T11C-0719
The Beginning of the Wilson Cycle, a Result of an Inherent Instability in Supercontinents Caused by the Lithosphere-Mantle System?
The Wilson Cycle, the repeated creation and destruction of supercontinents, may be the result of mantle flow and volume changes driven by slab movement and slab contributions to the upper mantle. While it is clear based on the images of a few slabs penetrating the 660 km boundary that the lower mantle must contribute to the upper mantle system, it seems likely that much of the tectonics we see at the surface is mostly interacting with the upper mantle system with occasional input from the lower mantle. Based on this we have derived a series of calculations to determine the volume of mantle change under Pangea caused by slab input and determined that between 300 and 200 Ma the material input greatly exceeds the possible output by the small ridge present in the Tethys during this time. Because Pangea was generally surrounded by slabs during this time we can assume that most of the material input from the slabs was retained under Pangea. Further, because many slabs, especially if they are rolling back as many of the Pangean slabs were, appear to lie on the 660 km boundary (rather than penetrate it) we assume that the volume of material contributed by the slabs was contributed in the upper rather than lower mantle. Because the input of material under Pangea is greater than the output at the Tethys ridge we propose that the excess mantle almost immediately began to push the slabs back and to push out against the root of the continents, effectively "breaking up" the supercontinent. Thus, the lithosphere is driven apart by the change in mantle volume beneath the supercontinent, however, the volume change and confinement of that mantle is driven by the lithosphere.
T11C-0720
Asthenospheric Flow and Origin of Volcanism in the Baikal Rift Area
The origin of low-volume, hotspot-like volcanism often observed in continental rift areas is debated, as is the nature of the flow in the mantle beneath. In this study we assemble seismic constraints on the mantle flow below the Baikal Rift Zone. We combine new evidence from upper-mantle tomography and from a radially anisotropic shear-velocity profile measured beneath and SE of the rift with published shear-wave-splitting constraints on azimuthal anisotropy. Fast directions of azimuthal anisotropy are perpendicular to the rift on both of its sides. Stable Siberian-Platform lithosphere on the northern side extends down to 200 km; on the southern side, the lithosphere is only 60-70 km thick. This asymmetry, with the thick cratonic root forming a barrier to convective flow, inhibits a mid-ocean-ridge-like flow pattern. The radially anisotropic profile yields evidence for horizontal flow in the upper asthenosphere (60-100 km) SE of the rift but no evidence for vertical flow. These results are consistent with horizontal asthenospheric flow NW to SE perpendicular to the rift, from beneath the thick Siberian towards the thin Baikal-Mongolian lithosphere. When asthenosphere ascends from the 200-km depth beneath the craton to shallower depths beneath the rift, decompression melting of embedded veins of enriched rock can produce the hotspot-like basaltic volcanism observed. The occurrence of such veins of enriched mantle in sub-cratonic asthenosphere has been predicted by geochemical mantle models; we propose that sub-horizontal asthenospheric flow from beneath cratons is a common cause for hotspot-like volcanism in continental rift zones.
T11C-0721
Numerical Models of Lithosphere-Mantle Interaction at Continental Rift Zones
Extension of continental lithosphere results in thinning of both the crust and mantle lithosphere. As a result of the thinning, mantle material wells up beneath the rift zone. Replacement of the colder lithosphere material with warmer mantle material may result in a thermal anomaly beneath the rift zone that is possibly visible in tomographic images. Tomographic images of the upper mantle below the Rio Grande rift indeed show a low seismic wave velocity anomaly extending to about 100-180 km depth. They also indicate that small-scale convection occurs in the rift zone, and that colder (or high seismic wave velocity-) material appears to be present at both sides of the rift. Numerical models of mantle flow beneath a continental rift predict small scale convection in the rift zone. Material flows upward under the center of the rift and downward at its sides, resulting in small temperature variations. The mantle material that fills the rift zone is supplied from the asthenosphere and flows from both sides into the rift zone. Extending the lithosphere to the point of breakup shows that the upwelling velocity increases toward the end of the continental rift phase (late syn-rift). After continental break up, the upwelling decreases and a mid-ocean ridge system is formed. The varying upwelling pattern during the different phases has implications for magmatism during rifting; it predicts for example that most melt is formed at the late-syn-rift stage.
T11C-0722
Is Input From the Deep Mantle Needed to Feed the Global Mid-Ocean Ridge System?
Over the past 20 m.y. the slabs that surround the Pacific basin have been retreating towards the center of the basin, causing the basin to shrink in size. The volume of upper mantle under the Pacific bound by these retreating slabs has therefore decreased. Since the upper mantle is relatively incompressible, the mantle must flow out of the basin. There are a few tectonic settings where the mantle can escape the Pacific basin system such as, mid- ocean ridges, slab windows, or hot spot volcanism. We assume that density differences across the 660 km discontinuity are too great for mantle to flow down without a density increase or driving force. Using the overall volume change of the Pacific basin, we have determined that Pacific mid-ocean ridge output only accounts for 19 percent of the displaced mantle volume down to the 440 km discontinuity barrier. Because hotspot volcanism output in the Pacific for the last 20 m.y. is relatively small, it is assumed that large quantities of mantle are escaping out of the Pacific basin and into the Atlantic and Indian ocean basins. Therefore, global volumes of mid-ocean ridge vs. displaced mantle must be calculated to determine the extent of lower mantle input. There are two possible outcomes of determining the ratio of displaced mantle to mid-ocean ridge output: 1) If all displaced upper-mantle is accounted for at all mid-ocean ridges, then the global upper mantle/ridge system is in equilibrium and no input for the deep mantle is needed to feed mid-ocean ridges. 2) If mid-ocean ridge output is greater than the displaced upper-mantle then there must be a finite amount of input from the deep mantle need to feed the ridges, yielding a global ratio of upper mantle to lower mantle input for mid-ocean ridge volcanism.
T11C-0723
Mechanisms of Topographic Uplift for the Southern Rocky Mountains
Current research into the timing and dynamics of topographic evolution in the Southern Rockies suggests that the land surface has undergone significant (up to 2 km ) and recent (c. 10Ma) uplift. Lines of evidence supporting this theory include documentation of profound incision along the river gorges draining the area (Colorado, Gunnison, Arkansas), a radial drainage pattern indicative of centralized uplift, and young topographic landscape characterized by high topographic roughness. The geographic focus of this uplift appears to be centered on the Aspen anomaly of central Colorado - an observed upper mantle low velocity domain similar in terms of spatial scale and velocity contrast magnitude to Yellowstone. This anomaly underlies the highest and roughest topography in the Rocky Mountain region and is suggestive that upper mantle processes are driving active tectonic uplift. Recent seismic results from the CD- ROM and La RISTRA experiments have indicated that both active asthenospheric processes and more ancient compositional heterogeneity within the lithosphere are influencing the tectonics of the region. However, while the tectonic nature of the Yellowstone plume has been the focus of a number of recent studies including subsurface imaging of the upper mantle, the nature of the upper mantle beneath the southern Rockies is still poorly understood. This knowledge gap will be rectified by data collected by the recently NSF-funded CREST project that will provide detailed seismic images beneath the Southern Rockies by 2010. Until the availability of such data we are using a twofold geodynamical modeling approach to evaluate the plausibility of a similar tectonic origin for the Aspen anomaly and Yellowstone. First, from an isostatic perspective the estimated crustal thicknesses of 45-50 km in the central Rockies may be insufficient to fully support average elevations of 3-4 km implying a significant mantle contribution to isostatic equilibrium. Secondly, we have used a finite element model (CITCOM) to evaluate the predicted topographic uplift due to a shallow plume impinging on the base of the lithosphere as well as the uplift associated with delamination of the upper mantle. Preliminary results suggest that the observed topography can well be explained by dynamic topography from a small plume or thermal anomaly in the upper mantle and supporting the notion that Aspen and Yellowstone both have origins in the upper mantle. Such a thermal origin for the Aspen Anomaly is further supported by geochemical evidence from hot-springs in the region.
T11C-0724
Plate Motions and the Viscosity Structure of the Mantle
The viscosity structure of the Earth's mantle is likely to play an important role with respect to the motion of the lithospheric plates. A zone of low viscosity in the upper mantle has been proposed to facilitate plate motion. Another typical feature of the viscosity profile in the mantle is a significant viscosity increase at greater depth. We have employed a three-dimensional mantle convection model to explore the relation between the appearance of plate motion and the viscosity structure of the mantle beneath. With this model we further elucidate the mechanisms relevant for the formation of a low or high viscosity zone. The model allows for a complex rheology of the fluid (strong temperature, pressure and stress dependence of the viscosity) and can so account for the self- consistent formation of plates at the surface of the convecting mantle. As a general result we observe that a delicate balance between temperature, stress and pressure dependence of the viscosity is required to obtain stable plate motion and that within this balance the viscosity shows a local minimum at shallow depth (low viscosity zone) and a local maximum at greater depth.
T11C-0725
The Effect of Subduction Events on Rising Mantle Currents
Mantle convection systems with periodic events of subduction exhibit two different kinds of upwelling currents: plumes and thermals. Plumes are strong upwelling currents with a well-defined head and tail structure. Thermals having only little buoyancy do not develop this typical structure and are strongly influenced by the large- scale convection. In vigorous convection thermals are driven towards the locations of the plumes. In this study we numerically investigate the structure and evolution of the rising mantle currents in a 2D Rayleigh- Benard set-up. Periodic events of the subduction occur due to the application of a temperature- and stress- dependent rheology. Model calculations in a Cartesian geometry with large aspect ratios are run with variations in the Rayleigh number, the viscosity contrast (measure of the temperature dependence) and the yield stress (measure of the stress dependence). Our study indicates that a system with periodic subduction events is characterised by three phases. In the first phase the cold thick lid at the surface has no influence on the upwelling currents. Strong, stationary plumes drive the convection. In the second phase the subducted cold material suppresses upwellings. Only small transient currents rise in front of the cold slab. In the third phase the cold material at the bottom is heated up allowing for the formation of strong upwelling currents again.
T11C-0726
Introducing tectonically and thermo-mechanically realistic lithosphere in the models of plume head –lithosphere interactions (PLI) including intra-continental plate boundaries.
Plume-Lithosphere Interactions (PLI) in continets have complex topographic and magmatic signatures and are often identified near boundaries between younger plates (e.g., orogenic) and older stable plates (e.g., cratons), which represent important geometrical, thermal and rheological barriers that interact with the emplacement of the plume head (e.g., Archean West Africa, East Africa, Pannonian – Carpathian system). The observable PLI signatures are conditioned by plume dynamics but also by complex rheology and structure of continental lithosphere. We address this problem by considering a new free-surface thermo-mechanical numerical model of PLI with two stratified elasto-viscous-plastic (EVP) continental plates of contrasting age, thickness and structure. The results show that: (1) surface deformation is poly-harmonic and contains smaller wavelengths (50-500 km) than that associated with the plume head (>1000 km). (2) below intra-plate boundaries, plume head flattening is asymmetric, it is blocked from one side by the cold vertical boundary of the older plate, which leads to mechanical decoupling of crust from mantle lithosphere, and to localized faulting at the cratonic margin; (2) the return flow from the plume head results in sub-vertical down-thrusting (delamination) of the lithosphere at the margin, producing sharp vertical cold boundary down to the 400 km depth; (3) plume head flattening and migration towards the younger plate results in concurrent surface extension above the centre of the plume and in compression (pushing), down-thrusting and magmatic events at the cratonic margin (down-thrusting is also produced at the opposite border of the younger plate); these processes may result in continental growth at the "craton side"; (4) topographic signatures of PLI show basin-scale uplifts and subsidences preferentially located at cratonic margins. Negative Rayleigh-Taylor instabilities in the lithosphere above the plume head provide a mechanism for crustal delamination. In case of several cratonic blocks, the combined effect of subsidence and lithospheric thinning at cratons edges, while plume head material is being stocked in between the cratons, favours major magmatic events at cratonic margins. Numerous field evidence (West Africa, Western Australia) underline the trapping effect of cratonic margins for formation of (e.g.) orogenic gold deposits, which require particular extreme P-T conditions. Location of gemstones deposits is also associated with cratonic margins, as demonstrated by the Tanzanian Ruby belt. Their formation depend on particularly fast isothermal deepening processes, which can be reproduced by slab-like instabilities induced by plume head-cratonic margin interaction. On the other hand, absence of magmatic events should not be interpreted as evidence for the absence of plume: at surface, these events may not necessary have unambiguous deep geochemical signatures, as the hot source plume material stalls below Moho and forms a long-lasting (10 to 100 Myr) sub-Moho reservoir. This should induce strong crustal melting that may overprint deeper signatures since crustal melts are generated at much lower temperatures than mantle, and produce light low-viscous rapidly ascending magmas. Drip-like down- sagging of the lithospheric mantle and metamorphic lower crustal material inside the plume head may contaminate the latter and also alter the geochemical signature of related magmas.
T11C-0727
Faults induced by Slab Bending: A Consequence of Pressure-Dependent Plasticity Coupled with Dislocation Creep
Modeling of slab bending is of crucial importance in geodynamics and earthquake hazard issues, since this process is always associated with subduction and is related to the structural and seismic features inside the slab at the trench area. Of special interest is bending-related faulting of the incoming plate creating a pervasive tectonic fabric that cuts across the crust, penetrating deep into the mantle. Faulting is active across the entire ocean trench slope, promoting hydration of the cold crust and upper mantle surrounding these deep active faults. The along-strike length and depth of penetration of these faults are also similar to the dimensions of the rupture area of intermediate-depth earthquakes. We performed systematic numerical modelling of spontaneous bending process of incoming plate during subduction. In the beginning of this experiment two plates of different ages are juxtaposed together along the transform fault with low plastic strength creating favorable conditions for spontaneous initiation of subduction and the concurrent slab bending. Deformation pattern in slab bending area is quite distinct: the top of the slab is subjected to intense plastic deformation with localized faults zones while the bottom of the slab is deformed in a ductile way by dislocation creep with enhancement of the deformation due to high stresses in the bending area. The plastic deformation and dislocation creep fields are, respectively, characterized by extension and compression in a horizontal direction. These two fields are clearly separated by the narrow non-deforming area in the core of the slab characterized by low deviatoric stresses. Results of our experiments suggest that in case of free surface condition atop the slab it can be easily bent by its own weight triggering spontaneous retreating subduction. The bending process is facilitated (i) by lowered pressure in the extension region favouring deep penetration of faults and (ii) by high stresses in the compression region resulting in local lowering of slab viscosity due to the power-law nature of dislocation creep. The large depth of penetration of localized faults (10-50 km) and their uniform preferential dip are caused by pressure-dependent rather then depth-dependent plasticity. On the other hand dislocation creep in the deep portion of the slab promote lateral narrowing of bending zone ,causing the appearance of slab bending faults within localized 100- 150 km wide area atop the slab.
T11C-0728
Mantle Evolution Beneath Supercontinents Since the Archean
Supercontinents promote mantle warming and associated large-scale magmatism. Mechanisms for this feedback include the clustering of mantle plumes and lithospheric forcing of convective length scales. Modeling studies show that long-wavelength convection inherent to an internally heated mantle with no active plumes and a supercontinent produces mantle temperature anomalies of +100 °C. Such global mantle warming could have sourced the Central Atlantic Magmatic Province during the breakup of Pangea. However, the sensitivity of the mantle to smaller supercontinents that might have existed during the Archean when total continental area amounted to only 10--20% of the Earth's surface is unknown. Here we use 3D spherical mantle convection models with continents to show that convective length scales, and hence subcontinental temperatures, increase concurrently with continental area. Plumes magnify warming, leading to a temperature excess of ~200 °C. We also find that significant broadly distributed heating occurs beneath diminutive supercontinents appropriate to the Archean, even in purely internally heated models. This result could help to explain the pulse in continental growth at 2.7 Ga.
T11C-0729
Competing Effects of Ridge Proximity and Slab Strength on Slab Detachment
The detachment of lithosphere from a subducted plate causes observable changes in subduction zones by affecting mantle dynamics and plate tectonics. Knowledge of the dynamics of slab detachment and its signature in the overlying surface is necessary for a complete understanding of the deformation and fate of subducted slabs and may provide constraints on the rheology of subducting lithosphere and the upper mantle. Early evidence supporting slab detachment includes observations of gaps in hypocentral distributions and within tomographic images of subducted slabs. The process has also been called upon to explain observed magmatism, uplift, extension, and metamorphics, but few studies have explored or demonstrated a link between these processes and slab detachment. A proposed cause of slab detachment is the approach of buoyant features such as a spreading ridge to subduction zones, a scenario that has previously not been tested using fully-dynamic numerical models including a non-Newtonian rheology. We present dynamic two-dimensional models of ridge subduction exploring the dependence of detachment occurrence and resultant surface effects on subducted slab length, distance of the ridge from the trench, spreading rate, and lithospheric yield strength. We compare model results with surface observations in Baja California, where the approach of the East Pacific Rise to the trench may have led to detachment of the Cocos slab.
T11C-0730
Dynamics of Continental Plate Delamination: a Numerical Study Applied to the Northern Apenninic Belt.
Suduction zone are sites on the earth surface where a gravitationally unstable lithosphere sinks into the mantle. Normally, cold oceanic plates are good candidates for such behaviour, thought subduction of young slabs is sometimes observed (i.e., Cascadia and Nankai subduction zones). However, other ¢®¡Æanomalous¢®¡¾ subduction zones exist where the subducting plate is a gravitationally stable continent. How is it possible, hence, to subduct continets (or at least part of them)? What are the dynamical processes that occur when weak and buoyant material as the continental crust enter the trench? To answer to these intriguing questions we perform 2D numerical models of an oceanic plate with the attached continental passive margin subducting at 2 cm/yr under a fix continent. As collision occurs, weak (Wet Quartzite flow law) upper crust is deeply subducted and hydrated mantle wedges between the plates through deformation of the low viscosity material. The plates have now lost completely coupling and the lower plate, pulled by the oceanic slab, starts to retreat. The buoyant and weak upper crust is scraped off by the passively advancing astenospheric wedge so that the lower plate increases its negative buoyancy. The upper crust undergoes two deformational phases with early compressional structures being later dissected by extension. The compressional front is characterized by thick crust and low heat flux, while the extensional domain has shallow mantle and high heat flux. Results of the model are on good agreement with the geological and geophysical data collected in the Northern Apennnines, supporting, hence, the retreating and delaminating continental plate model for the last 20 Ma evolution of this belt.
T11C-0731
Ready to Detach: new insight into the geodynamics of northern Apennines.
Due to their near-alignment with the relative plate motion between African and Eurasian plates, the Apennines mountain range requires a local geodynamic mechanism. A subduction zone retreat scenario is often invoked, with the lithosphere under the Adriatic sea descending beneath the lithosphere of the Tyrrhenian sea. Many surface expressions of a retreating subduction zone are found in the Apennines, such as the seismic expression of a lithospheric slab at depth, juxtaposed belts of compression and extension in the crust, a difference in crustal thickness consistent with thinning the overriding plate, and evidence for lateral translation of the orogen over time. In northern Apennines the lack of active volcanism and deep seismicity suggest an interruption of the subduction process, or else encourage models of lithospheric removal that leave the upper crust in place. New seismological observations obtained in the northern part of the Apennines also encourage scenarios independent of steady-state subduction. The indicators of upper mantle deformation derived from two different seismological techniques (shear-wave splitting and surface-wave mode conversion) suggest a pattern of upper mantle deformation that differs in significant ways from that predicted by the subduction zone retreat scenario. We compare the inferred mantle flow pattern with tomographic images, trench position reconstructions and deep crustal seismicity, and argue for the incipient-detachment scenario in northern part of the Apennines. Specifically, we propose that upper mantle flow on the western side of northern Apennines (beneath the Tyrrhenian Sea) likely responds to the retreat of the Adriatic subduction zone further south, while deformation on the eastern side is better explained by an incipient detachment of the Adriatic mantle lithosphere. http://earth.geology.yale.edu/RETREAT/
T11C-0732
Seismic Strong Motion Array Project (SSMAP) to Record Future Large Earthquakes in the Nicoya Peninsula area, Costa Rica
The seismic strong motion array project (SSMAP) for the Nicoya Peninsula in northwestern Costa Rica is composed of 10 – 13 sites including Geotech A900/A800 accelerographs (three-component), Ref-Teks (three- component velocity), and Kinemetric Episensors. The main objectives of the array are to: 1) record and locate strong subduction zone mainshocks [and foreshocks, "early aftershocks", and preshocks] in Nicoya Peninsula, at the entrance of the Nicoya Gulf, and in the Papagayo Gulf regions of Costa Rica, and 2) record and locate any moderate to strong upper plate earthquakes triggered by a large subduction zone earthquake in the above regions. Our digital accelerograph array has been deployed as part of our ongoing research on large earthquakes in conjunction with the Earthquake and Volcano Observatory (OVSICORI) at the Universidad Nacional in Costa Rica. The country wide seismographic network has been operating continuously since the 1980's, with the first earthquake bulletin published more than 20 years ago, in 1984. The recording of seismicity and strong motion data for large earthquakes along the Middle America Trench (MAT) has been a major research project priority over these years, and this network spans nearly half the time of a "repeat cycle" (~ 50 years) for large (Ms ~ 7.5- 7.7) earthquakes beneath the Nicoya Peninsula, with the last event in 1950. Our long time co- collaborators include the seismology group OVSICORI, with coordination for this project by Dr. Ronnie Quintero and Mr. Juan Segura. The major goal of our project is to contribute unique scientific information pertaining to a large subduction zone earthquake and its related seismic activity when the next large earthquake occurs in Nicoya. We are now collecting a database of strong motion records for moderate sized events to document this last stage prior to the next large earthquake. A recent event (08/18/06; M=4.3) located 20 km northwest of Samara was recorded by two stations (Playa Carrillo and Nicoya) at distances of 25-30 km with maximum acceleration of 0.2g.
T11C-0733
Ultrahigh Pressure Metamorphic Terrane Evolution; Norwegian Caledonides
Rocks in Norway's Western Gneiss Region (WGR) experienced high pressure and ultrahigh pressure (UHPM) (4GPa., 800C) peak metamorphic conditions during the Scandian orogeny at 410Ma. Thermobarometric studies of exhumed ultramafic eclogite pods from the Nordfjord, Soroyane and Nordoyane areas place tight time constraints on subduction, UHP metamorphism and exhumation, with all but the final phase of exhumation occurring in ca. 12 million years. However, few structures apparently related to the descent phase of terrane evolution were observed during field studies. Rather, ubiquitous quartz-rod lineation and pervasive minor folding indicate top-to-the-west, relatively shallow unroofing of the subducted margin as indicated in a new bedrock map of a portion of the Norwegian coast. Many of the mapped units have been redescribed, with emphasis put on those features that are of interest to the geophysical community.. To address the ambiguous kinematics of UHPM evolution, numerical models are employed in this study to consider the trajectory of crustal materials during continental collision that concentrate on the delicate balance of forces driving and resisting the subduction of buoyant continental materials as a function of kinetically-controlled equilibration.. In the WGR, past stability of coesite and rarely, of diamond, is preserved in robust mafic eclogites as inclusions within zircon and garnet grains. However, the extent of UHPM equilibration of the volumetrically dominant quartzo-feldspathic gneisses and consequently the contribution of these lithologies to the overall subduction suystemare unclear. . As such, simple equilibrium- defined strength and density parameters are insufficient to define natural model behavior. (Meaning of this next sentence escapes me. How does the following sound?) Rather, numerical solutions involving end member and intermediate states between equilibrium and non-equilibrium assemblages are explored While UHP metamorphic reactions in the oceanic slab and lithospheric mantle serve primarily to increase the magnitude of driving forces, complete UHPM equilibration within continental gneisses causes runaway subduction, with no UHP rocks returning to the surface. Model material softening during simulated eclogitization of key boundary layers serves to decouple buoyant crustal material from the sinking slab and continental mantle. This decoupling allows for the initiation of crustal exhumation during continued convergence and subduction, as observed in the WGR.
T11C-0734
Subduction of Subarc Mantle Peridotites: Evidence From the Higashi-akaishi Garnet- peridotites in the Sanbagawa Metamorphic Belt, Japan
Garnet-bearing peridotites have been reported from continental collision zones, but are very rare in oceanic subduction complexes. So far reported examples are boulders in a stream in northern Dominican Republic (Abbot et al., 2006) and a kilometer-scale unit of garnet peridotites (2 x 5 km) at Mt. Higashi-akaishi in the Sanbagawa belt in the southwestern Japan. The Sanbagawa belt formed during the oceanic subduction along the eastern Asian margin in Cretaceous time. The metamorphic history of the peridotites in the Sanbagawa belt has been well established by previous work, but the protolith is contentious. Proposed origins include cumulates of mafic magmas, mantle wedge peridotites, and the base of an oceanic plateau. The Higashi-akaishi peridotite body is composed mostly of anhydrous Ol-rich rocks with minor lenses of Cpx- rich rocks. Ol-rich rocks show a refractory geochemical signature with high Cr (>3000 ppm), MgO(>43 wt%)and Ir-type PGE(>5ppb) with low Al2O3(<0.5 wt%), and TiO2 (<0.05 wt%). Their refractory nature is further supported by high Cr# (>0.75) in spinel and high Mg (Fo=90-93) in Ol, which plot in the refractory part of the Ol-Sp mantle array. Cpx-rich rocks contain low Ni (<1000 ppm), low Ir-type PGE (< 2ppb) and show a typical gsubduction-related signatureEwith high fluid-mobile elements and low HFSE, suggesting that Cpx- rich rocks are cumulates of arc magmas. Considering that the rocks had underwent a sharp increase in P from less than 2 to greater than 4 Gpa during the prograde metamorphism (Enami et al., 2004), we suggest that they originated in the root of an arc at the depth greater than 50 km. The sliver of the subarc mantle peridotites was entrained by a mantle flow towards the trench, incorporated into the Sanbagawa subduction channel and subducted to a deeper than 100 km depth before exhumation. Such major mantle corner flow is compatible with the lack of a lubricating serpentinite layer during early high-temperature subduction. The condition may have been maintained by massive eastward flow of asthenospheric mantle from the northeastern Asia in Cretaceous time when the Sanbagawa belt began to form.
T11C-0735
The Nature of Mantle Sources for Perm-Triassic Traps of Western Siberian Plate and Siberian Platform
This work deals with the consistent dynamics for the development of melting zones due to convection in the upper mantle and the adjacent melting zones in inhomogeneous lithosphere. This model is studied for magma systems of Western Siberian Plate and Siberian Platform (SP). The results of numerical modeling allow us to conclude the following: 1. Structural and petrochemical zoning for Perm-Triassic trap magmatism of the Siberian Platform can be explained due to two-level melting of mantle rocks over hotspots (McKenzie, 1984). An alternative approach to explain mechanism of trap magmatism because of hydrodynamic evolution of magma under inhomogeneous lithosphere (King, Anderson, 1995) does not allow us to obtain voluminous melting of the mantle rocks for the lithosphere. 2. Main volumes of tholeiitic magmas (which compose both the intrusives in the SP platform cover and SP profiles of lava shield) are due to melting of mantle rocks in the thinned parts of the lithosphere over a number of spread hotspots. 3. The area of partial melting in the lithosphere is appearing after the zone of decompressing melting developed to its maximum extent, under the conditions of convection in the upper mantle. The melting zone appearing at the depth about 70-100 km has a plate-like shape, thickness up to ~20-30 km and horizontal extent coexistent with the sizes of the lava shield. This research was supported by the President's grants NSh-1573.2003.5, and by the Russian Ministry Science and Education grant RNP.2.1.1.702.
T11C-0736
Development of three dimensional Eulerian numerical procedure toward plate-mantle simulation: accuracy test by the fluid rope coiling
Reproducing a realistic plate tectonics with mantle convection simulation is one of the greatest challenges in computational geophysics. We have developed a three dimensional Eulerian numerical procedure toward plate-mantle simulation, which includes a finite deformation of the plate in the mantle convection. Our method, combined with CIP-CSLR (Constrained Interpolation Profile method-Conservative Semi-Lagrangian advection scheme with Rational function) and ACuTE method, enables us to solve advection and force balance equations even with a large and sharp viscosity jump, which marks the interface between the plates and surrounding upper mantle materials. One of the typical phenomena represented by our method is a fluid rope coiling event, where a stream of viscous fluid is poured onto the bottom plane from a certain height. This coiling motion is due to delicate balances between bending, twisting and stretching motions of fluid rope. In the framework of the Eulerian scheme, the fluid rope and surrounding air are treated as a viscosity profile which differs by several orders of magnitude. Our method solves the complex force balances of the fluid rope and air, by a multigrid iteration technique of ACuTE algorithm. In addition, the CIP-CSLR advection scheme allows us to obtain a deforming shape of the fluid rope, as a low diffusive solution in the Eulerian frame of reference. In this presentation, we will show the simulation result of the fluid rope coiling as an accuracy test for our simulation scheme, by comparing with the simplified numerical solution for thin viscous jet. http://www.es.jamstec.go.jp/esc/research/Solid/index.en.html
T11C-0737
80-Myr history of buoyancy and volcanic fluxes along the trails of the Walvis and St. Helena hotspots (South Atlantic)
Walvis and St.~Helena are the only long-lived hotspot chains in the South Atlantic. Therefore, their characterization is important to constrain the processes associated with mantle plume formation, their temporal evolution, and the interaction with plate and mantle dynamics in the region. We study the temporal evolution of plume buoyancy and magma production rate along both hotspot chains, which are constrained from the swell and volume of volcanic materials emplaced along the chain. The regional depth anomaly is calculated by correcting the 2' bathymetry grid of Smith & Sandwell (1997) for thermal subsidence and sediment loading. We separate the topography associated with volcanism and the swell surrounding the hotspot chains using the MiFil filtering method (Adam et al., 2005). We then estimate the temporal variations associated with both parameters by computing volumes along the hotspot tracks. Neither Walvis nor St.~Helena show a 'classical' hotspot behavior. We find that two plumes are at the origin of the St.~Helena chain. This study also shows a swell associated with the Circe seamount, supporting the existence of a hotspot NW of the St.~Helena trail. The variation in swell and volcanic fluxes suggests temporal variability in the plume behavior at time scales of 10-20~m.y. and 5~m.y., which may be related to oscillations and instabilities of the plume conduit, respectively. Cumulative fluxes in the area are largest for Walvis and weakest for Circe, and all are significantly lower than that reported for the Hawai'i hotspot.
T11C-0738
Supercontinent Formation in 3-D Spherical Mantle Convection Models With Multiple Continental Blocks
Much of the large-scale tectonics on the Earth in the last Ga is predominated by the assembly and breakup of supercontinents Rodinia and Pangea. However, the mechanism that is responsible for supercontinent formation remains poorly understood. Zhong et al [2007] recently showed that mantle convection with moderately strong lithosphere and lower mantle is characterized by a largely degree-1 planform in which one hemisphere is predominated by upwellings while the other by downwellings. They further suggested that the downwellings should attract all the continental blocks to merge in the downwelling hemisphere, thus leading to supercontinent formation there. However, Zhong et al. [2007] did not consider drifting and collision processes of continents. In this study, we explore the supercontinent formation mechanisms by including drifting and collision processes of multiple continental blocks in 3-D spherical mantle convection models. We use thermochemical CitcomS code to model 3-D spherical mantle convection with continental blocks. In our models, particles are used to represent continents and to track their motions. We found that for models with mantle viscosity (i.e., moderately strong lithosphere and lower mantle) that leads to degree-1 convection as reported in Zhong et al. [2007], initially evenly- distributed continental blocks always merge to form a supercontinent on a time-scale of about 6 transit times (i.e., corresponding to about 300 Ma). The hemisphere where a supercontinent is formed is predominated by downwellings as continents merge towards there, while the other hemisphere by upwellings. However, after the supercontinent formation, upwellings are generated beneath the supercontinent. This scenario is qualitatively consistent with what Zhong et al. [2007] proposed. We also found that while some convection models with intrinsically small-scale planforms may also lead to formation of a supercontinent, some other models may fail to produce a supercontinent. For these models with intrinsically small-scale planforms, the merged continental blocks promote long-wavelength mantle structure near the continents. However, in non-continental regions, convective wavelengths remain relatively small. We suggest that time-scales for supercontinent formation and convective wavelengths in non-continental area are important parameters that help constrain mechanisms for supercontinent formation.
T11C-0739
Dynamic Effects of Long-Wave Density Gradients due to Internal Heating
Though it is estimated that 50--80% of the heat lost from the earth is generated internally within the mantle, our conceptual picture of mantle flow is based on simplified boundary layer models for bottom heated convection, in which the temperature is adiabatic outside thin boundary layers, and the flow is driven purely by the excess mass of cold plumes (subducting slabs). That picture underlies simplified models in which mantle flows are calculated from horizontal density gradients inferred from seimology. As a reminder of the assumptions underlying that approach, we give a solution of the coupled fully nonlinear Boussinesq equations describing horizontal flow in a horizontal layer of viscous but thermally non--conducting fluid with uniform internal heating. The shear stress and heat flux both vanish at the top and bottom of the layer; the flow is driven purely by the small-- amplitude long--wave baroclinicity due to internal heating, and heat generated internally is removed by the purely horizontal flow. This solution models flow near the centre of a large--aspect ratio convection cell, like that occurring beneath the Pacific plate. (There, heat generated internally is first transported horizontally over the cell length, then removed at the cell end by mixing of the heated material with the cold subducted slab. Our solution models only the first of those processes.) Although our solution does not contain a cold plume, it predicts that the small--amplitude long--wave density field due to internal heating can, by itself, generate velocities close to those observed. Simplified mantle flow models that assume the motion to be driven purely by seismically observable density differences across slabs may be ignoring an essential part of the forcing. Morris, S. J. S. GRL, 34,doi:10.1029/2007GL030059