Publications

Export 11 results:
Sort by: Author Title Type [ Year  (Desc)]
2013
Crowell, BW, Bock Y, Sandwell DT, Fialko Y.  2013.  Geodetic investigation into the deformation of the Salton Trough. Journal of Geophysical Research-Solid Earth. 118:5030-5039.   10.1002/jgrb.50347   AbstractWebsite

The Salton Trough represents a complex transition between the spreading center in Baja California and the strike-slip San Andreas fault system and is one of the most active zones of deformation and seismicity in California. We present a high-resolution interseismic velocity field for the Salton Trough derived from 74 continuous GPS sites and 109 benchmarks surveyed in three GPS campaigns during 2008-2009 and previous surveys between 2000 and 2005. We also investigate small-scale deformation by removing the regional velocity field predicted by an elastic block model for Southern California from the observed velocities. We find a total extension rate of 11mm/yr from the Mesquite Basin to the southern edge of the San Andreas Fault, coupled with 15mm/yr of left-lateral shear, the majority of which is concentrated in the southern Salton Sea and Obsidian Buttes and is equivalent to 17mm/yr oriented in the direction of the San Andreas Fault. Differential shear strain is exclusively localized in the Brawley Seismic Zone, and dilatation rate indicates widespread extension throughout the zone. In addition, we infer clockwise rotation of 10 degrees/Ma, consistent with northwestward propagation of the Brawley Seismic Zone over geologic time.

2011
Smith-Konter, BR, Sandwell DT, Shearer P.  2011.  Locking depths estimated from geodesy and seismology along the San Andreas Fault System: Implications for seismic moment release. Journal of Geophysical Research-Solid Earth. 116   10.1029/2010jb008117   AbstractWebsite

The depth of the seismogenic zone is a critical parameter for earthquake hazard models. Independent observations from seismology and geodesy can provide insight into the depths of faulting, but these depths do not always agree. Here we inspect variations in fault depths of 12 segments of the southern San Andreas Fault System derived from over 1000 GPS velocities and 66,000 relocated earthquake hypocenters. Geodetically determined locking depths range from 6 to 22 km, while seismogenic thicknesses are largely limited to depths of 11-20 km. These seismogenic depths best match the geodetic locking depths when estimated at the 95% cutoff depth in seismicity, and most fault segment depths agree to within 2 km. However, the Imperial, Coyote Creek, and Borrego segments have significant discrepancies. In these cases the geodetically inferred locking depths are much shallower than the seismogenic depths. We also examine variations in seismic moment accumulation rate per unit fault length as suggested by seismicity and geodesy and find that both approaches yield high rates ( 1.5-1.8 x 10(13) Nm/yr/km) along the Mojave and Carrizo segments and low rates (similar to 0.2 x 1013 Nm/yr/km) along several San Jacinto segments. The largest difference in seismic moment between models is calculated for the Imperial segment, where the moment rate from seismic depths is a factor of similar to 2.5 larger than that from geodetic depths. Such variability has important implications for the accuracy to which future major earthquake magnitudes can be estimated.

2010
Wei, M, Sandwell DT.  2010.  Decorrelation of L-Band and C-Band Interferometry Over Vegetated Areas in California. IEEE Transactions on Geoscience and Remote Sensing. 48:2942-2952.   10.1109/tgrs.2010.2043442   AbstractWebsite

Temporal decorrelation is one of the main limitations for recovering interseismic deformation along the San Andreas Fault system using interferometric synthetic aperture radar. To assess the improved correlation properties of L-band with respect to C-band, we analyzed L-band Advanced Land Observation Satellite (ALOS) interferograms with a range of temporal and spatial baselines over three vegetated areas in California and compared them with corresponding C-band European Remote Sensing Satellite (ERS) interferograms. Over the highly vegetated Northern California forests in the Coast Range area, ALOS remains remarkably well correlated over a 2-year period, whereas an ERS interferogram with a similar temporal and spatial baseline lost correlation. In Central California near Parkfield, we found a similar pattern in decorrelation behavior, which enabled the recovery of a fault creep and a local uplifting signal at L-band that was not apparent at C-band. In the Imperial Valley in Southern California, both ALOS and ERS have low correlation over farmlands. ALOS has lower correlation over some sandy surfaces than ERS, probably due to low signal-to-noise ratio. In general, L-band interferograms with similar seasonal acquisitions have higher correlation than those with dissimilar season. For both L-and C-band, correlation over vegetated areas decreases with time for intervals less than 1 year and then remains relatively constant at longer time intervals. The decorrelation time for L-band is more than 2 years in the forest in California whereas that for C-band is less than 6 months. Overall, these results suggest that L-band interferograms will reveal near-fault interseismic deformation once sufficient data become available.

Wei, M, Sandwell D, Smith-Konter B.  2010.  Optimal combination of InSAR and GPS for measuring interseismic crustal deformation. Advances in Space Research. 46:236-249.   10.1016/j.asr.2010.03.013   AbstractWebsite

High spatial resolution measurements of interseismic deformation along major faults are critical for understanding the earthquake cycle and for assessing earthquake hazard. We propose a new remove/filter/restore technique to optimally combine GPS and InSAR data to measure interseismic crustal deformation, considering the spacing of GPS stations in California and the characteristics of interseismic signal and noise using InSAR. To constrain the longer wavelengths (>40 km) we use GPS measurements, combined with a dislocation model, and for the shorter wavelength information we rely on InSAR measurements. Expanding the standard techniques, which use a planar ramp to remove long wavelength error, we use a Gaussian filter technique. Our method has the advantage of increasing the signal-to-noise ratio, controlling the variance of atmosphere error, and being isotropic. Our theoretical analysis indicates this technique can improve the signal-to-noise ratio by up to 20%. We test this method along three segments of the San Andreas Fault (Southern section near Salton Sea, Creeping section near Parkfield and Mojave/Big Bend section near Los Angeles), and find improvements of 26%, 11% and 8% in these areas, respectively. Our data shows a zone of uplift to the west of the Creeping section of the San Andreas Fault and an area of subsidence near the city of Lancaster. This work suggests that after only 5 years of data collection, ALOS interferograms will provide a major improvement in measuring details of interseismic deformation. (C) 2010 COSPAR. Published by Elsevier Ltd. All rights reserved.

2009
Wei, M, Sandwell D, Fialko Y.  2009.  A silent M-w 4.7 slip event of October 2006 on the Superstition Hills fault, southern California. Journal of Geophysical Research-Solid Earth. 114   10.1029/2008jb006135   AbstractWebsite

During October 2006, the 20-km-long Superstition Hills fault (SHF) in the Salton Trough, southern California, slipped aseismically, producing a maximum offset of 27 mm, as recorded by a creepmeter. We investigate this creep event as well as the spatial and temporal variations in slip history since 1992 using ERS-1/2 and Envisat satellite data. During a 15-year period, steady creep is punctuated by at least three events. The first two events were dynamically triggered by the 1992 Landers and 1999 Hector Mine earthquakes. In contrast, there is no obvious triggering mechanism for the October 2006 event. Field measurements of fault offset after the 1999 and 2006 events are in good agreement with the interferometric synthetic aperture radar data indicating that creep occurred along the 20-km-long fault above 4 km depth, with most of the slip occurring at the surface. The moment released during this event is equivalent to a M-w 4.7 earthquake. This event produced no detectable aftershocks and was not recorded by the continuous GPS stations that were 9 km away. Modeling of the long-term creep from 1992 to 2007 creep using stacked ERS-1/2 interferograms also shows a maximum creep depth of 2-4 km, with slip tapering with depth. Considering that the sediment thickness varies between 3 km and 5 km along the SHF, our results are consistent with previous studies suggesting that shallow creep is controlled by sediment depth, perhaps due to high pore pressures in the unconsolidated sediments.

2008
Sandwell, DT, Myer D, Mellors R, Shimada M, Brooks B, Foster J.  2008.  Accuracy and Resolution of ALOS Interferometry: Vector Deformation Maps of the Father's Day Intrusion at Kilauea. Ieee Transactions on Geoscience and Remote Sensing. 46:3524-3534.   10.1109/tgrs.2008.2000634   AbstractWebsite

We assess the spatial resolution and phase noise of interferograms made from L-band Advanced Land Observing Satellite (ALOS) synthetic-aperture-radar (SAR) data and compare these results with corresponding C-band measurements from European Space Agency Remote Sensing Satellite (ERS). Based on cross-spectral analysis of phase gradients, we find that the spatial resolution of ALOS interferograms is 1.3x better than ERS interferograms. The phase noise of ALOS (i.e., line-of-sight precision in the 100-5000-m wavelength band) is 1.6x worse than ERS (3.3 mm versus 2.1 mm). In both cases, the largest source of error is tropospheric phase delay. Vector deformation maps associated with the June 17, 2007 (Father's day) intrusion along the east rift zone of the Kilauea Volcano were recovered using just four ALOS SAR images from two look directions. Comparisons with deformation vectors from 19 continuous GPS sites show rms line-of-site precision of 14 mm and rms azimuth precision (flight direction) of 71 mm. This azimuth precision is at least 4x better than the corresponding measurements made at C-band. Phase coherence is high even in heavily vegetated areas in agreement with previous results. This improved coherence combined with similar or better accuracy and resolution suggests that L-band ALOS will outperform C-band ERS in the recovery of slow crustal deformation.

2007
Wdowinski, S, Smith-Konter B, Bock Y, Sandwell D.  2007.  Diffuse interseismic deformation across the Pacific-North America plate boundary. Geology. 35:311-314.   10.1130/g22938a.1   AbstractWebsite

Crustal movements and deformation within the diffuse Pacific-North America (Pa-NA) plate boundary are dominated by the right-lateral motion between the two plates. By using the Pa-NA pole of rotation (PoR) spherical coordinate system, we decompose observed crustal movements into parallel and normal components to the Pa-NA plate motion. We transformed the 840 velocity vectors of the Southern California Earthquake Center (SCEC) 3.0 velocity field into the Pa-NA PoR system in order to characterize the interseismic velocity across the plate boundary. Our results show that despite the very different deformation styles occurring across the San Andreas fault, the fault trace follows the half plate motion contour. Deviation occurs in the southern section, where the half motion contour correlates with the San Jacinto and Imperial fault segments. Our analysis yields interesting asymmetric patterns in both parallel and normal components. The parallel component shows asymmetrical velocity gradients across the San Andreas fault, and the normal component indicates compression southwest of the Big Bend, but not northeastward. The observations are compared with viscoelastic modeling results, which show a similar velocity field. The main disagreements between the observations and the model are in a narrow band along the San Andreas fault and in the Mojave block, suggesting that crustal heterogeneities and additional unmodeled fault segments should be considered in future models.

2006
Smith, BR, Sandwell DT.  2006.  A model of the earthquake cycle along the San Andreas Fault System for the past 1000 years. Journal of Geophysical Research-Solid Earth. 111   10.1029/2005jb003703   AbstractWebsite

[1] We simulate 1000 years of the earthquake cycle along the San Andreas Fault System by convolving best estimates of interseismic and coseismic slip with the Green's function for a point dislocation in an elastic plate overlying a viscoelastic half-space. Interseismic slip rate is based on long-term geological estimates while fault locking depths are derived from horizontal GPS measurements. Coseismic and postseismic deformation is modeled using 70 earthquake ruptures, compiled from both historical data and paleoseismic data. This time-dependent velocity model is compared with 290 present-day geodetic velocity vectors to place bounds on elastic plate thickness and viscosity of the underlying substrate. Best fit models (RMS residual of 2.46 mm/yr) require an elastic plate thickness greater than 60 km and a substrate viscosity between 2 x 10(18) and 5 x 10(19) Pa s. These results highlight the need for vertical velocity measurements developed over long time spans (> 20 years). Our numerical models are also used to investigate the 1000-year evolution of Coulomb stress. Stress is largely independent of assumed rheology, but is very sensitive to the slip history on each fault segment. As expected, present-day Coulomb stress is high along the entire southern San Andreas because there have been no major earthquakes over the past 150 - 300 years. Animations S1 and S2 of the time evolution of vector displacement and Coulomb stress are available as auxiliary material.

2003
Smith, B, Sandwell D.  2003.  Coulomb stress accumulation along the San Andreas Fault system. Journal of Geophysical Research-Solid Earth. 108   10.1029/2002jb002136   AbstractWebsite

[1] Stress accumulation rates along the primary segments of the San Andreas Fault system are computed using a three-dimensional (3-D) elastic half-space model with realistic fault geometry. The model is developed in the Fourier domain by solving for the response of an elastic half-space due to a point vector body force and analytically integrating the force from a locking depth to infinite depth. This approach is then applied to the San Andreas Fault system using published slip rates along 18 major fault strands of the fault zone. GPS-derived horizontal velocity measurements spanning the entire 1700 x 200 km region are then used to solve for apparent locking depth along each primary fault segment. This simple model fits remarkably well (2.43 mm/yr RMS misfit), although some discrepancies occur in the Eastern California Shear Zone. The model also predicts vertical uplift and subsidence rates that are in agreement with independent geologic and geodetic estimates. In addition, shear and normal stresses along the major fault strands are used to compute Coulomb stress accumulation rate. As a result, we find earthquake recurrence intervals along the San Andreas Fault system to be inversely proportional to Coulomb stress accumulation rate, in agreement with typical coseismic stress drops of 1-10 MPa. This 3-D deformation model can ultimately be extended to include both time-dependent forcing and viscoelastic response.

2002
Lyons, SN, Bock Y, Sandwell DT.  2002.  Creep along the imperial fault, southern California, from GPS measurements. Journal of Geophysical Research-Solid Earth. 107   10.1029/2001jb000763   AbstractWebsite

[1] In May of 1999 and 2000, we surveyed with Global Positioning System (GPS) 46 geodetic monuments established by Imperial College, London, in a dense grid (half-mile spacing) along the Imperial Fault, with three additional National Geodetic Survey sites serving as base stations. These stations were previously surveyed in 1991 and 1993. The Imperial College sites were surveyed in rapid-static mode (15-20 min occupations), while the NGS sites continuously received data for 10 h d(-1). Site locations were calculated using the method of instantaneous positioning, and velocities were determined relative to one of the NGS base stations. Combining our results with far-field velocities from the Southern California Earthquake Center (SCEC), we fit the data to a simple elastic dislocation model with 35 mm yr(-1) of right-lateral slip below 10 km and 9 mm yr(-1) of creep from the surface down to 3 km. The velocity field is asymmetrical across the fault and could indicate a dipping fault plane to the northeast or a viscosity contrast across the fault.

Sandwell, DT, Sichoix L, Smith B.  2002.  The 1999 Hector Mine earthquake, southern California: Vector near-field displacements from ERS InSAR. Bulletin of the Seismological Society of America. 92:1341-1354.   10.1785/0120000901   AbstractWebsite

Two components of fault slip are uniquely determined from two line-of-sight (LOS) radar interferograms by assuming that the fault-normal component of displacement is zero. We use this approach with ascending and descending interferograms from the ERS satellites to estimate surface slip along the Hector Mine earthquake rupture. The LOS displacement is determined by visually counting fringes to within I kin of the outboard ruptures. These LOS estimates and uncertainties are then transformed into strike- and dip-slip estimates and uncertainties; the transformation is singular for a N-S oriented fault and optimal for an E-W oriented fault. In contrast to our previous strike-slip estimates, which were based only on a descending interferogram, we now find good agreement with the geological measurements, except at the ends of the rupture. The ascending interferogram reveals significant west-side-down dip-slip (similar to1.0 in) which reduces the strike-slip estimates by I to 2 in, especially along the northern half of the rupture. A spike in the strike-slip displacement of 6 m is observed in central part of the rupture. This large offset is confirmed by subpixel cross correlation of features in the before and after amplitude images. In addition to strike slip and dip slip, we identify uplift and subsidence along the fault, related to the restraining and releasing bends in the fault trace, respectively. Our main conclusion is that at least two look directions are required for accurate estimates of surface slip even along a pure strike-slip fault. Models and results based only on a single look direction could have major errors. Our new estimates of strike slip and dip slip along the rupture provide a boundary condition for dislocation modeling. A simple model, which has uniform slip to a depth of 12 km, shows good agreement with the observed ascending and descending interferograms.