U13A-0863
Computational techniques for kinematically constrained convection models
In order to unify plate kinematics with mantle dynamics we need to be able to apply plate-like boundary conditions to convection models in which the penalty for the model failing to satisfy the condition can be weighted according to the uncertainty in the kinematic reconstruction. Constraints typically include "rigidity" of oceanic plate interiors and prescribed plate motion vectors. An additional requirement is a kinematically-plausible dynamic model for the deformation at plate boundaries and regions where there is poor observational constraints on the kinematics. Numerically, the application of the rigidity constraint can be achieved by either penalizing the strain-rate in the plate interior or by prescribing that the velocities must be consistent with a rigid body motion. Plate reconstruction information is imposed upon the rigid plates by an additional penalty on the mismatch between the convection model plate velocities and those from the reconstructed plate motions. Here we first formulate the numerical techniques required to apply arbitrary and coupled constraints on mantle convection and to solve the resulting systems of equations efficiently. We then explore the effectiveness of each of these approaches.
U13A-0864 [WITHDRAWN]
A Hypothesis Explaining The Dynamics Of Plate Tectonics
Since the introduction of the plate tectonics theory, one of the unsolved fundamental problems has been the understanding of how and when plate tectonics began, as well as initial dynamics of plate tectonics although several mechanisms have been suggested (i.e. the hypothesis that asthenospheric convections drive plates, which is not supported by modern geodynamic theories). The kinematics behind the subduction zones and sea- floor spreading have been well understood and well described by the theory of plate tectonics. However, there is no consensus on the main driving mechanism for the plate tectonics and how it began. We are proposing a hypothesis that attempts to explain several fundamental questions: (1) Why is the planet Earth unique among the silicate planets of the solar system in terms of plate tectonics?; (2) How did plate tectonics start?; (3) What did drive the lithospheric plates in the first place?; (4) Why do mantle plumes form? What kinds of forces are responsible for the plumes?; (5) Why is the outer core fluid?; (6) Why do "ultra-low seismic velocity zones" exist at the base of the mantle?; and (7) Why does the magnetic field of Earth change? The other explanations and examples to be presented include three-dimensional animations of the Paleoproterozoic and Neoproterozoic Earth, and the kinematics and dynamics of the tectonic plates such as the Atlantic and Pacific plates.
U13A-0865
Stratigraphic and Geologic Constraints on Geodynamic Models of North America Since the Cretaceous
Stratigraphic and sea level data along with plate motions and mantle tomographic images are used as constraints on inverse and forward models of mantle convection beneath North America. Using plate reconstruction with continuously closed plate polygons (developed using the GPlates program), we compute forward and inverse models of mantle convection with the finite element method (CitcomS). With plate motions, seismic tomography is used to estimate mantle initial conditions using an adjoint of the energy equation, as described by Liu and Gurnis during this meeting. For North America, we developed a set of forward and inverse regional geodynamic models for last 100 million years, with plate movements consistent with global paleogeographic models. We used sediment isopachs, paleoshorelines, and tectonic subsidence curves as the primary stratigraphic and geologic constraints. Tectonic subsidence history for North America was the main constraint for inverse models. In forward mantle convection models, we used the prediction of dynamic topography, along with published eustatic curves, to model sediment isopachs, paleoshorelines, and tectonic subsidence. Modeling results are than compared to stratigraphic and geologic observations, and models are updated iteratively. In a fixed North American frame of reference, our models indicate that a dynamic topography low moved eastward from 100 million years to the present, due to of the descent of the Farallon slab as the plate moved westward. During the Cretaceous, the dynamic topography low was located in the middle of North American continent. During this period, eustatic sea-level reached its maximum and when combined with a dynamic topography low, caused a substantial inundation of North America, creating the Western Interior Seaway. At the present time, the dynamic topography low is located in the eastern United States, and we argue that this region is experiencing tectonic subsidence. This tectonic subsidence can explain significant departure of regional sea-level curves developed for the eastern United States, in comparison with eustatic sea-level curves. This presentation focuses on the details of the stratigraphic data and their integration with geodynamic models.
U13A-0866
Global Reconstruction of Plate Boundaries: 0 to 55 Ma in the Indo-Atlantic Hotspot Frame of Reference
Plate kinematic constraints are integrated into a near global model of the first-order plates, as well as many smaller crustal and lithospheric fragments that serve as the basis for a new set of plate boundary distribution maps as a function of time. At present the maps extend back to 55 Ma. The maps employ an Indo-Atlantic hotspot frame of reference of Müller et al. (1993) that is not statistically different from a mean paleomagnetic frame of reference derived from the major, relatively undeformed, continents over this interval of time. Total reconsruction poles and estimated approximately instantaneous (3 m.a. averages) poles provide the kinematic foundations of this new set of plate boundary maps. The underlying kinematic motions are consistent over 83 Ma with limited variations in mean global half spreading rate at 26.5 plus/minus 4.2 mm/a, with a significant fraction of this variation occurring between C13 (approx. 33.06 Ma) and C6 (approx. 20.13 Ma) reflecting increased rate of Nazca-Pacific divergence. The is no simple secular trend evident in global half spreading rates, as has been inferred in the past. An interesting aspect of these reconstructions is variation in rate of plate boundary migration in this frame of reference. These variations imply potentially deeper mantle buoyancy controls on certain boundaries, and may allow independent identification of active versus passive segments of the mid- oceanic ridge system.
U13A-0867
Cenozoic Tilting of the Australian Continent due to Dynamic Topography
We investigate the possibility of a mantle-dynamic origin to account for the observed pattern of inundation of the Australian continent in the Cenozoic. Since the Paleocene, the Australian continent has experienced a series of regional marine incursions and regressions, which are inconsistent with the expected flooding history due to changes in eustatic sea level alone. During this time, the Australian continent has undergone no major episodes of mountain building or rifting which might account for these patterns of inundation. Since the Eocene, the Australian plate underwent rapid northward motion as the spreading rate at the South East Indian Ridge increased. As it moved northwards, the Australian plate moved away from a dynamic topography low caused by the sinking Gondwanaland slab beneath the South East Indian Ridge, and towards a dynamic topography low caused by subducted slab material in South East Asia. It is thought that these dynamic topography features at the southern and northern extremes of the Australian plate produce an underlying static and long wavelength dynamic feature over which the Australian plate has migrated through the Cenozoic. This dynamic feature should be expressed by an increase in the latitudinal asymmetry of the Australian dynamic signal. Estimates of the dynamic motion of the Australian plate since the Paleocene are made by matching observed patterns of marine incursion with models of marine inundation. Models of inundation are created by backstripping sediment from present-day topography and dynamic motion is quantified by the displacement needed to approximate the observed flooding according to eustatic sea level. We explore the trend of these displacements according to their paleo-position. Preliminary analysis suggests that the continent is influenced by a dynamic feature that is both temporally and spatially varying. We attempt to interpret the evolving dynamic topography field of Australia in the context of kinematic and 3-D dynamic models of the Australian region which provide an integrated explanation for the patterns of marine inundation in the Cenozoic.
U13A-0868
Which Earth for its True Polar Wander? Deciphering Subducted Plates and Upwellings Contributions
On geological timescales, True Polar Wander (TPW) is due to surface loads and mass deficits or mass rearrangement within the mantle, which disturbs the inertia of the planet. Earth presents a complex and changing distribution in density. Subducted plates are evident producers of these heterogeneity in the mantle. The history of density heterogeneities due to cold plates diving through the mantle has been reconstructed for the past 120 Myr by Ricard et al. (1993). After this work, we calculate the induced perturbations of the inertia tensor and their effect on the spin axis. But the computed wander appears large and lacking features comparing with paleomagnetic observations (Besse & Courtillot, 2002). Greff-Lefftz (2004) showed that the doming regime is the only upwelling structure able to produce large TPW as subducted plates do. We add the inertial perturbations due to simple models of the African and the Pacific superplumes to the ones due to subducted plates in order to see if they can stabilize the rotation pole at relatively high latitudes (paleomagnetism stands the pole between 60 and 90 degrees of latitude during the last 200 Myr). But the distribution of superplumes (one under each pacific and african hemisphere, in the lower mantle) imposes a position for their axis of maximum inertia roughly close to northern Atlantic. That is to say the superposition of the inertial effects due to subducted plates and doming regime would carry the pole in the atlantic area, towards the equator. Besides, a large amount of small plumes rises through the upper mantle on top of the superplumes. They should induce TPW along a meridian joining the two pacific and african hemispheres. We check off these plumes during the last 200 Myr. Modelling them simply, we take into account the perturbations of the inertia they induced in order to know if they are the essential component to stabilize the rotation pole as they are to explain the current geoid.
U13A-0869
Constraining Absolute Plate Motion by 40Ar/39Ar Age Dating and Geochemical Fingerprinting of Linear Volcanic Chains
Determination of reliable Absolute Plate Motion (APM) models remains a major frontier in global tectonics and geodynamics. However, APM models have been constructed with mixed success in the past, based on the geometry of linear volcanic chains and assuming a mantle reference frame of "fixed" hotspots. These reconstructions have been negatively impacted by the facts that the number of continuous volcanic chains decreases significantly back through geological time, become discontinuous and cannot be associated with active or zero-aged volcanoes. In reality, most APM models start to deteriorate for plate motion stages older than 40 Ma and become largely unconstrained prior to 70 Ma. In addition, the few seamount trails that were studied in detail often show disturbed age progressions, indicating that their volcanic evolutions were affected by processes other than hotspot volcanism or that their mantle plumes were not fixed in the Earth's mantle. Recent age data shows that the volcanoes themselves require build-up times of several millions of years, adding yet another complication in the reconstruction of APM models. The above difficulties show that the relationships between the morphology of linear volcanic chains, their age systematics and absolute plate motions are more complex than originally envisioned by Morgan in 1971. However, there are a series of observations and recent developments that make us confident that linear volcanic chains remain attractive targets for detailed geochronological and geochemical investigations that will help us to improve APM models. In fact, they may re-establish the basic assumption of an absolute reference frame with respect to which we can determine past plate motions, in particular, for periods older than 70 Ma. Recent denser sampling of seamounts has shown that some seamount chains retain very distinct geochemical fingerprints over tens of millions of years. Good examples are the EM2-type Samoan shield building lavas and the HIMU-type hotspots that formed the Cook-Austral chain. The latter hotspots can be extended into the Tertiary seamounts of the Gilbert Ridge and Tokelau Seamounts, and even the West Pacific Seamount Province which includes HIMU-type seamounts as old as 120 Ma. Correlating geochemistry with geochronology thus allows us to join segments of seamount trails that otherwise could not be connected, when based on their morphologies and ages alone. In turn, we have found an effective tool that we can apply in deciphering the history of intra-plate volcanism in the oceans, allowing us to reconstruct longer portions of their APM histories, up to at least 120 Ma.
U13A-0870
Plate circuit reconstructions and the consistency between Pacific and non-Pacific paleomagnetic reference frames in Late Cretaceous to Paleogene times (80--50 Ma)
Reconstructions of the Pacific plate relative to those bordering the Atlantic and Indian oceans have traditionally relied on the use of the fixed hot spot reference frame. However, ~13° of Hawaiian hot spot southward migration in the Late Cretaceous and Paleogene (81--47~Ma) recorded by paleomagnetic inclination data from the Emperor Seamounts, plate circuit reconstructions and simulations of convective mantle flow consistently suggest rapid motion between the Pacific and Indo-Atlantic hot spot groups during this time interval. This provides motivation to reexamine the history of Pacific plate motion with respect to the plates of the Indo-Atlantic realm using plate circuit reconstructions. Comparisons of Late Cretaceous to Paleogene paleomagnetic data from the Emperor Seamounts with non-Pacific reference poles reconstructed to the Pacific plate provide valuable insights both into veracity of plate circuit reconstructions and the resolution of paleomagnetic data sets. Here we present two alternative kinematic models based on the most recent and reliable plate-pair reconstructions linking the Pacific and Indo-Atlantic plates through a series of active and fossil spreading ridges: one running through West to East Antarctica, and another through the Lord Howe Rise to Australia. We find good agreement between Late Cretaceous (~75--80~Ma) Pacific and non-Pacific data that attests to the accuracy of the paleomagnetic estimates and plate circuit reconstructions. However, paleomagnetic data lack the resolution necessary to distinguish between the two alternative plate circuit models. In contrast, small discrepancies were observed in some comparisons of Paleogene data sets (~55--60~Ma). It will be argued that these differences reflect a systematic bias in average non-Pacific poles and/or minor non-dipole field contributions rather than errors in the plate circuits. The possible sources of these biases will be discussed.
U13A-0871
New Ar/Ar ages from the Ninetyeast Ridge, Indian Ocean: Beginning of a robust Indo-Atlantic hotspot reference frame
The Indo-Atlantic hotspot reference frame is one of the pillars of current plate reconstruction models. Most of the geodynamic literature (including the description of this Union session) assumes that this reference frame is relatively well-determined, at least for the last 80 Ma. Geochronological data sets -- principally Ar/Ar analyses of seamount and ocean island basalt -- are critical inputs to plate motion models. However, the quality of the existing Indo-Atlantic hotspot data sets, although sufficient for demonstrating general age progressive volcanism, is insufficient for robust determinations of absolute plate motion in general and tests of fixed hotspot frames in particular. Most of these Ar/Ar age determinations were made over 20 years ago; significantly better quality age data are now possible with the current generation of rare gas mass spectrometers and improved sample preparation techniques. Only after a significant number of new analyses on existing sample collections, and the collection of new sample suites where more detailed geographic coverage proves necessary, will we be able to accurately characterize the Indo-Atlantic hotspot reference frame, even for the last 80 Ma. Full development and testing of geodynamic models dependent on well-determined hotspot reference frame(s) -- fixed or moving -- will not be possible until this task is complete. Our research group is focusing on the Indian Ocean hotspot system, starting with the 5000 km long Ninetyeast Ridge (NER). Early work on NER basement cores recovered during DSDP Legs 22 and 26 and ODP Leg 121 showed generally age progressive volcanism, from greater than 80 Ma at Site 758 in the north to less than 40 Ma at Site 254 in the south. We have determined reliable Ar/Ar ages based on internally consistent incremental- heating age spectra from at least three aliquots of at least two distinct samples of volcanic basement from five of the six sites re-examined. The precision of our results is better than 0.5 m.y. at each site, i.e., less than the life span of typical hotspot shield volcanoes. Our new Ar/Ar ages show a remarkably linear age progression, from 77 Ma at Site 758 in the north to 42 Ma at Site 254 in the south. Most significantly, the overall rate of volcanic propagation along the ridge is 35 percent faster than previously suggested, necessitating a significant re- evaluation of current plate and hotspot motion models. The spatial resolution of the existing drill sites (6 sites along a 5000 km long ridge) is insufficient to test for likely spreading ridge jumps and microplate capture events, which also must be taken into account when deriving accurate models. However, we have recently recovered basalt from 23 of 33 new dredge sites along 3500 km of the NER, with a geographic distribution specifically designed to enable such tests. Further tests of hotspot models for the Indian hotspot reference frame as a whole will require a similar re-analysis of other volcanic lineaments, especially the Laccadives-Chagos Ridge.