T31C-1313 INVITED 0800h
Recent Tomographic Results in Southern California and their Tectonic Implications
The Broadband Seismic Network in Southern California, developed in the last fifteen years, represents one of the most dense, permanently-instrumented network in the world. Along with its high-quality broadband data comes an opportunity for various waveform analyses for retrieval of seismic parameters in the crust and mantle. We will review the state of the tomographic studies in this region with special reference to various tectonic features that have been identified. Review will be on various body wave tomography including both P- and S-wave tomographic results, shear-wave splitting data and surface wave results for S-wave velocity and anisotropy. In this paper, we will be concerned with the large-scale tectonic framework in this region and thus focus on larger-scale features in tomographic results. We will not go into details of small-scale features, such as shallow crustal structures, that are important for ground motion prediction and hazard mitigation. Some of the tectonic features that we will discuss are: (1) Slow S-wave velocity mantle anomaly under the Southern Sierra: this was confirmed by body wave study (e.g. Savage et al., 2003) and surface wave study (Prindle-Sheldrake and Tanimoto 2004, hereafter PT). (2) Deep velocity contrast across San Andreas. This was seen in body wave results (Helmberger et al., 2001) and surface wave results (PT) (3) Deep fast velocity anomaly in the Transverse Range. This was reported first by the well-known paper by Humphereys et al. (1990) and confirmed by others (e.g., Kohler et al. 1998) using P-wave travel time data. Surface wave study found similar features in S-wave velocity model but the location of the fast velocity root is not necessarily the same. (4) Deep root-like structure seems to exist under the Western Transverse Range, that have rotated 90 degrees approximately in the last 15 Ma. There seems to be a hint for the existence of similar root-like structure also under the Peninsular Ranges from which the Western Transverse range broke off and rotated 90 degrees. (5) S-wave splitting data and surface wave data generally match on the pacific plate side of this region but differences seem to exist on the eastern side. Surface wave results seem to indicate a much stronger anisotropic structure on the Pacific plate side. On the Pacific plate side, the fast direction of azimuthal anisotropy is along the direction of the Pacific plate motion or (equivalently) parallel to the faults in general. (6) Under the Salton Sea region, longer period Rayleigh wave results at about 20 mHz indicate the fast direction perpendicular to the fault. If there is spreading in this region, the fast direction coincides with the spreading direction. This direction rotates 90 degrees for higher frequency maps. This result seems to stand our various resolution tests and indicates an interesting situation that the fast direction of S-waves differ for the crust and the underlying mantle.
T31C-1314 INVITED 0800h
The role of the mantle and lower crust in driving Transverse Range convergence
Convergence in the Transverse Ranges is a consequence of lithosphere converging and sinking beneath the Transverse Ranges. By lying below the Pelona schist subduction complex, this lithosphere is not North America; it probably is a fragment of the Farallon oceanic plate. The schist outcrops on low-angle faults near the Transverse Ranges, suggesting that the aseismic lower crust is made largely of schist. The descending lithosphere pulls southern California crust toward the Transverse Ranges, maintaining the San Andreas as the dominant fault south of the Transverse Ranges. Lithospheric convergence is not matched by the crust, which avoids most convergence. This is a dynamic result of the mountains themselves, whose gravitational potential energy pushes back on the converging crust. Hence the upper crust and mantle are simultaneously dynamically coupled and kinematically decoupled by the viscous lower crust.
T31C-1315 INVITED 0800h
Will SCIGN Fulfill its Biggest Promise, to Test Regional Tectonic Models?
A primary objective of SCIGN from the outset has been to test regional tectonic models, especially for compressional deformation in the Los Angeles region. Papers by Walls et al. (Nature, 1998) and Argus et al. (Geology, 1999) espoused nearly opposite interpretations based on their initial results from SCIGN. There were several clear reasons for differences between their results and interpretations. At that time SCIGN velocity field results differed considerably between the several groups analyzing the data. Also, very different methods of post-processing analysis and modeling were used. Despite historical prevalence and initial InSAR results indicating land subsidence in the basins of southern California (Galloway et al., Water Resources Research, 1998), neither of these studies considered its effects. Then, Bawden et al. (Nature, 2001) accounted for the impact of land subsidence on SCIGN horizontal velocity component estimates. They first imaged the problem areas using InSAR and then selectively deleted velocity estimates from affected SCIGN sites, allowing them to derive a cleaner horizontal tectonic velocity field across Los Angeles (after having eliminated the problematic data). Argus et al. (AGU abstract, 2002) then re-evaluated their earlier study, and included results based on the SCEC survey-mode GPS data base. Meanwhile, much progress was made on other fronts. The SCIGN Analysis Comparison Committee improved and combined SCIGN data products. More stations were installed using improved SCIGN methods, and data accumulated to lengthen the SCIGN GPS time series. In addition, the quality and robustness of the InSAR image stacks improved (Peltzer et al., Geology, 2001). Remaining problems include; 1) the C-band InSAR available from the ERS missions have not yet quite allowed coherent imaging of deformation crossing the San Gabriel range north of Los Angeles and out into the Mojave desert, and 2) GPS stations in the mountains have noisier data than their flatland counterparts. The latter problem arises, in part, because several of the crucial stations (e.g., CHIL, WLSN, TABL) were early prototype installations, each with individual characteristics such as monuments and radomes that are undesirable by today's standards. The noise at higher elevation stations may also be caused by sky view blockage due to the presence of trees and occasional snow. For the time being, work continues on understanding subsidence and other error sources, and how best to handle these in quantitative models of regional deformation. Much more could be done now in terms of modeling the SCIGN combined velocity and time series data, and also the SCEC CMM3 horizontal velocity data. Enough SCIGN stations have been built to the new standards in the San Gabriel mountains so that we will eventually obtain reliable estimates of the vertical motions there and across all of southern California. With the passage of time, especially in the absence of earthquakes (that confuse the velocity picture), and with the addition of 170 more PBO GPS stations throughout the region over the next 5 years, there will surely be excellent geodetic data to constrain and test the future's geodynamic models of the Transverse Ranges.
T31C-1316 0800h
Progressive Basin Reorientation in the Northern Gulf of California
The northern Gulf of California is one of a few regions where evidence of active rifting can be seen in a shallow marine setting. This activity has been studied using a 150 x 200 km grid of MCS data collected by the R/V Ulloa in 1999. Four basins have begun opening to accommodate highly oblique right-lateral transtensional motion between the Pacific and N. American plates; all basins exist in different phases of development. The two northernmost basins, the Consag (CB) and Wagner Basins (WB), have axes and bounding faults striking N-S. These basins are shallower than those farther south (~200m vs. 350-900m) because of their proximity to the mouth of the Colorado River, whose sediment supply has more nearly filled the available accommodation space. The two southern basins, the Upper Delfin (UDB) (with its two sub-basins the Northern Upper Delfin (NUDB) and Southern Upper Delfin (SUDB)), and the Lower Delfin (LDB) have axes and bounding faults striking roughly NE-SW. High sedimentation rates in all four basins have provided a record of deformation that we have examined in detail. Seisworks/2D, a seismic interpretation program by Landmark Graphics, was used to identify and correlate stratigraphic layers and faults, and construct a history of basin development for the northern Gulf. This study showed that sediment currently buried as much as 1500 msec beneath the seafloor near the three southern basins accumulated in precursor basins with axes and bounding faults oriented N-S. These earlier basins were smaller and more numerous than the modern basins. The southernmost basin has reoriented so that its axis strikes NE-SW and the bounding faults now show both normal and strike-slip offsets, as opposed to the earlier faults, which were almost exclusively oblique-slip. The SUDB and NUDB are in intermediate stages of development between the smaller and N-S oriented WB and the larger, NE-SW and more mature LDB. These SUDB and NUDB sub-basins are currently undergoing reorientation accompanied by alternating areas of active faulting that we expect will eventually resolve into a single extensional basin. Throughout this period of competition for dominance, a "see-saw" pattern of sediment accumulation has developed, characterized by alternating dip directions of major reflectors and the locations of depocenters switching from one sub-basin to the other. Thus, basins of the northern Gulf appear to undergo an evolution of clockwise reorientation that is progressing from south to north. We suggest the delayed reorientation in the northernmost Gulf is due to locally more ductile crust kept warm by the thick deposit of Colorado River sediments maintaining a high geothermal gradient.
T31C-1317 0800h
Mid-Crustal Decollement(s) Beneath the Transverse Ranges of Southern California
Composite images of the fault systems associated with the 1971 M 6.7 San Fernando, 1987 M 5.9 Whittier Narrows, and 1994 M 6.7 Northridge earthquakes were constructed using industry reflection and oil-test-well data in the upper few km of the crust, relocated aftershocks in the seismogenic part of the crust, and LARSE reflection and refraction data in the middle and lower crust. In the vicinity of the San Fernando Valley, a newly interpreted San Fernando fault zone dips ~25 degrees northward and extends over a 50-km interval from near the surface in the northern San Fernando Valley to the San Andreas Fault (SAF) in the middle to lower crust. This fault zone, or decollement, follows a prominent, aseismic reflective zone below and northward of the San Fernando hypocenter. The inferred fault zone for the Northridge earthquake extends 20 km southward at a dip of ~33 degrees from the point where it is overridden by the San Fernando decollement, beneath the northern San Fernando Valley, to ~20 km depth beneath the Santa Monica Mountains. It follows a weak, aseismic reflective zone below and southward of the Northridge hypocenter. A prominent mid-crustal reflective zone is also imaged beneath the San Gabriel Mountains 70 km east of the San Fernando Valley. The reversed polarity and large amplitudes of reflections returned from the top of this zone seem to require the presence of fluids in this zone. This zone appears to connect the Puente Hills Fault system, in the northern Los Angeles basin, to the SAF at a gentler angle than the San Fernando decollement. A segment of the Puente Hills Fault system slipped in the 1987 M 5.9 Whittier Narrows earthquake. The Sierra Madre fault, between the Puente Hills Fault system and the SAF, appears to be an upward splay from the decollement. The San Fernando decollement appears to follow a suture between a block of the Peninsular Ranges terrane, underlying the Santa Monica Mountains and San Fernando Valley, and the central Transverse Ranges terrane. The San Gabriel decollement appears to extend across the suture between the Peninsular Ranges terrane, underlying the Los Angeles basin, and the central Transverse Ranges terrane. The similarities and differences between the San Fernando and San Gabriel reflective zones merit further field investigations to determine whether these zones are laterally connected or structurally related.
T31C-1318 0800h
Crustal and Sub-Crustal Structure Beneath the Central Transverse Ranges of S. California
We have constructed receiver functions for the SCSN stations in the Central Transverse Ranges of S. California. In addition to the Moho and some crustal interfaces, there is a prominent north-dipping interface at approximately 90 to 105 km in depth (Ps arrival time is 10-12 seconds relative to direct P). The reflector does not have the correct arrival time or ray parameter move-out to be a multiple of any crustal phase, and is a clear arrival on a number of stations. The structure could be part of the subducted Farallon Plate, but if so, it would contradict arguments that the slab has been stripped away in this area (Atwater and Stock, 1998). The depth of the interface also excludes the possibility of its being part of the captured Monterey plate by Pacific plate (Nicholson et al., 1994). The reflector also cross-cuts the drip structure reported by Humphreys and Clayton (1984). The reflector is only seen consistently for events from SW back-azimuth, for events from NW or SE, the interface is not clear. The reflector is also seen in several neighboring stations of LARSE II passive data, which complements the image from SCSN data. We have constructed three N-S profiles across the Central Transverse Ranges. The east one runs through the Precambrian basement rock of the east San Gabriel Mountains and only has a single mid-crustal interface. The central profile runs through the Mesozoic granite rock, and has two mid-crustal interfaces, which indicates significant lateral variations between the two profiles. On the west profile, the Moho depth has an abrupt change with the north end about 5-7 km deeper than the south end. These indicate that the crustal structure and Moho interface here are still closely related to the tectonic evolution history of this area, and that topography is not isostatically supported.
T31C-1319 0800h
Structure and Deformation in the Transpressive Zone of Southern California Inferred from Seismicity, Velocity, and Qp Models
We synthesize relocated regional seismicity and 3D velocity and Qp models to infer structure and deformation in the transpressive zone of southern California. These models provide a comprehensive synthesis of the tectonic fabric of the upper to middle crust, and the brittle ductile transition zone that in some cases extends into the lower crust. The regional seismicity patterns in southern California are brought into focus when the hypocenters are relocated using the double difference method. In detail, often the spatial correlation between background seismicity and late Quaternary faults is improved as the hypocenters become more clustered, and the spatial patterns are more sharply defined. Along some of the strike-slip faults the seismicity clusters decrease in width and form alignments implying that in many cases the clusters are associated with a single fault. In contrast, the Los Angeles Basin seismicity remains mostly scattered, reflecting a 3D distribution of the tectonic compression. We present the results of relocating 327,000 southern California earthquakes that occurred between 1984 and 2002. In particular, the depth distribution is improved and less affected by layer boundaries in velocity models or other similar artifacts, and thus improves the definition of the brittle ductile transition zone. The 3D V$_{P}$ and V$_{P}$/V$_{S}$ models confirm existing tectonic interpretations and provide new insights into the configuration of the geological structures in southern California. The models extend from the US-Mexico border in the south to the Coast Ranges and Sierra Nevada in the north, and have 15 km horizontal grid spacing and an average vertical grid spacing of 4 km, down to 22 km depth. The heterogeneity of the crustal structure as imaged in both the V$_{P}$ and V$_{P}$/V$_{S}$ models is larger within the Pacific than the North America plate, reflecting regional asymmetric variations in the crustal composition and past tectonic processes. Similarly, the relocated seismicity is deeper and shows a more complex 3D distribution in areas exhibiting compressional tectonics within the Pacific plate. The V$_{P}$ values are 0.2 to 0.4 km/s too high to support an abundant occurrence of schist beneath the Mojave Desert and the San Gabriel Mountains. The models reflect mapped changes, from east to west, in the lithology of the Peninsular Ranges. The interface between the shallow Moho of the Continental Borderland and the deep Moho of the continent forms a broad zone to the north beneath the western Transverse Ranges, Ventura basin and the Los Angles Basin and a narrow zone to the south, along the Peninsular Ranges. Similarly, the 3D Qp model includes several features that correspond to regional tectonic features and possibly the thermal structure of the southern California crust. A clear low Qp zone extends from the San Bernardino Basin, across the Chino Basin, San Gabriel Valley, into the Los Angeles Basin. This zone is consistent with the geology and decreases with depth from east to west. The Peninsular Ranges have a high Qp zone consistent with the high velocities in the 3D V$_{P}$ model. There are also zones of high Qp in the southern Mojave and southern Sierras. Several clear transition zones of rapidly varying Qp, extend across major late Quaternary faults and connect regions of high and low Qp. The strongest low Qp zone coincides with the Salton Trough where near-surface low Qp is associated with the sediments and the deeper low Qp may be associated with elevated mid-crustal temperatures.
T31C-1320 0800h
Tomographic Imaging of Vp/Vs Along the LARSE II Profile in Southern California
In situ shear-wave velocities were measured during the Los Angeles Region Seismic Experiment, Phase II (LARSE II). Shear (S) waves are observed from most shots along the profile, including shots in the three Cenozoic sedimentary basins (San Fernando, Santa Clarita, and Antelope Valleys). Velocity models were derived from P-wave and S-wave arrivals using standard inversion techniques. Vp/Vs ratios were calculated from the P-wave and S-wave models. In the Santa Monica Mts and the central Transverse Ranges, where crystalline rocks outcrop, S-wave velocities range between 1.3 and 2.0 km/s at the surface and increase rapidly to 2.5-3.0 km/s below 1.5 kilometers. In the central Transverse Ranges, the Vp/Vs ratio is ~ 2.1 near the surface but decreases to 1.6-2.0 at depths $>$1-2 km. Surface Vp/Vs ratios are higher in the Santa Monica Mts. (2.4-2.7), but within a kilometer decrease below 2.0 and at 2 kilometers ratios are 1.6-1.8. In the sedimentary basins (the San Fernando Valley, the Santa Clarita Valley, and the Antelope Valley) and the over-thrust sedimentary rocks of the Santa Susana Mts., S-wave velocities are much slower at the surface, ranging between 0.6 km/s in the southern San Fernando Valley to 1.2 km/s at basin edges in the Santa Clarita and the Antelope Valleys. Vp/Vs ratios are high at the surface of the basins: as high as 3.0 in the Santa Clarita and the Antelope Valleys and higher in the San Fernando Valley. These ratios decrease rapidly with depth and probably reflect crack and pore closure in the sediments due to increased load and cementation of the sediments. At the San Andreas fault, a Vp/Vs ratio high (2.7) is visible at the surface, but we do not see a down-dip extension defining the fault at depth. However, the average custal Vp/Vs ratios are perturbed at the major faults, which include the San Gabriel, San Francisquito, Clearwater, and San Andreas faults. A zone of increased Vp/Vs ratios correlates with the active San Gabriel fault and a step in depth is visible at the San Francisquito/Clearwater and San Andreas faults.
T31C-1321 0800h
Role of Tectonics in the Tomography and Anisotropy of the Southern California Uppermost Mantle
SKS splitting measurements in southern California have been interpreted (a) normal to the direction of most compressive stress (b) related shearing of the asthenosphere as the lithosphere moves over the mantle and (c) shearing along the San Andreas fault. No single mechanism appears to explain all the data. Splitting results from the LARSE II seismic array that crossed the San Andreas reveal it has minimal influence splitting. Surface waves show that azimuthal anisotropy may extend to depths of 200 km, that is, well into the asthenosphere. The recognition of the high velocity body under the San Gabriel Mountains seen in P-wave tomography has led to the suggestion that the mantle lithosphere has delaminated and is sinking in a sheet like structure that has developed a twist with depth. We compare the three dimensional distribution of the tomographic anomaly with that expected if delaminaton began at the time of the transition of the margin from transtension to transpression and the formation of the San Gabriel Mts. We assume the North American and Pacific plates are moving over a relatively stagnant mantle into which the mantle lithosphere sinks. We examine whether the finite shear strains that are necessarily associated with the delamination process might explain the distribution of SKS splitting by orienting olivine crystals in a three-dimensional flow. Such alignments may also be important in interpreting the tomography, that hitherto has been assumed to be due to isotropic changes in the velocity structure.
T31C-1322 0800h
Horizontal Displacement of the Hector Mine Earthquake, (California, 16/10/99, Mw 7.1), Derived from Aerial Photography Intercorrelation
Surficial slip distribution is a key parameter to understand earthquake source processes. GPS, InSAR and field works are the most common methods used to derive the slip function. However, each of those methods bears severe limitation, respectively the small number of measurements, the decorrelation close to the rupture, and the difficulty to estimate the amount of distributed deformation in the rupture zone. Here we apply optic image sub-pixel intercorrelation technique with high resolution aerial photography from USGS to derive the slip function of the Hector Mine earthquake (California, 16/10/99, Mw 7.1). This new technique provides independent measurement of the horizontal displacement every 16 m along the entire rupture length with no near-field saturation problem. We used 15 low altitude air photo pairs (images acquired between 1989 and 2002) to produce a map of the rupture. This map is almost identical to the field survey data where available; the map also images segments that had been assumed from seismologic data but not identified in the field, probably due to distributed deformation in unconsolidated sediments. Slip distribution along the rupture was measured from profiles perpendicular to the fault, and averaged on 500 m long patches. The measurement accuracy varies between 10 and 50 cm according to the area, allowing us to image slip variation at the scale of few km. A systematic correspondence can be established between slip variations and changes in the geometry of the rupture, like bifurcation or change in azimuth. Particularly, two peaks of slip are observed: the first one about 5 m of slip, located 2 km south of the epicenter, and the second one about 4 m of slip, located about 10 km north of the epicenter, with a local minimum of 3 m between the two peaks. This local minimum is probably due to the existence of the second rupture where the epicenter is located. We measured horizontal slip on this branch up to 1 m, whereas almost no slip has been measured in the field. In detail, slip measurements in bedrock areas are in good agreement with field data whereas in sediments, our data are up to 2 m larger. We interpret this discrepancy as the difficulty to estimate the distributed deformation during field measurements when our technique has no difficulty to integrate this deformation. Finally, due to the accuracy of our technique, the slip-curve we propose reconcile surficial displacement measurements with the slip distributions modeled from seismologic, InSAR and GPS data, especially concerning the existence and location of the peaks of larger slip.
T31C-1323 0800h
Crust-Mantle Dynamics of the Salton Block, Eastern Transverse Ranges, Southern California
Active crustal deformation in the Eastern Transverse Ranges and Salton Trough region of southern California is a combination of horizontal strike-slip faulting, mountain building, and extensional basin formation. The stresses necessary to drive the relative motion of crustal blocks and vertical deformation are either transmitted laterally across faults or from below by an actively deforming mantle, or by some combination of the two mechanisms. The tomographically imaged high-velocity anomaly beneath the Transverse Ranges presumably represents converging and downwelling lithospheric mantle that, if strongly coupled to the upper crust, would drive upper crustal deformation. We find evidence for a relatively strong, high viscosity lower crust from numerical kinematic models of geodetic data and present preliminary dynamic models of the forces acting on the Salton block which address the roles of stresses due to topography, fault strength and crust-mantle coupling.
T31C-1324 0800h
Fault Trends and the Evolution of the Pacific-North America Transform in Southern California
The Pacific-North America (PAC-NOAM) transform boundary evolved during the past 30 Ma, lengthening more than 1000 km and spanning a zone exceeding 200-km across southern California. The relative plate motion vector has been estimated using seafloor magnetic anomaly patterns. Orientations of major transform fault segments within this boundary provide direct evidence of the relative motion at the time these faults formed, where the faults preserve their original orientations. Avoiding areas of known vertical-axis block rotations, we find at least three major fault trends that document past and present tectonic kinematics. A northwest trend of 330 degrees is related to subduction trends in the forearc region that defined the late Mesozoic and early Tertiary coastline and has subsequently controlled the orientation of oblique rifting during the Neogene initiation and growth of the PAC-NOAM transform. This trend is manifest in the San Diego Trough and adjacent coastal rifts and associated fault zones including the Coronado Bank and Newport-Inglewood. The middle Miocene transform orientation appears to be 300-310 degrees, which imparted extensional character to faults reactivated with older subduction trends. Major faults inferred to represent Neogene transform fault segments with this trend include the Whittier, Palos Verdes Hills, Santa Cruz-Catalina Ridge, Catalina Escarpment, and possibly the Mojave segment of the San Andreas fault. In late Miocene time, the plate motion vector rotated clockwise eventually achieving its modern orientation of about 320 degrees. Active faulting showing pure strike-slip character on the San Clemente - San Isidro fault zone and the Imperial Fault show this trend, as do transform faults in the northern Gulf of California. An intermediate trend is apparent in some areas along the San Clemente fault zone in the Borderland, and along the Elsinore and San Jacinto fault zones, which transect the Peninsular Ranges. The intermediate trends may be due to the transition in the plate motion vector, or possibly fault refraction across different crustal rheology. Important consequences of this systematic change in plate motion vector and fault trends regarding the tectonic evolution of southern California include: 1) transpression and the preponderance of restraining bends along northwest-trending right-slip faults; 2) structural inversion of Miocene "pull-apart" basins; 3) tendency to "capture" large elongate blocks of North America crust by the Pacific plate. The latter process continues following the capture of Baja California with the northward rift propagation into the Owens Valley aong the Eastern California Shear Zone.