G11A-0760 0800h
A Preview of the SHIFT Global Seismicity Forecast
Project SHIFT (Seismic Hazard Inferred From Tectonics) will produce global maps of forecast long-term-average (Poissonian) seismicity, at a variety of magnitude thresholds, as well as derived maps of shaking hazard. The basic assumption is that shallow seismicity is proportional to fault slip rates and/or anelastic strain rates in the shallow frictional part of the lithosphere. Historical seismicity is used only to calibrate this relationship. The kinematic surface-velocity model is a composite of (1) global plate model PB2002 of Bird [2003, G$^{3}$] for simple boundaries between the quasi-rigid parts of the 52 plates; (2) local high-resolution models of 13 non-rigid orogens, fit to geologic, geodetic, and stress data using kinematic F-E program NeoKinema; (3) anelastic strain rate estimates in plate interiors, based on a realistic nonlinear rheology and a computed stress field from dynamic F-E program Shells. Tectonic motion is converted to long-term seismicity (earthquakes/m$^{2}$/s above some threshold) based on the calibration study of Bird & Kagan [2004; BSSA], which used 20$^{th}$-century seismicity to determine apparent boundary half-widths, coupled lithosphere thicknesses, beta values, and corner magnitudes for each of 7 plate-tectonic settings. A preview map of the global seismicity forecast above threshold magnitude 5.663 (moment 3.5x10$^{17}$ N m) will be displayed for discussion and comment, although only 2 of 13 orogens (Persia-Tibet-Burma and Gorda-California-Nevada) have been modeled to date. Future plans include (a) replacement of the current "Earth3" Shells model for strain-rates in plate interiors with an "Earth5" model having 4x better resolution and optimized boundary conditions; (b) modeling of the other 11 orogens with NeoKinema; (c) computation of shaking hazard based on seismicity, by traditional methods. It is hoped that the SHIFT long-term seismicity forecasts will provide a firm foundation for more advanced studies of the time-dependence of seismicity, and of the spatial distribution of creeping fault patches.
http://element.ess.ucla.edu
G11A-0761 0800h
Mechanical modelling of oblique convergence in the Zagros, Iran
Recent GPS surveys indicate that Zagros kinematics corresponds to an oblique convergence between a rigid central iranian plateau and the arabian plate at 6-8 mm/yr at the longitude of the Persian gulf. Convergence is almost frontal in central Zagros and oblique (45°) in northern Zagros. It has been proposed that internal deformation of northern Zagros occurs in a partitionned mode. In such a view, the main recent fault bordering the iranian plateau accomodates all the tangential motion, while shortening happens by pure thrusting within the fold and thrust belt. We test this tectonic concept using a 3D mechanical finite element model of Zagros which accounts for the horizontal rheological layering of the lithosphere (strong upper crust, weak lower crust, strong or weak uppermost mantle) as well as the possible vertical heterogeneity of the main recent fault. Surprinsingly, a partitionning mode only occurs when the collision is very oblique. As this is not the case in northern Zagros, we find that 25 percent only of the tangential motion can be taken by the main recent fault. As the GPS based tangential motion across Zagros is 5 mm/yr, the present-day geodetic slip rate of this fault should be as low as 1.2 mm/yr. Also, our model predict no strike-slip activity of a vertical fault when convergence is more frontal. This last result appears consistent with the distributed strain occuring in central Zagros. We discuss the implications of our findings with respect to seismic and geological data available in Zagros.
G11A-0762 INVITED 0800h
Investigating the mechanics of the seismic cycle along plate boundaries
The seismic cycle concept offers a convenient kinematic framework to integrate geodetic and geological observations. Generally the integration is made through elastic dislocation modeling, eventually based on Savage's backslip model. This sort of kinematic description is valid insofar as there is no permanent strain accumulating off the main fault zone. This approximation applies well to subduction zones and strike-slip faults but can also be applied to orogenic contexts providing some caution regarding the modeling of vertical displacements. The kinematic description of strain over the seismic cycle in orogenic contexts such as the Himalaya or Taiwan seems to make sense in view of the rheology of continental rocks and it dependence on temperature. In this context the thermal structure might be the key factor determining the transition with depth from a seismogenic fault portion obeying rate-weakening friction, where motion is presumably dominantly stick-slip, to a zone undergoing predominantly strain strengthening brittle creep, which slip steadily during the interseismic period or produce transient afterslip in the postseismic period, and finally to a zone of aseismic ductile flow at depth. For subduction zone the physical factors for the variations with depth of fault properties is more enigmatic. In particular, the fact that the plate interface generally creeps at depth below about 50km does cannot be easily explained by thermally enhanced ductility nor by serpentinization of the mantle wedge. Some simple model of stress transfers during the seismic cycle is used to analyze jointly seismicity rate and crustal strain in the interseismic period and during post-seismic relaxation. We show that reloading of the upper brittle crust, due to postseismic afterslip and viscous relaxation is a viable mechanism to explain jointly geodetic data and the decay rate of aftershocks. This is substantiated by the analysis of a few cases such as the Chi-Chi 1999 earthquake, or the 2001 Peru earthquake. The model is also used to assess the possibility of non-stationary strain in the interseismic period. It turns out that depending on the viscosity and thickness of the viscous shear zone at depth stress transfer during the seismic cycle may induce significant variation of interseismic strain that could be measured from geodetic techniques, and explain possible discrepancies between geological slip rates and geodetic slip rates.
G11A-0763 0800h
Co-Seismic Gravity Modeling of a Fault Network in Southern California
The mitigation of seismic hazards within the highly populated area of Southern California is a difficult problem which is dependent upon such observational parameters as plate motion and strain accumulation. It has been suggested [\textit{Okubo}, 1992] that one can use Bouguer anomalies as a proxy for permanent strain deformation at depth. To date, there are no large-scale models that successfully model the temporal gravity signal over extended spatial regions encompassing a network of faults, such as that which exists in Southern California. We present here a computational model for the temporal evolution ofcoseismic gravity changes based on the analytical solutions of \textit{Okada} [1985]and \textit{Okubo} [1992], as inferred through the slip histories (both aseismic and unstable) generated by the general earthquake model simulations of the southernCalifornia fault network [\textit{Rundle} 1988]. These long-period gravity signals are well within the range of land based observations. Currently, several remote sensing satellites designed specifically for gravity data collection (e.g. GRACE, CHAMP, GEOS) are ideal candidates for the observation of these signals once the integration times of the data collection are long enough to produce the required resolution [\textit{Sun and Okubo}, 2004]. This technique should be applicable to any mechanism in which deformation occurs such as volcanic activity or glacial rebound.
G11A-0764 0800h
Macro- and Micro-deformation Features of the Creeping Strand of the San Andreas Fault, and Evidence for Changes in Historic Creep Rate, Flook Ranch, Bitterwater Valley, CA
Evidence of active right lateral creep on the San Andreas Fault (SAF) at Flook Ranch (latitude $36.3982\deg$) includes offset cultural features, prominent tectonic geomorphology, macro- to micro-scale deformation of late Holocene alluvium, and historic variations in rate of activity. At Flook Ranch, the fault zone consists of two $N35\deg$W-striking faults that bound a fault-parallel, linear swale that is bordered by two 0.5-m-high scarps. Trenches excavated at the site exposed massive silt and clay loam overlain by $<$1,100-year-old thin-bedded silt and gravel, and two (western and eastern) fault zones bounding, in part, the margins of the linear swale. Creep-related structures at the fault zones include 4-m-wide zones of abundant fractures extending across the eastern and western fault zones, broad synformal warping, and truncated alluvial deposits. Initial observations of faulted sand samples at the microscopic scale show many distinct, narrow ($<$0.5mm) faults that consist of ultra-fine grained (clay size) material and bound silt and silty sand units. Sand units immediately adjacent to faults are notable for their lack of textural features that might record faulting; no porosity reduction, grain size reduction, or localized preferred grain orientation is observed at fault contacts. At the eastern fault zone: (1) apparent vertical offsets range from $<$0.2 m to $\sim$1.0 m, (2) Reidel shears dip steeply to the east, strike $10\deg$-$30\deg$ clockwise from the main SAF zone, and flower upward into pervasive extensional fractures, and (3) Reidel shears, rotate clockwise into extensional fractures as they approach the ground surface. In comparison, the western fault zone is characterized by: (1) apparent vertical offsets of as much as 2 m, (2) steep west dipping Reidel shears that strike $0\deg$-$50\deg$ clockwise from the fault zone, and (3) fractures extending to the ground surface show no progressive clockwise rotation up section. A March 2003 survey of an offset $\sim$95 year-old fence line yielded a dextral offset of 1.19-1.51 m and a creep rate of 22-28 mm/yr. This creep rate is: (1) lower than a 1966 reported creep rate of 32-33 mm/yr of the same fence line (Brown and Wallace, 1968), and (2) lower than a 1983 reported historic creep rate of 28-32 mm/yr of the fence line (Cotton et al., 1986), but (3) greater than the average 26 year (1969-1995) creep rate of 16-17 mm/yr of a U.S.G.S creep meter located $<$10 m south of the fence (Schulz, 1989). In summary, structural relations exposed in trenches, coupled with survey data, indicate that creep is concentrated primarily within the two fault zones, and that these fault zones record different faulting styles and variability in activity and deformational histories.
G11A-0765 0800h
Slow slip events estimated from ground deformation and seismic wave velocity structure estimated from seismic wave tomography in the Tokai region, Japan
Ohota et al. (2004) estimated the fault model of the 2001 Tokai slow slip event based on the re-processing GPS data comparing the ground deformation in 1997-2000. From their discussion, in the period of slow slip event, interplate coupling is tightly remained in the depth of 10-25 km in the plate boundary in the Tokai region. Forward slip consisting of slow slip event is ongoing with a rate of 3 cm/yr at the depth of 25-40 km in the plate boundary. The slow slip event fault locates outside of the 1944 Tonankai earthquake fault model by Tanioka and Satake (2002). It suggests that slow slip event is occurred outside of asperity area of historical large earthquakes along the Nankai Trough. Recently dense seismological network is constructed and seismic wave velocity structure are discussed with detail from the seismic wave tomography method (Matsubara et al.,2004; Kamiya and Kobayashi, 2004). According to their resolution, the low Vp and Vs layer is estimated in the 25-40 km depth of the plate boundary between the Philippine Sea plate and Eurasian plate in the Tokai region. Precisely we discus the location of low Vp and Vs layers, it is located upper part pf the plate boundary with width of 10km, which is corresponded to the lower continental crust..It is located just under northern area of Lake Hamana and Mikawa Bay. Kodaira et al.(2004), they conclude that high Vp/Vs ration distributed in subducting upper ocean crust around 30 km depth of the plate boundary is a occurrence mechanism of the 2001 Tokai slow slip event. From our research, the rate of Vp/Vs in upper part of the subducting oceanic crust is changing with the depth of the plate boundary. Until 25 km depth, velocities of P-wave and S-wave are not low, and the rate Vp/Vs is normal. At depth of 25 - 40 km, velocities of P-wave and S-wave are decreasing, and the rate Vp/Vs is increasing. The depth of high rate is correspond the slow slip fault estimated by Ohota et al.(2004) and hypocenters of unvolcanic low frequency tremore observed by Obara et al.(2002). If high Vp/Vs rate is depend on the dehydration of the subducting oceanic crust, it is suggested that around Lake Hamana, an upper part of subducting crust is dehydrating at the 25 km depth of plate boundary, and dehydrating is triggered the slow slip events and tremors.
http://members8.tsukaeru.net/kimata/presen_at%20_meeting.html
G11A-0766 0800h
Uplift "potential" derived from GPS horizontal measurements and its geological implications in Taiwan
A mathematic model based on conservation equation was constructed to evaluate a theoretical uplift rate based on the published GPS measurements around Taiwan, which is undergoing a collision between the Luzon arc and the Eurasian continental margin. With this simple model, "potential for uplift" is calculated. The potential uplift rate is generally higher in the east and decreases westward. Segmentation of the tectonic regimes onland Taiwan can be clearly observed from the result. For example, the northeastern Taiwan (Ilan Plain) has negative values of potential uplift, which means the whole Ilan Plain is undergoing subsidence. In fact, the Ilan Plain is the westernmost extension of the opening Okinawa trough. The calculated uplift rates were then compared with the topography in order to evaluate whether areas with higher theoretical uplift rates are areas with higher elevation. The coherence between calculated uplift rate and the topographic data were done by calculating their cross-correlation. The results show that areas with high conherence are areas with higher seismicity. The underlying rationales of this observation are not fully understood, however, we speculate that the differences in time scale of deformation or whether surface processes are involved might be responsible for this observation.
G11A-0767 0800h
Extensional Strain Accommodation by Diking in the Transitional Main Ethiopian Rift
Strain localizes as rifting proceeds to continental breakup, but the partitioning of strain between faults and magmatic intrusion remains controversial. We integrate new local seismicity data with fault slip, numerical models and geodetic data from the volcanically active Main Ethiopian Rift (MER) to evaluate strain partitioning in a rift transitional between continental and oceanic in style. Geodetic data show that strain has localized to 20 km-wide, 60 km-long 'magmatic segments' marked by aligned eruptive centers, dikes, and small offset normal faults within the central MER (Bilham et al, Geophys. Res. Lett., 27, 1999). From October 2001 to February 2003 over 2000 local earthquakes were recorded on 170 broadband seismic instruments deployed within the Northern MER and its uplifted rift flanks. Earthquake locations are estimated with the best 1-D and 3-D velocity models obtained from local earthquake tomography. The catalogue of earthquakes is complete above Ml - 2. Earthquakes cluster within the magmatic segments and the majority of events occur at depths of 6-10 km. Epicentres of small magnitude events (Ml $<$ 4) parallel faults that cut Quaternary - Recent lavas and aligned eruptive centres within the magmatic segments. Large offset border faults bounding the MER are inactive, excluding a dense cluster of events concentrated at the intersection of the 29 My Red Sea Rift and the 11 My MER, where the older Red Sea Rift flank is flexing into the younger MER. Historical events follow the same spatial patterns. Fault plane solutions and fault slip data show N100 extension, with some local variations for events beneath shield volcanoes. Thus, the distribution and mechanism of local and historical seismicity, as well as existing geodetic data show that strain is concentrated within the narrow magmatic segments, and not along border fault detachments. Local earthquake tomography shows 20 km-wide high velocity zones extending to the base of the seismogenic layer beneath magmatic segments, which we interpret as basaltic intrusions that feed dikes, fissures, and shallow magma chambers. We test these interpretations with elastic dislocation models of deformation above dikes, and compare results with seismicity, fault, and dike distributions. This integrated study shows that strain is accommodated by magmatic intrusion and small displacement faults. Current geodetic data is restricted to 1 profile and thus a dense GPS survey is needed to quantify and simulate strain partitioning.
G11A-0768 0800h
Present day Fault Kinematics of Northern Baja California, Mexico
Although the major plate boundary between the North American and Pacific plates runs through the Gulf of California, the northern portion of Baja California accomodates a significant amount of plate motion. Partitioning of slip among the different faults, as well as the behavior of blocks between the faults, is still unclear. The interpretation of the surface velocity field measured by campaign GPS is highly model dependent. In particular the amount of slip accomodated by each fault can be quite sensitive to assumption about rheological properties, fault geometry, and stage in the earthquake cycle. We recently collected new campaign GPS measurements in the region that improve our knowledge of Northern Baja Kinematics. Here we present the new velocity field. We analyze the effect of different modeling assumptions on slip partitioning along the different faults using finite element models. In particular we show that it is important for the models to incorporate geological observation such as earthquake cycle.
G11A-0769 0800h
New Constraints on Models for Time-Variable Displacement Rates on the San Jacinto Fault Zone, Southern California
Existing geodetic, geomorphic, and geologic studies yield apparently conflicting estimates of fault displacement rates over the last 1.5 m.y. in the greater San Andreas fault (SAF) system of southern California. Do these differences reflect biases in one or more of the inference methods, or is fault displacement really temporally variable? Arguments have been presented for both cases. We investigate the plausibility of variable-rate fault models by combining basin deposit provenance, fault trenching, seismicity, gravity, and magnetic data sets from the San Bernardino basin. These data allow us to trace the path and broad timing of strike-slip fault displacements in buried basement rocks, which in turn allows us to test weather variable-fault rate models fit the displacement path and rate data through the basin. The San Bernardino basin lies between the San Jacinto fault (SJF) and the SAF. Isostatic gravity signatures show a 2 km deep graben centered directly over the modern strand of the SJF, whereas the basin is shallow and a-symmetric next to the SAF. This observation indicates that stresses necessary to create the basin have been centered on the SJF for most of the basin's history. Linear magnetic anomalies, used as geologic markers, are offset $\sim$25 km across the northernmost strands of the SJF, which matches offset estimations south of the basin. These offset anomalies indicate that the SJF and SAF are discrete fault systems that do not directly interact south of the San Gabriel Mountains, therefore spatial slip variability combined with sparse sampling cannot explain the conflicting rate data. Furthermore, analyses of basin deposits indicate that movement on the SJF began between 1.3 to1.5 Ma, yielding an over-all average displacement rate in the range of 17 to 19 mm/yr, which is higher than some shorter-term estimates based on geodesy and geomorphology. Average displacement rates over this same time period for the San Bernardino strand of the SAF, on the other hand, are inferred to be low, consistent with some recent short-term estimates based on geodesy, but in contrast with estimates based on geomorphology. We conclude that either published estimates for the short-term SJF displacement rate do not accurately reflect the full SJF rate, or that the SJF rate has decreased over time, with implications for rate changes on other faults in the region. We explore the latter explanation with models for time-variable displacement rate for the greater SAF system that satisfy all existing data.
G11A-0770 0800h
Recent Crustal Deformations In Kii Peninsula, Southwest Japan Derived From Dense GPS Observations: Interplate Coupling And 2004 Earthquake Sequence SE Off Kii Peninsula
Kii peninsula, southwest Japan is located close to the Nankai trough and suffers from the subduction of the Philippine Sea plate. Deeper part of the source region of interplate earthquakes, such as 1946 Nankai is beneath this peninsula. Therefore it is important to reveal detailed crustal deformations in this area for the purpose of understanding of generation process of interplate earthquakes and long-term forecast of their occurrence. We established 10 GPS observation sites filling the gaps of the GEONET operated by GSI along two lines nearly parallel to the relative motion between the Philippine Sea and Amurian plates. Resultant average spacing is 5~10km. We have repeated the campaign survey of this traverse using dual-frequency receivers since March 2001. So far, we have collected data from 4 campaigns done every March and can discuss displacement/velocity field during the recent 3 years. We calculate coordinates of our campaign sites with GSI_fs permanent sites in ITRF2000. We use the GIPSY/OASIS II software with JPL precise ephemeredes in the analysis. Velocity of each site is calculated referring to the first campaign and converted to the relative velocity to the Amurian plate using the Euler vector by Heki et al. (1999). Velocities derived from 2001 and 2002 campaigns are about 20mm/yr in the middle part of Kii peninsula and 35mm/yr at its southern tip, respectively, relative to the Amurian plate. Their directions are WNW. The gradient is almost linear up to 33.8N and there is no significant change in gradient in northern part. Velocities along the eastern line are a little smaller than those on the western line. We apply a model with multiple fault segments to this velocity field and estimate their slip deficit rates. We adopt the 9 segments around Kii peninsula from the fault model of Sagiya and Thatcher (1999). Estimated slip deficit rates are as large as 70 mm/yr that is slightly larger than the relative plate motion. If we add deeper extension to the modeled fault, slip deficit rates of about ~60 mm/yr are obtained. If we subtract a 10mm/yr nearly westward block motion, which proposed by several researchers, from the observed velocities, slip deficit rates are significantly reduced. However we obtain slip deficit rate larger than 40mm/yr in the transition zone estimated by Hyndman et al.(1995). These results imply that significant strains are being accumulated in the transition zone as well as the main locked zone. Two earthquakes of Mjma larger than 7 occurred southeast off Kii Peninsula on September 5, 2004. GEONET of GSI revealed southward motion up to 5cm in the eastern part of Kii peninsula. We have started reoccupation of our sites and will hopefully observe co- and postseismic deformations due to this activity.
G11A-0771 0800h
Observation of Pluri-Decadal Movement at the NE-Border of the Adria-Plate by Subsurface Long-Base Geodetic Instrumentation and Modeling
The movement of the Adria plate is important in the understanding of the seismicity of Italy and the Dalmatian region. Its northern margin delineates the most seismic area of the Alps and has been hit by destructive earthquakes. The deformation of the margin has been observed over four decades in a natural cave with long-base highly stable and sensitive instruments. We discuss the observed deformation on the decadal time scale. We show that the movement is the superposition of a steady state deformation rate, and a time varying deformation with half-period of about twenty years. The station is near to the Adriatic Sea, wherefore a sea loading effect is present. The change of sea-level in time involves also a steady state as well as decadal scale variations. We model the geodetic observations in terms of the deformation due to the plate-tectonic movement and discuss the deformation due to the sea-loading.
http://www.units.it/~geodin/strumentipag.html
G11A-0772 0800h
Stream Channel Offset and Preliminary Slip Rate on the San Andreas Fault, at the Van Matre Ranch Site, in the Carrizo Plain, California
To understand the spatial and temporal variation in fault slip it is important to improve the spatial coverage of slip and slip rate measurements along major active faults. A set of well-preserved channels are offset across the San Andreas fault at the Van Matre Ranch (VMR) site (35.154N, 119.700W) in the Elkhorn Hills area of the Carrizo Plain. The fault zone and offset channels at VMR were exposed by excavation in 1993 and 2004. This study included one fault-perpendicular and 5 fault-parallel trenches that exposed the buried thalwegs of several offset channels. Seventeen samples were collected from channel margin deposits for $^{14}$C dating and survey data was taken for accurate offset measurement of the buried thalwegs and geomorphic channels. The geomorphic history of the site is well manifested in the excavations with clear evidence for initial incision of the channels into Plio-Pleistocene fan units that were typically heavily bioturbated. The channels then back filled and the stratified channel sediments grade laterally into clayey silts. The buried thalweg of the currently active channel is offset 24.8 m, while the geomorphic offset is 27.6 m (qualitatively defined conservative uncertainties on offsets are $\pm$ 1m). The thalweg of the first beheaded channel is offset 48.8 m with a geomorphic offset of 51.8 m. The geomorphic offset of the second beheaded channel ranges from 71.9 to 79.0 m. There are no ages associated with these channels. The median dates of samples from the clayey silts in the currently active channel margin range between A.D. 1221 and 1108, implying a 34.7 mm/yr slip rate. The significance of the samples ages is dependent upon interpretation of the sediments in which they were collected. They were collected from clayey silts which are either colluvium, washed down from adjacent hill slopes, or autochthonous alteration of the channel deposits by pedogenic processes (largely burrowing). If the samples were derived from colluvial processes, the ages of the samples would provide a maximum slip rate. However, if the samples were derived from older channel sediments, then they would indicate a slip rate minimum. This preliminary slip rate is consistent with the measured slip rate at Wallace Creek, approximately 18 km to the northwest where Sieh and Jahns documented a late Holocene slip rate of approx. 33.9 $\pm$ 2.9 mm/yr, and with the regionally assumed 35-mm/yr rate derived from decadal time-scale geodetic measurements.
G11A-0773 0800h
Cosmogenic Be-10 ages of Angel Lake and Lamoille moraines and late Pleistocene slip rate of the rangefront normal fault, Ruby Mountains, Basin and Range, Nevada
We use Be-10 cosmogenic radionuclide (CRN) exposure dating to quantify the timing of late Pleistocene glacial advances and to estimate the rangefront normal fault slip rate along the Ruby Mountains in the Basin and Range, Nevada. Ten Be-10 CRN exposure ages from the Angel Lake terminal moraine in Hennen Canyon limit deposition to between 15.4-23.1 ka (average = 18.2 ka; SD = 2.5 ka), an interval that overlaps with the Tioga glacial advances in the Sierra Nevada and Pinedale advances in the Rocky Mountains during MIS-2. The termination of the Angel Lake glaciation at ~15.4 ka is nearly synchronous with the final highstand and subsequent rapid desiccation of Lake Lahontan. Previous relative age dating studies in the Ruby Mountains have inferred deposition of Lamoille moraines during MIS-4 (~59-74 ka) or MIS-6 (~130-190 ka). However, fifteen Be-10 CRN exposure dates obtained from a Lamoille lateral moraine in Hennen Canyon (average = 30.7 ka; SD = 11.5; range = 19.3-66.5 ka) do not support these assignments and instead suggest that the moraine may have been deposited during MIS-3. Assuming the average Be-10 model age from the Lamoille surface represents a minimum age constraint, we obtain a maximum fault slip rate of 0.40-0.60 mm/year for a $60\deg$ dipping fault and 11-16 m of vertical separation across faulted Lamoille moraines. Accommodation of 0.4-0.6 mm/year of slip by the Ruby Mountains fault zone implies an average 0.2-0.3 mm/year of horizontal strain accumulation across the fault during the last ~31 ka, and thus characterization of the central Basin and Range as a geodetic microplate may be at odds with late Pleistocene horizontal displacement rates along the Ruby Mountains fault zone. If correct, the observed disparity between the CRN exposure dates and geologic correlations underscores the importance of quantitative age constraints for climate reconstruction, landscape evolution, and tectonic geomorphology.
G11A-0774 0800h
Numerical Modeling of Deformation in the Los Angeles Basin, Southern California
Deformation in the Los Angeles Basin is characterized by a narrow band of shortening between downtown Los Angeles and the San Gabriel Mountains. Geodetic measurements indicate that there is $\sim$4.5 mm/yr shortening across this band; however, the mechanism by which this shortening is accommodated is poorly characterized and understood. It is unclear whether shortening is accommodated by elastic strain accumulation and release along a series of sub-parallel thrust faults, including the Sierra Madre fault (the frontal fault of the San Gabriel Mountains) or whether it is accommodated by anelastic processes taking place in the low rigidity sediments within the basin. We will present the results of numerical modeling of the region using both 2D and 3D finite element models from the GeoFEST (Geophysical Finite Element Simulation Tool) code and the QuakeSim Computational Portal. These results will be compared with geodetically observed deformation being recorded by the Southern California Integrated GPS Network (SCIGN). Previous modeling indicates that the low rigidity basin sediments play a key role in accommodating the observed shortening, but those models were unable to predict the observed deformation. We have further refined our 2D models by testing model sensitivity to changes in the modeling domain (i.e. creating a box with greater depth) and to the degree of mesh refinement in areas characterized by large material contrasts. The results of these model sensitivity tests indicate that a modeling domain that is 400 km x 600 km with a mesh refinement of $\sim$1.25 km node spacing near the transition between basin sediments and bedrock (and 5 km node spacing elsewhere) produce velocities that do not reflect errors caused by numerical artifacts within the model. We will compare the newly refined 2D results with kinematic models constructed by other workers. We will also construct a simple 3D model of the Los Angeles Basin in order to characterize the problem and to see if we are better able to predict observed deformation with a more realistic geometric and mechanical model.
G11A-0775 0800h
Incorporation of Complex Rheologies Into Models of Interseismic Displacements
The standard viscoelastic model of interseismic displacements near strike-slip faults was first proposed by Savage and Prescott in 1978 (SP78). This model has been widely used to model data and is a powerful tool to guide intuition about interseismic displacements. The model of SP78 provides the displacements throughout a seismic cycle due to an infinitely long strike-slip fault periodically breaking an upper elastic layer overlying a Maxwell linear viscoelastic half-space. The model of SP78 has fairly stringent assumptions, such as 1) that the upper seismogenic layer is elastic, 2) that the aseismic lower region can be considered a half-space, 3) that the anelastic half-space behaves simply, 4) that the shear moduli in the upper layer and half-space are identical, 5) that successive ruptures are infinite in length and 6) they occur regularly in time with constant coseismic displacement, and 7) that the system is in a cycle-invariant state, where the displacements throughout the interseismic period do not depend on the particular cycle (often referred to as a mature or steady-state). These assumptions are likely not applicable to the crust and upper mantle. For example, the shear modulus of the seismic and aseismic regions is not uniform and recent studies of postseismic relaxation have indicated that the anelastic region of the Earth has a much richer time dependence than the single exponential decay predicted by Maxwell viscoelasticity. Additionally, recent paleoseismological studies have indicated that rarely do earthquakes occur periodically. When earthquakes repeat irregularly the assumption that crustal deformation is cycle-invariant is likely inappropriate. We have extended the model of SP78 to more general earthquake rupture histories and a wider suite of linear viscoelastic rheologies of the seismic and aseismic layers. While the general model we propose is still too simple to model geodetic data with high spatial coverage, it is rich enough to be used to build intuition about interseismic displacements assuming irregular earthquake histories and complex rheologies. Finally, the method we propose can be applied to existing 3D models of finite ruptures that utilize the Correspondence Principle.
G11A-0776 0800h
Slip rate of the Calico fault: Implications for anomalous geodetic strain accumulation across the Eastern California shear zone
Recent earthquake activity and high geodetically derived fault-slip rates across the Eastern California shear zone motivate comparisons with long-term geologic deformation rates to test for transient strain accumulation. We report new geologic slip-rate results from a transect at 34.8$\deg$N across the central Mojave Desert where six dextral faults (Helendale, Lenwood, Camp Rock, Calico, Pisgah-Bullion, and Ludlow) accommodate all late Quaternary right-lateral displacement. High-resolution LIDAR topography data have been successfully acquired across all six faults as part of a project to measure a complete budget of long-term geologic fault slip rates. Field investigations of the northern Rodman Mountains conducted with the aid of the new topography data identified several surfaces dextrally offset by the Calico fault. A preliminary slip rate of 1.3$\pm$0.3 mm/yr is calculated from an 800$\pm$ 200 m offset of alluvial fan deposits containing clasts of the ca. 600 ka Pipkin basalt flow. Cosmogenic surface exposure age dating of offset geomorphic surfaces and refined Ar/Ar dating of the basalt flow, in progress, will provide multiple constraints of this fault slip rate. The slip rate of the Calico fault is more than twice that of the Blackwater fault, located on strike with the Calico fault in the northwest Mojave Desert. This discrepancy supports that strain is transferred away from the Calico fault and other adjacent northwest-striking dextral faults onto domino-style rotating blocks bounded by sinistral faults in the Fort Irwin region. A newly identified active thrust fault and fault-related fold bounding the northern Rodman mountains accommodates shortening east of the Calico fault that may be caused by space problems at the intersection of these conjugate fault systems. Overall, slip rate on the Calico fault, together with existing paleoseismic histories on adjacent faults, does not account for more than 5 mm/yr of strain accumulation across the Eastern California shear zone. Geodetic strain rates of 10 mm/yr or more across this region indicate either province-wide transient strain accumulation or that paleoseismic data consistently underestimate long-term fault slip rates. Additional measurements of offset features for each fault that average several earthquake cycles are required to test these alternative hypotheses.
G11A-0777 0800h
Hydrological Changes Induced by the 2003 Tokachi-oki Earthquake at Wells in Hokkaido, Japan
Thirty hydrological responses induced by the M8.0 Tokachi-oki earthquake in 2003 were observed at hot spring wells and an undersea coal mine in Hokkaido, Japan. The purpose of measuring the groundwater level or discharge rate of the wells is mainly to monitor hot spring resources. Nineteen wells are observation wells to monitor the static groundwater level. Groundwater of two wells flows out without pumping, and that of nine wells is pumped out and used. Twelve coseismic increases and 18 decreases in groundwater levels or discharge rates were observed in 29 wells and one coal mine in response to the Tokachi-oki earthquake. Twenty-seven of the 30 hydrological anomalies can be explained as a poroelastic responses to the earthquake-induced contractional or dilatational strain inferred from a fault model determined by dense static GPS observation. In five wells, changes in groundwater levels induced by four large earthquakes, the Kushiro-oki (Mw 7.7), Hokkaido-nansei-oki (Mw 7.7), Hokkaido-toho-oki (Mw 8.3) and Sanriku-haruka-oki (Mw 7.6) earthquakes in 1993 -- 1994 were proportional to the inferred volumetric strain steps. Observed groundwater-level changes and inferred volumetric strain steps induced by the Tokachi-oki earthquake are consistent with the proportionality. Strain sensitivities estimated by $M_2$ tidal constituent are consistent with those determined by coseismic responses in three of the five wells. The strain sensitivities estimated by $M_2$ at the three wells are plausible, because extracted phase shifts in groundwater level are almost the same as theoretical phase shifts in volumetric strain. We observed many increases in groundwater-levels at wells in eastern Hokkaido induced by the 2003 event, whereas many water-level decreases induced by the 1952 Tokachi-oki earthquake (Mw 8.2) were reported at wells in the same area. It suggests that strain induced by the 1952 event was dilatational but the strain induced by the 2003 event was contractional in eastern Hokkaido.
G11A-0778 0800h
Mapping Active Fault Zones in Southern California Using Master Multispectral Imagery Data
Recent studies of active fault zones using the GPS and InSAR techniques have revealed slip rates that often differ from the slip rates determined from geological observations. This discrepancy is principally due to the different time windows over which surface movements are integrated in both approaches. If surface velocities near faults vary over cycles of several hundreds of years, it becomes important to document the slip history along faults over various time scales as it has been recorded in the Quaternary deposits along the fault. To this endeavor, we have acquired sets of images of the major active faults in Southern California using the MODIS/ASTER airborne simulator (MASTER) instrument. The lines are flown at low altitude above the ground to provide 4 to 5 m spatial resolution in the 50 spectral bands (0.5 to 13 microns) of the instrument. A preliminary set of data was acquired in the summer 2003 over the Garlock and the Blackwater faults in the Mojave. A more extensive campaign carried out in September 2004 covered more than 1000 km of fault lines from the central section of the San Andreas fault to the Salton Sea area. The data are being processed to extract reflectance and emissivity information. Preliminary analysis of the 2003 data confirmed the strong potential of the MASTER thermal bands to identify changes in surface emissivity due to subtle variations of the mineral composition of the deposits. Additional information on the near surface structure of the fault zones can be obtained by combining day and night surface temperature maps, as buried sections of faults are revealed by thermal capacity contrasts between the two sides of a given fault. The paper will present the data set acquired during the 2003 and 2004 campaigns and the status of the raw data processing into geo-referenced emissivity and reflectivity maps of the fault zones.
G11A-0779 0800h
Crustal Deformation of Luzon Island, Philippines from GPS-based Geodynamic Models and Structural Analyses of Satellite Imagery
We develop a comprehensive model of plate boundary deformation, based on satellite-based structural analyses combined with geodetic and seismic evidence of present-day deformation of Luzon Island, Philippines. The region comprises the northern part of the Philippine Mobile Belt, a plate boundary zone situated between two active, opposing subduction zones. Correlation of geodetic and seismic evidences with the structural map generated from the Shuttle Radar Topography Mission (SRTM) DEM and Landsat images reveal the complexity of recent deformation style of the island. Geologic structures were mapped using co-registered sets of SRTM DEM shaded relief images and Landsat 7 images. The SRTM shaded reliefs were processed at different illumination directions and overlaid with cloud-free Landsat 7 mosaic. We then use the satellite-based identification of active faults to develop a microplate model for active deformation in the Philippine arc. We apply an elastic block model to describe active deformation within the region, utilizing simultaneous inversions of GPS-observed site velocity vectors and earthquake slip vectors to solve for block rotations, elastic fault strain accumulation on block boundaries, and internal block strain. Campaign-based GPS observations acquired from 1996-2002, and focal mechanism data from the Harvard-CMT catalogue from 1972-2003 were primary inputs for deriving the block models. Findings for slip across the major faults show 35-44 mm/y sinistral slip along the Philippine fault system, 19-28 mm/y along the Digdig fault and 21-28 mm/y motion along the northern Cordillera fault. Slower slip rates are detected along the Macolod Corridor System , with 11-13 mm/y left lateral, trans-tensional motion. Strain across the Philippine fault is modeled as a near vertical strike-slip fault, with elastic strain accumulation on a fully locked fault to 15 km. depth.
G11A-0780 0800h
Repeated Historic Surface Ruptures of the Denali Fault at Delta River, Alaska During Large Earthquakes in 1912 and 2002
The Denali fault ruptured through stands of mature spruce trees on the Delta River valley floor during an Ms 7.2-7.4 earthquake on July 6, 1912 and again during the Mw 7.9 earthquake of November 3, 2002. In both events, most trees on the surface rupture along 2 km of the fault trace were damaged by splitting and tilting. Displacements in both events were dextral with a subordinate dip-slip component (south side down). Tree ring counts from older damaged trees on the fault trace closely date the age of the penultimate event at 1912. The only earthquake that fits the requirements for timing, location, and size to have caused the pre-2002 tree damage is an Ms 7.2-7.4 event on July 6, 1912, the epicenter of which had been located 40 km southwest of the Delta River fault crossing by Boyd and Lerner-Lam (1988). Intensity data for the widely felt 1912 earthquake are compatible with unilateral westward rupture on the Denali Fault. Empirical data for the estimated magnitude range of the 1912 earthquake suggest a surface rupture length of 60-84 km, average horizontal displacement of 140-210 cm, and maximum horizontal displacement of 240-390 cm (Wells and Coppersmith, 1994). The 2002 surface rupture was marked by large dextral surface slip (to 800+ cm) and variable dip slip along 240 km of the Denali Fault and 65 km of the Totschunda fault in the central Alaska Range. The Delta River valley, 90 km east of the 2002 epicenter, lies within a transition zone about 10 km wide in which dextral slip diminishes from 600 cm or more east of the valley to less than 450 cm west of the valley. Geodetic data for the 2002 rupture along the TAPS oil pipeline in the Delta River valley indicate a total of 580 cm dextral slip and 130 cm dip slip distributed over a zone 1,000 m wide. Surface trace of the fault is poorly developed on the valley floor because of distributive deformation in the thick underlying unconsolidated deposits; offsets on individual fissures are less than 130 cm dextral and 60 cm vertical. Paleoseismologic data from test pits at the west bank of the Delta River indicate two earlier surface faulting events on the Denali Fault probably comparable in size to the 2002 earthquake. Preliminary 14C dates of these paleo-events suggest a recurrence interval of about 350-400 years and an average long term slip rate of about 15 mm/yr. Rupture during the smaller 1912 earthquake contributes to reduce the size of any 2002 slip deficit that may exist along the Denali fault in the Delta River valley and probably also in the low-slip segment extending 70 km west of the valley to the end of the 2002 Denali fault surface trace.
G11A-0781 0800h
Bayesian Inference of Lower-Crustal Viscosity Near the Kunlun Fault Based on Geologic, Geomorphic, and Geodetic Data
Bayesian methods provide the means of integrating geologic, paleoseismic, seismic, and geodetic data to improve estimates of fault zone and lower-crustal properties that are important for predicting the long-term deformation of plate boundary zones, understanding deformation and stress transfer following earthquakes, and estimating future seismic hazard. We developed a Bayesian methodology that integrates geologic, geomorphic, and geodetic data to provide probabilistic estimates of fault-zone and lower-crustal properties. Unlike previous studies that have used {\it a priori} information to constrain optimizations, and bootstrapping to assess uncertainty in model parameters, we explicitly employed Bayes' rule in our analysis and estimate model parameters using simulation methods. Our Bayesian methodology provides a means of straightforwardly assessing covariance between and uncertainty within model parameters, and allows geologic and geomorphic information to be used to quantitatively constrain fault-zone and lower crustal properties. Because the methodology casts model parameter estimates in a probabilistic framework, inversions obtained from these methods provide far more information about the behavior of the model parameters inferred from data than do conventional geophysical inversions. Finally, these methods may be modified to incorporate other a priori information (e.g., fault zone geometry from microseismicity, and paleo-eathquake timing and recurrence from paleoseismic excavations), and thus may serve as important tools in seismic hazard assessments that strive to formally incorporate different data types. We applied this methodology to the Kunlun Fault in northern Tibet, where quantitative estimates of lower-crustal viscosity are lacking. Here, geologic and geomorphic information constrains the range of permissible long-term slip rates and coseismically generated offsets. We combined these a priori estimates with GPS velocities using a Bayesian implementation of an elastic-viscoelastic earthquake cycle model to estimate fault-zone and lower crustal properties in the area. We found that the non-dimensional relaxation time varies between 0.04 and 3.25 (95% bounds), implying plastosphere viscosity between $3.8 \times 10^{19}$ Pa s and $6.7 \times 10^{21}$ Pa s. These viscosities are large compared to those required by models of mid-crustal channel flow, perhaps suggesting that the mid- to lower-crust underneath northern Tibet is significantly more viscous than previously envisioned.
G11A-0782 0800h
Comparison of the Plio-Pleistocene and Quaternary transtensive deformation structures along the Pacific-North American Plate boundary, Eastern Sierra region, southwestern Inyo County, California
This paper provides a description of the Late Neogene and Quaternary Pacific-North American of plate boundary deformation along the southeast margin of the Sierra Nevada block and southwest margin of the Great Basin province, Inyo County, California. Offset flood basalts, lacustrine deposits and peneplain remnants provide excellent marker horizons for constraining the timing, magnitude and orientation of crustal deformation. In addition, Quaternary fault scarps are well exposed in the arid basins of the region and there has been extensive work on Quaternary slip rates during the last 25 years. Progressive deformation, and successive abandonment of structures has been recognized between the early Pliocene (~6 Ma) and the Holocene. Several poorly organised shear zones with apparently low deformation rates resulting in poorly defined and discontinuous fault zones have been identified within the intermontane regions. The orientation of small-scale topographic basins within the intermontane regions provides useful indicators of the geometry of intrablock strain and locally show pervasive kilometer-scale intrablock deformation of Plio-Pleistocene age. The more laterally continuous NW trending faults of Middle and Late Quaternary age, even those near and bounding the range fronts are commonly steeply dipping, about 75-85 degrees, with mostly strike-slip (~60-80%) to dip-slip displacement, consistent with the regional transtensive plate boundary setting.
G11A-0783 0800h
Displacement and Shear Discontinuities in GPS and Other Survey Networks
It seems we can detect the orientation and nature of faults and shear zones using a method for analyzing deformation with survey data. Survey data are in terms of arbitrary coordinate systems---the initial $(X,Y)$ and current $(x,y)$ systems---rather than in terms of natural coordinate systems $(S,N,s,n)$ of the fault or shear zone. The angle $\alpha$ is the clockwise angle between {\it X} and {\it S}. A method for detecting faults and shear zones involves computing the deformation gradient tensor $F_X_Y$ with the survey data, and then determining the natural coordinates by maximizing the partial derivative $\frac{\partial s}{\partial N}$ as a function of $\alpha$. Results are presented from survey measurements made on a landslide surface, near earthquake ground ruptures, and in the vicinity of a tectonic plate boundary. I studied displacements of points on the Slumgullion landslide in Colorado and contrasted the results with mapped structures. I find sudden, large, jumps in the partial derivative $\frac{\partial s}{\partial N}$ where there are faults with a strike-slip component parallel to $\alpha$. Such discontinuities in the derivative are absent where strike-slip shift is absent. Likewise, I studied part of the ground rupture associated with the 1999 Ducze-Bolu earthquake in Turkey. The rupture passed beneath the Bolu-Kaynasli viaduct, damaging the structure. The piers of the viaduct were permanently displaced, and were surveyed before and after the earthquake. This analysis shows that the earthquake rupture, at the ground surface, is a belt of right-lateral shear approximately 50 meters wide in the vicinity of the viaduct. This method can also be used with GPS data. Preliminary results from 16 continuously operating stations near San Francisco suggest portions of the San Andreas, Hayward and Calaveras fault zones were active at different times during 2003. For example, during the interval April to May all three fault zones were possibly active. From August to September, it seems the Hayward and Calaveras fault zones were active, but only in a region north of San Francisco Bay.