T23B-2011
Slope failures in subduction zones revealed from analogue models
Submarine landslide is a common feature along subduction margins and recently regarded as one of the most serious geohazards. This study performed a series of shortening physical experiments and time-lapse digital pictures of deformation are analysed with an image correlation technique; PIV. A number of slope failures were observed in the models, and were concentrated at active deformation structures, particularly at the frontal thrust. The slope failures can be classified into two types, Type I developing at the lower half of the slope, and Type II covering the entire slope. Each event of slope failures may correspond to cyclic activity of thrust fault, thus the pile of the failed sediments stored in the footwall should records the history of thrust activity. Slide deposits that can be explained by these models have been observed in recent and Miocene prisms at the Nankai Trough, Japan.
T23B-2012
Investigating the Role of Extensional Deformation at Convergent Margins Using a Combined Analog and Numerical Approach.
Understanding why extensional stress occurs and how it is accommodated at convergent margins remains fundamental to understanding the structural and mechanical evolution of wide array of contractional orogens. We investigate the parameters that control the occurrence of extension coeval with contraction, specifically, variations in rheology (laterally and with depth) and initial plate geometries, (i.e., plate motion obliquity). Understanding the specific contribution of these parameters to the development of contractional orogens is fundamental to a better understanding of the tectonics and seismic hazards of many convergent margins. To this end, we have used a multifaceted modeling approach that employs 3-dimensional analog modeling with quantitative analysis in conjunction with 2 and 3-dimensional finite element numerical modeling. We find that a relatively small amount of extensional stress will play an important role in the long-term structural evolution of convergent systems. However, the manner in which this stress will be translated to deformation (localized on discrete accommodation structures), and in particular how the style of strain (extension or contraction) will evolve, is a strong function of rheology at depth and the initial geometry of the margin.
T23B-2013
Influence of the Convergence and Stepover Angles in the Structural Style of Strike-Slip Systems: Analogue Models
The presented results of analogue models analyze the influence of the relation between the convergence angle and the stepover angle on strike-slip systems. The experiments include 2 stepovers arrays to generate both transpressional and transtensional zones. These experiments were prepared using 5 cm thick sandpack (sand diameter <500 μm, internal friction angle 30°, density 1,400 kg/m3) to simulate brittle deformation; the base of the model was formed by thin zinc base plates, one of them mobile, cut in such a way so as to produce restraining and releasing strike-slip stepovers; the rate convergence was constant. We carried out 3 series of experiments in which the convergence angles (0°-60°) and stepover angles (30°-60°) were varied systematically. Preliminary results indicate that by increasing the angle between the stepover and the convergence vector: a) the restraining area generated a positive flower structure that is progressively wider; b) progressively more reverse faults which absorb more shortening were generated. Locally, strike-slip faults in the positive flowers were accommodated by the geometry of the stepover base. Areas most complex involved the development of normal faults, which subsequently were inverted. In conclusion, the relation between the convergence and stepover angles is a main factor that determines the structural style of flowers structures on strike-slip systems. This work was done through the ACT-18 PBCT project.
T23B-2014
How to exhume an arc-continent collision ophiolitic mélange: An Example from Taiwan
We used time-space equivalence concepts to study the evolution and exhumation of the Lichi ophiolitic mélange in Taiwan in an oblique convergent zone. By analyzing several cross sections from younger subduction zone to mature collision zone, we studied how the mélange was formed and exhumed to the surface. Using previously unpublished seismic reflection profiles and side-scan sonar images, we have documented strong evidence of complex mass-wasting and faulting processes. For the first time, we interpret these observations using taper wedge model. The pre-mélange setting is driven by a major backthrust that propagates arcward on top of the forearc basement and incorporates the forearc basin strata into the rear of the accretionary prism in the subduction zone. In the initial collision zone, the backthrust steps down into the basement and incorporated the forearc basement materials into the rear of the accretionary prism. In the process the basement materials were incorporated into the hanging wall of the backthrust and transported arcward, as suggested by our previous gravity modeling results. The mass wasting processes and faulting are in response to evolving dip angle of the backthrust between the irregular forearc basement topography and bathymetry. These multi-phase tectonic deformations and surficial processes work together to form the combined olistostromal and tectonic features of the Lichi Mélange. And similar combinations of such processes might have generated some ophiolitic mélanges in other collision zones.
T23B-2015
Unraveling the tectonic processes behind the contemporary observed deformation rates in southern Alaska
Southern Alaska forms part of the complex tectonic boundary between the North American and Pacific plates where the interplate boundary transitions from strike-slip to flat and oblique subduction associated with microplate collision to normal subduction. The deformation of this broad plate boundary has been extensively observed by a large array of GPS receivers that illuminate a wide variety of current (1996-2002) deformation characteristics. These include northwestward directed velocities that diminish rapidly across the Denali Fault and a region of southeastward directed velocities near the site of the great 1964 Alaska earthquake. Here we attempt to explain the major trends of the deformation by using a 3-D viscoelastic Lagrangian finite element model that incorporates the complex geometry of the Pacific slab as it subducts beneath North America, the major earthquakes in the region during over the past half century (just prior to the 2002 Denali Fault earthquake), and postseismic relaxation of a mobile lower crust and mantle associated with these events. Results suggest that the deformation field is dominated by convergence of the subducting Pacific plate. However, in order to explain the rapid drop-off in velocities across the Denali Fault, the regions to the south must be substantially weaker mechanically. This would be consistent with distributed brittle behavior throughout these accreted allochthonous terranes, each of which has undergone extensive internal deformation in the past. In addition, on-going viscoelastic relaxation associated with the 1964 earthquake is consistent with southeastward directed velocities in the vicinity of the western Kenai Peninsula. And postseismic relaxation associated with earthquakes in 1949, 1958, and 1972 along the Queen Charlotte - Fairweather fault system explains well the lack of northwestward directed velocities east of the Fairweather fault. The locking depth of the megathrust and the viscoelastic structure also play important roles in the response of the region to plate convergence.
T23B-2016
A Mechanical-Thermal Coupled FEM Model for Tectonic Deformation in Subduction Zones
In plate subduction zones mechanical and thermal processes interact with each other on a geological time scale in the following way. Oceanic plate subduction induces internal flow, which transfers heat in advection and changes temperature distribution in subduction zones. The temperature change leads to the alteration of lithosphere-asthenosphere boundaries. We modeled such mechanical-thermal interacting processes in subduction zones with FEM. First, we developed a mechanical FEM model to compute long-term internal velocity fields due to plate subduction. Next, we developed the thermal FEM model to compute temperature changes due to thermal diffusion and advection. Then, combining these models, we constructed a mechanical-thermal coupled model, and numerically simulated the evolution process of mechanical and thermal structure in plate subduction zones. In this simulation, at a certain time step, we compute internal velocity fields due to plate subduction. Then, using the computed results, we evaluate temperature changes, and update the lithosphere-asthenosphere boundaries to compute internal velocity fields at the next time step. From numerical simulations with the mechanical-thermal coupled model we revealed that the mechanical and thermal structure in subduction zones develops with time as follows. In the early stages of plate subduction (0-2 Myr), cooling the mantle wedge leads to the thickening of the lithosphere near the plate boundary, which increases the uplift rates of the island-arc lithosphere. As time goes on, the thinning of the island-arc lithosphere due to heat transfer in advection proceeds, which develops a localized high uplift-rate zone in the island-arc.
T23B-2017
Kinematic Model of the Izu Peninsula in the Northern Margin of the Philippine Sea Plate
The Izu Peninsula in the northern margin of the Philippine Sea plate (PH) is the end of the Izu-Bonin arc that is colliding with the central Japan (CJ). Nationwide continuous GPS array has shown that the motion of the peninsula is significantly deviated compared with the stationary motion of the main part of PH. We propose a new kinematic model to explain a convergence process of the northern margin of PH. We consider the Izu Peninsula and its surrounding form a crustal block that has been detached from PH. At the same time we assume the Izu block suffers from significant internal deformation between PH and CJ. At first we determine a rigid rotation (the Euler pole and rotation rate) of the Izu block relative to PH and CJ. Then rigid motions predicted from the above rotation pole and rate are subtracted from the observed GPS velocities. The residuals show systematic strain patterns that are characterized by no remarkable deformation near the collision boundary with CJ and clear NW-SE contraction inside the block. Next we combine a collision at the root of the peninsula (Ishibashi and Itani, 2004) and a horizontal detachment beneath the peninsula (Seno, 2005) to explain the internal deformation. Thus the surface deformation can be expressed by a summation of two effects, collision and detachment, which can be evaluated by tensile and shear faults in an elastic half-space, respectively. However, relative magnitudes of these effects are unknown. Therefore we search an optimum model that reproduces the observed deformation field varying relative magnitudes of two effects. One possible model is that about 60% of the relative motion between PH and CJ is absorbed by the collision and the rest by the slip on the detachment. The remaining problems are exact location and configuration of the boundary between the Izu block and PH, strength of interaction at that boundary, and a horizontal extent of the detachment. We will refine a model using GPS data obtained at islands and reefs located near the possible boundary between the Izu block and the main part of PH.
T23B-2018
How does lateral variation in basal friction control the evolution of accretionary wedge as well as interactions between fault kinematics and surface processes?
In the last few decades, the kinematics and mechanics of accretionary wedges have been extensively studied utilizing different numerical and analogue simulation techniques. In accordance to the critical taper theory, all theses studies revealed that internal and basal friction are key factors controlling the kinematics and mechanics of accretionary prisms. Particularly internal deformation mode, mass transfer pattern, and wedge geometry, e.g. slope taper, dip angle of accretionary faults and out-of-sequences thrusts, are controlled by (a) bulk strength of undeformed and accreted strata and (b) friction conditions at the basal décollement. However, to date only few 3D modeling studies have dealt with the deformation behavior and fault kinematics as a function of lateral variations along the deformation front of the coefficient of basal friction. In addition, effects of the kinematics of accreted thrusts and faults on the evolution of slides along the slope surface ave not yet been accomplished. Main purpose of this study is to determine the influence of lateral variations of friction conditions at the décollement on the mechanics and kinematics of accretionary wedges. In particular, we focus on fault kinematics and surface processes as a function of basal friction. To achieve this, the position and geometry of thrusts and faults, their re-activation behavior in time and space, as well as interactions between fault activity and slope destabilization are analyzed as a function of spatial distribution of basal friction along the décollement. For these studies, we use a numerical particle-based method - the Discrete Element Method (DEM), which allows investigation of such complex deformation processes in the upper brittle crust in three dimensions. This technique enables the testing of a wide range of material parameters and model configurations. Furthermore, detailed information about wedge evolution, internal structures and mass transfer modes at arbitrary horizontal or vertical slices could be retrieved. This results in a high temporal and spatial resolution of deformation processes and mass transfer modes. Hence, these 3D models reveal a deeper understanding of interactions between long-term accretion modes and short-term slope destabilization events.
T23B-2019
Modeling of the seismic response of subducting Cocos plate in Central Mexico
The subducted Cocos plate beneath Central Mexico as imaged with receiver functions from the MASE (Meso- American Subduction Experiment) array tectonically underplates the continental crust for a distance of approximately 300 km from the trench. In this study we construct a velocity model to match the impedance changes at the interface between the crust and slab as well as mid-crustal features. The receiver functions are modeled with a 2D finite-difference code, and the result shows that the need for a low-velocity zone (lower than normal oceanic crustal velocities), corresponding to a thin serpentinized layer on the top of the subducting slab, to reproduce the impedance contrasts. We also find that in order to match the weak multiples in the receiver functions, the interface needs to be rough on the scale-length of 3 km. We also include random heterogeneities in the crust to explain the short-scaled time/depth variations that are seen on the crust-slab interface.
T23B-2020
Three-Dimensional Particle Dynamics Simulations of Oblique Plate Convergence Using Bonded Particle Assemblages
Many convergent boundaries around the world exhibit significant curvature (e.g., Himalayas, Caribbean Arc, Aleutian arc), resulting in spatially varying convergence directions. The along-strike transitions from normal convergence to oblique convergence can produce a range of deformation structures reflecting the degree of strain partitioning along the boundary, including strike-parallel extension and strike-slip faulting. To explore the geometry and controls on these structures in more detail, and to better resolve the stresses and strains responsible for such deformation, we have carried out three-dimensional particle dynamics simulations using the discrete element method (DEM). Simulations are carried out in a numerical sandbox scaled to 20 km wide by 60 km long. A curved vertical wall of particles defines the backstop, which emerges through a fixed planar wall of particles into a layer of particles sedimented under gravity. Furthermore, particles have been bonded together to simulate cohesive rock, allowing brittle failure to occur and persistent faults to develop. The geometries and slip directions of these deformation structures can be resolved in detail using visualization tools provided by Generic Mapping Tools (GMT) and Earth Vision (EV). We extract and display specific particle properties (e.g., displacements and strains) and assemblage attributes (e.g., stress components) to fully interpret and analyze the deformation history and stress evolution of the growing wedge. Spaced imbricate thrust faults develop at the head of the backstop, where convergence is normal, whereas distinct strike-slip faults develop along the lateral margins. As expected, the oblique convergent zones exhibit more complicated deformation, including dip-slip faults and en-echelon extensional fault arrays. We compare the emergent structures with examples from analogous plate boundaries to assess the validity of the models.
T23B-2021
Relocation of Central Oregon Subduction Zone Earthquakes using hypoDD
Historically, the subducting Juan de Fuca plate has produced very large thrust earthquakes along the Cascadia subduction zone. These events occur every few hundred years on average, with very little documented seismic activity in the interim. Since 2003, about 40 earthquakes have been detected in the nominally "locked" zone offshore central Oregon. Analysis of the two largest earthquakes (MW 4.8 and MW 4.9) by Trehu et. al. (Geology, 36, 127-130, 2008) suggests that they were low angle thrust events on the plate boundary. Due to the station coverage in this area, many smaller events were not well recorded, and their catalog locations have large uncertainties. Using a double-difference algorithm (hypoDD), we relocated 30 cataloged events from 2003 to 2008 relative to one another. Our data consist of earthquake catalogs and station waveforms from the Pacific Northwest Seismic Network and the USArray Transportable Array deployed in Oregon during 2006-2007. Relocation confirms that these earthquakes occurred in distinct clusters around the two largest events (12 July and 19 Aug. 2004). The northern cluster, at 44.6N, appears to be split into two separate clusters, with the 19 August event and its three immediate aftershocks occurring ~5 km northeast of the tightly clustered group of 2006-2007 events. The southern cluster, at 44.3N, remains elongated in a NE direction. The new hypocenter distribution decreases the depth range within each cluster, though relative depths between clusters, as well as absolute event depths, depend on the velocity model and initial locations. We are currently operating an onshore/offshore seismic array (COLZA – Central Oregon Locked Zone Array) to better constrain microseismic activity in this region. Preliminary analysis of ML 2.1 and 2.9 events that occurred on 29 Apr. and 1 May 2008, in the southern cluster indicates that depths are shallower than reported in the ANSS catalog. Three-dimensional velocity heterogeneity, as determined from active source seismic experiments and potential field modeling, will also be incorporated into the analysis.
T23B-2022
Where Does the Seattle Fault End? Structural Links and Kinematic Implications
The Seattle fault is one of several east-trending compressional structures in the Puget Lowland (PL), seemingly at odds with the northeasterly oriented compression along the Juan de Fuca subduction zone. The existence of these faults is thought to be related to the northward movement of a strong Oregon forearc block. A weaker PL block accommodates north-south shortening between Siletzia and the slower-moving Canadian Coast Mountains to the north. The northward movement of the PL requires either the Cascade and Olympic Mountains to move northward and shorten at nearly the same rate as the PL, or the existence of strike-slip accommodation zones bounding the PL. We use results from three study areas along the Seattle fault to constrain its behavior: the westward terminus at the foot of the Olympic Mountains, its central reach near Bainbridge Island and its eastward terminus in the Cascade foothills near Fall City, WA. Geologic map data, trench observations across faults, Lidar topographic scarp observations, seismic reflection profiles and potential field anomalies are integrated to determine fault structure. These data indicate that the Seattle fault extends further east and west than previously thought. This suggests its connection to strike-slip fault zones bounding the east (Rattlesnake Mountain fault zone, right-lateral) and west (Saddle Mountain deformation zone, left-lateral) ends of the fault. Gravity and aeromagnetic anomalies along the Seattle fault are best modeled by a fairly simple, reverse fault (dipping south 35-50 degrees). The strike of the proposed PL-block bounding faults suggests a kinematic explanation for the existence, position and dip of the Seattle fault and other east-striking compressional structures in the region. An analog (clay) model illustrates the growth of both the Seattle uplift and the Kingston arch resulting from these proposed kinematic bounding conditions. The strike-slip faults form a crustal "funnel", narrowing to the north, which squeezes the PL block from the east and west as it moves northward. The reverse Seattle fault results from a buckling of the PL which accommodates both north-south shortening and east-west space accommodation problems. East-west compression is supported by transpression on both the Rattlesnake Mountain and Saddle Mountain deformation zones. The latitude of the Seattle fault corresponds with the most abrupt narrowing of the "funnel" at the southward terminus of the Olympic Mountains-bounding fault zone, which suggests a correlation between the position of the east-striking thrust faults in the PL and strain accommodation in the Olympic Mountains.
T23B-2023
Processes Controlling Spessartite Generation beneath Mt. Rainier, Washington
Mt. Rainier is a composite volcano located in the Cascade volcanic arc about 60km SE of Tacoma, Washington (Crowley et at., 1997). The volcanism at Mt. Rainier is a consequence of subduction of the Juan de Fuca plate under the North American plate at an average rate of 50mmyr-1 (Lowrie, 2007). Based on a compilation of data representing basaltic andesites through dacites (Sisson & Vallance, submitted, Sisson, personal communication) 125 MELTS (Ghiorso & Sack, 1995) models have been run to constrain the mechanisms of formation for the typical liquid line of descent. Using a variety of parental magmas, models were run through pressure space of 1-10 kbars, with initial water contents from 1-5 wt. %, and with fO2 buffers of QFM or NNO. The best fit model for the Mt. Rainier rocks involves crystallization at 1-1.5 kbars from a melt with an initial H2O content of 3.5 wt. % and oxygen fugacity of QFM. Comparing the best fit MELTS models with the suite of associated rocks from Mt. Rainier provides evidence for fractional crystallization and magma mixing as the dominant processes, consistent with the work of Sisson & Hankins (2007). While most Mt. Rainier lavas typically contain phenocrysts of plagioclase and pyroxene, one particular lava flow classified as a spessartite andesite possesses phenocrysts of amphibole and pyroxene. The spessartite also has a high concentration of K2O relative to typical andesites and is enriched in LREEs and LILEs. Fractional and equilibrium crystallization MELTS models have not successfully reproduced the trends exhibited by the spessartite, suggesting that the magma evolved by processes different than simple crystallization; MELTS modeling of more complex petrogenetic scenarios is currently underway. Electron microprobe data of amphibole and pyroxene from the spessartite demonstrate a complex history of crystal growth as well. The amphiboles range in composition from ~2 to 15 wt. % Al2O3 and ~6 to 12 wt. % FeO*. Pyroxene phenocrysts have Mg# ranging from ~63 to 84. Most phenocrysts possess thin mafic reaction rims, and many amphiboles have rims of opaque oxides. The additional MELTS models, coupled with detailed examination of core to rim phenocryst data, will explore the importance that processes such as magma recharge and wallrock assimilation have on the formation of the spessartite magma.
T23B-2024
Spatial Temporal Relationships and Rhyolite Geochemistry of the High Lava Plains, Oregon: Compositional Variations and Relationships to Other Volcanic Provinces
The High Lava Plains (HLP) are a Cenozoic bimodal basalt-rhyolite province located in southern and central Oregon, east of the Cascade Range and in conjunction with the Columbia River basalts, Steens basalts and Yellowstone - Snake River Plain (SRP) track, are part of the largest Cenozoic volcanic province in the world. New 40Ar-39Ar dates show two east to west time-transgressive patterns to the rhyolitic volcanism. Chemically, the <12 Ma rhyolites of the HLP separate into three distinct groups in major and trace elements. One age progression, which has been described by previous workers, runs sub parallel to the Brothers Fault Zone, a set of small, diffuse, northwest-striking normal faults with offsets generally less than 10 m. This trend spans from ~12 Ma to Recent eruptions at Newberry Volcano. Typically, these rhyolites are high silica (SiO2 > 75 wt. %) and follow a tholeiitic differentiation trend. The other age progression occurs in a north-south belt 60 to 110 km east of and parallel to the Cascade arc and ranges in age from is ~7.5 to ~4.5 Ma. These rhyolites are typically lower in silica and follow a calc-alkaline differentiation trend. These two age progressions overlap in the northwest corner of the HLP. Peralkaline rhyolites are also distributed throughout both age progressions and are high silica (SiO2 ~76.5 wt. %), high FeO/ast (> ~2 wt. %), high Zr, Nb, Ta and are significantly depleted in elements with a strong affinity for plagioclase (Al, Ba, Sr, Eu). The basalts of the province show no time- transgressive behavior and are primarily primitive HAOT (high-alumina olivine tholeiites) that have undergone variable amounts of fractionation. The tholeiitic differentiation trend rhyolites of the HLP are chemically similar to Icelandic rhyolites and both provinces are underlain by thin, mafic crust. The crust of the HLP is covered by extensive Cenozoic volcanism and sedimentary cover and likely consists of stitched Mesozoic accreted terranes. It is less mafic and slightly thicker (~35 km) than that of Iceland. These differences in crustal character may explain the differences in eruptive products between the two provinces, specifically the basalt to rhyolite ratio (HLP 2:1, Iceland 10:1) and higher iron at similar silica in Icelandic rhyolites. The tholeiitic trend rhyolites are also similar in composition to the Quaternary rhyolites of the SRP.
T23B-2025
Preliminary Analysis of Multibeam, Subbottom, and Water Column Data Collected from the Juan de Fuca Plate and Gorda Ridge Earthquake Swarm Sites, March-April 2008.
Two oceanographic expeditions were undertaken in the northeast Pacific during April and September of 2008 to collect a variety of scientific data at the sites of intense earthquake swarms that occurred from 30 March to 9 April 2008. The earthquake swarms were detected by the NOAA/PMEL and US Navy SOSUS hydrophone system in the northeast Pacific. The first swarm occurred within the central Juan de Fuca Plate, ~280 km west of the Oregon coast and ~70 km north of the Blanco Transform Fault Zone (BTFZ). Time history of the events indicate this swarm was not a typical mainshock-aftershock sequence, and was the largest SOSUS detected swarm within the intraplate. This intraplate swarm activity was followed by three distinct clusters of earthquakes located along the BTFZ. Two of the clusters, which began on 10 and 12 April, were initiated by MW 5+ earthquakes suggesting these were mainshock-aftershock sequences, and the number of earthquakes on the BTFZ were small relative to the intraplate swarm. On 22 April, another intense earthquake swarm began on the northern Gorda Ridge segment adjacent to the BTFZ. The Gorda swarm produced >1000 SOSUS detected earthquakes over a five-day duration, with activity distributed between the mid-segment high and the ridge-transform intersection. This swarm was of special interest because of previous magmatic activity near its location in 1996. Overall, the March-April earthquake activity showed an interesting spatio-temporal progression, beginning at the intraplate, to the transform, then to a spreading event at the ridge. This pattern once again demonstrates the Juan de Fuca plate is continually moving and converging with North America at the Cascadia Subduction Zone. As the initial swarm was not focused on the ridge crest, it was not interpreted as a significant eruptive event, and we did not advocate a large-scale Ridge2000 response effort. The earthquake activity, however, did have an unusual character and therefore a short (four-day) cruise was organized using the R/V Wecoma in April (support via NOAA Vents Program and NSF). While this cruise was underway, the Gorda Ridge swarm began and therefore another day was added to also sample the Gorda site. A total of 11 CTD casts were completed, covering the significant areas of earthquake activity. Measurements for helium isotopes have been completed on all 11 casts, and for methane and CO2 on one of the Gorda Ridge casts. A second response cruise aboard the R/V Melville will take place in September, funded by the NOAA/Vents, providing 2 days of multibeam survey time. The cruise plan is to collect EM120 multibeam bathymetry and backscatter data, as well as 3.5 kHz subbottom in the area of the initial swarm. The northern Gorda Ridge will also be surveyed, with the goal of comparing this bathymetry with previously collected data to see if there is evidence of depth anomalies and therefore recent seafloor eruptions.
T23B-2026
Geology and Geophysics of the High Lava Plains and Associated Geologic Provinces of the Northwestern United States: Insights into the Evolution of the Continental Lithosphere
In an attempt to better understand the processes and mechanisms responsible for the geological evolution and structure of the eastern portion of the Pacific Northwest, we have undertaken an integrated analysis of gravity, magnetic, seismic profiling, petrologic, receiver function and structural data from an area comprising parts of Washington, Oregon, California, Idaho, Utah and Nevada. Geochemistry, geochronology, structural geology and GPS data are also being used to constrain the interpretation of observed geophysical features. The current state of our analysis suggests a mixed influence of mutually related subduction, extension, transform displacements, and asthenospheric upwelling. Of particular interest is the occurrence of a surprisingly low gravity anomaly across the volcanic trend leading to Newberry volcano in central Oregon from southeasternmost Idaho. This is unusual for a region with such voluminous mafic igneous magmatism in general and more so when compared to the contemporaneous volcanics of the Eastern and Western Snake River Plain, which show very significant high gravity anomalies as would be expected. The anomalous low gravity values encompass the entire area of the 17 Ma to 5 Ma aged volcanics east of the Cascades that constitute the High Lava Plains that partly overlie the northern reach of the Basin and Range. These volcanics along the High Lava Plains do not appear to have a clear relation with the dominant northwest trending fault structures common to this region. Along the Brothers fault system for instance, local gravity anomalies partially align with this dominant northwest fault trend in some areas, and thus appear to be structurally controlled while other areas along the fault system show no such alignment. Also observed in the gravity data is a linear series of gravity highs stretching from southeast Washington southwards into central Nevada approximately along the Oregon – Idaho border. This gravity "lineament" appears to represent the eastern limit of the Precambrian craton of North America and coincides with the 87Sr/86Sr > 0.706 line. A low gravity anomaly is also observed within the Harney basin of eastern Oregon. This gravity low is encircled by a local relative gravity high to form a ring defining the basin limits. Northwest and north- northeast oriented faults also form a ring around this area in fairly close agreement with the gravity anomalies. Magnetic anomalies also appear to outline the bounds of the Harney basin. However, while the volcanics of the Western and Eastern Snake River Plain are clearly defined in the magnetic data, the volcanics of the High Lava Plains are much less clearly defined.
T23B-2027
Variability in New Shortening Estimates from Southern Peru (12-14S); Implications for Mass Balance of the Andean Plateau.
One of the fundamental questions of interest with regards to the Andean Plateau is the mass balance of material needed to create and sustain a 3-4 km high plateau. Is crustal shortening sufficient to support an isostatically compensated crust of 60-70km? We present new estimates of shortening across the northern margin of the Andean Plateau. The cross section extent, from the eastern edge of the volcanic arc to foreland basin, is approximately one half of the physiographic width of the Andean Plateau in Peru. Cross sectional shortening estimates in southern Peru (12-14°S) provide a best estimate of 123 km or 40% shortening with an absolute minimum estimate of 86 km or 30% and absolute maximum estimate of 275 km or 60%. We determined the maximum and minimum shortening estimates using the cross sectional area and possible variations in assumptions made about the amount of erosion, detachment dip, involvement of basement thrusts and displacement along faults. The best estimate of shortening is well short of the required 240-300km of shortening needed in order to account for a 60-70km thick crust under the entire plateau. This suggests that for an isostatically equilibrated crust either 1) there is a significant amount of shortening (~150km) in the western half of the plateau which, is hidden by the volcanic arc or 2) crustal material is being added to the Peruvian section of the Andean Plateau either through lower crustal flow or a process of magmatic underplating followed by differentiation and delamination.
T23B-2028
Orogenic Float Model: an Explanation for the Dynamics of the Venezuelan Andes
The Venezuelan (or Merida) Andes are a NE-trending intracontinental orogen that started to rise from the late Miocene due to the E-W far field convergence between the Maracaibo block to the northwest and the Guyana shield to the southeast. Oblique convergence is responsible for strain partitioning with thrusting along both foreland basins and right-lateral strike-slip faulting along the NE-SW Bocono fault cutting the Venezuelan Andes along-strike. The central part of the belt is also cut by the N-S left-lateral strike-slip Valera fault that branches the Bocono fault in the triple junction geometry, favoring the crustal escape of the Trujillo triangular block towards the NE. Onset of strike-slip motion along major faults and their geometry at depth remains a matter of debate. Our work, based on the integration of geologic and geophysical data aims to better constrain both the geometry and the tectonic evolution of the major tectonic structures. We use the orogenic float model (Oldow et al., 1990) as a first hypothesis to construct two NW-SE trans-Andean crustal scale balanced sections. The late Neogene-Quaternary shortening varies from 40 km in the south to 30 km in the north across the Trujillo block, indicating that a quarter of the deformation seems to be absorbed by the tectonic escape process. The cross-sections served also as the basis for the building of a 3-D geologic model of the Venezuelan Andes, permitting to clearly understand the link and geometry of the faults at depth. The decollement level used for the orogenic float model, located at 20 km depth, is crucial for the motion of the Trujillo block. Both the Bocono and Valera faults have listric shapes connecting to the decollement level. The connexion of the two fault surfaces forms a hinge line dipping towards the north in a geometry favoring the escape of the Trujillo block and allowing the gravity forces to play an important role in the process. Oldow J. S., Bally A. W., Ave Lallemant H. G., 1990, Transpression, orogenic float, and lithospheric balance. Geology, v.18, pp. 991-994. ECOS-Nord project V06U02
T23B-2029
Gravity Anomaly Between Immature And Mature Subduction Zones In The Western Pacific And Its Implications For Subduction Evolution
From immature to mature subduction zones, the western Pacific is a key area to investigate subduction evolution. Among them, the Yap and Mussau trenches located in the boundary of Caroline plate and the Hjort trench to the south of New Zealand are considered as immature subduction zones. The common geological features of immature subduction zones include: (1) a short trench-arc distance, (2) the lack of Wadati-Benioff zone, and (3) the absence of arc volcanism triggered by subducting slab. On the other hand, the Izu-Bonin- Mariana and Tonga-Kermadec trenches are well-developed or mature subduction zones characterized by active arc volcanism and deep earthquakes. We compare two end-members, immature and mature subduction zones, with gravity anomaly derived from satellite altimetry which has sufficient accuracy for this kind of regional study. The isostatic residual gravity anomalies show that the width of non-isostatically- compensated region of the mature subduction zone is substantially wider than that of immature ones. Moreover, when we removed the gravitational effects due to the seafloor from the free-air gravity anomaly, a large difference was found between the immature and mature subduction zones in the overriding plate side. In the mature subduction zones, a low gravity anomaly of ~200-250 mGals can be found in the overriding plate which differs from the immature subduction zones. We discuss the possible causes of the low gravity anomaly including: (1) serpentinization in the upper mantle; (2) presence of partial melt in the mantle wedge; (3) difference in the density structure between the overriding and subducting plates in terms of slab age and cooling history; and (4) thickened crust or anomalous crustal structure beneath the arc. Serpentinization cannot explain the low gravity anomaly at ~150-200 km from the trench. Also, the difference of gravity anomaly due to the difference of the slab age is insufficient to account for the total anomaly. In this study, we explore various causes of low-gravity-anomaly development associated with the evolution of subduction zones systematically with special attention on the dynamic topography and thermal models.
T23B-2030
The Tonga-Vanuatu Subduction Complex -- a Self-Optimized 3D Slab-Slab-Mantle Heat Pump
Recently published geophysical and geochemical data and increasingly actualistic free subduction models
prompted a fresh look at 2 classics hinting, in combination, that a coupled 3D slab-slab-upper mantle
interaction (Scholz and Campos, 1995; full citations at URL below) might power the prodigious surface heat
dissipation (Lagabrielle et al., 1997) characterizing one of Earth's most remarkable tectonomagmatic
systems, the Tonga-Vanuatu Subduction Complex (TVSC). The 3D TVSC includes (1) the kinematically,
magmatically, and bathymetrically distinct North Tonga (NT, 14-26° S) and South Vanuatu (SV, 16-23°
S) trenches and slabs, (2) the shared NT-SV backarc, and (3) entrained mobile upper mantle (MUM). That
Earth's greatest convergence, rollback, and spreading rates; most disseminated spreading (the North Fiji
Basin (NFB) ridge swarm); and greatest concentration of aggregate active ridge length coincide in a 1,500
km TVSC can't be accidental. To the north and south, the respective active NT and SV trenches swing
abruptly 90° counterclockwise into continuity with the Vitiaz and Hunter fossil trenches, both active in the
Late Miocene but now sinistral strike-slip loci standing over long exposed PA and AU slab edges. These 2
active-fossil trench pairs bracket a hot, shallow and geophysically and geochemically exceptional TVSC
interior consisting of 2 rapidly spreading backarcs set back-to-back in free sublithospheric communication:
The Lau-Havre NT backarc on the east and the ridge-infested SV backarc (NFB) on the west. The NFB and
adjacent North Fiji Plateau make up the unplatelike New Hebrides-Fiji Orogen (Bird, 2003). As in the western
Aleutians, the NT-Vitiaz and SV-Hunter subduction-to-strike-slip transitions (SSSTs) stand above toroidal
fluxes of hot, dry PA and AU MUM driven along-trench and around the free NT and SV slab edges from
subslab to supraslab regions by dynamic pressure gradients powered by slab free-fall and induced viscous
couplings. These edge flows must converge and mix beneath the shared TVSC backarc, which must then
shed a huge advected subslab heat load by maximizing ridge length in the area available. Found at both
SSSTs are adakites indicative of a TVSC source laced with slab-edge melt and boninites consistent with flux-
melting of hot, dry subslab MUM on entry to the supraslab wedge. Isotopics reveal widespread source mixing
of Pacific and Indian MOR end-members. Diverging NT and VS trenches rotate clockwise at extremely high
rates about pinning points at and Euler poles near trench-floater intercepts: Louisville Ridge on PA, and
West Torres Plateau-D'Entrecasteaux Ridge on AU. In this configuration, the spinning, free-falling NT and SV
slabs form a highly coupled self-organized gravity-powered pump pulling hot subslab MUM beneath the
TVSC with enough left-over head to power severe transition zone buckling of an 80+ Ma NT (PA) slab also
actively extending toward its free edge. Several nonlinear couplings (e.g., temperature-dependent viscosity
and slab damage at tightening upper hinges) feedback positively to pump efficiency. The TVSC is but one
possible slab-mantle pump partaking of the strong self-optimizing tendency characteristic of all natural flow
systems (e.g., Bejan and Lorente, 2006). Slab-mantle pump natural history is now under investigation, as
such pumps may have allowed a shrinking post-Pangean Pacific with an unrelenting sublithospheric room
problem to relieve itself of excess MUM by making efficient use of available circum-Pacific slab curtain
porosity -- a commodity that may have been in very short supply through most of the Cretaceous.
http://www.cliffshade.com/agufm08/addl-info-agu-abstract-14187.pdf
T23B-2031
Petrogenesis of Age Equivalent John Day Ash Flows near Prineville, Oregon
Early Oligocene volcanism in central Oregon underwent a chemical shift from the chiefly calc-alkaline intermediate rocks of the Clarno Formation (~54Ma - ~33Ma) to the primarily bimodal alkali-basalt and rhyolite assemblage of the John Day Formation (~39Ma - ~20Ma). The western facies of the John Day Formation between the Blue Mountains uplift and the Cascade Range contains several rhyolitic ash flow tuffs (members A-I) around Ashwood and Antelope as well as several age equivalent units to the south near Prineville. Suggested sources for these units range from (1) vents at or near the present Cascade Range, (2) isolated rhyolite domes west of the Blue Mountains, and (3) the Crooked River Caldera. Preliminary whole rock data of these age equivalent rhyolitic ash flows reveal two chemical types that are characterized by relative Zr contents. Both groups plot within the A-type rhyolite range on a Ga/Al diagram, though the high- Zr group has a stronger A-type affinity with higher concentrations of HFS elements (Zr, Hf, Ti), Fe and Ba, but is less enriched in Th and light rare earths. The low- Zr group tends to be less enriched in HFS elements, Fe and Ba, and is slightly more enriched in Th and light rare earths. Neither group shows significant Na-Ta depletions, and LIL/HFS element ratios are not decoupled to a degree (except for Ba, Sr, and Rb from feldspar fractionation) that would suggest a subduction related origin for these units. These data support petrogenesis of the rhyolites in an intraplate extensional setting, rather than subduction related arc volcanism typified by the Clarno Formation. Extension may have developed in a back-arc regime as a result of westward migration of the subduction zone.
T23B-2032
Interaction of the Walker Lane and the Cascade Volcanic Arc, Northern California
We utilize modern geologic and gravity data sets to examine the interaction between the Walker Lane and the southernmost Cascade Volcanic Arc. The Cascade Volcanic Arc in the Lassen region is the product of eastward subduction of the Juan de Fuca plate beneath the North American plate in northern California. To the southeast, the Walker Lane is a structural zone that takes up ~15-25% of the dextral movement between the Pacific and North American plates. The intersection of these two tectonic features in northeastern California is presumed to migrate northward roughly parallel to the northward migration of the Mendocino Triple Junction. Several workers have inferred that NW-trending dextral faults of the Walker Lane intersect and interact with the southernmost Cascade Volcanic Arc in the Lassen region. In the southernmost Cascade Arc, a pronounced unconformity separates <3.5 Ma volcanic rocks from underlying pre-Tertiary sedimentary and metamorphic rocks. The southern limit of volcanism has contracted northward from 40°7.5' at ~3.5 Ma to 40°22.5' at 0.5--0 Ma. Volcanism along the axis is dominated by volcanic centers---large, long-lived, composite, calc-alkaline edifices erupting the full range of compositions from basaltic andesite to rhyolite. Older volcanic centers (3.5--1 Ma) show no correlation with residual gravity, whereas a negative gravity anomaly (-15 mGal) does coincide with <800 ka focused volcanism at the Lassen Volcanic Center (LVC) and the Caribou Volcanic Field (CVF). The most negative part of this gravity anomaly (-30 mGal) coincides with the LVC, a locus of major silicic volcanism, including a significant caldera eruption at 610 ka and a 300--0 ka silicic domefield. Faults are conspicuous E, SE and N of the LVC, but are nearly absent within the LVC. Faults and gravity gradients SE of the LVC strike ~315°, sub-parallel to Walker Lane trends farther to the SE, whereas vent alignments and gravity gradients north of the LVC strike ~345°. Pronounced linear alignments of volcanic vents just east of the LVC suggest significant ENE extension at that latitude. We conclude that the gravity low encompassing the LVC and the CVF reflects the presence during the last 800 ka of a large, crustal-scale tensional regime. This regime was most pronounced under the LVC, where enhanced basaltic intrusion from the mantle produced voluminous silicic magmas through interaction with the crust. We suggest that this tensional regime and the resulting enhanced magmatism were produced by translation of right-lateral strike-slip on the Walker Lane into ENE transtension in the Lassen region. We further note that changes in tectonic activity along parts of the Walker Lane at ~3.5 Ma appear to correlate with the inception of volcanism along the Cascade Volcanic Arc in the Lassen region. We speculate that, beginning at 3.5 Ma, the northern Walker Lane increasingly interacted with the Cascade subduction zone to produce transtensional environments favorable to the development of major volcanic centers.
T23B-2033
Lithospheric Structure Beneath Coastal and Continental Croatia Determined from Broadband Teleseismic Waveform Modeling
Due to intricate interaction of the Apennines, the Adriatic microplate (Adria), the Alps, the Dinaric Alps (Dinarides) and the structural units in the Pannonian basin, the Croatian Adriatic coast is geologically and tectonically one of the most complicated regions in the Mediterranean. Most seismicity along the Adriatic coast is caused by the interaction of Adria with Dinarides, which sit on the Eurasian plate. Although this general scheme is not disputable, there are open questions regarding the nature of this interaction and the structure of the lithosphere in the Circum-Adriatic region. Seismicity of the continental part of Croatia is less intense: strong events there are relatively rare, but can reach magnitudes exceeding 6, which is typical of intraplate seismicity. Although there were a number of seismic exploration studies, broadband teleseismic data have not yet been used to increase constraints on shallow and deep structure of the area. In the last decade Croatian seismological network has been expanded and modernized, and currently there are 13 digital broadband stations in operation. Using the high-quality seismograms that have been accumulated over the last decade we have computed and analyzed receiver functions for a number of Croatian seismological stations, which helped us gain the first comprehensive insight into the structure of lithosphere beneath Croatia. Most of the stations are located along the coast of the Adriatic Sea and only one station is located deeper inland. Unprocessed data for receiver functions analysis consists of three-component broadband seismograms for selected earthquakes with good signal to noise ratio and epicentral distances between 60° and 90°, mainly confined in the western part of the Pacific Rim, from the Kuril Islands in the north to the Indonesia in the south. To determine lithospheric structure and Moho depth from receiver functions, we used several methods and introduced various improvements in data analysis. We have constrained our velocity/thickness grid search to four layers in the crust using both empirical and best-fit values for Vp/Vs ratio. The values we found for the depth of Moho range from about 30 km for one inland station and exceed 40 km for some stations along the Adriatic coast. We compared our results with Moho maps inferred from interpolated reflection and refraction profiles and found that the agreement is good only for some areas. The interpretation of the 1D profiles that we derived is work in progress. These structural models will also be used to elucidate source mechanisms of both tectonic and volcanic earthquakes that occur in this part of the world.
T23B-2034
Historical Nautical Charts and Hydrographic Surveys as Recorders of Vertical Displacement Above Shallow Portions of Subduction Zones
Islands and coastlines above megathrusts in shallow subduction zones commonly show patterns of uplift and subsidence accompanying great earthquakes consistent with predictions from elastic dislocation theory for large displacements on the megathrust. Shallow faulting in the upper plate commonly modifies the details of the pattern, but does not overwhelm the pattern resulting from displacement on the megathrust. During large or great earthquakes the most trenchward locations have shown uplifts of several meters. Examples (with approximate uplifts) include Isla Santa Maria ( 3 meters), Chile, 1835; Isla Guafo (4 meters) and Isla Guambino (6 meters), Chile, 1960; Montague Island (10 meters) and Middleton Island (4 meters), Alaska, 1964; and Nias Island (2 meters), off Sumatra, 2004-5. In favorable circumstances, these uplifts and the resulting changes in coastal morphology are recorded on historical nautical charts and hydrographic surveys. These observations can potentially supplement geologic and geodetic observations of sea level to extend and fill gaps in the record of vertical motions. The uplifts of Montague and Middleton Islands, Alaska, are clearly resolved in the soundings of hydrographic surveys carried out before and after the 1964 earthquake and are of the same order as those determined from other means. The uplift of Isla Santa Maria, Chile, famously described by Captain Robert Fitz-Roy of the HMS Beagle, is also reflected in the coastal morphology shown on the nautical chart prepared by the crew of the Beagle in 1835. Comparison of the 1835 chart with subsequent charts and aerial images suggests that the uplift at the time of the earthquake may have been largely reversed in the subsequent 173 years. This interpretation would be consistent with the average uplift rates over the last several thousand years determined by Bookhagen and others, and with the notion that a cycle of uplift during earthquakes and submergence during strain accumulation in the interseismic period is superimposed on a much slower, long-term tendency for uplift in this location.
T23B-2035
Subduction Zone Diversity and Nature of the Plate Contact
We recently showed that the overall dynamics of subduction and initial collision depends on whether the plate contact is a fault or a channel. Here, we combine results of our numerical experiments with a re-analysis of published observations. Overall, our synthesis connects seismic moment release with back-arc deformation and tectonic processes at the margin. It leads us to identify four classes of subduction zones. The first two classes results directly from our numerical experiments. In class 1, subduction zones are characterized by a plate contact that is largely fault-like with an accretionary margin. In class 2, the plate contacts are largely channel-type and have an erosive margin. Class 3, where the plate contact is entirely channel-like, consists of accretionary margins with a high sediment supply. Subduction zones of class 4, mostly characterized by an erosive convergent margin (northern Chili, Peru, Honshu and Kuril), are more complicated. They can be explained by incorporating regional observations.
T23B-2036
Vp/Vs Ratio and Depth to Moho and the Subducting Cocos Slab across Northern Costa Rica estimated from Receiver Function Analysis
Costa Rica is located near the southern end of the Middle American Trench (MAT) in a complicated tectonic setting controlled by the interaction of the Cocos, Caribbean, and Nazca plates. The oceanic Cocos plate subducts to the northeast underneath the Caribbean plate creating a volcanic arc located 150 km away from MAT. In Northern Costa Rica the arc basement is represented by part of Caribbean Plateau that includes flood basalts, mafic oceanic rocks, serpentinized peridotites, and silicic sediments. For this study, P and PP wave receiver functions have been calculated using teleseismic earthquakes recorded in Northern Costa Rica by broadband stations of the CRSEIZE, Pocosol, and Corisubmod experiments, and stations JTS and HDC from the Global Seismology Network and the Geoscope Project, respectively. The goal of this work is to constrain the major boundaries such as the base of the continental crust and the top of the subducting Cocos slab, as well as Vp/Vs ratios to estimate the composition and physical state of the lithosphere. These calculations are relevant as they provide a velocity structure that directly improves earthquake locations, gives insights into the tectonic evolution of the region, and are useful to describe the extent of the serpentinized forearc mantle wedge. Receiver functions are computed using an iterative pulse stripping time domain deconvolution technique. The depth and average Vp/Vs ratio to the discontinuities are estimated using a stacking algorithm that sums receiver function amplitudes of direct Ps and its multiples. Our results show a thick crust of 41 km underneath the volcanic arc and a thinner crust underneath the backarc and forearc, where the Moho discontinuity is visible at depths of 33-38 km. Moho is observed as a weak signal beneath stations located in the forearc region, which is consistent with previous studies that suggested serpentinization of the mantle wedge. The descending Cocos slab is observed at depths from 20 to 40 km beneath the Nicoya Peninsula in good agreement with the contours of the top of the Cocos slab from previous studies in this region and between 50-78 km underneath forearc and arc. The average Vp/Vs to Moho is 1.90 underneath the volcanic arc and varies from 1.72 to 1.88 in the forearc and backarc regions.
T23B-2037
Seamount Accretion Inferred From Deformation Structures of Greenstones of the Funabuseyama Unit in the Mino Jurassic Accretionary Complex, Central Japan
The effects of seamount subduction on the deformation of overriding accretionary wedge were commonly studied by seismic-reflection profiles, bathymetric data or sandbox experiments. Although subducted seamounts have been considered to work as an asperity of large earthquakes, seamount-accretion processes are still ambiguous. Because Funabuseyama Unit in the Mino Jurassic accretionary complex, central Japan is characterized by abundant distribution of Permian greenstone and limestone, which are interpreted to be of seamount origin, this example gives us a great opportunity to study the seamount subduction/accretion processes that might to be an essential work to make asperities of large earthquakes. Deformation structures within greenstones suggest the two main stages during seamount-accretion processes. The first stage is related to the detachment of seamount took place at the part of volcanic rocks capped by carbonate rocks. The second is related to the fragmentation of volcanic rocks in response to shearing or brecciation, and to the incorporation of fragments into muddy matrix of melange. The low-grade metamorphism in the greenstones and the illite crystallinity (diagenetic zone) from pelitic rocks around the greenstones indicate that a series of seamount- accretion processes occurred in the depth of several kilometers during subduction.
T23B-2038
Consequences of a Large Basement High Subduction on the Makran Accretionary Wedge Growth and Slope Stability, off Pakistan
The Makran accretionary complex results from offscrapping of sediment belonging to the Arabian plate added to the Eurasian plate since Paleocene. The wedge growth has been controlled through time by large sediment inputs both from erosion of Himalayan Ranges and from the Makran prism itself. The latter source Up to 7 km of detrital sediments fill the trench known as the Oman basin, a triangular-shaped sedimentary basin bounded by the Makran deformation front to the N, the passive Arabian margin to the SW and the Murray Ridge to the SE, a transtensional feature corresponding to the Arabian-Indian plate boundary. At present time, the wedge toe is mostly fed from Makran prism erosion through a large canyon network cutting the offshore wedge. We use the analysis of canyon incision and the mapping of knickpoints in their pathway to realize a semi-quantitative estimate of the recent activity of the thrust faults and their variations along the accretionary wedge. The Little Murray Ridge, a large SW-NE trending basement high with an "en echelon" pattern, enters the subduction near 64.55°E; 24.50°N. This bathymetric an irregular and partly buried feature is probably partly of volcanic origin but also of tectonic origin as shown by normal faults bounding some elements of the ridge. Using seismic lines from our CHAMAK cruise as well a older seismic lines, we better define the shape of the Little Murray Ridge and we show that the consequences of the basement high underthrusting in the wedge morphology and structures. We also map the distribution of discrete slump scars along the margin and estimate the eroded volumes in order to discuss the link between the basement high subduction and the slope instabilities.
T23B-2039
VEOX: A new Seismic Line in Mexico to Reveal Cocos Behavior
In 2005, the MesoAmerican Subduction Experiment (MASE) seismic network was deployed across central Mexico and succeeded in imaging the horizontal slab subduction of the Cocos Plate under North America. In 2007, 46 of the stations were deployed some 400 km to the east across the Isthmus of Tehuantepec (VEOX line) where the slab is dipping at 20 degrees. Preliminary results from this experiment will be presented and compared to the MASE line. The goal is to determine which physical parameters are controlling the angle of subduction and why the process is different in the two locations. Receiver functions show the Moho is thinning toward both coasts and there is a poorly resolved image of the slab on the Pacific side. Attenuation results indicate a concentrated zone of attenuation, not unlike the zone that was found on the MASE line. The dynamical implications for the difference will be discussed.