U43B-0048
Geochemical Tracing of Asthenosphere Flow
Geochemically, the asthenosphere can be viewed (at least beneath oceanic lithosphere) as divisible into isotopically distinct domains, with a systematic variability within these domains. Both between-domain and within-domain variations provide evidence for asthenosphere flow fields. Between mantle domains, tracing the domain boundary through time is possible by isotopic fingerprinting of lavas that vary in space and time. Detailed studies to which the author has contributed include the SE Indian Ocean (the Australian-Antarctic Discordance), the SW Atlantic (Drake Passage), and the SW Pacific (Lau Basin); another good example is Central America. In these cases, mantle from one isotopically-defined domain has invaded mantle of another domain through a gateway formed by rifting of continental lithosphere or by creation, by collision or ridge subduction, of gaps in the subducted oceanic lithosphere. Many domain boundaries can be recognised along the ocean ridge system but, without extensive oceanic drilling cannot be traced through time. Within mantle domains, geochemical gradients in asthenosphere-derived lavas provide a basis for tracing of asthenosphere flow. Conceptually, as asthenosphere flows beneath progressively thinner lithosphere, decompression produces small melt fractions which escape to produce volcanoes or contribute to metasomatism of the lithosphere. This in turn leads to progressive depletion of the asthenosphere in incompatible elements. Moreover, as the more fusible components melt first, the isotope composition can also change systematically. Thus, the asthenosphere changes in composition by flow, for example from plume-fed regions to ocean ridges, from converging continents to ocean basins and from outside arc-basin systems towards the arc front. Tracing asthenosphere flow by geochemical mapping of arc-basin systems reveals clear geochemical gradients which point to a variety of flow patterns depending on tectonic configuration. Inferences such as trench-parallel flow along the Central Lau Basin, or trench-orthogonal flow behind the Izu arc, support the results of shear-wave splitting experiments and emphasise the value of combining geochemical and geophysical approaches.
U43B-0049
The asthenosphere as a small-scale convection and plume graveyard
The asthenosphere is often associated with the low (seismic) velocity zone and/or the low-viscosity zone located under the colder and more viscous lithosphere. One important question is if the asthenosphere characterics (low seismic velocity and low viscosity) are due exclusively to the pressure-dependence of mantle mineral properties (such as water content, or partial melting), or if they could be the signatures of mantle flow. In terms of mantle convection, the lithosphere constitutes the upper cold thermal boundary layer (TBL) of the mantle, while the asthenosphere would be the region right underneath it. Is there any fluid mechanics reason why this region should be different from the rest of the convective mantle ? To try to answer this question, we studied the convective patterns developping at high Rayleigh numbers in a layer of sugar syrup, a fluid with strongly temperature-dependent viscosity (and therefore a good analog to mantle material). A new method allowed us to simultaneously measure the temperature and velocity fields. We found that in the "sluggish lid" regim, three different scales of convection develop. The largest convective scale is cellular, with cold downwelling sheets of viscous fluid encasing hotter, less viscous, parts of the tank. Within each of those cells develop several (typically 3 to 7) hot 3D upwelling plumes. Last, small- scale instabilities develop from the bottom of the upper cold TBL, especally (but not only) when hot upwelling plumes impact and flatten under the cold TBL. Hence there is a region under the cold TBL where all hot plumes end their lives as well as generate small-scale cold localized instabilities. This laboratory "asthenosphere" therefore presents an horizontally-averaged temperature hotter and an horizontally- averaged viscosity lower than both our experimental mantle at mid-depth and the cold TBL. This is reminiscent of the mantle asthenosphere. However, the experiments further show that such an asthenosphere should be "patchy", since constituted of cold small-scale instabilities as well as hotter plumes pancakes.
U43B-0050
Mantle Convection, Plate Tectonics, and the Asthenosphere: A Bootstrap Model of the Earth's Internal Dynamics
Several studies have highlighted the role of a low viscosity asthenosphere in promoting plate-like behavior in mantle convection models. It has also been argued that the asthenosphere is fed by mantle plumes (Phipps- Morgan et al. 1993; Deffeyes 1972) and that the existence of the specific plume types required for this depends on plate subduction (Lenardic and Kaula 1995; Jellinek et al. 2002). Independent of plumes, plate subduction can generate a non-adiabatic temperature gradient which, together with temperature dependent mantle viscosity, leads to a low viscosity near surface region. The above suggests a conceptual model in which the asthenosphere can not be defined solely in terms of material properties but must also be defined in terms of an active process, plate tectonics, which both maintains it and is maintained by it. The bootstrap aspect of the model is its circular causality between plates and the asthenosphere, neither being more fundamental than the other and the existence of each depending on the other. Several of the feedbacks key to the conceptual model will be quantified. The implications for modeling mantle convection in a plate-tectonic mode will also be discussed: 1) A key is to get numerical simulations into the bootstrap mode of operation and this is dependent on assumed initial conditions; 2) The model implies potentially strong hysteresis effects (e.g., transition between convection states, associated with variable yield stress, will occur at different values depending on whether the yield stress is systematically lowered or raised between successive models).
U43B-0051
Low Viscosity Channels and the Stability of Long Wavelength Convection
Mantle convection simulations with a low viscosity channel, akin tothe Earth's asthenosphere, are characterized by long wavelength flow structure. Boundary layer theory predicts that as the viscosity of the channel decreases, the wavelength that maximizes heat transfer increases. As a pattern selection criterion, this analysis is not complete. It provides no mechanism to relate the optimal heat transfer wavelength to the wavelength that is realized or preferred in nature. We present numerical simulation suites, for bottom and internally heated end-members, to demonstrate that the cell wavelengths that maximize heat transfer are also the most stable. This does not rule out the possibility of multiple wavelengths being realizable but it does imply that wavelengths near the stability peak will be preferred and, for the configurations we explore, the stability peak corresponds to the energetically most efficient flow configuration.
U43B-0052
2D and 3D Numerical Experiments Assessing the Necessary Conditions for a Plume-fed Asthenosphere
In past years we have presented observation evidence which suggests to us that in Earth's mantle there exists a buoyant asthenosphere layer fed by upwelling in mantle plumes, and consumed by accretion and transformation into overlying lithosphere by ridge upwelling and melt-extraction (which creates a ~60km-thick layer of compositional lithosphere at mid-ocean ridges), by plate cooling (which accretes a further ~40km of asthenosphere after 100 Ma of near-surface cooling), and by dragdown by subducting slabs (which drags a further ~20km sheet of buoyant asthenosphere on either side of the subducting slab). This scenario has been recently reviewed in Yamamoto et al (GSA Vol. 431). We believe that the reason this mode of mantle convection has not yet been seen in numerical models of mantle convection is due to the inability of current models to model the correct upwelling rates in focused lower-viscosity plumes (i.e. that, due to numerical resolution problems they currently underpredict plume upwelling) and to correctly model the magnitude of downdragging of a more buoyant but lower viscosity asthenosphere layer by subducting slabs (which they currently overpredict, cf. Phipps Morgan et al., Terra Nova, 2007). Here we present results from a suite of 2D and 3D calculations that include the effects of ridge accretion, plate cooling and well-resolved asthenosphere dragdown by subducting slabs. In the 2D experiments we do not let mantle plumes spontaneously form at the hot base of the mantle. Instead we extract mantle at a prescribed rate from a single region near the bottom of the mantle (the base of the 'plume stem') and inject this hot material into the uppermost mantle using a local dilation element 'source'. The point is to bypass an incorrect 2D treatment of plume upwelling (plumes should be pipes that only slightly disrupt surrounding flow instead of sheets that break 2D mantle flow), in order to explore what upwelling flux is needed to form a persistent plume-fed asthenosphere. We find that in order to create a global sub-oceanic plume-fed asthenosphere: 1) the plume-flux should be more than about 1.2 times the slab-flux; 2) the asthenosphere should be at least 0.5 percent more buoyant than underlying mantle when its viscosity is ~ 10e19 Pa-S. We also vary the location of the plume top from near-ridge to near-trench using dilation elements, the observation is that a counter flow will always form and therefore make a persistent plume-fed asthenosphere when these conditions are satisfied. For the 3D experiments, we choose a different approach than our 2D experiments. This time we allow a real plume form within our domain. For our initial experiments, we place a ridge-centered plume at one corner of the computational region to take advantage of the two resulting symmetry planes that allow us to model the plume with only one fourth the nodes needed for similar resolution for a plume within the interior of the 3D computational region. We use an unstructured grid with high grid resolution within the plume conduit and slab entrainment regions of the box. We will discuss the flow pattern and assess the necessary conditions for a plume-fed asthenosphere in these 3D runs.
U43B-0053
Implications of the Kinked Boyd Kimberlite Geotherm for the Thermal Evolution Beneath Continents
For many years, the standard model of oceanic and continental lithosphere is that both are underlain by similar asthenosphere, the 'Low Velocity Zone'. Kimberlite xenoliths provide direct probes to continental material exhumed from as deep as 180km. Using standard techniques for the pressure and temperature- dependent partitioning of elements between mineral-pairs, they also provide a snapshot of the continental geotherm at the time of kimberlite eruption. We can accurately model kimberlite geotherms as being caused by a transient reheating event shortly (5-15Ma) before the kimberlite eruption, and have used this approach to model kimberlite geotherms from S. Africa, Siberia, and Canada. This interpretation of the Boyd geotherm is consistent with the conceptual framework of a plume-fed asthenosphere in which the >200km thick cratonic roots of continents are typically underlain by 'normal mantle' at approximately 1100-1200C, and only underlain by hotter plume material at approximately 1400-1500C during transient reheating events associated with kimberlite eruptions. This analysis also implies that continental cratons do not cause underlying mantle to appreciably warm on a 200Ma-Ga timescale -- i.e. they do not act as effective 'insulating blankets' for underlying mantle. Note that the non-plume geotherm inferred by this temperature modeling is also consistent with seismic and heat flow inferences of the differing temperature structure beneath cratons and ocean basins. The existence of a normal non-plume and transient hotter plume geotherm also provides a natural explanation for the striking dichotomy between volcanic (i.e. plume-influenced) and non-volcanic (non-plume-influenced) volcanic margins that form from the rifting of continents (Reston and Phipps Morgan, Geology, 2004).
U43B-0054
Electrical Conductivity of Partially Molten Peridotite Analogue Under Shear: Supporting Evidence to the Partial Melting Hypothesis for the Oceanic Plate Motion
So far, two hypotheses have been proposed to explain softening of the oceanic asthenosphere allowing smooth motion of the oceanic lithosphere. One is partial melting, and the other is hydraulitic weakening. Although the hydraulitic weakening hypothesis is popular recently, Yoshino et al. [2006] suggested that this hypothesis cannot explain the high and anisotropic conductivity at the top of the asthenosphere near East Pacific Rise observed by Evans et al. [2005]. In order to explain the conductivity anisotropy over one order of magnitude by the partial melting hypothesis, we measured conductivity of partially molten peridotite analogue under shear conditions. The measured samples were mixtures of forsterite and chemically simplified basalt. The samples were pre- synthesized using a piston-cylinder apparatus at 1600 K and 2 GPa to obtain textural equilibrium. The pre- synthesized samples were formed to a disk with 3 mm in diameter and 1 mm in thickness. Conductivity measurement was carried out also at 1600 K and 2 GPa in a cubic-anvil apparatus with an additional uniaxial piston. The sample was sandwiched by two alumina pistons whose top was cut to 45 degree slope to generate shear. The shear strain rates of the sample were calibrated using a Mo strain marker in separate runs. The lower alumina piston was pushed by a tungsten carbide piston embedded in a bottom anvil with a constant speed. Conductivity was measured in the directions normal and parallel to the shear direction simultaneously. We mainly studied the sample with 1.6 volume percent of basaltic component. The shear strain rates were 0, 1.2x10(-6) and 5.2x10(-6) /s. The sample without shear did not show conductivity anisotropy. In contrast, the samples with shear showed one order of magnitude higher conductivity in the direction parallel to the shear than that normal to the shear. After the total strains reached 0.3, the magnitude of anisotropy became almost constant for both of the strain rates. The magnitude is thus independent of the strain rate. This study demonstrates that the anisotropy at the top of the asthenosphere can be explained based on the partially molten asthenosphere sheared by the plate motion.
U43B-0055
The Birth, Life, and Death of the Oceanic Asthensphere from Seismic Anisotropy
Oceanic asthenosphere represents one of the basic components of plate tectonics. It permits the surface plates to move rapidly by partially decoupling them from the mantle below. The existence of such a mechanical asthenosphere is best demonstrated by seismic anisotropy. Indeed, models for mantle flow including such an asthenosphere, driven by known mantle density heterogeneity and the motions of the surface plates, predict surprisingly well the orientation of anisotropy from both shear-wave splitting fast polarization directions and azimuthal anisotropy in surface waves. What is the fate of asthenosphere upon subduction of the oceanic plate? A global survey of below-slab anisotropy from splitting (Long and Silver, 2008) suggests that all but two subduction zones exhibit trench-parallel fast polarization directions that argue for the dominance of trench-parallel flow, rather than the asthenospheric shear flow found beneath oceanic plates, or the broader slab-entrained flow that would be expected in the absence of any asthenosphere. Such a result is most consistent with the existence of a thin (of order 10 km) decoupling zone beneath the slab throughout the upper mantle, which we take to be an attenuated asthenosphere. The two subduction zones that exhibit below-slab trench-normal fast polarization directions are Cascadia and Mexico. It is significant that these are subduction zones where ridge-trench separation is the smallest. These two cases can be accounted for if asthenosphere has a formation stage that requires a minimum amount of finite strain. If this minimum strain is not reached prior to subduction, then slab-entrained flow, rather than trench-parallel flow, would dominate the below-slab flow field and lead to trench-normal fast polarization directions. We thus seek a mechanism for the asthenosphere that explains its formation stage, its maintenance beneath an oceanic plate, and upon subduction, its persistence throughout the upper mantle but in an attenuated form. One popular explanation for a weak asthenosphere is that it marks a maximum in homologous temperature and a corresponding minimum in viscosity. Yet, such a mechanism, being solely dependent on pressure (depth), does not explain the observed anisotropic characteristics, because it predicts that the asthenosphere would be fully developed at the ridge, and would completely disappear upon subduction when the bottom of the slab reaches the base of the asthenosphere. A more promising mechanism is strain- weakening by shear heating. While shear-heating is known to be a negligible source of heat for typical asthenospheric viscosities (e.g. 1019 Pa s, Turcotte and Schubert, 1982), if asthenosphere formed from shear heating of ambient mantle possessing an order of magnitude higher viscosity, then this could produce, by itself, an increase in temperature of order 100°C and a reduction in viscosity of about an order of magnitude. This mechanism would explain the persistence of asthenosphere at greater depth in subduction zones, due to the advection of this heat, and the attenuation of the subducted asthenosphere, since the positive buoyancy of the warm asthenosphere would resist subduction. This hypothesis has several testable predictions. For example, it predicts that midplate asthenosphere is warmer than near-ridge asthenosphere by about 100°C. Second, it predicts that other subduction zones with very small ridge- trench separation would have below-slab trench-normal splitting fast polarization directions.
U43B-0056
Mapping the Lithosphere and Asthenosphere Boundary in Southern Africa from S receiver function
Southern Africa, which contains the Kaapvaal and Zimbabwe cratons, is an ideal place to address how deep cratonic lithosphere extends. In this study, we image the lithosphere and asthenosphere boundary (LAB) in southern Africa from S receiver functions that are computed from seismograms recorded at the southern Africa seismic experiment array. We choose teleseismic events with M≥5.6 and epicentral distances between 58° and 105° that generate S or SKS to P conversions at receivers. Individual S receiver functions are stacked by station with a whole depth moveout correction relative to the reference slowness of 6.4 s/degree. We found that the depth of the LAB roughly correlates with geological provinces, from ~85 km at the Cape-Fold Belt in the southern end of Africa to 155 km at the northern Kaapvaal craton. Receiver functions are more complicated in the vicinity of the Bushveld Complex, showing two S to P conversions from the upper mantle, which corresponds to ~80 km and ~130 km, respectively. We interpret that the deep one is the LAB and the shallow one is from a velocity reduction layer in the lithosphere due to iron enrichment. S to P conversion phase from the LAB becomes weaker towards north to the interior of the cratons, indicating the velocity contrast at the LAB is reduced. The trend of the LAB depth variation from S receiver functions largely agrees with that from surface wave tomography, while the absolute LAB depth is generally thinner in this study. Forward modeling of S receiver function from tomography models will be conducted to reconcile the LAB depth from the two different methods.
U43B-0057
High Resolution Imaging of the Aspen Anomaly - CREST and USArray
The Aspen anomaly is a low velocity (-4.5% Vs and -2.5% Vp) upper mantle feature that approximately underlies the highest elevations of the Colorado Rocky Mountains. It is geographically, and possibly causatively, associated with Proterozoic structure in the lithosphere underlying the Colorado Mineral Belt. The coincidence of high topography, evidence of Cenozoic uplift, slow mantle velocities, magmatism, and possible inherited Proterozoic lithospheric structure and associated rheological weakness, together suggest that the Aspen anomaly may be a significant mantle geodynamic influence on the evolution of the central Colorado Rocky Mountains through to the present day. Competing end-member models for the origin of the anomaly are: 1) upwelling asthenosphere associated with Cenozoic mantle modification, and 2) thermal, rheological, and/or compositional variations in the lithosphere resulting from reactivation of much older lithospheric structures. Hybrid models involving interaction between recent mantle reorganization and older lithospheric flaws are also possible. In August of 2008, we deployed 59 IRIS PASSCAL broadband seismographs above the Aspen anomaly with a mean station spacing of 26 km, which will remain in place until October 2009. This IRIS PASSCAL deployment was embedded within the 70-km spacing USArray Transportable array and 4 NEIC sites. In total, this composite array is 94 seismic sites which will provide improved resolution to transition zone depths beneath the Colorado Rockies. We report on resolution tests for seismic and joint seismic/gravity inversion and results from early data from this experiment in the context of Aspen anomaly hypotheses and predictions for three-dimensional upper-mantle velocity heterogeneity.
U43B-0058
Seismic properties in the asthenosphere beneath the petit-spot region inferred from BBOBS data
The petit-spot is a term for submarine volcanoes recently discovered by Hirano et al. [2006]. They are young (0-10 Ma) volcanic micro-knolls in very old (~130 Ma) NW Pacific plate (about 500 km offshore from the Japan Trench). Although the estimated activity distributes widely in time (0-10 Ma) and in space (over 600 km), the volume of each volcanic edifice is small (several orders of magnitude less than the previously known seamounts and knolls). The mechanisms of melt production and magma eruption process of the new type of volcanism are still unknown. Hirano et al. [2006] proposed that small fraction of melt came from the asthenosphere through fractures in the lithosphere which were induced by flexure around the outer-rise. Local seismicity located in petit-spot region may relate to the magma eruption process. In order to investigate nature of the petit-spot, we promote a comprehensive survey based on geophysics, geochemistry, geology, and numerical modeling, which is called petit-spot multidisciplinary project. As a part of the petit-spot multidisciplinary project, seismological observation using Broad-Band Ocean Bottom Seismometers (BBOBSs) is conducted. We deploy three BBOBSs in the petit-spot region with about 100 km spacing. We arrange BBOBS array to enclose both Yukawa-kaikyu and recent seismicity around the petit-spot. The sites are equipped with Guralp CMG-3T sensors recorded at 100 Hz. The BBOBSs were deployed in May 2007 by R/V KAIREI (JAMSTEC) and were recovered July 2008 by R/V YOKOSUKA (JAMSTEC). All the BBOBSs were successfully recovered with high quality data. We establish two main goals of this observation, precise earthquake location of local events, and estimation of seismic properties in the asthenosphere beneath the petit-spot. The first and second one will constrain the mechanism of the magma eruption and the melt production, respectively. In this study, we focus on the melt production process. We plan to measure seismic attenuation and travel-time anomalies in the asthenosphere beneath petit-spot region to infer physical properties of the region. The results are expected to constraint especially on the melt production process of the petit-spot. In this presentation, we will introduce the overview of the earthquake observation in the petit-sport region, example of BBOBS data, and preliminary results of the estimation of seismic properties of the asthenosphere beneath the petit-spot region.
U43B-0059
Characterizing Lithospheric Thickness in Australia using Ps and Sp Scattered Waves
The purpose of this study is to constrain the morphology of the lithosphere-asthenosphere boundary throughout Australia using scattered waves. Prior surface wave studies have shown a correlation between lithospheric thickness and the three primary geologic provinces of Australia, with the shallowest lithosphere located beneath the Phanerozoic province to the east, and the thicker lithosphere located beneath the Proterozoic and Archean regions. To determine lithospheric thickness, waveform data from twenty permanent broadband stations spanning mainland Australia and the island of Tasmania were analyzed using Ps and Sp migration techniques. Waveform selection for each station was based on epicentral distance (35° to 80° for Ps and 55° to 80° for Sp), and event depth (no greater than 300 km for Sp). For both Ps and Sp a simultaneous deconvolution was performed on the data for each of the twenty stations, and the resulting receiver function for each station was migrated to depth. Data were binned with epicentral distance to differentiate direct discontinuity phases from crustal reverberations (for Ps) and other teleseismic arrivals (for Sp). Early results in both Ps and Sp show a clear Moho discontinuity at most stations in addition to sharp, strong crustal reverberations seen in many of the Ps images. In the eastern Phanerozoic province, a strong negative phase at 100-105 km is evident in Ps for stations CAN and EIDS. The negative phase lies within a depth range that corresponds to the negative velocity gradient between fast lithosphere and slow asthenosphere imaged by surface waves. We therefore think that it is the lithosphere- asthenosphere boundary. On the island of Tasmania, a negative phase at 70-75 km in Ps images at stations TAU and MOO also appears to be the lithosphere-asthenosphere boundary. In the Proterozoic and Archean regions of the Australian continent, initial results for both Ps and Sp migration indicate clear crustal phases, but significantly more complicated signals at mantle depths. However, at some stations along the southern edge of the thick sub-cratonic lithosphere (previously imaged by surface waves) phases exist which may represent a lithosphere-asthenosphere boundary at depths of 110-115 km. Constraining the relationship of lithospheric thickness to the age and tectonic history of the overlying crust in Australia is important for better understanding the long term evolution of the continent.
U43B-0060
Inference of Shallow, Weak, and Stress-Dependent Asthenosphere at Active Continental Margins from Postseismic Observations: Confirmation of Experimental Power- Laws
Rheological flow laws of mantle minerals (most importantly olivine) can be determined in the laboratory, but results require large extrapolations to tectonic strain rates, raising questions as to their applicability. Furthermore, laboratory experiments can say little of in situ environments (temperature, water content, strain rates), thus effective viscosities within the mantle are estimated with uncertainty. Here we use large earthquakes as giant natural rock squeezing experiments to infer the rheological properties of the asthenosphere in active continental deformation zones and the environment in which it flows. Our objective is to compare inferred in situ flow laws to those derived in the laboratory and to help understand the environment of the asthenosphere inboard of active oceanic/continental margins. In these natural experiments, the earthquakes impart significant stress through the lithosphere and into the asthenosphere, inducing flow in the latter that causes transient postseismic surface displacements. GPS observed displacements are then used to constrain finite element models of postseismic relaxation to determine the rheological properties of the asthenosphere. We consider the postseismic response following the 2002 M7.9 Denali Fault, Alaska earthquake and the 1999 M7.1 Hector Mine, California earthquake. The former event occurs inboard of an active subduction zone while the latter event occurs inboard of a recent (<12 ma) subduction zone (currently the San Andreas Fault). In both cases a wide array of continuous GPS stations were in place to record postseismic surface displacements from immediately after the earthquake. We find that most (if not all) postseismic displacements that occurred beyond about a half-rupture length away from each event were caused by flow within a weak asthenosphere below a strong 40-50 km thick lithosphere. In both cases transient postseismic displacement time series could not be explained by a Newtonian rheology, but are well predicted when using a stress-dependent power-law rheology as suggested by laboratory experiments of samples deformed by dislocation creep. Postseismic responses are explained by power-laws with parameters (A, n, Q) within the uncertainties of experimental flow laws for wet olivine. These asthenospheres are wet, hot (1200-1300°C), and experience a background strain rate of the order of 0.1-1.0 υstrain/year. Perhaps most interesting is the transient nature of the effective viscosity of these asthenospheres associated with their stress dependence. Immediately after each earthquake the effective viscosities of the asthenosphere directly beneath the rupture are as low as 1017 Pa s. Within a year the effective viscosities have risen to more than 1018 Pa s and within a decade they have recovered to background levels of the order of 1019 Pa s.
U43B-0061
SdP receiver function images of the lithosphere-asthenosphere boundary beneath the Western U.S using USArray data
S-wave receiver functions from the USArray transportable array data have been made to examine the lithosphere and upper mantle structure beneath the western U.S. The project was initiated at the CIDER 2008 workshop in Santa Barbara, CA. Our preliminary results image the volume between the west coast of the US and 107°W and between the Canadian and Mexican borders to approximately 200 km depth. The large area allows us to investigate the structure of a variety of tectonic regions: the San Andreas plate boundary region, Mendicino Triple Junction, Cascadia subduction zone, Sierra Nevada, Snake River Plane, Basin and Range, Colorado Plateau, and the Rocky Mountains. Data from 59 events between 2005 and 2008 at epicentral distances of 55 to 80 degrees from the 556 stations in the transportable array database were processsed. To start we depth converted, spatially repositioned, and common conversion point stacked the SdP receiver functions using a 1D reference model based on averages of Crust2.0 and MC35. We are currently recalculating the depth and lateral positioning of the receiver functions using linear tomography corrections for laterally variable structure. The depth converted and preliminary CCP stack of the SdP receiver functions image a number of interesting lithospheric features including its base, the lithosphere- asthenosphere boundary (LAB). The LAB is a negative amplitude feature that has significant topographic variation, and cannot be described as a single surface. In the Basin and Range and particularly in the southern Sierra Nevada drip region the SdP LAB depths agree well with those determined from PdS receiver functions (Levander et al., 2008), however elsewhere large discrepancies exist between LAB estimates from the two different types of receiver functions We see a particularly strong correlation between calculated equilibration pressures of 911 primitive basalt whole rock samples from across the western United States, extracted from the NAVDAT database (http://navdat.kgs.ku.edu/), and the LAB estimate from the SdP images beneath the southern Basin and Range and the Colorado Plateau. The depth estimates from the geochemistry data and comparison with the PdS receiver function images for the same region allows us to interpret the lithosphere-asthenosphere boundary and its relation to the tectonic provinces in the western United States. In addition, we are able to describe possible links between the changes in LAB topography and tectonic evolution.
U43B-0062
Linking Upper Mantle Processes and Long-wavelength Topographic Swells in Cenozoic Africa
The topography of present day Africa is influenced by two different wavelengths of dynamic support. The South African Superplume sits beneath Sub-equatorial Africa and is thought to be supported by a lower mantle thermo-chemical anomaly. On a smaller scale a series of topographic domal swells, 1000km in diameter, occur across the continent. The swells are characterised by elevated dynamic topography, a positive long-wavelength gravity anomaly and a negative velocity perturbation from a higher mode surface wave tomography model. In addition, where the lithosphere is thinner than 100km, the swells are capped with volcanic products, erupted periodically since ~30 Ma. These areas include the Cameroon Volcanic line, Hoggar, Tibesti and Darfur in North Africa, and the Ethiopian Plateau and the Kenyan dome found along the East African Rift system. The given relationships suggest the domal swells result from and are supported by upper mantle convection. In order to investigate these relationships a database of 3000 geochemical analyses has been assembled for Cenozoic African volcanism, from both literature search and by new analyses of samples collected from the Al Haruj volcanic field in Libya. Incompatible trace element ratios and REE trends from primitive basalts (>7wt% MgO) erupted less then 10Ma, representing the products of mantle melting, are compared with the upper mantle velocity structure. At depths of 75-100km the greatest velocity perturbation is associated with the Afar/Ethiopia region, where as smaller perturbations are found beneath the North African swells of Hoggar, Tibesti and Darfur. The comparison of absolute velocities, taken from the higher mode tomography model, with trace element ratios has found the low seismic velocity Afar/Ethiopia region to have shallow melting at high melt fractions (La/Yb~9) whereas North African swells with faster seismic velocities at 100 km depth, show deeper melting with smaller melt fractions (La/Yb~30). This positive correlation continues to depths of 150km and is believed to represent variations in mantle potential temperature beneath the African continent. With further modelling of major, trace and REE data we hope to provide insights into variations in mantle potential temperature, melt fraction and velocity structure beneath the topographic swells across the African continent.
U43B-0063
Sharp Lithosphere-asthenosphere Boundaries of Oceanic Plates
P- and S-receiver function (RF) analysis of borehole broadband ocean bottom seismic data (Kumar et al., 2008, this meeting) and the high-resolution RF image of the subducting Pacific plate beneath the northeast Japan (Kawakatsu, 2008, this meeting) both show the presence of sharp lithosphere-asthenosphere boundaries (LABs) of oceanic plates which appear to show dependence on the plate age. The apparent plate-age dependence of the thickness of the oceanic plate is consistent with a thermally controlled origin for the oceanic LAB, but the fact it is observed in short period (~3s) indicates a sharp boundary (the transition thickness of less than 10-15km), thus a chemical or fabric origin. The observed amplitude of the LAB signals, on the other hand, requires a rather large S-wave speed reduction of ~7%, similar to the observation beneath the eastern North America (Rychert et al., 2007, JGR). One possibility to explain these features is the presence of partial melting in the asthenosphere. The depth of partial melting of the model of Mierdel et al. (2007, Science) estimated using a thermal model incorporating pressure and thermal effect on the thermal diffusivity (Honda&Yuen, 2001, GRL) reproduces the basic trend in the data. For a texturally equilibrated partially molten region, however, a 7% S-wave speed reduction translates into ~3.5% of melting (Takei, 2002, JGR) which may be unrealistically large. The presence of the rather strong LAB signal of oceanic plates reported here may be partly attributed to other mechanisms such as the presence of shear zone of partially-molten region in the asthenosphere (e.g., Holtzman et al., 2003, Science).
U43B-0064
Microstructural and rheological evolution in naturally deformed peridotite mylonites
We have conducted a microstructural study of a highly deformed mantle shear zone from the Josephine Peridotite, Oregon, US. The goal of this study is to understand how microstructural evolution at large strains leads to transitions in rheological behavior. The particular shear zone we investigate here exhibits a higher degree of localization than previously studied shear zones in the Josephine Peridotite. The margin of the shear zone is characterized by a single microstructural domain, which contains moderately strong olivine and orthopyroxene fabrics oriented obliquely to the shear zone. The highly deformed samples from the center of the shear zone contain two distinct microstructural domains—a coarser grained domain (~800 microns) that contains only olivine, and a finer grained domain (~200 microns) that contains both olivine and orthopyroxene. The coarser grained domain has a very strong and still highly oblique LPO. However, within the finer grained domain, the olivine has only a modest LPO and the orthopyroxene has no LPO at all, suggesting that a transition to grain-boundary sliding has occurred. Olivine fabric strength increases towards the center of shear zone, however unlike the results in Warren et al. (2008) the obliquity of the fabric is unchanged. We hypothesize that the existence and persistence of the oblique olivine fabric is due to multiple generations of deformation. The olivine fabric was generated during the first deformation phase. A second phase of deformation with moderately different kinematics was primarily accommodated by deformation in the newly recrystallized fine grained orthopyroxene rich domain, allowing the strong and oblique olivine fabric to remain unmodified. This suggests that the recrystallization of orthopyroxene, while rare, may significantly modify the rheological behavior of peridotite (cf Skemer and Karato 2008). These samples provide a crucial microstructural link between modestly deformed shear zones (Warren et al., 2008) and highly deformed ultramylonites (Warren and Hirth, 2006).
U43B-0065
Low-Velocity Zone Structure Beneath the Kaapvaal Craton From S-wave Receiver Functions
The southern African Plateau is marked by anomalously high elevations, reaching 1-2 km above sea level. There is much debate as to whether this topography is compensated by a lower mantle source or by elevated temperatures in the upper mantle, which may be manifested as either an anomalously thin lid and/or as an anomalously slow low-velocity zone (LVZ). We use S-wave receiver functions (SRFs) to estimate the lithospheric thickness and sublithospheric mantle velocity structure beneath the Kaapvaal craton, which forms the core of the Plateau. To fit the SRF data, a LVZ is required below a 160 km thick lid, but neither the lid thickness nor the shear velocity decrease (4.5%) associated with the LVZ is anomalous compared to other cratonic environments. Therefore, we conclude that elevated temperatures in the sublithospheric mantle contribute little support to the high elevations in this region of southern Africa and that other compensation mechanisms are likely.
U43B-0066
Melt Could Create a Sharp Lithosphere-Asthenosphere Boundary Below Eastern North America
Recent studies have imaged a sharp drop in shear wave velocity (5-10%) over less than 11 vertical km at the lithosphere-asthenosphere boundary below eastern North America. While the magnitude of this velocity drop is too large to be explained by purely thermal gradients, it could be matched if a small amount of partial melt was trapped in the shallow asthenosphere. Here we present numerical experiments and melting calculations, which examine the range of mantle flow patterns, composition, and water-contents necessary to produce a small melt fraction in the region of slow shear wave velocities observed at the lithosphere- asthenosphere boundary below eastern North America. A parameterization of peridotite melting with varying H2O-content was created for this study and can be used with any finite-element model of mantle convection which tracks temperature and velocity with time. A step-shaped decrease in lithospheric thickness observed at the continental margin of eastern North America, together with lateral plate motions, is ideal to produce edge-driven convection and asthenospheric upwelling at the continental margin. We predict 1- 2% melting at a depth of 90-150 km for an asthenosphere with a mantle potential temperature of 1350°C and 450-900 ppm H2O or for a mantle at 1400°C with 150-300 ppm H2O. If the asthenosphere has potential temperature less than 1300°C or contains less than 150 ppm H2O, no melting will occur. Small lenses of remnant subducted slabs, CO2-rich mantle, or hydrated mantle wedge within the asthenosphere will have lower melting temperatures relative to anhydrous peridotite and are likely to produce melting. Because the asthenospheric melting is highly dependent on mantle potential temperature and H2O-content, the onset and degree of melting below the lithosphere- asthenosphere boundary may fluctuate with time and space. Thus if small degrees of asthenospheric melting are indeed responsible for the sharp drop in shear wave speed at the lithosphere-asthenosphere boundary, the magnitude of the shear velocity drop is also likely to vary.
U43B-0067
Lithospheric Discontinuity in the Central Slave Craton: A New Perspective From Seismological, Electrical and Petrological Constraints
It has been shown in recent years that lithospheric discontinuities may hold the key to better understanding the processes responsible for the formation and evolution of continents. Archean cratons have remained stable for billions of years, and therefore provide a unique window into the tectonic processes that took place during the time when they were formed. In this study, we present a new receiver function (RF) image of the lithospheric mantle beneath the Slave craton. We analyze P-to-S (Ps) converted waves from a dataset consisting of 135 events recorded at a linear array of 30 broadband seismic stations spanning ~400 km across the Slave craton. Our RF image shows a consistent positive velocity gradient across the array at depths from 34-41 km, indicating the location of Moho. Furthermore, we observe a pronounced southward- dipping negative gradient at ~103-134 km depths beneath central Slave. This low velocity layer is spatially coincident with an electrical conductive anomaly derived from previous magnetotelluric experiments, and with the upper boundary of the petrologically-constrained ultra-depleted region. One possible explanation for this geophysical/petrological boundary is that it represents a compositional interface marked by alteration minerals that cause the low seismic velocities (e.g., phlogopite) and inter-granular graphite films that cause the conductive anomaly. We speculate that this front may have originated through metasomatism associated with a subduction event that played an important role in the assembly of the Slave craton.
U43B-0068
Metastable transformations of eclogite to garnetite in subducting oceanic crust
Changes of mineralogy and density of mid ocean ridge basalt (MORB) due to high-pressure transformations are crucial to the dynamics of subducting oceanic crust in Earthfs deep mantle. Previous studies have discussed this issue assuming the equilibrium transformation in MORB, and mechanisms and kinetics of transformations have not been examined so far. In order to clarify metastable mineralogy of subducting oceanic crust, mechanisms of the eclogite-garnetite transformation were examined at 11 GPa and 1273-1823K using synthesized eclogite with MORB composition as a starting material. We found that processes of the eclogite-garnetite transformation in MORB are kinetically divided into two stages. The 1st stage in the transformation is the formation of majoritic garnet from clinopyroxene. The 2nd stage is the formation of majoritic garnet from original pyropic garnet by absorbing clinopyroxene component. In the present study, the 1st stage in the transformation proceeds promptly above 1273K, whereas the 2nd stage was not observed even at the highest temperature of 1823K for 180 minutes heating. The differences in kinetics are due to contrasting atomic diffusivities between clinopyroxene and garnet. The kinetic effect would change the mineralogy and rock texture in MORB, which provides important implications for dynamics of subducting oceanic crust by changing the density and viscosity relation. Density of the equilibrium MORB largely increases with increasing pressure by the eclogite-garnetite transformation. On the other hand, the metastable MORB has much lower densities than the equilibrium model. Metastable transformations of eclogite to garnetite possibly affect the rheology of the subducting oceanic plates. If the eclogite-garnetite transformation is kinetically inhibited, the post-pyroxene transformation may lead to rheological weakening of the subducting oceanic crust through the grain-size reduction.
U43B-0069
Crust and upper mantle of Europe, Greenland, and the North Atlantic region: An overview
We review the structure of the crust and the upper mantle in the area which covers about 1/8 of the globe and encompasses most of Europe, Iceland, Greenland, and Svalbard. Using the results from seismic (reflection and refraction profiles, P- and S-wave body-wave and surface-wave tomography), thermal, gravity, and petrologic studies (based both on the results of the authors and on literature compilations), we propose an integrated model of the structure and physical properties of the upper mantle in the entire region down to a depth of 250-300 km. The results are summarized in a series of maps of lateral variations in crustal and lithospheric thicknesses, seismic shear wave velocity at different depth slices, heat flow and lithosphere temperatures, as well as density and compositional variations in the lithospheric mantle. Our primary attention is to the lithosphere structure of the onshore parts of the region, although the less well constrained deep structure of the North Atlantic is also discussed.
U43B-0070
Is the olivine recrystallized grain size piezometer sensitive to water concentration?
Paleostresses in naturally-deformed peridotites are measured using an experimentally-determined recrystallized grain size-stress piezometer. This olivine piezometer was determined at low pressure (<300 MPa) using gas-confining media rock deformation apparatuses (gas rigs), which have high stress resolution, but the samples have low water contents due to low water fugacity. However, results from a recent experimental study by Jung and Karato (2001) performed at high pressure (2 GPa), thus high water fugacity, contend that for the same stresses, the recrystallized grain size of olivine is considerably larger at water concentrations >450 H/106 Si. Jung and Karato (2001) used a solid media assembly in a Griggs rock deformation apparatus (Griggs rig), which has very poor stress resolution and used a dislocation density piezometer to infer the stresses during deformation. In this study two sets of experiments were performed to 1) test the observation that the recrystallized grain size of olivine aggregates is dependent on water content in addition to stress and 2) calibrate the accuracy of the molten salt cell. The first series of experiments was performed using the molten salt cell in the Griggs rig on olivine aggregates in a general shear geometry at high confining pressure (1.6 GPa), temperatures (1100-1200°C), strain rates of 1x10-5 and 1x10-4/s and low water contents (<100 H/106 Si), which produced a wide range of differential stresses (60-550 MPa). The second series of experiments were performed at low pressures (300 MPa) in axial compression using pure metals (Ni and Fe) and dolomite to calibrate the molten salt cell in the Griggs rig to a gas rig over a wide range of stresses (40-250 MPa). The results of the shear experiments on olivine aggregates indicate that, after reaching high peak stresses at low shear strains, the aggregates strain weaken as the LPO becomes more uniform and reach a constant stress at high shear strain. The recrystallized grain size and stress of each sample fit a line with the same slope as the piezometer derived from stresses measured at low pressures in gas rigs, but all of our stresses were slightly lower for the same grain size. Stresses from these experiments were also inferred using the dislocation density piezometer; these inferred stresses are higher than the measured stresses at stresses <250 MPa and lower than the measured stresses at stresses >250 MPa. These inferred stress- recrystallized grain size data fit a line that directly overlaps the Jung and Karato (2001) data. The inferred stress-recrystallized grain size relationship has a considerably shallower slope than the measured stress-recrystallized grain size relationship. Initial results of the calibration experiments indicate that the molten salt cell in the Griggs rig measures stresses similar to the gas confining media (+/- 20 MPa). The hit points are very well defined and friction is lower than in solid media assemblies. However, steady state stresses are not reached until higher strains in the Griggs rig (5-10%) than in the gas rig (1%), most likely due to differences in compliance between the two rock deformation apparatuses. These results indicate: 1) stresses measured in the molten salt cell at high strains yield the same slope, but are lower than those measured in the gas rig, 2) the measurement of stress using the dislocation density piezometer overestimates the stresses relative to the gas apparatus and 3) the recrystallized grain size piezometer is likely not sensitive to water content. Use of the dislocation density piezometer by Jung and Karato (2001) resulted in overestimation of stresses. class="ab'>