S U M M A R YWe use data from the Chile Argentina Geophysical Experiment (CHARGE) broad-band seismic deployment to refine past observations of the geometry and deformation within the subducting slab in the South American subduction zone between 30 • S and 36 • S. This region contains a zone of flat slab subduction where the subducting Nazca Plate flattens at a depth of ∼100 km and extends ∼300 km eastward before continuing its descent into the mantle. We use a grid-search multiple-event earthquake relocation technique to relocate 1098 events within the subducting slab and generate contours of the Wadati-Benioff zone. These contours reflect slab geometries from previous studies of intermediate-depth seismicity in this region with some small but important deviations. Our hypocentres indicate that the shallowest portion of the flat slab is associated with the inferred location of the subducting Juan Fernández Ridge at 31 • S and that the slab deepens both to the south and the north of this region. We have also determined first motion focal mechanisms for ∼180 of the slab earthquakes. The subhorizontal T-axis solutions for these events are almost entirely consistent with a slab pull interpretation, especially when compared to our newly inferred slab geometry. Deviations of T-axes from the direction of slab dip may be explained with a gap within the subducting slab below 150 km in the vicinity of the transition from flat to normal subducting geometry around 33 • S.

S U M M A R YThe Pampean flat slab of central Chile and Argentina (30 • -32 • S) has strongly influenced Cenozoic tectonics in western Argentina, which contains both the thick-skinned, basementcored uplifts of the Sierras Pampeanas and the thin-skinned Andean Precordillera fold and thrust belt. In this region of South America, the Nazca Plate is subducting nearly horizontally beneath the South American Plate at ∼100 km depth. To gain a better understanding of the deeper structure of this region, including the transition from flat to 'normal' subduction to the south, three IRIS-PASSCAL arrays of broad-band seismic stations have been deployed in central Argentina. Using the dense SIEMBRA array, combined with the broader CHARGE and ESP arrays, the flat slab is imaged for the first time in 3-D detail using receiver function (RF) analysis. A distinct pair of RF arrivals consisting of a negative pulse that marks the top of the oceanic crust, followed by a positive pulse, which indicates the base of the oceanic crust, can be used to map the slab's structure. Depths to Moho and oceanic crustal thicknesses estimated from RF results provide new, more detailed regional maps. An improved depth to continental Moho map shows depths of more than 70 km in the main Cordillera and ∼50 km in the western Sierras Pampeanas, that shallow to ∼35 km in the eastern Sierras Pampeanas. Depth to Moho contours roughly follow terrane boundaries. Offshore, the hotspot seamount chain of the Juan Fernández Ridge (JFR) is thought to create overthickened oceanic crust, providing a mechanism for flat slab subduction. By comparing synthetic RFs, based on various structures, to the observed RF signal we determine that the thickness of the oceanic crust at the top of the slab averages at least ∼13-19 km, supporting the idea of a moderately overthickened crust to provide the additional buoyancy for the slab to remain flat. The overthickened region is broader than the area directly aligned with the path of the JFR, however, and indicates, along with the slab earthquake locations, that the flat slab area is wider than the JFR volcanic chain observed in the offshore bathymetry. Further, RFs indicate that the subducted oceanic crust in the region directly along the path of the subducted ridge is broken by trench-parallel faults. One explanation for these faults is that they are older structures within the oceanic crust that were created when the slab subducted. Alternatively, it is possible that faults formed recently from tectonic underplating caused by increased interplate coupling in the flat slab region.

[1] We present the first results from a dense network of 36 campaign and 46 continuous GPS stations located in the Eastern Transverse Ranges Province (ETR), a transition zone between the southernmost San Andreas fault (SSAF) and eastern California shear zone (ECSZ). We analyzed the campaign data together with available data from continuous GPS stations for the period 1994-2009. We used the GPS velocity estimates to constrain elastic block models to investigate fault-loading rates representing four hypotheses characterized by different fault-block geometries. Fault-block scenarios include blocks bounded by the east-striking left-lateral Pinto Mountain, Blue Cut, and Chiriaco faults of the ETR; blocks bounded by a right-lateral north-northwest striking structure (the "Landers-Mojave earthquake line") that cuts obliquely across the ETR and mapped Mojave Desert faults; and combinations of these end-member hypotheses. Each model implies significantly different active fault geometries, block rotation rates, and slip rates for ETR and ECSZ structures. All models suggest that SSAF slip rate varies appreciably along strike, generally consistent with rates derived from tectonic geomorphology and paleoseismology, with a maximum of ∼23 mm/yr right-lateral along the southernmost Coachella Valley strand, decreasing systematically to <10 mm/yr right-lateral through the San Gorgonio Pass region. Slip rate estimates for the San Jacinto fault are ∼12 mm/yr for all models tested. All four models fit the data equally well in a statistical sense. Qualitative comparison among models and consideration of geologic slip rates and other independent data reveals strengths and weaknesses of each model.

[1] We examine shear wave splitting in teleseismic phases to observe seismic anisotropy in the South American subduction zone. Data is from the CHARGE network, which traversed Chile and western Argentina across two transects between 30°S and 36°S. Beneath the southern and northwestern parts of the network, fast polarization direction (j) is consistently trench-parallel, while in the northeast j is trench-normal; the transition between these two zones is gradual. We infer that anisotropy sampled by teleseismic phases is localized within or below the subducting slab. We explain our observations with a model in which eastward, Nazca-entrained asthenospheric flow is deflected by retrograde motion of the subducting Nazca plate. Resulting southward flow through this area produces N-S j observed in the south and northwest; E-W j result from interaction of this flow with the local slab geometry producing eastward mantle flow under the actively flattening part of the slab.

[1] The San Bernardino basin is an area of Quaternary extension between the San Jacinto and San Andreas Fault zones in southern California. New gravity data are combined with aeromagnetic data to produce two-and three-dimensional models of the basin floor. These models are used to identify specific faults that have normal displacements. In addition, aeromagnetic maps of the basin constrain strike-slip offset on many faults. Relocated seismicity, focal mechanisms, and a seismic reflection profile for the basin area support interpretations of the gravity and magnetic anomalies. The shape of the basin revealed by our interpretations is different from past interpretations, broadening its areal extent while confining the deepest parts to an area along the modern San Jacinto fault, west of the city of San Bernardino. Through these geophysical observations and related geologic information, we propose a model for the development of the basin. The San Jacinto faultrelated strike-slip displacements started on fault strands in the basin having a stepping geometry thus forming a pull-apart graben, and finally cut through the graben in a simpler, bending geometry. In this model, the San Bernardino strand of the San Andreas Fault has little influence on the formation of the basin. The deep, central part of the basin resembles classic pull-apart structures and our model describes a high level of detail for this structure that can be compared to other pull-apart structures as well as analog and numerical models in order to better understand timing and kinematics of pull-apart basin formation.

We assess the ability of predicted seismic velocities (Vp, compressional wave; Vs, shear wave; and Vp/Vs) to identify regions in the upper mantle that are enriched in orthopyroxene relative to normal melt-depleted peridotite compositions. Orthopyroxene enrichment has been found in mantle xenoliths from a number of locations, including the Colorado Plateau and the Kaapvaal craton. We fi nd that the Vp/Vs ratio is very sensitive to orthopyroxene concentration, but it is not sensitive to depletion level. We compare these predicted velocities and Vp/Vs ratios to the high Vs, low Vp/Vs anomaly found above the central Chile-Argentina fl at slab. Within error, the predicted velocities of some of the orthopyroxene-enriched xenoliths match the observed velocities from the Chile-Argentina upper mantle. Because the anomaly above the central Chile-Argentina fl at slab does not conform to known terrane boundaries but does align with the downgoing Juan Fernandez ridge track, we suggest that the orthopyroxene enrichment in this area may be related to fl at slab processes, which would have implications for our understanding of lithospheric silica enrichment in the western U.S. and elsewhere.

[1] The Pampean flat-slab region, located in central Argentina and Chile between 29 and 34 S, is considered a modern analog for Laramide flat-slab subduction within western North America. Regionally, flat-slab subduction is characterized by the Nazca slab descending to $100 km depth, flattening out for $300 km laterally before resuming a more "normal" angle of subduction. Flat-slab subduction correlates spatially with the track of the Juan Fernandez Ridge, and is associated with the inboard migration of deformation and the cessation of volcanism within the region. To better understand flat-slab subduction we combine ambient-noise tomography and earthquake-generated surface wave measurements to calculate a regional 3D shear velocity model for the region. Shear wave velocity variations largely relate to changes in lithology within the crust, with basins and bedrock exposures clearly defined as low-and high-velocity regions, respectively. We argue that subduction-related hydration plays a significant role in controlling shear wave velocities within the upper mantle. In the southern part of the study area, where normal-angle subduction is occurring, the slab is visible as a high-velocity body with a low-velocity mantle wedge above it, extending eastward from the active arc. Where flat-slab subduction is occurring, slab velocities increase to the east while velocities in the overlying lithosphere decrease, consistent with the slab dewatering and gradually hydrating the overlying mantle. The hydration of the slab may be contributing to the excess buoyancy of the subducting oceanic lithosphere, helping to drive flat-slab subduction.

Basement-cored uplifts are observed globally and remain an enigmatic feature of plate tectonics due to the fact that, in many cases, they occur distant from a plate boundary. The Laramide Bighorn Arch in Wyoming is an archetypal basement-involved foreland arch and provides an excellent setting for the investigation of such structures. Previous studies proposed diverse arch formation models; each of which predicts a unique crustal geometry. We use high-resolution crustal imaging from teleseismic P wave receiver functions to test these models. We obtained our data from 239 three-component seismometers deployed as part of the Bighorns Arch Seismic Experiment as well as coeval regional Transportable Array stations. A sequential, two-layer thickness V P /V S (H-κ) stacking algorithm constrains sediment and crustal structure. Receiver function Common Conversion Point stacking results in 2-D transect images across the arch. Our results define an upwarp of the crust beneath the central and northern arch that extends into the Powder River Basin, north-northeast of the arch. The lack of Moho-cutting faults or a Moho geometry mirroring the arch rules out most shortening models except a crustal detachment model where shortening was accomplished by fault-propagation folding on a thrust splay ramping off a midcrustal detachment fault. The mismatch between gentle, symmetric Moho and asymmetric Laramide arch geometries and their trends suggests a pre-Laramide origin for at least a part of the Moho high. This high, perhaps in combination with a lesser degree of Laramide lithospheric buckling, may have caused emergent Laramide thrusting and thus nucleated the Bighorn Arch. Our results suggest that midcrustal detachment can form basement-involved foreland arches and suggest the hypothesis that preexisting undulations in the Moho may have nucleated individual arches.