U53A-01
Non-closure of the geologically instantaneous global plate motion circuit: Implications for plate non-rigidity
Global plate motion models such as NUVEL-1 provide rigorous tests of the assumption of plate rigidity and appear, at first glance, to strongly validate that assumption. But the complex interactions of the plate circuits in such global models can hide important inconsistencies. Here we make a straighforward comparison of Pacific-North America motion estimated through three non-overlapping sets of data: (1) plate motion data (spreading rates, fault azimuths, and slip vector azimuths) along the Pacific-North America boundary, (2) space geodetic data, and (3) plate motion data along the Pacific-Antarctic-Nubia-North America plate motion circuit. We find that all three estimates are inconsistent with one another and, in particular, the plate circuit results are highly inconsistent with the two directly estimated results, being both of the wrong magnitude and in the wrong direction. The inconsistency is large and described by an angular velocity of non-closure with a magnitude 0.11 ± 0.03 degs/Myr (95% confidence limits). Projected into the Gulf of California, the non-closure is a velocity of 12 ± 3 mm/yr toward 196. If not owing to an unrecognized systematic error, this non-closure is presumably caused by non-rigidity of one or more of the North American, Nubian, Antarctic, and Pacific plates. We consider several hypotheses to explain the inconsistency. We can conceive of no plausible diffuse plate boundary to explain the difference. In particular, deformation in western North America is excludable as the cause of the difference as both direct plate motion data and space geodetic data differ highly significantly from the predictions from the circuit. The bias in rates of seafloor spreading caused by outward displacement of magnetic anomaly reversal boundaries (DeMets and Wilson, 2005) is in the right direction but can account for only ~15% of the inconsistency. Much of the inconsistency may be explainable, however, by horizontal thermal contraction of oceanic lithosphere, especially of the young Pacific plate lithosphere flanking the Pacific-Antarctic Rise and East Pacific Rise. This young lithosphere includes a continuous band connecting Baja California to the Pacific-Antarctic Rise. The expected contraction rate of this young lithosphere is the right order of magnitude to explain the discrepancy. If this hypothesis is correct, simple plate tectonic predictions may need to be corrected by models fully incorporating the horizontal thermal contraction of the lithosphere to obtain the accuracy needed for some plate recontructions and for comparisons with some space geodetic data.
U53A-02
Three Dimensional Modeling of Lithosphere and Mantle Dynamics Elucidating Lithosphere-Mantle Coupling
We present a global study of joint modeling of lithosphere dynamics and three dimensional mantle circulation. Our study provides insight into the lithosphere-mantle coupling problem emphasizing the role of lateral viscosity variations in the asthenosphere that affect basal tractions associated with density buoyancy driven mantle convection. Forces driving lithospheric motion can be categorised into: (1) forces due to gravity acting on density variations within the lithosphere and (2) those acting at the base of the lithosphere (basal tractions) generated by density variations deeper than the lithosphere. Density variations within the lithospheric thin sheet give rise to horizontal gradients in gravitational potential energy (GPE) which in turn produce deviatoric stresses. These stresses coupled with the stresses from basal tractions contribute to a total deviatoric stress field that should match stress field indicators, such as the style of strain and directions of principal axes of strain from the Global Strain Rate Map (GSRM), as well as stress directions from the World Stress Map (WSM). We test different mantle density and viscosity models in order to find the best mantle-lithosphere coupling model, which, when added to the GPE solution gives a best fit to the stress field indicators from GSRM and WSM. The vertically averaged deviatoric stress field from GPE differences is constrained by gravity, topography, geoid and crustal thickness data. The mantle circulation models, which provide the basal traction field, are constrained by long wavelength observations of geoid, topography and plate motion, together with seismic tomography and the inferred history of subduction. Our modeling results indicate that the lithosphere-mantle coupling is laterally varying. In some areas, the GPE variations play a dominant role in generating the observed stresses and the lithosphere-mantle coupling is inferred to be weak, while in other areas stresses from basal tractions are required in addition to the stresses from GPE in order to satisfy the observed stresses. We will present observationally-constrained circulation and rheological models that would generate global traction fields capable of explaining the lateral variation of lithosphere-mantle coupling and the high-resolution observations of strain rate and stress at the Earth's surface.
U53A-03
Exploring Constitutive Relationships for Plate Boundaries Using Kinematic Models as Constraints.
Given the relentless struggles of mantle convection modelers over the years to produce emergent plate-like behaviour from their computational models, it is easy to conclude that the constitutive models of the lithosphere which have been employed in the past have been overly simple. That being said, is it not rather surprising that the kinematic picture of plate tectonics is such a good description of the oceanic lithosphere ? One feature of the plate kinematic picture is that it draws an absolute distinction between the plate interiors and the plate boundaries. It also distinguishes the 3 plate boundary types and ascribes entirely different behaviour to each one. In a modeling sense, the constitutive behaviour of each of the different plate boundary types is distinct. This is no surprise since constitutive relationships describe how the nature of small scale physical processes is manifest at some significantly larger scale; mid-ocean ridges, transforms and subductions zones are themselves complex physical systems with complicated responses to the changes in the tectonic stress field. In order to improve our understanding of the constitutive behaviour of the plate boundary zones we explore some simple systems of plates / plate boundaries where the kinematic evolution is straightforward. We use such systems to examine possible constitutive rules for different plate boundary types in isolation and then in combination
U53A-04
Subduction hinge migration: The backwards component of plate tectonics
There are approximately 50 distinct segments of subduction zones in the world, of which 40% have oceanic lithosphere subducting under oceanic lithosphere. All of these ocean-ocean systems are currently experiencing hinge-rollback, with the exception of 2 (Mariana and Kermadec). In hinge-rollback, the surface trace of the suduction zone (trench) is moving in the opposite direction as the plate is moving (i.e. backwards). Coincidentally, the fastest moving plate boundary in the world is actually the Tonga trench at an estimated 17 cm/yr (backwards). Although this quite important process was recognized soon after the birth of plate tectonic theory (Elsasser, 1971), it has received only a limited amount of attention (Garfunkel, 1986; Kincaid and Olson, 1987) until recently. Laboratory models have shown that having a three dimensional experiment is essential in order to build a correct understanding of subduction. We have developed a numerical model with the neccessary 3-D geometry capable of investigating some fundamental questions of plate tectonics: How does hinge-rollback feedback into surface tectonics and mantle flow? What can we learn about the forces that drive plate tectonics by studying hinge-rollback? We will present a quantatitive analysis of the effect of the lateral width of subduction zones, the key aspect to understanding the nature of hinge-rollback. Additionally, particular emphasis has been put on gaining intuition through the use of movies (a 3-D rendering of the numerical models), illustrating the time evolution of slab interactions with the lower mantle as seen in such fields as velocity magnitude, strain rate, viscosity, as well as the toroidal and poloidal components of induced flow. This investigation is well-suited to developing direct comparisons with geological and geophysical observations such as geodetically determined hinge retreat rates, geochemical and petrological observations of arc volcanics and back-arc ridge basalts, timing and distribution of metamorphic core complexes in backarc basins under extension, paleostress observables such surface movements and block rotations, observations of seismic anistropy determined by shear wave splitting, and the emerging studies of regional tomographic models of seismic anistropy.
U53A-05
Multiscale seismic tomography of mantle plumes and subducting slabs
Multi-scale (local, regional and global scales) tomographic studies are made to determine the 3-D velocity structure of the deep Earth, particularly for imaging mantle plumes and subducting slabs. Plume-like slow anomalies are clearly visible under the major hotspot regions in most parts of the mantle, in particular, under Hawaii, Iceland, South Pacific and Africa (Zhao, 2001, 2004). The slow anomalies under South Pacific and Africa have lateral extensions of over 1000 km and exist in the entire mantle, representing two superplumes. The Pacific superplume has a larger spatial extent and stronger slow anomalies than that of the Africa superplume. The Hawaiian plume is not part of the Pacific superplume (Zhao, 2004). The slow anomalies under hotspots usually do not show a straight pillar shape, but exhibit winding images, suggesting that plumes are not fixed in the mantle but can be deflected by the mantle flow. Strong mantle plumes under significant hotspots may originate from the CMB. However, there are some small-scaled, weak plumes originating from the transition zone or mid mantle depths (Zhao, 2004). Clear images of subducting slabs and magma chambers in the upper mantle wedge beneath active arc volcanoes are obtained, indicating that geodynamic systems associated with arc magmatism and back-arc spreading are related to deep processes, such as convective circulation in the mantle wedge and dehydration reactions of the subducting slab (Zhao et al., 1997; Zhao, 2004). Evidence also shows that arc magma and slab dehydration may also contribute to the generation of various types of earthquakes in subduction zones (Zhao et al., 2000, 2002). Most of the slab materials are stagnant in the mantle transition zone before finally collapsing down to the CMB as a result of large gravitational instability from phase transitions. The active intraplate volcanoes in East Asia continent (such as Changbai and Wudalianchi volcanoes) are not plume-related hotspots, but are a kind of back-arc volcanoes whose formation was closely related to the deep subduction of the Pacific slab and its stagnancy in the mantle transition zone (Zhao, 2004; Zhao et al., 2004; Lei and Zhao, 2005). The active Tengchong volcano in Southwest China is related to the subduction of the Burma microplate. Zhao, D. (2001) Seismic structure and origin of hotspots and mantle plumes. Earth Planet. Sci. Lett. 192, 251-265. Zhao, D. (2004) Global tomographic images of mantle plumes and subducting slabs: insight into deep Earth dynamics. Phys. Earth Planet. Inter. 146, 3-34.
U53A-06
Slab pull, slab weakening and their influence on great earthquakes, deep earthquakes and surface deformation
Viscous flow in the mantle, by coupling to Earth's lithosphere, ultimately determines plate motions, lithospheric stresses, and patterns of seismicity. To constrain the mantle's influence on these varied observables, we synthesize recent predictions of plate motions and lithospheric stresses determined from self-consistent models of mantle flow. We use these models of mantle flow, together with global models of lithospheric heterogeneity, to quantify the origin of the lithospheric stress field. These models suggest that the mantle signal is the strongest component of observed lithospheric stresses, but with large geographic variability. Strong lateral heterogeneity in mantle viscosity or in the rheology of the lithosphere may explain most of these geographic variations. We find in fact that strong cratonic regions couple directly to mantle flow. By accounting for the effects of viscous flow induced by slab buoyancy (lower mantle slab suction) and the slab directly transmitting its excess weight to the plate (upper mantle slab pull), we can entirely explain the difference in speed between subducting and non-subducting plates. The strength of slab pull has increased during the Cenozoic, resulting in an increase in the relative speed of oceanic plates. If the degree of pull and suction varies across subduction zones, we can explain, to first order, the observed variations in seismic moment release for large underthrusting earthquakes: at seismically uncoupled subduction zones, slabs are typically attached to plates and transmit nearly their entire upper mantle weight to the plate directly. At seismically coupled subduction zones that produce great earthquakes, however, slabs are nearly completely detached from their subducting plates. This suggests that slabs subducting in a compressional environment experience stress-induced weakening that prevents the effective transmission of the slab pull force. Indeed, we find strong correlations between strong plate-slab attachment and the presence or absence of back-arc basins. If we further compare moment release at depth with the degree of plate-slab attachment, we find positive correlations with the seismic moment released from intermediate and deep earthquakes. This implies that shallow slab weakening that occurs at trenches where compressive stresses (and great earthquakes) dominate, not only detaches slabs from plates, but is also maintained as the slab descends, discouraging deep seismicity. Our integrated view suggests that self- consistent models of mantle flow, plate driving forces and lithospheric stresses can lead to a greater understanding of interior dynamics and surface deformation.