U51B-0024
How lateral viscosity variations control the dynamics of western North America
We investigate the effects of lateral viscosity variations (LVVs) and long-range force transmission on the dynamics of the deforming North American lithosphere, with focus on the western region close to the plate boundary. We address the question how basal shear tractions as generated by mantle convection affect the stress field in North America on top of the inherited gravitational potential energy variations. Our efforts are partially motivated by the finding of Humphreys & Coblentz (2007) that an inversion for plate forcing requires cratonic pull and mantle tractions, but at a reduced amplitude from the predictions of an earlier model without LVVs by Becker & O'Connell (2001). The presence of old high viscosity cratonic lithosphere with underlying continental keel, along with weak plate boundaries, likely creates large LVVs, which will potentially have long- range effects on the deformation of the North American plate. We quantify these effects by computing basal tractions globally using a high resolution finite element mantle convection code, CitcomS, that can take into account several orders of lateral viscosity variations. Density anomalies within the lithosphere produce gravitational potential energy (GPE) differences, which also give rise to a stress field. The observed stress field is the combined effect of stresses from GPE differences and mantle tractions acting at the base of the lithosphere. One of our goals is to incorporate the effect of the shallow lithosphere within our model of mantle convection. The combined stress field will then be compared with stress observations such as the World Stress Map and strain-rates from the Global Strain Rate Map. Our goal is to find a global model that self-consistently explains both kinematic surface constraints and internal deformation indicators that is adapted to the North American plate.
U51B-0025
Crustal Buoyancies Dominate over Plate Boundary Effects in the U.S. Basin and Range: The Requirement of Weak Faults
We model the long-term dynamics of the seismogenic layer within the plate boundary zone of western North America, where the relevant time scale is 105 -– 106 years. We utilize a forward dynamic modeling approach, where the body force distributions, inferred lateral variations in effective viscosity, power-law behavior, and the known far-field velocity boundary conditions are defined. Body forces are the differences in gravity potential energy per unit area (GPE), calculated by performing depth integration of vertical stress from the surface down to a common depth reference (20 km below sea level). In our treatment of the seismogenic layer, depth-integrated viscosities are proportional to the assumed long-term friction on faults (expected deviatoric stress at fault failure) and inversely proportional to the long-term strain rates (the known fault slip rates). The velocity boundary conditions are defined using PA-NA, CO-NA, RI-NA, and JF-NA plate motion estimates. Self-consistent dynamic strain rate tensor solutions to the force-balance equations were solved for and scored with kinematic strain rate tensor and velocity fields of western North America, obtained from a large set of highly detailed Quaternary fault observations (e.g., United States Geologic Survey Quaternary fault and fold data base of the United States). We use fault data from present-day to 750 ka to investigate models using a range of long-term fault friction coefficients from 0.02 -- 1.0 under hydrostatic pore pressure conditions. Evaluation of fitness of the dynamic solutions to deformation indicators is achieved using two different measures. In one measure, the forward dynamic strain rate tensor styles are scored by misfit to the kinematic strain rate tensor styles inferred from Kostrov summation of Quaternary fault observations. In a second measure, the dynamic model velocity fields are scored via reduced chi-square misfit with the long-term kinematic model velocity field defined by Quaternary fault observations. Models constructed with low fault friction coefficients (μ < 0.20) achieve a vastly superior fit to Quaternary fault observations than do models with intermediate or high fault friction coefficients. Successful models require a weak and distributed fault fabric in which the deviatoric stresses associated with internal crustal buoyancies in the Basin and Range Province dominate over deviatoric stresses associated with velocity boundary conditions. The deformation field in western North America therefore suggests and supports the premise of a sufficiently dense fabric of faults that possess low long-term friction coefficients of 0.10--0.20.
U51B-0026
Something Old, Something New: EarthScope, Heat Flow and Gravity Too
EarthScope's transportable array (TA) of seismograph stations has the potential to vastly improve our
knowledge of internal mass distributions needed to constrain geodynamical models of the conterminous US.
Moho estimates are especially useful because they directly constrain the internal mass load plus lithospheric
deflection at that depth, placing stringent limits on possible loading scenarios. The nominal 70-km spacing of
the TA is ideal for flexural sampling but poses a problem for error processes. Iterative deconvolution and H-
K stacking used in EARS (Crotwell & Owens, 2005) receiver function estimates, for example, estimates
Moho from a grid search for crustal thickness H and Vp/Vs ratio K that maximizes stack amplitudes
at predicted phase arrival times. The method assumes a single uniform layer-over-halfspace, but the multiple
impedance contrasts and complicated 3D structure of the real Earth often yield several similar-amplitude
maxima, particularly if earthquake sources are few or poorly distributed. At 70 km spacing, standard spatial
correlation approaches do not help to distinguish which maximum corresponds to the Moho. I will present a
methodology for improved estimation of crustal thickness from H-K stacks. Confidence intervals from
optimal interpolation and gravity modeling of the receiver function estimates are used to calculate likelihood
ratios in H-K parameter space from the χ2 statistics of fit; these in turn are used to re-weight H-K
stacks. Importantly, this does not "force" the seismic data to fit some a priori expectation but rather
discriminates between more- and less-likely stack maxima. This bootstrap approach depends on (possibly
erroneous) values at nearby sites so is applied iteratively over all sites in the region of interest, beginning
with those deemed most improbable. Regressed gravity modeling parameters and semivariograms used for
OI evolve along with the estimates. For example, Moho density contrast Δρmoho changes from
75 to 300 kg m-3 over the course of six iterations, while the TA-scale RMS gravity residual drops from
71 to 50 mGal. Resulting estimates of crustal thickness variations are geodynamically intriguing, and the
gravity residual appears to be dominated by the geothermal mass field. I am currently evaluating strategies to
independently constrain geothermal mass in the modeling procedure using heat flow measurements, which
should both improve the resolution and further illuminate western US mass structure. The presentation will
also evaluate possible geodynamical interpretations of the results.
http://anquetil.colorado.edu/~arlowry/Crustal.htm
U51B-0027
Heat flow, crustal differentiation and lithospheric strength in North America
In stable North America, thermal models based on heat flow and heat production measurements suggest that the mechanical resistance of the lithosphere on a regional scale is greater in provinces of elevated heat flow. This is contrary to the general belief that higher surface heat flow means less stable lithosphere. We show that crustal differentiation is equally important to determine lithospheric strength. The degree of crustal radiogenic differentiation may be described using the average surface and crustal heat flows, and is quantified through the differentiation index. This index is obtained as the ratio between regional average values of heat production at the surface and in the bulk crust. The differentiation index is calculated with the bulk average heat production, suggesting that crustal differentiation processes are largely driven by internal radiogenic heat. We show that the most stable of lithospheres may be characterized by relatively high surface heat flow, simply a result of the distribution of heat sources through the crust. This may have important implications for the thermo-mechanical evolution of stable continental interiors, and for the vertical distribution of crustal heat sources through time.
U51B-0028
Measuring flexural rigidity over individual geological provinces: On the design of optimal windows for spectral data analysis on flat two-dimensional domains
The effective elastic thickness of the lithosphere, its spatial variations, and its possible directional dependence, are geophysical quantities that provide much important information about the structure and evolution of the continents -- especially when combined with other observables such as heat flow, Moho depth, seismic anisotropy, and crustal age. The decades-old picture from the study by Bechtel et al. (Nature, 1990), that elastic thickness in North America is controlled largely by the thermal state of the lithosphere, is due for a revision, and for two reasons. The first is the new wealth of geophysical data being collected from the EarthScope suite of experiments; the second is substantial improvements in the measurement and modeling of gravity and topography and the inversion of their cross-power-spectral properties for flexural rigidity, including its anisotropy. In this presentation we focus on the latter, namely on the development of a new dedicated method to map elastic thickness variations in complex geologic terrains. To reach our goal of matching the tremendous data quality available for North America with measurement and modeling sophistication as far as elastic thickness is concerned, we initially focus on the somewhat more abstract problem of measuring wavelength-dependent properties (such as power spectra, admittance or coherence functions) of data collected over two-dimensional (2-D) geologic provinces of arbitrary description. Thomson's multi-taper method, which uses spatio-spectrally concentrated "Slepian" data windows has been widely used for estimating the spectral characteristics of geological data (gravity and magnetic anomalies, topography, etc). However, these "traditional" 2-D tapers suffer from problems both in the space and spectral domain. In the space domain, they have been restricted to the estimation of power spectra for data defined on rectangular regions. It is clear, however, that typical geologic provinces have irregular boundaries. In the spectral domain, the power spectra of the tapers themselves are not isotropic; this introduces anisotropy as artifacts of the measurement. In recent work, Simons, Wieczorek, and Dahlen have solved the spatiospectral concentration problem on the surface of the sphere; the resulting "Slepian" tapers are ideal to measure power spectral properties on irregularly shaped spherical domains. Here, by analogy, we have solved the Slepian concentration problem in a 2-D flat Cartesian domain: the solutions of this problem yield the tapers suitable for a true 2-D multitaper method, and allow the analysis of 2-D regions of arbitrary geometry (in the space domain); with a spectral response that is nevertheless isotropic (in the spectral domain). We illustrate the theory and numerical implementation of this new methodological development in the analysis of geophysical data by applying it to the study of anisotropy in the elastic thickness of the continental lithosphere, with an initial focus on the North American lithosphere.
U51B-0029
Data-Derived Coulomb Stress Rate Uncertainties of the San Andreas Fault System
Interseismic stress rates of the San Andreas Fault System (SAFS), derived from the present-day geodetic network spanning the North American-Pacific plate boundary, range from 0.5 - 7 MPa/100yrs and vary as a function of fault locking depth, slip rate, and fault geometry. Calculations of accumulated stress over several earthquake cycles, consistent with coseismic stress drops of ~3-7 MPa, also largely depend on the rupture history of a fault over the past few thousand years. However, uncertainties in paleoseismic slip history, combined with ongoing discrepancies in geologic/geodetic slip rates and variable locking depths throughout the earthquake cycle, can introduce uncertainties in stress rate and in present-day stress accumulation calculations. For example, a number of recent geodetic studies have challenged geologic slip rates along the SAFS, varying by as much as 25% of the total slip budget; geodetically determined locking depths, while within the bounds of seismicity, typically have uncertainties that range from 0.5 - 5 km; uncertainties in paleoseismic chronologies can span several decades, with slip uncertainties on the order of a few meters. Here we assess the importance of paleoseismic accuracy, variations in slip rates, and basic stress model components using a 3-D semi-analytic time-dependent deformation model of the SAFS. We perform a sensitivity analysis of Coulomb stress rate and present-day accumulated stress with respect to the six primary parameters of our model: slip rate, locking depth, mantle viscosity, elastic plate thickness, coefficient of friction, and slip history. In each case, we calculate a stress derivative with respect to a parameter over the estimated range of uncertainty, as well as any tradeoffs in parameters. Our results suggest that a 25% variation, or exchange, of slip rates between the primary SAFS and faults of the Eastern California Shear Zone (ECSZ) yields a respective decrease (SAFS) and increase (ECSZ) of stress rate by only 0.5 MPa/100yrs. Alternatively, variations in locking depth spanning respective fault depth uncertainties can increase/decrease stress rates by as much as 2-3 MPa/100yrs. We also examine stress accumulation changes based on a suite of plausible historical faulting scenarios; for example, when only 70% of paleoseismic slip events are accounted for in the model, we calculate a ~0.5-3.0 MPa increase in earthquake cycle stress accumulation spanning several segments of the SAFS. Stress variations of these magnitudes have critical implications for seismic hazard analyses given that modeled stress accumulation levels of the southern San Andreas appear to be approaching those of historically great events (~7+ MPa).
U51B-0030
Upper Mantle Velocity Structure of the Cascadian Back-arc and Implications for the Tectonomagmatic Evolution of the High Lava Plains, Oregon
We present new S-wave and P-wave tomographic images of the upper mantle beneath the back-arc region of
the Cascade subduction zone based on the combined datasets of the EarthScope Transportable Array (TA)
and the high-density 106 station broadband array deployed as part of the High Lava Plains (HLP) Project.
Data from the TA stations provide a well-constrained background model of the entire Pacific Northwest
whereas data from the HLP array, coupled with active source data, provide high-resolution images of fine-
scale crustal and upper mantle structures in the HLP region. With the notable exception of the Snake River
Plain/Yellowstone "hotspot" track, the HLP terrane of central and eastern Oregon has been the most active
intra-continental region of North America in late Cenozoic time. Most strikingly, moreover, the High Lava
Plains also appears to be a region in which in excess of 100,000 km2 of new continental lithosphere was
added to the continent in late Cenozoic time.
We will discuss the formation of the HLP in the regional context of Cascadian back-arc processes, with focus
on competing plume-related and post-Laramide plate reconstruction models. Key phases of lithospheric
evolution have been variously ascribed to the post-Laramide (55-30 Ma) "sweep" from flat to normal
subduction, impact of a mantle plume, back-arc spreading, Basin and Range extension, lithospheric
delamination, and/or asthenospheric inflow along the northward migrating edge of the descending Juan de
Fuca (formerly Farallon) plate. Aspects of each of these processes are either mutually exclusive or
apparently inconsistent with observations. The current tomographic images are quite compatible with much
of the surface geology, e.g. significantly reduced velocities beneath the SRP and HLP time-progressive
volcanic tracks and higher velocities beneath stable North America. Yet a number of prominent features do
not have ready explanations, and these include: (1) a major low-velocity anomaly in the mantle wedge
extending from Newberry volcano in the south northward into north-central Oregon and well into the Columbia
River Basin of southern and central Washington; (2) the Owyhee Plateau forming an apparent lithospheric
block, but ringed by lower velocity mantle and post-17 Ma volcanism; (3) a NE-SW lineament of reduced
velocities that extends from the Snake River Plain/Yellowstone hotspot track to the southern terminus of the
Cascades, roughly coincident with the edge of the descending plate; and (4) a band of elevated velocity
mantle that extends across northern Washington eastward into Proterozoic North America, from the
uppermost mantle to a depth of at least 300 km. The western SRP, characterized by a major topographic
depression and the presence of voluminous 0-4 Ma basalts erupted along much of its length, is conspicuous
among the surface features that do not exhibit a corresponding upper mantle signature. Apart from the very
prominent SRP low-velocity anomaly, we find no evidence of remnant plume signatures in the mantle west of
about 112 degrees west longitude.
http://www.dtm.ciw.edu/research/HLP
U51B-0031
Shear wave splitting, mantle flow, and young tectonomagmatism in the High Lava Plains of Oregon
The High Lava Plains (HLP) of southeastern and central Oregon represents a young (< 15 Ma), bimodal volcanic province that exhibits an age progression in rhyolitic volcanism towards the northwest, along with widespread basaltic volcanic activity. The age progression in the rhyolites is oblique to plate motion and approximately mirrors that of the presumed Yellowstone hotspot track, as inferred from Snake River Plain volcanism. Several models have been proposed to explain volcanic activity in the HLP; these variously invoke interactions between the tail of the inferred Yellowstone plume and asthenospheric corner flow, the rollback and steepening of the Cascadia slab, and/or significant lithospheric extension, perhaps associated with the Basin and Range province to the south. A temporary array of broadband seismometers (the High Lava Plains experiment, jointly operated by the Carnegie Institution of Washington and Arizona State University) was deployed in the region beginning in 2006. As of fall 2008 118 sites have been occupied as part of the HLP seismic experiment (with 104 currently operating), in addition to the broadband stations operated here as part of the USArray Transportable Array (TA). Here we present a data set of shear wave splitting measurements for SKS phases recorded at ~ 150 stations during the period 2006-2008. SKS splitting is used to characterize upper mantle anisotropy and when properly interpreted can place constraints on the geometry of upper mantle flow associated with tectonic processes. Stations in the HLP exhibit significant shear wave splitting, with average split times between ~ 0.8 seconds and ~ 3.0 seconds and average fast directions near N80°E. Although the backazimuthal coverage for SKS is not ideal, there is little evidence for backazimuthal variations in splitting parameters that might suggest complex anisotropic structure. The average split time in the HLP is ~ 1.8 sec, well above the global average of ~ 1 sec for continental regions. The measured fast directions are remarkably consistent and exhibit very little lateral variation in the HLP proper. We infer from the large split times and homogeneous fast directions that there must be significant active flow in a roughly E-W direction in the asthenosphere beneath the HLP. We discuss several different scenarios for HLP formation and evolution in light of the constraints on mantle flow provided by SKS splitting observations.
U51B-0032
Constraints on the causes of the High Lava Plains volcanism from inversions of surface wave phase and amplitude data
The High Lava Plains (HLP) of southeastern and central Oregon is a province of bimodal volcanic activity which postdates the voluminous flood basalts of the Columbia River area. The north-westerly trend of the rhyolitic volcanism of the HLP lies at an oblique angle to the north-easterly trend of the silicic volcanism associated with the Snake River Plain and Yellowstone caldera, though both volcanic tracks originated at the same place and time in southeastern Oregon some 15 Ma. While the Snake River Plain's volcanism has often been associated with the possible existence of a Yellowstone hot spot due to its alignment with North American plate motion, the High Lava Plains volcanism defies such a simple explanation. Possible contributing factors in the formation of the High Lava Plains include the subducting Juan de Fuca plate to the east and its effects on mantle dynamics, the effects of varying lithospheric thickness throughout the area, and the debated existence of a Yellowstone plume head and/or tail. We present results of a preliminary tomographic inversion of Rayleigh wave phase and amplitude using the method of Forsyth and Li [2005] to constrain phase velocities and shear wave velocities in the area of the High Lava Plains. Data for this inversion are provided by a temporary array of broadband seismometers deployed in the region beginning in 2006 by Arizona State University and the Carnegie Institution of Washington. As of summer 2008, 104 stations were operating as part of the HLP seismic experiment, in addition to the broadband stations operated here as part of the USArray Transportable Array (TA). Preliminary results include waveforms from ~50 large events recorded at an average of >90 stations. The method of Forsyth and Li [2005] used in this inversion employs a 2-plane-wave approximation of the incoming surface waves. The improvements in lateral resolution of this methodology over previous phase velocity inversions is significant, and helps constrain the effects of various mantle structures on the formation of this unusual tectonic province.
U51B-0033
Plume-lithosphere interaction: Implications for magmatism associated with the Yellowstone hotspot track
The relative contributions of lithosphere vs. upwelling sub-lithospheric mantle to formation of Snake River Plain-Yellowstone (SRPY) basaltic magmas remains an ongoing debate. The association of this province with initially thick and cold Archean lithosphere (Wyoming craton) poses a problem in that this lid will probably hinder and possibly prevent melting of rising plume hot material. However, petrologic modeling indicates that SRPY primitive basalts last segregated from mantle at ~1450°C and ~100 km depth, suggesting that their source is only slightly warmer than MORB-source mantle and significantly cooler than sources of most oceanic hotspot magmas. In the light of such evidence plume melting can only occur if the lid can be substantially thinned over geologically reasonable time. We developed a series of 3D time- dependent geodynamic models to study lithospheric thinning processes and to investigate the extent and rate of lithosphere thinning assuming an initial structure representative of the Wyoming craton. Modeling results show that thermal erosion by plume impingement alone appears incapable of providing the required lithospheric thinning for initial lithosphere thicknesses exceeding 150 km, even assuming a wide rage of plume parameters (chemical buoyancy Δρ=30-50 kg/m3 and temperature contrasts ΔT=150-300°C). On the other hand, alternative models (e.g., involving low-angle Laramide subduction, delamination) are considered unlikely because they conflict with geochemical evidence that SRPY basalts contain a dominant contribution of old, isotopically evolved mantle material - presumably derived from subcontinental lithospheric mantle (SCLM). Thus, we conclude that SCLM is likely to be preserved, that the SCLM lid thickness probably prevents substantial melting of rising plume material (tomographically imaged), and SRPY basalts are predominantly derived by melting of lithospheric mantle.
U51B-0034
HuBLE-UK, the Hudson Bay Lithospheric Experiment: Insights into the formation of the Canadian Shield
Hudson Bay lies in the Precambrian core of North America, which is comprised of the Canadian Shield and contiguous platform regions. The region is underlain by one of the largest lithospheric keels on Earth; it is also the site of one of the largest negative geoid anomalies. We have deployed 10 broadband seismic stations in the northern part of the bay that complement the existing POLARIS, CHASME and CNSN network stations in the region. Here we present preliminary SKS shear wave splitting analyses and independent tomographic inversion of P- and S-wave travel-time data in order to: 1) understand better the origin and evolution of the Hudson Bay cratonic interior basins; 2) to illuminate possible relationships between the lithospheric keel, sub-lithospheric mantle flow and formation of the Hudson Bay basin; 3) to improve understanding of postglacial isostatic rebound; 4) to map the lithospheric structure of the Trans-Hudson orogen in a region characterized by extreme salient-reentrant geometry, possibly analogous to the western syntaxis of the Himalayan front. SKS delay times vary from 0.5-1.2s, which indicate a lithospheric-scale anisotropic layer up to 150km thick. However, SKS fast directions and preliminary tomographic images do not relate simply to the structural trends of the Trans Hudson Orogen and neighboring Archean terranes. Our work complements ongoing HuBLE studies that focus on receiver function analyses, dispersion analysis of teleseismic Rayleigh waves, and applications of ambient noise tomography that extract more information about lithospheric structure of the Hudson Bay basin.
U51B-0035
Hudson Bay Lithospheric Experiment: Constraints on Lithospheric Thickness From Surface Waves
HuBLE (Hudson Bay Lithospheric Experiment) is an international initiative to study the structure, dynamics and evolution of the Hudson Bay region. In particular, we seek to understand the interaction between the Archean cratons surrounding the region and the underlying Paleoproterozoic Trans-Hudson Orogen, which formed during the collision of the Superior and Churchill Provinces at 1.9-1.8 Ga. Global and continental- scale tomographic models indicate a thick, cold and refractory lithosphere beneath Hudson Bay. Most tomographic models suggest that this region is associated with the highest velocities and thickest seismological lithosphere of the Canadian Shield. The HuBLE project commenced in 2006, with the deployment of a number of telemetered broadband seismograph stations on the east and west coasts of Hudson Bay. Along with existing stations from the POLARIS/FedNor initiative in northern Ontario, and permanent Canadian stations, the deployment ringed Hudson Bay on three sides. A second phase of deployment in 2007, using non-telemetered broadband stations, completed the coverage of the region. A considerable number of large teleseismic earthquakes have been recorded by the array since its installation, and the data are generally of high quality. We measure Rayleigh wave phase velocities for paths crossing Hudson Bay, using the two-station cross-correlation method of Meier et al. (2004). Average phase velocity dispersion curves are constructed using data from multiple earthquakes along each path, resulting in a set of reliable dispersion measurements in the period range ~15—-250~seconds. The data set therefore permits constraint of lithospheric shear wave velocity structure from mid-crustal to asthenospheric depths beneath the continent. Preliminary 1D shear wave velocity models of path-averaged structure are estimated using a smooth linearised inversion technique (Maupin & Cara, 1992). The models show a typically 'shield- type' signature, with a high-velocity mantle lid overlaying a weak negative velocity gradient, below which the velocities become similar to those of global reference models. The preliminary results suggest that the seismological lithosphere beneath Hudson Bay is thicker than beneath the surrounding cratonic regions, reaching a maximum of ~260~km.
U51B-0036
Structure of the Crust and Uppermost Mantle Beneath Hudson Bay Based on Ambient- Noise Tomography
Hudson Bay is a vast inland sea that penetrates deeply into north-central Canada, forming a conspicuous element of the North American coastline. The Bay conceals several fundamental tectonic elements of North America, including most of the Paleoproterozoic Trans Hudson orogen, the Paleozoic Hudson Bay basin and a large part of the lithospheric root beneath the Precambrian core of North America. This study is focused on regional crustal structure based on ambient-noise tomography. Twenty-one months of continuous ambient- noise recordings have been acquired from 31 broadband seismograph stations that encircle Hudson Bay. These stations are part of the Hudson Bay Lithospheric Experiment (HuBLE), an international project that is currently operating more than 40 broadband seismograph stations around the periphery of Hudson Bay. Following established processing procedures that include trace normalization and spectral whitening, cross- correlations are computed for all possible station pairs. The resulting waveforms are treated as Green functions, from which group velocity dispersion measurements can be made. Since Hudson Bay freezes during winter months, there is a pronounced asymmetry to the Green functions indicative of noise sources along the Atlantic seaboard. Preliminary results indicate shield-like conditions in most areas, but reduced crustal velocities beneath the Hudson Bay basin.
U51B-0037
Discontinuity Structure Beneath the North American Craton and the Signature of Continental Roots
Cratons are vast areas of continental lithosphere that have remained relatively intact and stable since the Precambrian. They tend to be underlain by deep roots, characterized by anomalously fast seismic velocities (~3%), which tomographic images show can extend to depths of ~350 km. The process by which these roots are formed is presently not well understood, but is believed to be associated with the accretion and imbrication of subducted oceanic material during the Archean. In order to test hypotheses on the timing and evolution of the root beneath the North American craton and its influence on the intracratonic Hudson Bay basin, we undertake a study of P-to-S receiver functions to obtain estimates of crustal thickness and Vp/Vs ratio. Our data come from the POLARIS, CHASME and CNSN networks, and the recently deployed HuBLE-UK broadband network of ten stations presently recording in northern Hudson Bay. Our results show Moho depths in Archean domains are consistent with other cratons worldwide (35-40km), with significant crustal thickening occurring in the areas adjacent to and within the Proterozoic Trans-Hudson Orogen (up to 47km). Receiver functions from stations on Archean basement exhibit a sharp Moho Ps phase and clear subsequent reverberations, in contrast to those in the Trans-Hudson, which show much more complexity. There is evidence of deeper discontinuities within the tectospheric root on the tangential receiver functions, which is indicative of dipping and/or anisotropic character. Results from this line of investigation, combined with other complementary studies (body-wave tomography, surface wave analyses, SKS splitting analyses and ambient noise tomography), will delineate the shape of the root, and also indicate whether shallow subduction played a role in cratonization. The study will shed light onto the evolution of the continent during its early history.
U51B-0038
Low heat flow and deep thermal structure of the lithosphere near the center of the Superior Craton.
We present new heat flow measurements in the central part of the Superior Province, near the core of the North American craton, where one expects the lithosphere to be thickest. Measurements in 21 deep boreholes yield heat flow values that never exceed 33 mW m-2. For all the sites, there is no relationship between the heat flux and the value of heat production measured on surface samples. Heat flux variations occur at wavelengths <100km, and are of crustal origin. Averages on 200 × 200km windows yield 29 and 31 mW m-2, the lowest values at this scale found so far in the Canadian Shield. In order to elucidate the deeper thermal structure of the Superior Province, we have studied the relationship between heat flux and S wave travel time delays. We find no significant correlation between the two, giving futher indication that heat flux variations are of crustal origin and do not imply large variations of deep thermal structure. For consistency with travel time variations, the Moho heat flux and the heat flux at the base of the lithosphere can not vary by more than 2 mW m-2, indicating that the heat supply at the base of the craton is remarkably uniform.
U51B-0039
Travel-time Tomography of the Abitibi-Grenville Region, Eastern Canada
Seismic studies of the Canadian Shield have indicated certain structural anomalies within the cratonic lithosphere. A low-velocity anomaly has been imaged near the Ontario-Quebec border, in the Abitibi- Grenville province, but its 3D geometry was poorly-defined due to a lack of seismograph station coverage on the Quebec side of the border. With the help of the 5 new seismograph stations installed in western Quebec in 2007, 26 others belonging to the POLARIS project and the Canadian National Seismograph Network (CNSN), and a data set of travel time picks from the ABI-96 teleseismic experiment (Rondenay et al., 2000), we analyse the P-wave velocity structure of the lithosphere in order to better understand the complexity of the region and the interaction of the lithosphere with thermal anomalies in the underlying mantle. Several analysis steps have been carried out. We first measured the relative arrival times of teleseismic P waves across the array, using the cross-correlation method of VanDecar & Crosson (1990). We present the results of an analysis of azimuthal variations of these arrival times for representative stations across the array. We have also calculated maps of relative arrival time residuals across the array for earthquakes coming from different back-azimuths, in order to examine systematic patterns of travel-time anomalies resulting from mantle heterogeneity. Finally, we have inverted the travel time data to estimate a preliminary model of the 3D P-wave velocity structure beneath the region, using a standard tomographic inversion technique.
U51B-0040
Crustal Thickness and Vp/Vs Beneath the Florida to Edmonton Broadband Seismometer Array
Ps receiver functions were used to constrain crustal thickness (H) and Vp/Vs beneath the 28 stations of the Florida to Edmonton Broadband Seismometer Array and 6 nearby permanent stations (IRIS/GSN, USNSN, CNSN, and NMSN). The stacking algorithm of Zhu and Kanamori (2000) was applied to receiver functions for the 9 stations exhibiting clear Pms, PmpPms, and PmpSms/PmsPms converted phases, thus constraining both H and Vp/Vs. For the remaining stations where only high quality Pms phases were evident, Vp/Vs was prescribed based on interpolation of well-constrained neighboring values, and the receiver functions were stacked to constrain H. Because a number of FLED stations overlie significant sedimentary layers, potential bias in H and Vp/Vs estimates from intracrustal phase interference was assessed via deconvolution and stacking of synthetic waveforms. Even for models where sediment layers have a high Vp/Vs and intracrustal phases are strong, accurate estimates of H and average crustal Vp/Vs can be retrieved from receiver function stacking. Along the FLED array, roots of thickened crust were observed beneath the Paleozoic Appalachian and Proterozoic Trans-Hudson orogens. In both cases, the ratio of surface elevation to root thickness and crustal root buoyancy inferred from gravity data decrease with orogenic age, thus fitting global trends. Crustal thickening was also observed beneath the Proterozoic Mid-Continent Rift in Iowa, consistent with compressional inversion of the rift during the Grenville orogeny. Crustal thickness decreases beneath the highest elevations at the NW end of the array. Because Vp/Vs also decreases at these stations, the increase in elevation could be compensated by a reduction in average crustal density, although contributions from mantle buoyancy cannot be ruled out.
U51B-0041
Multidisciplinary Observations of Subduction (MOOS) Experiment in South-Central Alaska
Seismic and geodetic data are being collected in the Kenai Peninsula and surrounding area of south central Alaska as part of the PASSCAL experiment MOOS. A total of 34 broadband seismic stations were deployed between the summers of 2007 and 2008. Seventeen of these stations continue to operate for an additional year and are scheduled to be removed in the summer of 2009. Numerous GPS campaign sites have and will be visited during the same time period. The MOOS seismic deployment provides coverage across the interplate coupled zone and adjacent transition zone in the shallow parts of the Alaskan subduction zone. It is a southern extension of an earlier broadband deployment BEAAR (Broadband Experiment Across the Alaska Range) to the north. When integrated with the previous BEAAR experiment, these data will allow high-resolution broadband imaging along a 600 km long transect over the Alaska subduction zone, at 10-15 km station spacing. The MOOS deployment allows us to test several hypotheses relating to the postulated subduction of the Yakutat Block and the nature of the coupled zone which ruptured in the great 1964 earthquake. The seismic and geodetic stations cover an area that includes part of the 1964 main asperity and the adjacent, less coupled, region to the southwest. Data gathered from this experiment will shed light on the nature of this boundary from both a geodetic and seismic (or earth structure) perspective. Shallow seismicity recorded by this network greatly improves the catalog of events in this area and helps to delineate active features in the subduction complex. Preliminary results from this project will be presented.
U51B-0042
The Fate of Unsubducted Fragments of the Farallon Plate
The Monterrey, Guadalupe, and Magdalena microplates all ceased subducting before the Farallon-Pacific spreading centers disappeared beneath the North American continent, leaving behind plate fragments that became incorporated into the Pacific plate as spreading ceased. Spreading directions changed dramatically on the Monterrey and Magdalena spreading centers, suggesting that the Farallon slab detached from the microplate fragments before or during the cessation of spreading. Although the sinking Farallon plate must have detached from all three microplates, it is not known where the detachments occurred and how much of the young slab was left behind still attached to the unsubducted surface microplates. After the microplates were incorporated into the Pacific plate, the plates moved relative to North America on predominantly strike- slip faults, as much as 820 km in the case of the Monterrey microplate. Did the slab fragments move along with the fossil microplates and coastal crustal blocks? We address the fate of the unsubducted fragments by looking for associated seismic velocity anomalies using teleseismic Rayleigh wave tomography and the stations of the USArray, NARS-Baja and RESBAN arrays. There are high velocity anomalies landward of each of the fossil oceanic microplates, indicating that slab fragments translated along with the surface plates. At least in the case of the Guadalupe plate, high shear velocity anomalies continue to depths exceeding 100 km, suggesting that the leading edge of the microplate slab was shoved downward into the mantle or that thermally induced density contrasts induced downward convection. We estimate how much slab may have remained attached to the surface microplate by integrating the velocity anomaly over the volume of the anomalous region.
U51B-0043
Evidence for a slab remnant beneath the Gulf of Califonia from surface wave tomography
The Gulf of California is a tectonically young and active rift system that links the East Pacific Rise to the San Andreas transform fault. Its formation is associated with the cessation of subduction of the Farallon plate beneath the North American continent which occurred approximately 12 Myr ago at the latitudes of central Baja California. Previous studies of the region imaged very low mantle shear velocities down to 250 km, but these studies could not resolve the lateral variations along the gulf that are needed to explain the variations in rifting (oceanic spreading in the south, diffuse continental deformation in the north), nor the diversity of magmas (with and without a slab signature) in central Baja California. We used data from the recent NARS-Baja project and other networks to measure Rayleigh- and Love-wave dispersion along interstation paths across and around the Gulf of California to obtain a detailed 3-D model of the crustal and upper mantle shear-velocity structure. We computed (anisotropic) phase velocity maps and inverted these for shear-velocity structure. Our most important result is the finding of a relatively high shear- velocity anomaly in the mantle beneath the central part of the Gulf of California that is interpreted as a remnant of the subducted (Farallon-derived) Guadalupe slab. The absence of such a slab remnant beneath the northern part of the gulf is in agreement with the presence of a 'slab window' as suggested from the tectonic evolution of the area. The patterns of azimuthal and radial anisotropy point to differences in mantle flow between the slabless region and the region with the slab remnant. It is inferred that the presence of a slab remnant beneath the central gulf can account for the different styles of rifting along the Gulf of California as well as for the variability of the magmatism (<12 Ma) in central Baja California.
U51B-0044
Constraints on lithology of downwelling lithosphere from the Sierra Nevada, California
A large high-wavespeed seismic anomaly has long been observed between about 70 and 250 km under the eastern San Joaquin Valley and westernmost Sierra Nevada. This body (the Isabella Anomaly) has been inferred to represent lithospheric material from the lower part of the Mesozoic Sierra Nevada Batholith that has become unstable and is now sinking. This inference is mostly based upon a shift from garnet-rich mantle xenoliths erupted through the Sierra before 8 Ma to garnet-free spinel peridodites erupted since 3 Ma; direct evidence on the composition and origin of the downwelling has been absent. New images of the Isabella anomaly derived from inversion of P-wave arrival times recorded by the Sierra Nevada Earthscope Project (SNEP) and surrounding EarthScope Transportable Array stations reveal a narrower feature than earlier work. We use the observed subsidence along about 180 km of the western margin of the Sierra Nevada (Saleeby and Foster, 2005) and the accumulation of 1 km (post 3 Ma) to 2.5 km (post 8 Ma) of sediment to infer that a load of about 2-5 x 1017 N was applied to lithosphere with flexural rigidity about 3.2 x 1022 N m. Dividing this load by the tomographic volume estimate of 5.6 x 105 km3 for the Isabella anomaly suggests a density contrast with surrounding mantle of about 40-100 kg m-3. Such a density contrast for a body with an average ~3% vp velocity anomaly is high for a purely thermal effect (which would produce a density anomaly of about 40 kg m-3); we infer that this body is at least in part compositionally different than the surrounding mantle, inferring that it contains substantially more garnet derived from the eclogitic root of the Sierran granitoids. We also show that this is consistent with constraints from gravity profiles across this body. This interpretation suggests that the entire dripping body is present still in the upper mantle and supports the idea that compositional variations near the crust-mantle boundary are important in driving convective instabilities.
U51B-0045
A Mantle Xenolith Window Into the Grenville Orogeny of Southern Laurentia
The creation and isolation of the craton, or stable SCLM, is intimately connected to orogenesis. However, the nature of the lithospheric mantle beneath orogenic belts is incompletely understood due to the general lack of mantle xenolith-bearing basaltic magmas in such regions. One such place where we are afforded the opportunity to study the deep lithosphere beneath an orogenic belt is in central Texas, United States. Mantle xenoliths occur in late Cretaceous alkali basaltic magmas erupted through the remnants of the Appalachian – Ouachita structural belt of eastern and southern Laurentia. The Appalachian – Ouachita structural belt, which is buried beneath most of the Gulf Plain, represents two dissimilar cycles of orogenesis. The earlier cycle was the culmination of a long period of Proterozoic juvenile crust formation along Laurentia's southern and eastern margin. The more recent (Paleozoic) cycle created the fold and thrust belts currently exposed in the Appalachian and Ouachita Mountains, but involved mainly thin-skin tectonics and accretion of terranes, rather than continental suprasubduction settings. We are interested in identifying the process which emplaced mantle lithosphere beneath this ancient orogenic belt, and whether the original lithospheric mantle has been preserved there. Here, we show that the xenoliths beneath west-central Texas are of continental origin. These samples also have geochemical signatures suggestive of a suprasubduction zone setting in the form of enrichments in fluid-mobile trace (e.g. La) elements over fluid-immobile trace elements (e.g. Nb). These observations imply that the original continental lithosphere created in the Proterozoic suprasubduction zone setting was likely preserved during continent-continent collision and did not undergo wholesale delamination over a billion year period. During this period, the mantle convectively resisted two episodes of supercontinent rifting. However, the lithospheric mantle may have been subsequently partially thinned or modified by extension associated with opening of the Gulf of Mexico or post-orogenic relaxation.
U51B-0046
Anisotropic Ground Truth: Integrating Petrological and Seismic Observations from the Crust of the Cheyenne Belt Continental Suture
We present initial results from an integrated geological and seismic study of crustal fabric in the Cheyenne Belt continental suture in Wyoming and Colorado. The geological constraints stem from petrological and electron backscatter diffraction (EBSD) analysis of lower crustal xenoliths and observations of mid-crustal surface exposures, the seismic constraints from azimuthal variations in teleseismic receiver functions observed at Earthscope USArray Transportable Array stations as well as previous temporary networks. The suture between the Archean Wyoming craton and the Proterozoic central Rocky Mountains crops out as a roughly 10 km wide shear zone exhumed from mid-crustal depths with well developed foliation and lineation. In addition to surface geological observations, we have determined the composition, fabric, and elasticity tensors of lower crustal xenolith samples from Leucite Hills, southern Wyoming. The xenoliths show Vp and Vs anisotropy of up to 12%, mostly due to highly aligned hornblende and biotite. This magnitude of anisotropy produces a clear signal in receiver functions if the anisotropy persists on a scale of kilometers. Rather than using the splitting of converted shear phases, we look for more robust characteristic patterns as a function of azimuth in the receiver functions to detect anisotropy. Depending on the azimuthal coverage and noise conditions at the station, we can use three different approaches to detect and measure anisotropy: 1) Determination of a crustal models including anisotropic parameters using synthetic seismograms; 2) harmonic analysis of converted phase power as a function of depth to distinguish between isotropy, horizontal or plunging axis anisotropy, and plunging interfaces; 3) Hypothesis testing based on geological obervations. Input from geology and petrology is used in all three approaches. We interpret the results in terms of the evolution of the suture and the physical properties of the deep crust in the study area, including the nature of the enigmatic high-velocity lower crust in the central and northern Wyoming craton.