T21B-0578
Statistic Characteristics And Applications Of BATS CMT Catalogue For Regional Earthquakes Around Taiwan
We collected 1130 regional centroid moment tensor (CMT) solutions, which were obtained by the Institute of Earth Sciences, Academia Sinica for earthquakes occurred in the Taiwan region in 1995-2005 by inverting the BATS broadband waveforms, to characterize their statistic properties and tectonic implications. The moment magnitude ranges from 3.3 to 7.1 in this catalog with a complete magnitude (Mc) of ~4.2. Predominant thrusting events are common in the colliding boundary, fold-and-thrusting belt, and subducting interface, whereas the normal faulting presents mainly in the Okinawa Trough, northern Hoping Basin, and the eastern Central Range. To evaluate the robustness of BATS CMT solutions, we compare 96 CMT solutions that are in common between BATS and Harvard catalogs. Results show that the BATS solutions are well constrained for events with misfit < 0.5, which are roughly comparable to Harvard¡¦s solution with moment error < 0.2. The BATS Mw is generally less than Harvard Mw by an average of ~0.17. This is very likely due to the fact that the BATS is using relatively shorter frequency band (0.1-0.02 Hz). Horizontal seismic strain directions, which are shown in terms of well constrained BATS CMT dataset, agree well with the surface geodetic GPS observations. Our results provide seismological constraints on the mode of deformation and stress field in the Taiwan region.
T21B-0579
The multiple inverse method applied to heterogeneous focal mechanisms in Ryukyu arc, Japan: determination of stress province in active island arc
Deformation and stress state in active island arcs are important for understanding the dynamics. Especially, forearc stress and deformation are key to understand mechanical interaction of the backarc and subduction. Present crustal stress states are derived from focal mechanism solutions of earthquake. Otsubo and Yamaji (2006 JPGU meeting) proposed a new method which is proposed to separate stresses from heterogeneous data of focal mechanisms of earthquakes. In this study, we apply the technique into the focal mechanisms in Ryukyu arc. Result in Tokara Islands area, northern part of Ryukyu arc is typical one because stress province was likely to be distinguishable using the P-/T-axis of the solutions in domain which sandwiched volcanic line. As a result, present study shows different stress province from one found by P-/T- axes. Two extensional stresses (NNE-SSW and NW-SE) were detected from the solutions. Under regional radial extensional stress state, spatial variation of state stress allows was to divide the Tokara Islands area into two directions normal faulting regimes of stress. The boundary of the stress provinces corresponds to the volcanic front. In backarc area A1, arc- perpendicular (NW--SE) extensional stress is dominant. The dominant focal mechanism in the backarc is strike- slip faulting although the focal mechanisms have a few NE--SW trending normal faults. Stress ratio is low (Φ=0.2). On the other hand, the arc-parallel (NNE--SSW) extensional stress province A2 is found in forearc area. Stress ratio is high (Φ=0.8). The dominant focal mechanism in the forearc is NW--SE trending normal faulting. Many active faults indicate the stresses in the area (e.g. Founier et al, 2001, JGR). Kubo and Fukuyama (2004, EPSL) indicated that the extensional stress province extends to a wide area in the Ryukyu forearc by using P-/T- axes of focal mechanisms. Such forearc deformation and stress provinces are formed in relation to the increasing curvature of arc as a result of backarc extension (Wessel et al., 1994, JGR; Ten Veen and Kleinspehn, 2002, Tectonics).
T21B-0580
Transitional pattern of stress field between Honshu arc and Ryukyu arc in Japan revealed by the multiple inverse stress analysis using the focal mechanisms
We analyzed stress state in the connective region between Honshu and Ryukyu arc to discover the transitional stress between with and without backarc opening of the arcs around southwestern Japan. We used multiple inverse stress analysis method developed in structural geology. This technique can reduce artificial effect to make dataset for stress tensor inversion and provide opportunities to detect spatio-temporal heterogeneities of the stress field. We used focal mechanism solutions (best double couple components of regional moment tensors derived by Fnet, NIED, Japan) as input data of stress inversion. We analyzed three datasets which divided spatially along arc (from Ryukyu to Honshu). Because seismicity with moderate magnitude in forearc side is low, we obtained results only in the region of backarc side. Each result show strike-slip regime having two distinct clusters with different stress axes (30°). In perspective view, Shmax and Shmin directions rotate gradually from Ryukyu to Honshu. This result constrains pattern of the fan shaped stress trajectory between compressive Honshu arc and extensional Ryukyu arc, which is important issue to the understand geodynamics of subduction zone island arc. On the other hand to interpret multiple stress solutions and these stress ratio variations, locally heterogeneous stress state are required.
T21B-0581
Thinking deeply: Is there a contribution to Asian tectonics from large-scale mantle flow?
Recent GPS observations in central and southeast Asia indicate that the entire eastern margin of Asia is moving eastward, independent of Tibet motion. This result argues that the eastward motion of south China is not a result of the India-Eurasia collision. Therefore, we investigate an alternative source that may be responsible for driving the observed eastward motion. Specifically, we investigate stresses associated with (1) lithospheric density variations and (2) basal tractions resulting from large-scale density buoyancies, with interest in the effects resulting from the eastern subduction systems. Using a self-consistent set of GPS observations and Quaternary fault slip data we first determine a continuous model strain rate and velocity field for central and southeast Asia. Deviatoric stresses associated with lithospheric density variations are determined from the CRUST2.0 data set. These calculations are performed globally to avoid boundary effects but only considered within the region of interest. We then solve for stress field boundary conditions in this region, which represent the effects of large- scale mantle flow associated with deeper density buoyancies. Separately, we use CitcomS combined with published tomographic models and history of slab subduction to determine mantle flow and associated stress field at the base of the lithosphere. By detailed comparisons between the stress field boundary conditions and the deviatoric stress fields resulting from mantle convection, we isolate the relationship between the eastward motion of central Asia and surrounding subduction systems.
T21B-0582
A Best-fit Lithosphere-Mantle Coupling Model Constrained by Plate Motions and the Velocity Gradient Tensor Field in the Plate Boundary Zones
Lateral variation in lithosphere-mantle coupling is caused by lateral variation in asthenosphere viscosity and distribution of density buoyancies in the mantle. The goal of the present study is to achieve a best-fit lithosphere- mantle coupling model for the Earth. That is, a model with appropriate radial and lateral viscosity variations that would successfully predict not only plate moions, but also deformation indicators along the Earth's plate boundaries. The convection model used is a whole mantle model driven by density buoyancies within the mantle with free slip boundary conditions at the surface and at the core-mantle boundary. We test viscosity structures by varying the thickness and viscosity of the asthenosphere layer and by incorporating lateral viscosity variations generated by major geological features of the Earth, such as the continent-ocean divide, the presence of cratonic roots as well as age differences in the oceanic lithosphere. For each structure, we predict the deviatoric stress field, the pattern of poloidal and toroidal flow and the toroidal-poloidal partitioning ratio. The predicted deviatoric stress field is added to the deviatoric stresses generated by lithosphere buoyancies (gravitational potential energy or GPE differences due to density buoyancies within the lithosphere), calculated based on the Crust 2.0 dataset, and the total stress field is compared with strain rate tensor information along the Earth's deforming plate boundary zones from the Global Strain Rate Map (GSRM). The best-fit model has to satisfy both the constraints of matching the plate motions and the deviatoric stress field simultaneously. We find that combined models with 1-2 orders of lateral viscosity variations within the lithosphere and 2-3 orders of lateral viscosity variations within the asthenosphere are able to match plate motions, the observed toroidal-poloidal ratio and the strain rate tensor data from GSRM. The viscosity variation in the lithosphere arises because of the higher viscosity continents as well as because of higher viscosity of old oceanic lithosphere compared to young oceanic lithosphere. The viscosity variations in the asthenosphere are mostly due to the presence of high viscosity keels below the Archean cratons. A weak (low viscosity) asthenosphere is crucial in matching the plate motions and giving rise to sufficient toroidal velocity. However, the weakness, if uniform, would decouple the deeper density buoyancies and would prevent the stresses from these deeper density buoyancies to be transmitted up to the lithosphere. Presence of strong asthenosphere below the Archean cratons enables transmission of these stresses to the overlying lithosphere and hence controls the match to the strain rate tensor data. These imply that major geological features play an important role in explaining the first order features of plate tectonics.
T21B-0583
Crustal response to the formation of the Hálslón reservoir in Iceland.
A major 2.4 km3 water reservoir (Hálslón) is being filled north of the Vatnajökull ice cap, Iceland located at the eastern edge of the plate boundary deformation zone in North Iceland. It overlies a fault system that has been active during the Holocene. When full the reservoir is 198 m deep with an area of 57 km2. The filling began on September 28, 2006 and was 7 m from being full at the end of August 2007. The load that is consequently being applied on the crust gives a good opportunity to study the characteristics of the crustal response. An extensive network for crustal deformation research was established in 2005. A total of 35 benchmarks were measured in GPS-campaigns in August, 2005, 2006 and 2007. Eight continuously measuring GPS-stations where established for resolving temporal changes in the Hálslón area in near real time. The first station was established in October, 2004, three in fall 2005 and the other four in September 2006. Known crustal movements in the area prior to filling process were uplift associated with decreasing load of the nearby Vatnajökull ice cap and seasonal variations due to the annual change in snow load. Since the filling of the reservoir began in September 2006, a local deformation have been detected in the part of continuous GPS-network closest to the reservoir, including a subsidence of 10-20 mm, horizontal movements of 5-10 mm to the northwest in part of the network and movement of 5-10 mm to the south east in other parts. The time series show that the movements have not been continuous. In the first 2- 3 months of the filling the GPS-stations moved, but then ceaset until May 2007 when movements continued. This is in accordance with the increase in the water level of the reservoir. Levelling over the most prominent fault in the area near the shore of the reservoir shows that site closer to the reservoir has subsided ~1 mm Since 2005. The horizontal movements of the GPS-stations cannot be solely attributed to the direct effect of the water load. Both the movements of the GPS-sites and the levelling results across the fault suggest that a part of these movements could be due to opening of fractures. Model calculations using a Young's modulus of 30 GPa and a crustal thickness of 10 km give a maximum expected elastic crustal subsidence of about 0.1 m. Considering additional ductile movements total subsidence of about 0.3 m is expected. The results from the GPS-campaigns compared with model calculations will give additional constraints on the crustal characteristics in the area.
T21B-0584
Role of Metamorphic Reactions and Related Density Changes in Mountain Building
It is well known that the rocks of the crust experience a wide range of pressure and temperature conditions, especially in tectonically active zones. These conditions are registered by imprints of different petrological variations, and lead to the significant variations of density from low to high grade metamorphic facies. Here we use a 2-D finite element model, Cast3M, that was modified to incorporate metamorphic phase changes. The density of each element is derived from petrogenetical grids yielding density ρ as a function of temperature T and pressure P for a given composition, including thermal expansion and compressibility effects. To ensure local mass conservation, the exact density values are imposed in the model as volumetric variations, which are implemented using the tectonic stresses and the boundary conditions. The scheme is stabilized for different types of rheology, with elastic or non-linear visco-elastic behaviour. The numerical model is able to simulate convergence and to take into account erosion processes. We apply our model in a convergent context with an initial topography. The scale of the model is ~1000 km in length and ~150 km in thickness, including a mantle, a lower and an upper crustal layer. The duration of the convergence is in the order of a few million years. Our preliminary results show that density variations due to mineral phase changes can play an important role in mountain building processes. We suggest that deformation localization and topography evolution studies should account for realistic rheologies, erosion mechanisms, as well as metamorphic facies changes. In models where metamorphic reactions are not correctly implemented, the induced errors with regard to mass conservation and crustal deformation are significant.
T21B-0585
Models of Seismic Anisotropy and Upper Mantle Deformation at the San Andreas Fault System: Investigation of the Role of Crust-Mantle Coupling
Upper mantle seismic anisotropy patterns at the San Andreas fault system are uniquely suited to investigating the role of crust-mantle coupling within the lithosphere. Published interpretations of seismic anisotropy in northern California include a two-layer model with fault-parallel anisotropy in the upper layer and nearly fault- perpendicular, E-W anisotropy in the lower layer. This has been interpreted to suggest that plate motions play a role in generating seismic anisotropy in the upper mantle in this region. In contrast, in southern California, the dominance of E-W anisotropy has been interpreted as a primarily asthenospheric signal, with little influence of the lithosphere. In this project, an integrated seismic-geodynamic analysis is being developed and applied to better quantify the role of lithospheric (plate) motions on the seismic anisotropy patterns at the San Andreas fault system with the goal of understanding the rheologic scenarios that allow these patterns to arise. We model mantle flow for the Pacific-North America plate boundary as a viscous channel using the well-studied surface kinematic field and inferred deeper asthenospheric flow patterns as boundary conditions. The fluid within the channel is characterized by depth-stratified rheologic properties, representing upper mantle within the mantle lithosphere and in the asthenosphere beneath. We investigate how mantle flow would produce olivine lattice preferred orientation (LPO) fabrics with a 3-D Cartesian implementation of a published kinematic code to calculate LPO by dynamic recrystallization. Our geodynamic models focus on how rheologic parameters govern the fabric development and whether we can interpret the observed patterns of seismic anisotropy at the San Andreas fault system in terms of the relationship between the observed plate motions at the surface and flow in the asthenosphere. These models will serve as priors in a tomographic inversion of new observations of seismic anisotropy derived from the USArray Transportable Array.
T21B-0586
Correlation of shear deformation and heat flow along the Coast and Insular belts of Western British Columbia, Canada
Western margin of North American cordillera is characterized by a continuous band of low heat flow with varying magnitudes. These zones are associated with the outer arc regions of the present or recent subduction of Kula/Farallon/Juan de Fuca plates under North America. The subduction of Farallon/Kula plate ceased some 20 My ago north of Juan de Fuca plate and the cessation resulted in the formation of a transform boundary carried by the Quinn Charlotte and associated sub-parallel shear zones between Pacific and North American plates. The extinct outer arc covering the Coast and Insular belts of western British Columbia show thermal deficit constrained by surface heat flow measurements. A detailed study of heat flow variations and the locations of the shear failure zones show close correlations. Our study support that these remnant thermal deficit areas of extinct outer arc blocks in the western border of the Cordillera play an important role in the style of deformation of the transform boundary between Western North America and Pacific plates.
T21B-0587
Transtensional Crustal State of Stress in the Eastern California Shear Zone Derived From 3D Finite Element Analysis
Kinematic studies of deformation within the Eastern California Shear Zone conclude that the observed surface displacements are partitioned between a large dextral shear component of the plate boundary deformation and a smaller extensional component of the Basin & Range tectonics. Consequently, transtension is widely expressed in the ECSZ. The state of stress in the ECSZ can be generally characterised by NNE-SSW compression and WNW-ESE extension supporting the hypothesis of a current dextral transtensional stress regime in the ECSZ. However, information about relative or absolute stress magnitudes is sparse and when available, only for local areas. Furthermore, discrepancies exist as to whether the state of stress in the ECSZ represents a singular distinct regional state of stress with slip partitioning explaining different deformation styles, or if the state of stress variaes spatially and/or temporarily. This study presents an analysis of the regional state of stress in the ECSZ and especially in the Owens Valley-Airport Lake Fault Zone. Using multi-scale 3D finite element analysis we focus on how (1) gravitational potential energy differences (DGPE) from crustal thickness variations, (2) plate boundary forces and (3) fault geometries influence the 3D crustal state of stress. Comparison to independent stress orientations, magnitudes and observed tectonic regimes shows that the modelling approach reproduces the absolute crustal state of stress. The FE results are in agreement with independent data and yield a consistent state of stress in the ECSZ characterised by transtensional tectonic regimes with the major active faults being situated in corridors of strike- slip tectonics The geometrical impact of the bending segment of the SAF results in increased horizontal stresses in the Mojave Block and transfers strike-slip tectonics into the Eastern California Shear Zone to accommodate a part of the relative plate motion.
T21B-0588
The State of Stress in a Seismogenically Quiet Region: a Case Study in the Southeastern Korean Peninsula
We collected a number of in situ stress data (mostly from hydraulic fracturing stress measurements) and contemporary earthquake focal mechanism data to analyze the current state of stress in a tectonically young region of the southeastern Korean peninsula. The direction of the maximum horizontal principal stress estimated from 84 borehole measurements (maximum depth of 315 m) is N66°±31°E, close to the best-fit stress inversion using the focal mechanism results, N62°E. From the hydrofracturing test results, the vertical stress is the least principal stress, indicating a thrust faulting stress regime. We classify the study area into three divisions based on the magnitude of stress ratio K (\equivhorizontal stress/vertical stress). In the east and south divisions (respectively region I and II), the K value is relatively high, ranging between 2.5 and 3.2, while it is relatively low in the diagonally central division (region III), ranging between 0.8 and 1.4. It is interesting that the region of low K coincides with that of relatively high fracture (or lineation) density, due to the major fault zones including the Ulsan and Yangsan faults. This tacitly implies that the high-density fracture network might contribute to the stress release in the low K region. This is also corroborated by the fact that the moderate contemporary earthquakes (M≥3) are predominated in the proximity of eastern part of the study area. We extended our study to the analysis of mechanical stability of quaternary faults in this area. In some high K regions, faults oriented optimally with respect to the crustal stress field can be critically stressed based on the Byerlee frictional coefficient larger than 0.6. However, all known 23 quaternary faults are situated such that they are in the low K region. This fact again suggests a possible stress release due to the quaternary faults.
T21B-0589
Faults and Stress: Investigating Sources of Seismic Hazard in Southwest British Columbia.
The Cascadia subduction zone in Southwest British Columbia experiences both megathrust earthquakes at the subduction interface as well as large events from faults in the overlying crust. These crustal events, with recorded magnitudes of up to 7.3, pose significant hazard to major population centres in the area. Previous studies have shown this is a complex region of deformation located above a bend in the subducting Juan de Fuca plate. By investigating seismicity patterns and the state of stress, we aim to gain a better understanding of the forces driving earthquake activity and the resulting seismic hazard. We present evidence for new active structures in the region by using the double difference technique for earthquake relocation. These structures do not correlate with mapped surface faults, but have a similar strike to faults in the area. We also examine crustal anisotropy from shear-wave splitting measurements at both permanent and temporary stations in the vicinity of Vancouver Island. Past observations of crustal anisotropy in various regions around the globe have been interpreted as the fast direction being related to the orientation of the maximum horizontal compressive stress. However, anisotropy can be complicated by crustal structure, as indicated by variations of the fast direction between nearby stations. Sheared and foliated rocks due to faulting and metamorphism can influence crustal anisotropy. This appears to be the case for some of the Vancouver Island stations. To use seismic anisotropy as a stress indicator requires identifying which stations are influenced by stress and which by structure. This can be done for most stations, as the fault strikes differ from the maximum horizontal compressive stress direction. Preliminary results indicate that most stations show margin-parallel fast directions related to the crustal stress; however, stations located within a few kilometres of major faults have fault-parallel fast directions. The stations that show stress-related anisotropy tell us about the orientation of the maximum compressive stress and could be used to monitor temporal changes in the stress over extended time periods. Stations with structure-related anisotropy are used to investigate how crustal faults effect shear-wave splitting results.
T21B-0590
Study On Aftershock Triggering In Consideration Of Tectonic Stress Field
Abstract: The occurrence of earthquake is related to the strength of rock and tectonic stress field. The seismic risk factor (SRF),D=\left|{τn }\right|/(μσn ) is proposed to describe the dangerous status of aftershock triggering in this paper. D<1 in a region means no aftershocks to be triggered there, D≥ 1, dangerous and aftershocks probably occur. A main-shock is considered as an unstable dynamic process of stress release induced by fault softening suddenly in mechanics. It may be simulated by reducing the material parameters of the fault which is loaded by the tectonic stress field. The fault of the main-shock is received as a fault zone and the material in it is supposed to be transversely isotropic. The main-shock is simulated by decreasing the shear modulus in the fault zone. The magnitudes and directions of regional tectonic stress filed far away from the epicenter area of the main-shock may be supposed as constant, its directions can be estimated by jointly considering the average directions of P and T axes based on the focal mechanism solution of earthquakes, velocity field from GPS as well as geological survey. As one order of approximation, the magnitudes of the regional tectonic stress field can be estimated by the Coulomb failure criterion. Finite element method (FEM) and the concept of the factor D are used to study the aftershock triggering of the 1976 Tangshan Ms=7.8 earthquake. The results show that: (1) Most of the aftershocks triggered by the Tangshan earthquake occurred in the two-leaf-shaped regions of D≥ 1 near the north-east end of the main-shock fault. The largest leaf is about 100km long and 40km wide. (2) The areas of aftershock triggering predicted by the seismic risk factorD and Δ CFS (the changes in the Coulomb failure stress) are almost the same near the fault. The difference between them is that the aftershock area predicted by Δ CFS≥ 0 is too large and the area predicted by the factor D≥ 1 is limited. The areas of aftershock triggering predicted by Δ CFS≥ 0.04 MPa are nearly the same as those of D≥ 1 obtained by the study. (3) Sometimes Δ CFS =0.01MPa is taken as a low threshold of aftershock triggering. However, Δ CFS≥ 0 only means the probability increase of the earthquake triggering, not means the earthquake will occur. The earthquake occurrence is not only related to Δ CFS, but also to the tectonic stress field before the main-shock.
T21B-0591
Paleostress Analysis Applied to Fault-Slip Data from the Southern Margin of the Central European Basin System (CEBS)
We investigate the paleostress fields which controlled the post-Variscan evolution of the Central European Basin System (CEBS). Therefore, field studies are carried out in the marginal areas of the CEBS where Late Palaeozoic and Mesozoic rocks of the basin fill are present in outcrops bearing the imprints of several deformation phases that affected the basin system since the latest Carboniferous. Field studies including structural analysis, measurement of fault-slip data and careful collection of kinematic indicators provide the data base for this study. In the case of polyphase tectonics, the chronology of successive events is deduced and the total fault population from each site is qualitatively divided into different subsets, each being consistent with one specific stress regime. Since the stratigraphy and evolution of the CEBS are well known, temporal and spatial correlations of paleostress orientations are possible. Beside cross-cutting relationships derived from outcrops, we apply different graphical and numerical methods to separate the faults into homogeneous subsets. Depending on (1) the nature of faults (i.e. neoformed or reactivated), (2) the distribution of fault-slip data and (3) the deformation style, the deviatoric stress tensor is calculated for each subset using either the Numeric Dynamic Analysis or the Multiple Inverse Method. The results are obtained in terms of a reduced stress tensor, consisting of (1) the orientations of the three principal stress axes Sigma1, Sigma2 and Sigma3 with Sigma1>Sigma2>Sigma3 and (2) the ratio of principal stress differences, R =(Sigma2-Sigma3)/(Sigma1-Sigma3) with 1>R>0.
T21B-0592
PyLith: A Finite-Element Code for Modeling Quasi-Static and Dynamic Crustal Deformation
We have developed open-source finite-element software for 2-D and 3-D dynamic and quasi-static modeling of crustal deformation. This software, PyLith version 1.1, combines the quasi-static viscoelastic modeling functionality of PyLith 0.8 and its predecessors (LithoMop and Tecton) and the wave propagation and spontaneous rupture modeling functionality of EqSim. The target applications contain spatial scales ranging from tens of meters to hundreds of kilometers with temporal scales for dynamic modeling ranging from milliseconds to minutes and temporal scales for quasi-static modeling ranging from minutes to hundreds of years. PyLith is part of the NSF funded Computational Infrastructure for Geodynamics (CIG) and runs on a wide variety of platforms, from laptops to Beowulf clusters. It uses a suite of general, parallel, graph data structures called Sieve for storing and manipulating finite-element meshes. This permits use of a variety of 2-D and 3-D cell types including triangles, quadrilaterals, hexahedra, and tetrahedra. Current features include kinematic fault interface conditions, Dirichlet (displacement or velocity), Neumann (traction), and absorbing boundary conditions, linear elastic, generalized Maxwell, and Maxwell linear viscoelastic materials, and quasi-static and dynamic time-stepping. Future releases will add dynamic fault interface conditions (employing fault constitutive models), additional viscoelastic and viscoplastic materials, and automated calculation of suites of Green's functions. We also plan to extend PyLith to allow coupling multiple simultaneous simulations. For example, this could include (1) coupling an interseismic deformation simulation to a spontaneous earthquake rupture simulation (each using subsets of the software), (2) coupling a spontaneous earthquake rupture simulation to a global wave propagation simulation, or (3) coupling a short-term crustal deformation simulation to a mantle convection simulation and an orogenesis and basin formation simulation. http://www.geodynamics.org
T21B-0593
The role of upper mantle in lithosphere deformation: Insights from stress/strain evolution of a strike-slip fault
It has been under continuous debate whether the upper crust or the upper mantle plays a major role in supporting tectonic stress and determining lithosphere deformation. The strength calculations from lithospheric mineral composition, thermal field and experimental rheology generally suggest a stronger upper mantle, while other evidences, including earthquake depth distribution and postseismic/postglacial relaxation, suggest a weak upper mantle. The role comparison between upper crust and upper mantle becomes even more non- straightforward as faults and shear zones are taken into account. To improve our understanding of the upper mantle's role in lithospheric stress and deformation, I explore how lithosphere rheology structures affect stress/strain evolution of a strike-slip fault using visco-elastic numerical models. The model incorporates a strike-slip fault in the upper crust, which experiences cyclical stick-slip events, and various lower crust and upper mantle rheological structures, including horizontally homogenous upper mantle and heterogeneous mantle with weak shear zones. The results show that a stronger lower crust may couple the upper crust and upper mantle and transfer the effect of upper mantle to the stress/strain evolution of a stick-slip fault. Horizontally homogenous lower crust and upper mantle exert little stress on the upper crust before fault slip events, and exert backward- drag force right after fault rupture. If weak shear zones exist, however, the lower crust and upper mantle exert active drag force on the upper crust before fault rupture. This may cause a relatively lower postseismic stress in the lower crust and upper mantle, thus lower postseismic surface velocities. The magnitude of the coupled stress depends on the shear zone rheology and bulk viscosity of lower crust and upper mantle. These results indicate that the role of upper mantle in lithosphere deformation not only depends on bulk mantle rheology, but also rheology of lower crust and shear zones.
T21B-0594
The Microboudin Method: a New Paleostress Analysis
The microboudin method has been recently established as a new paleopiezometer. The method is successfully applied to monomineralic metamorphic tectonites such as metacherts and marbles, which include the microboudinage structures of columnar minerals (tourmaline, epidote etc.). The absolute magnitude of the far- field paleodifferential stress σ0 is estimated from the two parameters: the dimensionless stress parameter λ, and the extensional fracture strength S0* of micrometer-scale columnar mineral grains. λ is derived from the mechanical model for microboudinage based on the shear-lag model and the Weibull theory. The value of S0* for tourmaline and epidote were obtained from directly measured flexural strength by considering the size and shape of these minerals and influence of time on fracturing. The values of σ0 estimated by this method for samples from Aksu (China), Eskisehir (Turkey), Wadi Tayin (Sultanate of Oman), Asemi (Japan) Greenbushes (Australia), and Sausar (India) are in the range of 1-15 MPa. These σ0-values correspond to the differential stress at mid-crustal levels (> 10 km).
T21B-0595
Uncertainty in the Relationship Between Lithospheric Structure and the Surface Stress Field
Variations in lithospheric thickness and density contribute to the Earth's surface stress field. The relationship between the continental lithosphere's isostatic state, sub-crustal structure and the surface stress field, however, remains unresolved due to the poor constraints on its thickness, composition and rheology. Here, we present calculations that systematically explore variations in the surface stress field as a function of isostatic compensation mechanism, dynamic vs. observed elevations, sub-crustal continental lithosphere structure and assumed "coupling depth". We compute the gravitational potential energy (GPE) of 2x2 degree lithospheric columns and then solve for the resulting stress field in a 3D elastic finite element model. Consistent with previous calculations, the stress regime and orientation varies significantly between continental lithospheric structures with compensated vs. uncompensated crustal layers. Imposing Pratt vs. Airy compensation, however, has comparatively minor effects on the overall stress orientations and regimes. Subtracting dynamic topography calculated from a mantle flow model results in additional large variations in stress orientation and regime in areas with large values of positive or negative dynamic topography. Varying the thickness of the continental mantle lithosphere as a function of crustal thickness or tectonic province does not produce significantly different results than models where the thickness is set by the difference between the crustal and compensation depths. To test the relationship between the surface stress field and decoupling between brittle and ductile portions of the lithosphere, we decrease the depth to which the GPE is calculated. We find that the stress field varies strongly as the global GPE integration depth decreases, which suggests that stresses may change significantly as the rheology of the lithosphere varies both vertically and horizontally. Finally, we address a number of assumptions inherent in the GPE method that oversimplify or fail to address rheological and structural complexities in the continental lithosphere, which may strongly influence the surface stress field.
T21B-0596
Basal Shear Traction, not Slab-Pull, may be the Dominant Plate-Driving Force
We previously reported [2006 AGU T11F-05] that we used thin-shell finite-element program SHELLS [Kong & Bird, 1995, JGR] to model the 52 plates of the PB2002 plate model [Bird, 2003, G3], and adjusted fault friction to the value of 0.1, at which we obtained the best possible fit to global geodetic velocities, stress directions, seafloor spreading rates, and seismic anisotropy. Here we analyze the forces and torques in the best model, in an attempt to better understand the balance which drives and resists plate motions. All boundary tractions acting on any plate are divided into 3 types: lithostatic pressure, side-strength, and basal- strength. ("Strength" is a modified stress tensor from which lithostatic pressure has been subtracted.) Side- strength is dominated by fault friction. Basal-strength includes distributed basal shear tractions (for all plates) and net slab-pull (for subducting plates only). Computed basal strength tractions are very large for some small plates, but we show that these large values are probably dominated by errors resulting from locally-incorrect (global-mean) fault friction used to compute side-strengths. Any possible correlations with physical characteristics of plates is concealed by these large errors in basal-strength tractions for small plates. However, results for large non-subducting plates (whose mean side friction will more closely approximate the global mean) suggest basal shear tractions of 1 MPa or less. Since Forsyth & Uyeda [1975, GJRAS], many have considered net slab-pull to be the dominant driving force on subducting plates. To investigate such a model, we compute net slab-pull for 11 plates with large subducting slabs by assuming negligible distributed basal shear tractions, and ascribing all basal-strength forces to net slab-pull (as line forces along trenches, acting only on the subducting plate). These model-dependent net slab- pull forces are mostly of plausible magnitude (<4x1012 N/m, except for outliers with large uncertainties). However, they show no clear correlation with: known sea floor age, relative or subduction velocity, trench depth, age-corrected trench depth, or trench length. While this lack of systematic relationships could be due to additional unknown errors in our model, we interpret these negative results as tending to disprove the initial assumption. That is, distributed basal shear traction forces are probably of comparable magnitude to, or greater than, net slab- pull forces, and the primary determinant of plate motion. Plate velocity, forward net slab-pull, and forward basal shear tractions are examined using Africa-fixed, no-net- rotation, and NUVEL1A-HS3 (hotspot) "absolute" reference frames. All analyses show basal tractions to be primarily forward, with the no-net-rotation reference frame leading to the most positive (forward) results for basal strength tractions. Our interpretation is that active mantle convection is typically driving plate motions with basal shear tractions, rather than resisting them. http://peterbird.name
T21B-0597
Timescale-dependent crustal deformation in the southwestern U.S.
Space-based geodesy has greatly refined crustal deformation in the southwestern United States, but how the short-term strain rates relate to long-term deformation remains uncertain. The crustal kinematics based on GPS measurements differs significantly from that derived from geological reconstruction. We have developed a three- dimensional viscous finite element model to explore the driving forces for present-day crustal deformation in the southwestern US, and to reconcile its discrepancy with geological records for the past few million years. Our results show that plate boundary traction can reproduce the GPS velocity fields but not the observed stress field. Conversely, gravitational buoyancy force produces a good fit to the stress field but contributes little to the GPS velocity field. Models with both forces combined can explain the observed crustal deformation and stress patterns. The optimal plate boundary traction constrained from GPS data is too high to fit geologically determined crustal kinematics. We suggest that the plate boundary traction responsible for long-term crustal deformation should be lower than that constrained from GPS data, which represent mainly interseismic strain rates. By lowering the plate boundary traction, our model produces close fits to the observed long-term surface velocity and stress orientations. The fits may be further improved with a slightly weakened lithosphere, presumably associated with a stronger mantle thermal perturbation in the geological past. http://www.missouri.edu/~yangyo
T21B-0598
Analysis of Stress Field in Caribean Continental Plate - Southern Costa Rica zone - from CMT Solutions
During the of period 1984 – 2007, 51 earthquakes were registered with magnitudes above 5.1 mb in southern Costa Rica. Depth distribution of hypocenters shows shallow (0 – 20 km) earthquakes located in continental area; intermediate depth hypocenters (20 – 60 km) corresponding to subduction of the Coco plate under the Caribean plate; and just one earthquake have focus under 60 km depth. We choose hypocenters shallower than 100 km because our goal is to study intraplate stresses. Centroid Moment Tensor solutions for subduction zone earthquakes have inverse fault mechanism. But, in continental area there are evidences of strike slip and normal fault mechanisms. Direction of main stress (σ3) at the southern continental zone is NE-SW; in the Pacific ocean border region, the main stresses are oriented parallel to the coast line; inside the valley region, mechanisms are predominantly strike slip with σ1 in the near N-S direction. This is a complex stress field, with rotation of main stresses in a short space. σ1 is vertical in some continental areas due to influence of elevated terrain. Stress axis paralell to coast line maybe also due to gravitational body force.
T21B-0599
French Polynesia Hotspot Swells Explained By Dynamic Topography
Situated on the South Pacific Superswell, French Polynesia is a region characterized by numerous geophysical anomalies among which a high volcanism concentration. Seven hotspots are required to explain the observed chains, volcanism ages and geochemical trends. Many open questions still remain on the origin of these hotspot chains: are they created by passive uplift of magma due to discontinuities in the structure of the lithosphere or by the ascent of mantle plumes? In this case, at which depth do these plumes initiate in the mantle? Many geophysical observations (bathymetry, gravity, magnetism, volcanism ages..) are used to understand the unique phenomenon occurring on this region. The most useful information may come from tomography models since they provide a 3D view of the mantle. Until recently, the tomography models over the region were quite inaccurate because of the sparse location of the seismic stations. The deployment of two new seismic stations networks (BBOBS and temporary island stations) has lately remedied this failing. The resulting tomography model obtained through the inversion of Rayleigh waves provides the most accurate view of the shallowest part of the mantle (depths ≤ 240 km) beneath French Polynesia. Indeed, for the first time the accuracy of a tomography model is good enough to provide information about plume phenomenology in this complex region. In order to quantify the plumes effect on the seafloor, we compute the dynamic topography through an instantaneous flow model. The general trend of the observed depths anomalies (highs and lows) is well recovered. For example the amplitude, location and extension of the swells associated with the Society, Macdonald and Rarotonga are accurately described by the dynamic model. We also find that dynamic uplift is associated with the Tuamotu archipelago which means that a part of the observed swell is due to the present day action of plumes. Since no volcanism ages are available over this chain, this new information may be quite important in understanding the archipelago origin. Another interesting result is that Arago, which is supposed to be an active hotspot along the Cook-Austral chain is situated on a bathymetric low which is well recovered by the dynamic model. Since this region is associated with downwelling flows, this makes us question its hotspot origin.