSE surveys and local earthquakes. Our new images are considerably improved over previous images in terms of resolution and reflector continuity. The new images show reflectors throughout the crust and suggest truncations in the Moho associated with the San Andreas Fault. A series of bright reflector segments, which are associated with the San Gabriel and San Andreas faults have been identified and might represent reflections from the fault zones. Our results suggest that the presence of high noise level, strong lateral velocity heterogeneity and wide angle geometry argue for, rather than against, the use of prestack depth imaging over the simple CMP stack techniques. As demonstrated in this study, it is now viable to conduct prestack depth imaging of crustal reflection data using a velocity model based on earthquake first arrivals thanks to the dense acquisition deployment.

Depth-calibrating seismic data in the presence of allochthonous salt

Seismic depth errors can be significant in PSDM-processed seismic data. The purpose of this paper is to discuss best practice techniques for depth-calibrating seismic volumes, particularly in the presence of allochthonous salt where lateral and vertical velocity heterogeneity is great. We will present the concepts of apparent seismic depth and true seismic depth, propose anisotropy as a first-order mechanism for seismic depth error, discuss the implications of depth error as it pertains to well planning and resource estimation, and present different corrective measures and discuss their implementation. Our paper will be supported by two case studies, one from the deepwater Gulf of Mexico and another from offshore West Africa.

Waveform tomography at a groundwater contamination site: Surface reflection data

We have applied acoustic-waveform tomography to 45 2D seismic profiles to image the 3D geometry of a buried paleochannel at a groundwater-contamination site at Hill Air Force Base in Utah. The paleochannel, which is incised into an alluvium-covered clay aquitard, acts as a trap for dense nonaqueous-phase liquids (DNAPLs) that contaminate the shallowest groundwater system in the study area. The 2D profiles were extracted from a 3D surface reflection data set. First-arrival traveltime tomography provided initial velocity models for the waveform tomography. We inverted for six frequency components in the band [Formula: see text] of the direct and refracted waves to produce 45 2D velocity models. The flanks and bottom of a channel with a maximum depth of about [Formula: see text] were well modeled in most of the 45 parallel 2D slices, which allowed us to construct a 3D image of the channel by combining and interpolating between the 45 image slices. The 3D model of the channel will be useful for siting extraction wells within the site remediation program. The alluvium that fills the channel showed marked vertical and lateral velocity heterogeneity. Traveltime tomography and waveform tomography can be complementary approaches. Used together, they can provide high-resolution images of complicated shallow structures.

Constraints on global maps of phase velocity from surface-wave amplitudes

We demonstrate that global maps of phase velocity can be obtained from a data set of fundamental-mode Rayleigh wave amplitudes in the period range 150–250 s. Although the maps are constructed without any information about the wave phase, they are highly correlated with phase-velocity maps derived entirely from measurements of phase delay. We consider there to be four factors that contribute to the observed amplitude anomalies: focusing by lateral velocity heterogeneity along the ray path, variable attenuation along the ray path, uncertainty in the strength of excitation and uncertainty in the response at the station. In our first analysis, we desensitize the data to attenuation, source uncertainty and receiver uncertainty by combining the amplitudes of four consecutive surface-wave arrivals. The resulting quantity is inverted for even-degree spherical-harmonic maps of phase velocity, which explain 50 per cent of the variance in an independent data set of great-circle phase-delay measurements. In our second analysis, the amplitude measurements are inverted simultaneously for global maps of phase velocity expanded to degree 20, global degree-12 maps of attenuation and source and receiver correction factors. The velocity maps obtained from the simultaneous inversion exhibit strong agreement with published phase-velocity maps. The success of our analysis, which is based on linearized ray theory, suggests that the underlying assumptions about the smoothness of heterogeneity in the mantle may be reasonable, and illustrates the utility of amplitude data for constraining elastic heterogeneity in the Earth.

Offshore E&P - Still More To Learn

Offshore, shallow or deep, is a realm unto itself in which technology and economics are entwined and constantly evolving. The technical agenda for the 2006 Offshore Technology Conference (OTC) May 1–4 in Houston promises to deliver the next chapter in offshore E&P. Approximately 40 technical sessions and more than 300 technical presentations are planned. Seven papers chosen by OTC 2006 program organizers emphasize that the one sure thing is that the pressure is on the industry to maintain a healthy balance between high risk, steep cost, and reasonable reward. The technical papers discussed in this article are listed below. Recent Findings in Anisotropy Estimation The authors of paper OTC 17866 estimate parameters of depth-variant vertical transverse isotropy (VTI) from the first-arrival times recorded in a 3D vertical seismic profile (VSP) survey over a deepwater Gulf of Mexico (GOM) project. They use both conventional slowness inversion and a recently developed travel-time extrapolation technique. The authors are not aware of any depth-dependent estimates of anisotropy from VSP data done before their study. Their work confirms that 3D VSP acquisition geometry is ideally suited for estimating in-situ seismic anisotropy; however, the locality of anisotropy measurements is achieved only when local quantities are used. This is not the case when the horizontal slowness components are involved in the inversion process because they are influenced by lateral velocity variations in the entire overburden. "Correcting the horizontal slownesses for lateral velocity heterogeneity requires knowing the seismic velocity model with accuracy hardly attainable in practice," the paper states. A VSP survey was shot in a spiral pattern, with waveforms recorded by 28 geophones placed in 100-ft increments at depths ranging from 17,528 to 20,228 ft. The first-break times were picked from the data. The authors apply both the conventional slowness and novel travel-time extrapolation techniques and obtain comparable estimates of anisotropy. The travel-time extrapolation method, however, better illustrates the accuracy of knowing the lateral velocity variations needed to achieve the locality of anisotropy estimates. "One subsurface feature—a salt body—profoundly influences our ability to estimate anisotropy from the VSP data," say the authors. "Its presence causes a decrease of the first-break times clearly seen in the northern portion of the survey. The salt outline is easily identified in the travel-time data (Fig. 1)." This case study reveals the problem with deriving anisotropy from the slowness surfaces; the presence of salt produces slowness distortions that are impossible to fit with any anisotropic model, forcing the authors to rely only on the data recorded outside the salt. Realistic constraints of parameters are discussed. "Our most interesting finding was the negative sign of the an ellipticity coefficient η," the authors say. "It remains to be understood whether the negative sign of η bears lithologic or any other significance."

Applying reflection tomography in the postmigration domain to multifold ground-penetrating radar data

Acquisition and processing of multifold ground-penetrating radar (GPR) data enable detailed measurements of lateral velocity variability. The velocities constrain interpretation of subsurface materials and lead to significant improvement in image accuracy when coupled with prestack depth migration (PSDM). Reflection tomography in the postmigration domain was introduced in the early 1990s for velocity estimation in seismic reflection. This robust, accurate method is directly applicable in multifold GPR imaging. At a contaminated waste facility within the U.S. Department of Energy's Hanford site in Washington, the method is used to identify significant lateral and vertical velocity heterogeneity associated with infilled waste pits. Using both the PSDM images and velocity models in interpretation, a paleochannel system that underlies the site and likely forms contaminant migration pathways is identified.

Site Amplification, Scattering, and Intrinsic Attenuation in the Mississippi Embayment from Coda Waves

We estimate site amplification, intrinsic Q ( Q I ) and scattering Q ( Q S ) values from three-component coda of broadband data recorded by Cooperative New Madrid Sesimic Network stations for center frequencies 0.2, 0.5, 1.0, 2.0, and 5.0 Hz. Near 0.5 and 1.0 Hz, site amplification factors for the transverse component recorded by stations in the Mississippi embayment are between 1.7 and 18, compared with stations outside the embayment. The amplification factor becomes smaller at frequencies lower than 0.5 Hz and higher at those greater than 1.0 Hz. The radial component shows a similar amplification; however, the vertical component shows relatively less amplification. Q I and Q S are estimated from coda using an energy flux model for two time windows after the direct S wave and at 0.5 and 1.0 Hz center frequencies. The two time windows include coda up to and after twice the S -wave travel time at each station. Coda parameters change with the time window and are attributed to a mixture of incoherent coda with coherent S-wave reverberations directly after the S wave. Q I and Q S inferred from stations outside of the embayment have higher values than those from stations within the embayment. A plane wave model qualitatively explains the lack of coda amplification at high frequencies compared with a low frequency without recourse to anelastic attenuation arguments. However, the plane wave model cannot explain the very large observed coda amplifications at 1 and 0.5 Hz seen in the data, suggesting that other wave-scattering effects must be occurring in the embayment velocity structure. This is consistent with the relatively low Q S values inferred from the energy flux model for stations within the embayment. In addition to coda amplification, we performed a synoptic noise study that revealed that pre-event ambient background noise behaves in a similar manner to the coda for stations within and outside of the embayment. The presence of large lateral velocity heterogeneities in the embayment may cause Q S values to be low there.

Time-domain Pure-state Polarization Analysis of Surface Waves Traversing California

— A time-domain pure-state polarization analysis method is used to characterize surface waves traversing California parallel to the plate boundary. The method is applied to data recorded at four broadband stations in California from twenty-six large, shallow earthquakes which occurred since 1988, yielding polarization parameters such as the ellipticity, Euler angles, instantaneous periods, and wave incident azimuths. The earthquakes are located along the circum-Pacific margin and the ray paths cluster into two groups, with great-circle paths connecting stations MHC and PAS or CMB and GSC. The first path (MHC-PAS) is in the vicinity of the San Andreas Fault System (SAFS), and the second (CMB-GSC) traverses the Sierra Nevada Batholith parallel to and east of the SAFS. Both Rayleigh and Love wave data show refractions due to lateral velocity heterogeneities under the path, indicating that accurate phase velocity and attenuation analysis requires array measurements. T he Rayleigh waves are strongly affected by low velocity anomalies beneath Central California, with ray paths bending eastward as waves travel toward the south, while Love waves are less affected, providing observables to constrain the depth extent of anomalies. Strong lateral gradients in the lithospheric structure between the continent and the ocean are the likely cause of the path deflections.

Elasticity of pyrope and majorite–pyrope solid solutions to high temperatures

Majorite–garnet solid solutions are major mineral phases in the Earth’s upper mantle and transition zone. Here we present the first Brillouin scattering measurements of the elasticity of majorite (Mj, Mg4Si4O12)–pyrope (Py, Mg3Al2Si3O12) solid solutions (Mj50Py50 and Mj80Py20) and single-crystal elasticity of pure synthetic pyrope at temperatures up to 800°C. The temperature derivatives of the adiabatic bulk (KS) and shear (μ) moduli for all compositions along the Mj–Py join are the same within the experimental uncertainties (−∂KS/∂T=14.0–14.5(20) MPa/K, −∂μ/∂T=8.3–9.2(10) MPa/K). The temperature dependence of the acoustic velocities for Mj–Py solid solutions is about half that of other major transition zone minerals. This implies that temperature variations in the transition zone, inferred from lateral velocity heterogeneity, can be significantly underestimated if the properties of majoritic garnet are not taken into account.

Estimate of inner core rotation rate from United Kingdom regional seismic network data and consequences for inner core dynamical behaviour

We analyse over 12 500 records of southwest Pacific earthquakes recorded by the UK network for PKP BC and PKP DF phases to seek constraints on the inner core’s rotation rate. Careful analysis and strict rejection criteria yield 655 PKP BC –PKP DF differential travel time residuals computed with respect to the AK135 reference model, densely sampling a small region of the inner core over a time period of 15 years. The data images both lateral and radial velocity heterogeneity in the inner core located geographically beneath the north Pacific between radii of 870 and 1080 km. The dominant feature is a longitudinal wave speed gradient centred at about 180° longitude. In combination with different earthquake catalogues and inner core anisotropy models, we explore different techniques to use this gradient to constrain the rotation rate. Our estimates range from 0.45±0.25 to 0.74±0.29°/yr relative to the mantle, with two of four estimates including zero rotation at the 95% (2σ) confidence level, indicating that rotation is marginally detectable. As an alternative to monotonic rotation, we find weak evidence for rotational oscillation of the inner core on time scales of about 280 days. In consequence, these observations suggest that: (1) viscous coupling between the outer and inner core simulated using viscous hyperdiffusivity is not appropriate for modelling inner core rotation; (2) temperature gradients between the interior and exterior of the tangent cylinder in the core are also small; (3) gravitational torques coupling the inner core to the mantle are large, on the order of 2×10 21 Nm; and (4) the inner core’s viscosity is high, 3.9×10 19 Pas.

Slowness Anomalies from Two Dense Seismic Arrays at Deception Island Volcano, Antarctica

In this article, we analyze the data collected by two short-period seismic arrays deployed at Deception volcano, Southern Shetland Islands, Antarctica. The field survey was conducted during the 1998-1999 austral summer and was aimed at a quantitative assessment of the complex wave fields associated with the magmatic and hydrothermal activity of the volcano. The two arrays had apertures of 320 m and 240 m and were separated by a distance of about 3 km. During the experiment, the arrays recorded several regional earthquakes related to the dynamics of the Brans- field Strait and adjoining areas and local volcano-tectonic earthquakes. Seismograms of earthquakes recorded at regional distances reveal a marked difference in the ap- parent velocities measured at the two array sites. We investigate the causes and implications of these anomalies by first comparing the effectiveness of estimating the horizontal slowness vector using three different techniques: the multiple signal classification (MUSIC) approach, the zero-lag cross correlation (ZLC) method, and plane-wave fitting to P-wave arrival times. While each technique provides the same horizontal slowness vector as the most likely estimates, the plane-wave fitting is associated with the most robust definition of measurement uncertainties. We then investigate the dispersive properties of Rayleigh waves in the 1-8 Hz frequency band at both arrays and invert the two dispersion curves for a shallow velocity structure. The results indicate a marked difference in the seismic velocities for the shallower 200 m beneath the two sites. This may be reconciled with the observed wave vector anomalies by assuming the existence of a sharp lateral velocity heterogeneity, the effect of which would be to bend downward rays impinging at the northernmost array. The reliability of this hypothesis is verified by computing finite-difference wave fronts in a 2D heterogeneous medium. Based on the morpho-structural char- acteristics of the volcano, the inferred velocity discontinuity maybe associated with the ring-fracture system bordering the collapsed caldera structure that extends over the inner part of the island.

Small‐scale lateral shear velocity and anisotropy heterogeneity near the core‐mantle boundary beneath the central Pacific imaged using broadband ScS waves

Core‐reflected ScS waves from 49 large (M≥5.1) deep Tonga‐Fiji events, recorded in western North America are used to study a localized region of the core‐mantle boundary (CMB) under the central Pacific Ocean. A total of 248 observations from the Berkeley Digital Seismic Network (BDSN), Caltech/United States Geological Survey (USGS) TERRAscope and the Incorporated Research Institutions for Seismology (IRIS) broadband arrays span epicentral distances of 73°–85°. ScS reflection points sample a CMB patch southeast of the Hawaiian islands at latitude 4° to 16° N and longitude −156° to −144° W, near the proposed source location of the Hawaiian plume. Highly variable ScS travel times, amplitudes, waveforms, and shearwave splitting indicate that the lowermost mantle in this region is heterogeneous both laterally and radially. ScSH‐SH differential travel times are on average 4 s larger than predicted by the Preliminary Reference Earth Model (PREM). This is due to delayed ScS arrivals and is largely accounted for by a model with a strong velocity decrease in D″. The ScSH‐SH residuals also show a spatial trend not accounted for by a radial model, indicating a lateral decrease in lower mantle shear velocity to the northeast. This lateral velocity gradient appears to cause focusing of ScS energy. ScSH/SH amplitude ratios are larger than predicted by PREM or by a model with a strong negative gradient with depth in D″, which enhances ScS. ScS splitting, corrected for lithospheric anisotropy beneath the receivers, indicates spatial variations in D″ anisotropy with a systematic change in the orientation of the fast polarization direction from transverse to the ray path (fast ScSH) to parallel to the ray path (fast ScSV) along a northeast traverse. These spatial trends suggest lateral gradients in the boundary layer shear flow on scale lengths of a few hundred kilometers, which may be related to dynamical flow near the near the root of the large Hawaiian plume.

Mantle discontinuity structure beneath the Colorado Rocky Mountains and High Plains

Analysis of mantle discontinuity structure using converted P to S (PaS) phases beneath Colorado from the Program for Array Seismic Studies of the Continental Lithosphere (PASSCAL) Rocky Mountain Front (RMF) experiment reveals significant topography at the 410 and 660 km depth discontinuities and corresponding transition zone thickness variations. A stack of all radial receiver functions resolves the 410 and 660 km discontinuities at average depths of 419 and 677 km, respectively. Imaging of lateral variations in mantle discontinuity structure is accomplished by geographically binning the Pds conversion points and then stacking the receiver functions in each bin to form spatial images, analogous to common depth point stacking. Corrections for lateral velocity heterogeneity are calculated using the local S wave tomographic model of Lee and Grand [1996] and a constant ∂lnVs/∂lnVP scaling of 1.3. This scaling value is determined from the relative scaling between teleseismic P and S wave travel time residuals measured from the Rocky Mountain Front deployment. Mantle discontinuity images using 150 km square bins show 20 km of 410 km discontinuity topography, 30 km of 660 km discontinuity topography, and up to 40 km of transition zone thickness variation. Features of the discontinuity structure include a 20 km depression of the 660 km discontinuity beneath western Colorado and a gradual 10 km dip of the 410 km discontinuity beneath the High Plains. The thickening of the transition zone beneath southwest Colorado is consistent with the presence of the subducted Farallon slab in this region as imaged by Van der Lee and Nolet [1997]. In general, our results show that the transition zone discontinuity structure is more complex than that predicted by the simple model of olivine phase boundaries modulated by vertically coherent thermal anomalies.

Analysis of reflection traveltimes in 3-D transversely isotropic heterogeneous media

A two‐point ray‐tracing technique for rays reflected from irregular, but smooth, interfaces in 3-D transversely isotropic heterogeneous media is developed. The method is based on Chebyshev parameterization of curved segments of the reflected rays, of the reflectors, and of the velocity and anisotropy distributions in the model. Chebyshev approximation also can describe the reflection traveltime surfaces to compress traveltime data by replacing them with coefficients of the corresponding Chebyshev series. The advantage of the proposed parameterization is that it gives traveltime as an explicit function of the model parameters. This explicitly provides the Frechét derivatives of the traveltime with respect to the model parameters. The Frechét derivatives are used in two ways. First, a two‐term Taylor series is constructed to relate variations in the model parameters to the corresponding perturbations in the traveltimes. This makes it possible, based on the results of a single ray tracing in a relatively simple model, to predict traveltimes for a range of more complicated models, without any additional ray tracing. Second, singular‐value decomposition of the Frechét matrix determines the influence of various model parameters on common‐source and common‐midpoint traveltimes. The singular‐value analysis shows that common‐source traveltimes depend mainly on the reflector position and shape. The common‐midpoint traveltimes also contain additional information about lateral velocity heterogeneity and anisotropy. However, both of these parameters affect the traveltimes in similar ways and so usually cannot be determined separately.

Crust and upper mantle shear velocity structure beneath the Tibetan plateau and surrounding regions from interevent surface wave phase velocity inversion

Average a priori shear velocity models are constructed for the Tien Shan, Tarim basin, Pamir‐Hindu Kush, Himalaya and NE India. These models are shown to account for most of the lateral velocity heterogeneity at wavelengths >1000 km. Interevent fundamental mode Rayleigh and Love wave phase velocities were measured in the period range 32–200 s and were inverted for path‐averaged shear velocity structures. These, in turn, were regionalized to estimate average, isotropic structures beneath the six regions listed above. Low upper mantle velocities are observed beneath the western and eastern Tien Shan. The Tarim basin structure is poorly constrained but exhibits very low upper crustal velocities (probably due to deep sedimentary accumulation), and either very high upper mantle velocities exist (>4.8 km/s) in a layer 70 km thick or the lithosphere is very thick (∼180 km). Low upper mantle velocities are observed beneath the Chang Thang region, the central, and the northeast plateau but not beneath the southern or southeastern plateau. The crustal velocity gradient with depth is zero in the east central plateau consistent with lower crustal basaltic intrusions there but is positive to the west and southeast of this zone. Upper mantle structures show that the Indian lithosphere could have subducted beneath the entire plateau only if either 1 mantle in the depth range 115–185 km has been altered (or removed) to reduce its average velocity or 2 the Indian lithosphere is only ∼85 km thick.

High‐resolution crosswell seismic experiment with a large interwell spacing in a west Texas carbonate field

A high‐frequency crosswell seismic dataset acquired in a west Texas carbonate field has demonstrated the feasibility of the technique with a large interwell spacing. Two crosswell profiles were acquired with a well spacing of 1500 ft (460 m) and over a depth interval from 7700 ft (2350 m) to 9600 ft (2930 m) using a piezoelectric bender source. The data quality is profile and depth dependent, with the ambient noise level at the receiver position being the most important factor. Noise levels and noise characteristics among three wells were significantly different. Tube waves and gas‐ and fluid‐movement in the borehole are the dominant noise sources found in the data set. Two lithologic properties, attenuation and transmission loss controlled the data quality. Good quality and high frequency (>1000 Hz) data were acquired over most of the survey interval which contains massive limestones. However, we could not acquire any useful data within the shale layers. Transmission losses and the effects of the source radiation pattern that occurred at interfaces with large impedance contrasts limited the aperture of the useful data. There were two critical issues encountered during the reflection imaging process: (1) sparse trace spacing and poor coherency of the reflection events in common‐source gathers degraded the image in a region near the receiver well; and (2) possible lateral velocity heterogeneity in the medium and limited aperture made it difficult to build an appropriate velocity model for reflection imaging.

Crustal tomography of the central kenya rift

During the Kenya Rift International Seismic Project 1990 (KRISP90) experiment the shots of the seismic refraction programme were recorded by a two-dimensional teleseismic network. This was installed in the central part of the Kenya Rift and on its shoulders, from Lake Baringo in the north to Lake Naivasha in the south P-wave first arrivals were used to determine lateral crustal velocity heterogeneities with a three-dimensional inversion programme. Mainly high-velocity zones were found which can be correlated with other geophysical anomalies and geological features. This comparison shows that the high-velocity bodies found within the crust are likely indicators for mafic cumulates or magma chambers. These evolved during the major magmatic episodes which are typical indicators for an active rift.

The preliminary interpretation of deep seismic sounding in western Yunnan

The preliminary interpretation of Project western Yunnan 86–87 is presented here. It shows that there obviously exists lateral velocity heterogeneity from south to north in western Yunnan. The depth of Moho increases from 38 km in the southern end of the profile to 58 km in its northern end. The mean crustal velocity is low in the south, and high in the north, about 6.17–6.45 km/s. The consolidated crust is a 3-layer structure respectively, the upper, middle and lower layer. P10 is a weak interface the upper crust, P20 and P30 are the interfaces of middle-upper crust and middle-lower crust respectively. Another weak interface P30′ can be locally traced in the interior of the lower crust. Interface Pg is 0–6 km deep, interface P109.2–16.5 km deep, and interfaces P20 and P30 respectively 17.0–26.5 km, 25.0–38.0 km deep. The velocity of the upper crust gradually increases from the south to the north, and reaches its maxmium between Nangaozhai and Zhiti, where the velocity of basement plane reaches 6.25–6.35 km/s, then it becomes small northward. The velocity of the middle crust varies little, the middle crust is a low velocity layer with the velocity of 6.30 km/s from Jinhe-Erhai fault to the north. The lower crust is a strong gradient layer. There exists respectively a low velocity layer in the upper mantle between Jinggu and Jingyunqiao, and between Wuliangshan and Lancangjiang fault, the velocity of Pn is only 7.70–7.80 km/s, it is also low to the north of Honghe fault, about 7.80 km/s. Interface P6/0 can be traced on the top of the upper mantle, its depth is 65 km in the southern end of the profile, and 85 km in the northern end.