Lateral variation in upper mantle viscosity: role of water

Differences in the viscosity of the earth's upper mantle beneath the western US (∼10¹⁸-10¹⁹ Pa s) and global average values based on glacial isostatic adjustment and other data (∼10²⁰-10²¹ Pa s) are generally ascribed to differences in temperature. We compile geochemical data on the water contents of western US lavas and mantle xenoliths, compare these data to water solubility in olivine, and calculate the corresponding effective viscosity of olivine, the major constituent of the upper mantle, using a power law creep rheological model. These data and calculations suggest that the low viscosities of the western US upper mantle reflect the combined effect of high water concentration and elevated temperature. The high water content of the western US upper mantle may reflect the long history of Farallon plate subduction, including flat slab subduction, which effectively advected water as far inland as the Colorado Plateau, hydrating and weakening the upper mantle.     

Influence of low viscosity asthenosphere on mantle flows

Numerical simulation in recent years has revealed that the cold lithosphere, whose viscosity is three to four orders of magnitude higher than that of the underlying mantle, behaves during mantle convection as a stagnant lid. On the basis of model calculations, this paper shows how convection changes over to this regime with increasing viscosity. Spatially fixed high viscosity inclusions and those moving with the convective flow have fundamentally different effects on the structure of convective flows. The model calculations indicate that anomalously low viscosity asthenospheric regions also lead to a specific regime of convection. With a decrease in the viscosity by more than three orders of magnitude, a further reduction in the viscosity of the region ceases to influence the structure of convection in the outer region. The boundary with this region behaves as a freely permeable boundary. In the low viscosity asthenospheric region itself, autonomous convection can arise under certain conditions.     

Geothermal gradient of the upper mantle beneath Jeju Island, Korea: Evidence from mantle xenoliths

Abstract Ultramafic xenoliths found in alkali basalts from Jeju Island, Korea are mostly spinel lherzolites accompanied by subordinate amount of spinel harzburgites and pyroxenites. The combination of results from a two-pyroxene geothermometer and Ca-in-olivine geobarometer yields temperature–pressure (T–P) estimates for spinel peridotites that fall in experimentally determined spinel lherzolite field in CaO-Fe-MgO-Al₂O₃-SiO₂-Cr₂O₃ (CFMASCr) system. These T–P data sets have been used to construct the Quaternary Jeju Island geotherm, which defines a locus from about 13 kbar at 880°C to 26 kbar at 1040°C. The geothermal gradient of Jeju Island is greater than that of the conventional conductive models, and may be as a result of a thermal perturbation by the heat input into the lithospheric mantle via the passage and emplacement of magma. Spinel–lherzolite is the main constituent rock-type of the lithospheric mantle beneath Jeju Island. Pyroxenites may be intercalated in peridotites at similar depth and temperature as re-equilibrated veins or lenses.     

A model for the layered upper mantle

Numerical modeling of mantle convection by Liu (1994, Science, 264: 1904–1907) favors a two-layer convection, if the results are reinterpreted for the correct phase relations in (Mg,Fe)₂SiO₄. The resulting chemical isolation of the upper and lower mantle suggests a highly differentiated and layered upper mantle to account for the discrepancy between the observed compositions of mantle xenoliths and the cosmic abundances of elements. It is shown that a layered upper mantle with a hidden reservoir can have a structure consistent with the observed seismic velocity profiles and an average bulk composition corresponding to the cosmic abundances. The evolution of the upper mantle and the origin of komatiites are discussed in the context of the proposed model.     

Evidence of electrical anisotropic structures in the lower crust and the upper mantle beneath the Rhenish Shield

We present the results of a long-period electromagnetic investigation of the crust and upper mantle beneath the Rhenish Shield in Western Germany. The magnetotelluric phase data reveal a frequency-dependent regional strike that varies only smoothly across the array. At short periods (t<100 s) the striking of the maximum phase splitting is N45°E, which can be explained with an electrical anisotropic lower crust. At long periods (t>1000 s) there is a consistent striking in WE direction, which provides strong evidence of an anisotropic structure in the upper mantle, too. Geomagnetic data were also used, whereas we reference the magnetic field components of all sites to an arbitrarily chosen field site. That provides a direct view of the anomalous current flow. We show that in case of a non-one-dimensional electrical substructure of the reference site all magnetic transfer functions of the other sites can be affected by lateral conductivity contrasts beneath the reference site. A simple method to remove such effects is introduced. Applying this method the magnetic data show a distinct anomaly in the northwestern part of the area. Finally, we present a 3D model of the conductivity structure beneath the Rhenish Shield which contains a superposition of two anisotropic structures in the lower crust (2000 S) and the upper mantle (20?000 S), respectively, and a local anomaly in the upper crust (4000 S).     

Anisotropic models of the upper mantle

Long period Rayleigh wave and Love wave dispersion data, particularly for oceanic areas, have not been simultaneously satisfied by an isotropic structure. In this paper available phase and group velocity data are inverted by a procedure which includes the effects of transverse anisotropy, anelastic dispersion, sphericity, and gravity. We assume that the surface wave data represents an azimuthal average of actual velocities. Thus, we can treat the mantle as transversely isotropic. The resulting models for average Earth, average ocean, and oceanic regions divided according to the age of the ocean floor, are quite different from previous results which ignore the above effects. The models show a low-velocity zone with age dependent anisotropy and velocities higher than derived in previous surface wave studies. The correspondence between the anisotropy variation with age and a physical model based on flow aligned olivine is suggestive. For most of the Earth SH > SV in the vicinity of the low-velocity zone. Neat the East Pacific Rise, however, SV > SH at depth, consistent with ascending flow. Anisotropy is as important as temperature in causing radial and lateral variations in velocity. The models have a high velocity nearly isotropic layer at the top of the mantle that thickens with age. This layer defines the LID, or seismic lithosphere. In the Pacific, the LID thickens with age to a maximum thickness of ～50 km. This thickness is comparable to the thickness of the elastic lithosphere. The LID thickness is thinner than derived using isotropic or pseudo-isotropic procedures. A new model for average Earth is obtained which includes a thin LID. This model extends the fit of a PREM, type model to shorter period surface waves.     

Shear localisation in upper mantle peridotites

Upper mantle peridotite bodies at the earth's surface contain relict structures and microstructures which provide direct information on the role and the mechanisms of shear localisation in the upper mantle. Deformation which occurred at high temperatures (T>950±50°C) is relatively homogeneous within domains ranging in scale from a few kilometres to a few tens of kilometres. Below 950±50°C strain is localised into centimetre to several hundred metre wide shear zones which commonly contain hydrated mylonitic peridotites. The microstructures developed in the peridotites suggest there is a correlation between the occurrence of shear localisation and the occurrence of strain softening and brittle deformation processes. The most important strain softening processes are inferred to be structural and reaction induced softening. Structural softening processes include dynamic recrystallisation and strain-induced transitions from dislocation creep to some form of grain-size-sensitive (GSS) creep. Reaction induced softening is related to the formation of fine grained polyphase reaction products which deform by GSS creep and the formation of weak sheet silicates such as phlogopite, chlorite, talc and antigorite. From experimental studies these softening processes and brittle deformation processes are inferred to occur mainly at temperatures less than about 910±160°C. This temperature range is inferred to be a significant rheological transition in the upper mantle. Below 910±160°C deformation during orogenesis may be accommodated by an anastomosing network of hydrated mylonitic shear zones with a distinct, perhaps weak, rheology. At higher temperatures strain is accommodated in much wider deformation zones.On the scale of the lithosphere the degree of localisation may be different to that determined at the scale of the periodotite massif. An anastomosing network of hundred metre wide mylonitic shear zones forming 0.05–0.3 by volume fraction of the mantle lithosphere atT<950°C could accommodate inhomogeneous or homogeneous bulk deformation depending on the spatial distribution and ordering of the mylonite zones. The higher temperature deformation at deeper levels in the mantle could be markedly inhomogeneous being concentrated in shear zones with widths in the range of 2–20 km, alternatively these zones may widen significantly during deformation, resulting in a decrease in the degree of localisation with increasing bulk strain.     

On stochastic near-critical oscillations in the upper mantle

The problem of the determination of thermomechanical conditions in the upper mantle under a moving lithospheric plate at a given shear stress does not have a unique solution. Given a fixed heat flux from the lower mantle, two types of motion are possible in the mantle: subcritical (slower and colder) and supercritical (faster and hotter). In this work, it is shown that, if these modes are rather close to each other (in the near-critical state of the mantle), transitions from one mode to another and backward are possible. The calculated period of a change in the mantle state amounts to ～0.3 Myr for the Pacific plate. The oscillatory regime of the mantle state can manifest itself in the activity pulsations of hotspots located near fast separating mid-ocean ridges.     

Models of electric conductivity in the upper mantle

Summary The accessible data of the apparent resistivity _T(T) were collected and supplemented by the spectral analysis of daily means of geomagnetic components observed at world geomagnetic observatories. The data were least-squares fitted; four models of the resistivity in the upper mantle which agree with the experimental data were obtained by solving the inverse problem using Marquardt's gradient-expansion algorithm.     

Fractionation of platinum group elements in upper mantle: Evidence from peridotite xenoliths from Wangqing

The peridotite xenoliths from Wangqing display a downward-concave PGE pattern with Pt at maximum due to Pt-Pd fractionation, which is different from the flat or negatively sloped patterns commonly observed for worldwide peridotite xenoliths and massifs. Taking into account the difference in partition coefficients of Pt and Pd between alloy and sulfide melt, it is suggested that some of the Wangqing peridotites must have been equilibrated with a melt in which alloy and sulfide coexisted. Segregation of Pt-rich alloy from Pd-sulfide melts due to density contrast and extraction of residual melts without sulfide saturation resulted in the fractionation of Pt from Pd. Project supported by the Ministry of Human Affairs and a part of an international collaboration program between the Chinese Academy of Sciences and Centre National de Recherche Scientifique (CNRS, France), and Australian Institute of Nuclear Sciences and Engineering (AINSE) grant to SJR.     

Space-time variations of the shear wave attenuation field in the upper mantle of seismic and low seismicity areas

This paper deals with characteristics of the short period S-wave attenuation field in the rupture zones of 37 large and great earthquakes with M _s = 7.0–8.6, as well as in low seismicity areas. We estimate the effective quality factor from Sn and Lg coda envelopes in two time intervals (Q ₁ and Q ₂). The quantity Q ₁ is a measure of shear wave attenuation in the uppermost mantle, at depths of down to approximately 200–250 km, while Q ₂ is relevant to deeper horizons of the upper mantle. We studied variations in the attenuation field in the rupture zone of the 1950 Assam earthquake. We examined the parameters Q ₁, Q ₂, and Q ₁/Q ₂ as functions of the time ΔT elapsed after a large earthquake. It is shown that the parameter Q ₂ in rupture zones is practically independent of ΔT, while the quantities Q ₁ and Q ₁/Q ₂ increase until ΔT ～ 20–25 years, especially rapidly for normal, normal-oblique, and strike-slip earthquake mechanisms. This analysis provides evidence that, as ΔT increases, so does the quality factor in the upper mantle for shear waves. It is supposed that this is related to the rise of mantle fluids to the crust. Geodynamic mechanisms are discussed that can support a comparatively rapid “drying” of the upper mantle beneath earthquake rupture zones.     

Pa andSa waves and the upper mantle

Summary This investigation is based on records of 96 earthquakes withPa andSa written by the Press-Ewing instruments at Uppsala in the interval June 1961–December 1962.Pa andSa waves are observed for all earthquake regions, irrespective of distance, focal depth or path properties. They have significantly higher velocities under continents than under oceans, which demonstrates corresponding differences in the upper mantle. Continental velocities are 8.35 km/sec (Pa) and 4.56 km/sec (Sa), oceanic velocities 8.01 km/sec (Pa) and 4.45 km/sec (Sa). The most frequent periods are 10 sec (Pa) and 20 sec (Sa). They are independent of distance forPa but increase with distance forSa. The best developedPa andSa are obtained for earthquakes at focal depths less than about 60 km. The particle motion ofSa may be anything from pureSV to pureSH motion and has high correlation to the particle motion ofS. The apparent angles of emergence (in average 51° forPa and 54° forSa) vary with distance. On the basis of our observations it is suggested thatPa andSa propagate by multiple reflections under grazing incidence under the Moho discontinuity.

---

Zusammenfassung Die vorliegende Untersuchung gründet sich auf die Registrierungen von 96 Erdbeben mitPa- undSa-Wellen, die mit Hilfe der Press-Ewing Instrumente zu Uppsala im Zeitraum Juli 1961–Dezember 1962 aufgenommen wurden. DiePa- undSa-Wellen sind für alle seismischen Regionen beobachtet worden, unabhängig von Entfernung, Herdtiefe oder Wellenweg. Sie haben bedeutend höhere Geschwindigkeiten unter den Kontinenten als unter den Ozeanen, was einen entsprechenden Unterschied im oberen Erdmantel beweist. Die kontinentalen Geschwindigkeiten betragen 8.35 km/sec (Pa) und 4.56 km/sec (Sa), die ozeanischen Geschwindigkeiten 8.01 km/sec (Pa) und 4.45 km/sec (Sa). Die am häufigsten vorkommenden Perioden betragen 10 sec (Pa) und 20 sec (Sa). Sie sind unabhängig von der Entfernung fürPa aber wachsen mit der Entfernung fürSa. Die am besten entwickeltenPa- undSa-Wellen werden für Erdbeben mit kleinerer Herdtiefe als rund 60 km beobachtet. FürSa wird jede beliebige Orbitalbewegung zwischenSV undSH beobachtet. Sie hat eine hohe Korrelation mit der Orbitalbewegung vonS. Die scheinbaren Emergenzwinkel (durchschnittlich 51° fürPa und 54° fürSa) variieren mit der Entfernung. Auf Grund unserer Beobachtungen wird die Hypothese aufgestellt, dass sich diePa-undSa-Wellen durch Mehrfachreflexionen, bei tangentialem Einfall, unter der Moho-Diskontinuität ausbreiten.

---

Resumen La presente investigación está basada en los sismogramas de 96 terremotos, que muestranPa ySa, registrados por los sismógrafos de Uppsala (Press-Ewing), durante el periodo Junio 1961 a Diciembre 1962. Las fasesPa ySa se observan en todas las regiones sismicas, independientemente de la distancia, de la profundidad focal ó de las propiedades de la trayectoria. Tienen velocidades significativamente mayores bajo continentes que bajo los océanos, lo que demuestra la existencia de diferencias en el manto superior. Las velocidades continentales son 8.35 km/seg (Pa) y 4.56 km/seg (Sa) y las oceánicas 8.01 km/seg (Pa) y 4.45 km/seg (Sa). Los periodos mas frecuentes son 10 segundos paraPa y 20 segundos paraSa. Tales periodos son independientes de la distancia en el caso dePa pero crecen con ella paraSa. Las fasesPa ySa mejor desarrolladas se obtienen para terremotos cuya profundidad focal es inferior a los 60 kilómetros. El movimiento de la particula del suelo debido aSa puede ser de cualquier tipo, desde puroSV a puroSH, y muestra gran correlación con el movimiento de la particula deS. Los ángulos de emergencia aparentes (de promedio 51° paraPa y 54° paraSa) varian con la distancia. Basándonos en nuestras observaciones sugerimos quePa ySa se propagan por reflexión múltiple, bajo incidencia rozante bajo la discontinuidad de Mohorovii.

     

Anomalous upper mantle beneath the Australian-Antarctic discordance

Within the Australian-Antarctic discordant zone, residual depth anomalies approach 1000 m. In sea floor younger than 10 Ma that is more than 500 m deeper than expected, Rayleigh wave phase velocities are significantly faster than in sea floor of comparable age in the Pacific. In this area, the shear wave velocity in the 20–40 km depth range is unusually fast, indicating that the lithosphere develops more rapidly than usual from an asthenosphere that is perhaps cooler than average. In sea floor that is older than 10 Ma, phase velocities are anomalously fast and independent of the residual depth. Beneath this older sea floor, the low-velocity zone in the oceanic mantle is much less pronounced than beneath sea floor of comparable age in the Pacific.     

Os isotope heterogeneity of the upper mantle: Evidence from the Mayarí–Baracoa ophiolite belt in eastern Cuba

In an attempt to quantify the extent of geochemical heterogeneity within a restricted and well dated portion of the upper mantle, 27 chromite separates from the 90 My old chromite deposits in the Mayarí–Baracoa ophiolite belt in eastern Cuba have been investigated for platinum group element (PGE) concentrations and Re–Os isotopic systematics. The samples are characterized by systematically subchondritic initial ¹⁸⁷Os/¹⁸⁸Os ratios and substantial heterogeneity. The initial ¹⁸⁷Os/¹⁸⁸Os ratios vary with chromite chemistry and with geographical distribution, reflecting differences in the Os isotopic evolution for the different upper mantle sections represented by the ophiolite. Accordingly, the Os isotope data might be divided into three groups. In the Moa–Baracoa district, where the chromite bodies are located in the mantle–crust transition zone, the calculated initial γOs values average − 0.97 ± 0.69 (n = 13). In the Sagua de Tanamo district, where chromite chemistry is highly variable and their location in relation the mantle sequence is less clear, the initial γOs values are intermediate, with an average of − 1.77 ± 0.80 (n = 7). In the Mayarí district, where the chromite bodies are located in the lower part of the mantle sequence, initial γOs values average − 2.66 ± 0.29 (n = 7). These subchondritic (i.e. negative) initial γOs values are most simply explained by Re depletion during ancient partial melting and/or melt percolation events.The Os isotope heterogeneity documented here indicates a high degree of geochemical complexity on small to intermediate length scales in the upper mantle. Our results, in combination with data on chromites from the literature, show that an “average present-day Os isotopic composition” for the hypothetical depleted MORB mantle (DMM) reservoir cannot be precisely established beyond the statement that it is “broadly chondritic”. Indeed, the upper mantle cannot be considered a sufficiently homogeneous geochemical “reservoir” to serve meaningfully as a baseline against which geochemical “anomalies” are evaluated. On the other hand, our findings are consistent with the “Statistical Upper Mantle Assemblage” or “SUMA”-concept, according to which a high level of geochemical heterogeneity is maintained in the upper mantle at all relevant length scales, as a result of the plate-tectonic cycle and intra-mantle processes such as melt-migration and metasomatism.     

Rounded mineral inclusions in upper mantle peridotite

Mantle-derived peridotite xenoliths contain abundant olivine with rounded spinel and orthopyroxene inclusions, and orthopyroxene with rounded olivine and spinel inclusions. The shape-change of spinel, olivine and orthopyroxene inclusions from the primarily polyhedral outline to the spherical outline is governed by interfacial diffusion of oxygen in spinel (the most sluggish atom) due to reduction of total interfacial energy of the host-inclusion system.The critical radius of maximum rounded inclusions of spinel in olivine is a function of temperature and annealing time. Assuming that the activation energy for the interfacial diffusion is 40–70 kcal mol^?1 and that the annealing time for the spinel lherzolite from Salt Lake Crater in Hawaii is 100 Ma, the annealing times for perioditites under the island arcs of Japan are estimated to be 1–10 Ma.     

MTS studies on the upper mantle conductivity in China

Based on the data from more than 200 MTS sites distributed within different areas of the Chinese continent, general characteristics of upper mantle conductivity have been described. At least two conductive layers have been found in the upper mantle of some areas. The first is thin with a resistivity of a modicum to few tens m; the second one is thicker with a resistivity of one to m. Nearly 300 heat-flow values indicate that there exists an exponential correspondent relationship between a depth of the upper mantle conductive layer with a thickness and an average value of heat flow. Based on the above results, the top depth map of this upper mantle conductive layer has been outlined for parts of the Chinese continent. This conductive layer is basically consistent with the low velocity zone in the upper mantle, and Cenozoic tectonism and current seismicity are significantly related to the variation of depth of the conductive layer in the upper mantle. The possible origins of the conductive layers in the upper mantle have been discussed here.     

宁夏地区上地幔地震各向异性特征

利用宁夏6个三分向宽频带数字地震台记录的震中距在85°—110°,震级大于6级的远震SKS波震相作了偏振分析,采用Silver-Chan方法求得了分裂参数,获得了宁夏地区上地幔各向异性图像.结果表明,宁夏地区上地幔各向异性普遍存在,本区盐池台(YCI)所在的鄂尔多斯块体下方的各向异性可归结为"化石"各向异性,其它地区的各向异性特征则反映了正在进行的构造运动,由此可为贺兰山脉东西两侧复杂的运动图像提供深部证据.最后推测本区的青藏高原东北缘—阿拉善块体及银川断陷区壳、幔的运动是耦合的,而鄂尔多斯块体地区壳、幔的运动型式尚不明朗.     

Dynamical stability of convection cells in the upper mantle

Summary A two-dimensional flow model of an incompressible fluid with constant viscosity has been used to study the changes in the large-scale flow pattern (aspect ratio 4). Implications for convection in the Earth's mantle are discussed.

---

a ¶rt; mn uuu ua u¶rt;mu nm m unm ¶rt; uu uu mu ama¶rt;a. ¶rt;am mam ¶rt; uu amuu u.

     

Changes in geomagnetic reversal frequency at times of tectonic change: Evidence for coupling between core and upper mantle processes

Coincidence of major plate tectonic changes about 42–45, 70–80, and 110–120 m.y. B.P. with changes in geomagnetic reversal frequency provides support for previous suggestions that processes in the upper mantle and those in the outer core are somehow coupled.