T51C-0675
Samples and Data Products from the San Andreas Fault Observatory at Depth (SAFOD)
The San Andreas Fault Observatory at Depth (SAFOD), a part of the National Science Foundation's EarthScope program, has recently completed the third phase of drilling and coring. With additional support from the USGS and the International Continental Drilling Program (ICDP), we have obtained over 30m of core from active deforming sections of the San Andreas Fault. In addition to the core, cuttings, and fluid samples collected at the drill site, there have been several successful instrument deployments in both the Pilot Hole and Main Hole, during the past year. This paper will discuss the acquisition, storage and distribution plan for the diverse types of data and samples that have been collected at SAFOD to date. The website at http://safod.icdp-online.org provides the detailed descriptions of all the physical samples (cuttings, core and fluid samples) available from SAFOD, together with sample request form and other general information on the SAFOD project. The samples are being archived at the , Integrated Ocean Drilling Program's Gulf Coast Repository in College Station, Texas. The geophysical monitoring program at SAFOD continues to expand. There is a real-time seismic data stream of 250 sample/sec data (downsampled from the 4000 sample/sec onsite data). Helicorder previews can be viewed at http://quake.wr.usgs.gov/cgi-bin/helipark.pl. The high sample rate data are available in SEED format from the Northern California Earthquake Data Center (http://quake.geo.berkeley.edu/safod/) and from the IRIS data center (http://www.iris.edu/data/data.htm). And, new this year, there are now both tiltmeter and strainmeter data flowing to the data centers on a daily basis, available at http://www.ncedc.org. As with all elements of EarthScope, these data and samples are openly available to members of the scientific and educational communities. http://www.earthscope.org
T51C-0676
Core Across the San Andreas Fault at SAFOD - Photographs, Physical Properties Data, and Core-Handling Procedures
Core samples were obtained that cross three faults of the San Andreas Fault Zone north of Parkfield, California, during the summer of 2007. The cored intervals were obtained by sidetracking off the SAFOD Main Hole that was rotary drilled across the San Andreas in 2005. The first cored interval targeted the pronounced lithologic boundary between the Salinian terrane and the Great Valley and Franciscan formations. Eleven meters of pebbly conglomerate (with minor amounts of fine sands and shale) were obtained from 3141 to 3152 m (measured depth, MD). The two conglomerate units are heavily fractured with many fractures having accommodated displacement. Within this cored interval, there is a ~1m zone with highly sheared, fine-grained material, possibly ultracataclasite in part. The second cored interval crosses a creeping segment of a fault that has been deforming the cemented casing of the adjacent Main Hole. This cored interval sampled the fault 100 m above a seismogenic patch of M2 repeating earthquakes. Thirteen meters of core were obtained across this fault from 3186 to 3199 m (MD). This fault, which is hosted primarily in siltstones and shales, contains a serpentinite body embedded in a highly sheared shale and serpentinite-bearing fault gouge unit. The third cored interval crosses a second creeping fault that has also been deforming the cemented casing of the Main Hole. This fault, which is the most rapidly shearing fault in the San Andreas fault zone based on casing deformation, contains multiple fine- grained clay-rich fault strands embedded in highly sheared shales and lesser deformed sandstones. Initial processing of the cores was carried out at the drill site. Each core came to the surface in 9 meter-long aluminum core barrels. These were cut into more manageable three-foot sections. The quarter-inch-thick aluminum liner of each section was cut and then split apart to reveal the 10 cm diameter cores. Depending on the fragility and porosity of the rock, the drilling fluid was removed either by washing with dilute calcium chloride brine (to approximately match the salinity of the formation fluids) or by gently scraping away drilling mud on the core surface. Once cleaned, each core section was photographed to very high resolution on a Geotek Multi- Sensor Core Logging (MSCL) system. This system was also used to determine the bulk density and magnetic susceptibility of each section. The 25 MB high-resolution photographs and the raw and processed physical properties data were then uploaded to the ICDP web server in Potsdam for public access (http://safod.icdp- online.org). The cores will be archived at the Gulf Coast Repository of the Integrated Ocean Drilling Program in College Station, TX. The MSCL photographs, physical property measurements, and other related data, such as geophysical logs, will be integrated using CoreWall, and will be on display at the meeting. All samples, data, and imagery are available to the science community.
T51C-0677
Active and Passive Seismic Imaging of the San-Andreas-Fault System at SAFOD
High-quality active and passive seismic data have been acquired in the vicinity of the San-Andreas-Fault system within EarthScope project SAFOD (San-Andreas-Fault-Observatory-at-Depth). We have processed parts of the available seismic data sets using newly developed techniques in order to derive a high-resolution image of the subsurface in the vicinity of the fault system. On one hand we applied Fresnel-Volume-Migration to the SAFOD2003 reflection seismic data set. Along the 50 km long profile we obtained a structural image of the region around the fault system from the surface down to about 10 km depth. The San-Andreas-Fault (SAF) appears as a strong subvertical reflector which can be followed from its surface trace down to about 4 km depth. At this depth it converges to a second similarly strong subvertical reflector which correlates with the Buzzard-Canyon-Fault (BCF) approximately 2 km southwest of the SAF. Also the Waltham-Canyon-Fault (WCF) about 9 km northeast of the SAF shows up clearly down to a depth of 5 km. Furthermore a bunch of subvertical reflectors appears at shallow depths over a distance of 10 km southwest of the SAF and another prominent reflector is visible approximately 15 km southwest of the SAF from almost the surface down to a depth of 10 km. Both latter features have no yet known surface expression, however a comparison with geodynamic modelling results shows that the existence of such faults is probably realistic. On the other hand we applied our novel migration-type location algorithm to passive seismic data recorded with an 80-level-3C-receiver array in the SAFOD main hole. We have located a number of local events in the vicinity of the fault system. A modification of the location algorithm also allowed the precise location of target events, which are the subject of the drilling activities in phase 3. The events are interpreted in terms of their relationship to the fault system structure. We have combined our results (target events, structural image) with additional available information (seismicity, borehole logs, shallow seismic images, etc.). Altogether this combined image provides a basis for a combined interpretation of the structure and the earthquake dynamics of this mega-shear zone on different scales and in particular in the vicinity of the SAFOD borehole.
T51C-0678
Variations Of Velocity Contrast Along The Rupture Zone Of The 2004 M6 Parkfield Earthquake On The San Andreas Fault
We systematically investigate the velocity contrast along the Parkfield section of the San Andreas Fault (SAF) that ruptured during the 2004 M6 Parkfield earthquake, using fault zone head waves (FZHW) that refract along the bimaterial interface. The analysis employs a total of 322 repeating earthquakes clusters identified from 8993 earthquakes in the relocated catalog of Thurber et al. (2006). The seismic data are recorded by 13 borehole stations in the High Resolution Seismic Network (HRSN) since 1987 and 23 surface stations in the Northern California Seismic Network (NCSN) since 1984, with normal distances to the fault less than 6 km. The study is part of a larger project on imaging bimaterial interfaces in the Parkfield region with multiple seismic networks. We stack waveforms of each repeating earthquake cluster, and align the peaks or troughs of the direct P waves assuming right-lateral strike-slip focal mechanisms. Clear FZHW are observed at surface and borehole stations that are within a few kms on the NE (slow) side of the SAF. The obtained velocity contrast is about 8% north of Middle Mountain, and decreases rapidly toward Gold Hill near the epicenter of the 2004 event. This implies an abrupt change of velocity contrast along the Parkfield section of SAF near Gold Hill. The observed variation of velocity contrast is consistent with 3-D tomography models of the Parkfield section, which include a high velocity body near Gold Hill on the NE side that produces a local reversal of the velocity contrast, and geological observations of a sliver of high-velocity rock immediately to the NE of the SAF associated with the Gold Hill fault.
T51C-0679
Analysis of fault zone head waves in the San Andreas and Southwest Fracture Zone around the hypocenter of the 2006 M6 Parkfield earthquake
We investigate the velocity contrast along the Parkfield section of the San Andreas fault (SAF) and the Southwest Fracture Zone (SWFZ) near the hypocenter of the 2006 M6 Parkfield earthquake by analyzing fault zone head waves (FZHW) that refract along the bimaterial fault interface (Ben-Zion, 1989). This study is a component of a larger project on imaging bimaterial interfaces in the Parkfield region with multiple seismic networks. The employed dataset has 1282 events recorded by the 2004 Parkfield Guided Wave Experiment (Michael et al., 2005) during the period of October 2004 to January 2005. This temporary network has 15 stations with L22 short-period sensors that are arranged in pairs on the main traces of the SAF and SWFZ. Due to the close proximity of the stations to the faults, complicated fault-zone-related phases are present in most waveforms. For accurate identification of FZHW and P wave arrivals, only waveforms with very high signal-to-noise ratios are used. By assuming right-lateral strike-slip focal mechanisms for all the events, we align the first peak or trough of the P wave arrival and examine the moveout of head wave with increasing along-fault propagation distance. So far we have identified clear FZHW signals at stations NE1O (near Gold Hill) and NE2O (near Parkfield Bridge), located within 100 m from the SAF, for propagation path along the SAF between SAFOD and Gold Hill. Analysis of head wave moveouts at these stations indicates small average velocity contrast in the immediate vicinity of the SAF between Middle Mountain and Gold Hill: ~4% contrast near Middle Mountain and ~3% near Gold Hill. Cross- section views of the seismicity reveal two concentrated bands of earthquakes: a shallower band around a depth of 5 km and a deeper band around a depth of 10 km. Head wave analysis of these two separate groups of events shows an increase of the velocity contrast with depth. The obtained results may reflect the velocity contrast between the NE bounding block and a damage zone generated by the 2004 event, as the employed stations are very close to the SAF. Analysis of the waveforms recorded at stations across the SWFZ is currently in progress.
T51C-0680
The Velocity Contrast Across the Parkfield Section of the San Andreas Fault Near the SAFOD Drill Site
We investigate the existence and strength of velocity contrasts in the Parkfield region of the San Andreas fault (SAF), focusing on the 15 km to the NW and SE of the SAFOD drill site. The research is based on observations of fault zone head waves (FZHW) that refract along an interface separating materials with different seismic velocities. When the fault separates contrasting materials the FZHW provide the best diagnostic and imaging tool of the properties of the fault interface. The study is part of a larger project on imaging bimaterial interfaces in the Parkfield region using data of multiple seismic networks. Here we examine waveforms from the temporary PASO deployment and nearby stations from the permanent High Resolution Seismic Network (HRSN) and Northern California Seismic Network (NCSN) over the time intervals July-December 2001 and April-August 2002. A total of approximately 500 events are located and recorded by 70 stations that are within 5 km of the SAF strike. We obtained accurate picks of head and direct P wave arrival times, with FZHW identified at 26 stations on the slow (NE) side of the fault. For stations on the slow block that are about 1-5 km from the fault there is a continuous head wave propagation for approximately 30 km along strike, from the creeping section NE of the SAFOD site to approximately 15 km SE of Middle Mountain. This encompasses the entire study region and implies a velocity contrast that is geometrically coherent throughout at least this portion of the fault. Stations within a few hundred meters of the fault show head waves only from events to the NW in the creeping section of the SAF. These stations are likely affected by small-scale structural complexities associated with the fault, such as damage zones, multiple fault branches or step-over. The average values of the velocity contrast, estimated from the arrival time moveouts between the head and direct P waves at the different stations, fall in the range of ~3-8%.
T51C-0681
Rapid Temporal Changes of Fault Zone Site Response Associated With Strong Ground Motion
We systematically analyze temporal changes in fault zone (FZ) site response along the Karadere-Düzce branch of the North Anatolian fault that ruptured during the 1999 \.{I}zmit and Düzce earthquake sequences. The study involves primarily comparisons of strong motion seismic data recorded by station VO inside the Karadere fault and station FP ~300 m away from the fault starting 8 days before and ending 72 days after the D¨¹zce mainshock. The spectral ratio between stations VO and FP is computed from the averaged spectra for the two horizontal components, and is used as a measure for FZ site response. The peak spectral ratio increases 80-150% and the peak frequency drops 20-40% at the time of the Düzce mainshock, and is followed by near-complete recovery with time scale of ~1 day. The observed temporal changes of FZ site response can be explained as reduction of seismic velocities by opening of pre-existing cracks inside the FZ due to strong ground motion, followed by logarithmic recovery. Our observations suggest nonlinear behavior of the fault zone material under strong ground motion of nearby major earthquakes. We also apply this method to the weak motion records generated by the 36 repeating earthquake clusters identified by Peng and Ben-Zion (2006) during the same period, but no clear temporal changes of peak spectral ratio or peak frequency is observed. This is likely because the first post-Düzce events in the repeating clusters occurred at least a few hours after the Düzce mainshock, when most of the rapid coseismic changes have been recovered.
T51C-0682
Temporal Changes of Shear Wave Velocity and Anisotropy in the Shallow Crust Induced by the 10/22/1999 M6.4, and M6.0, Chia-Yi, Taiwan Earthquake
We analyze temporal changes of seismic velocity and anisotropy in the shallow crust from similar earthquake clusters that are recorded at a 200-m-deep borehole station CHY of the Taiwan Central Weather Bureau Seismic Network. This station is located directly above the hypocenter of the 10/22/1999, M6.4 and M6.0, Chia-Yi, Taiwan earthquakes, and is immediately south of the rupture zone of the 09/20/1999 Mw7.6 Chi-Chi earthquake. We perform waveform cross correlation on three-component seismograms generated by 2571 local earthquakes with straight-line incident angle to station CHY less than 45 degrees. The similar earthquake clusters are identified with inter-hypocentral distance less than 5 km and the medium correlation coefficient (CC) greater than 0.90. The three-component seismograms show clear shear wave splitting (~0.16 s time delay and a near E-W fast polarization direction), and surface-reflected downgoing shear waves ~1 s after the direct upgoing waves [Liu et al. 2004, 2005]. We apply a 1-Hz high-pass filter, and compute the autocorrelation function for each horizontal component in a time window 2 s before and 3 s after the handpicked S arrival. The secondary peak in the autocorrelation function is used as a measure of the time delay between the upgoing and downgoing shear waves for the horizontal components of each event. The time delays for both E-W and N-S components measured from similar earthquake clusters increase by ~3 percent at the time of the Chia-Yi mainshock, and followed by a logarithmic recovery. However, the difference of time delay between the N-S and E-W components, which represents the shear-wave anisotropy in the top 200 m of the crust, remains essentially unchanged. We propose that opening of microcracks in the shallow crust with no preferred orientation due to the strong shaking of the Chia-Yi mainshock, followed by gradual closure, would explain the temoral changes of the seismic velocity for both horizontal components, but virtually no observable changes in shear wave anisotropy. We also use different median CC values to identify similar clusters, and separate earthquakes into different groups based on their hypocentral locations. The obtained results are esentially the same, suggesting that the observed temporal changes are not sensitive to the raypath between station CHY and the epicenter of each event, but are mostly controlled by variation in material property in the top 200 m of the crust.
T51C-0683
Seismic imaging of scatterer migration using waveform data of repeating earthquakes
Time-lapse seismic imaging (4D) is a new and rapidly-evolving technology. It has attracted wide attention in recent years, because accurate imaging of the evolving subsurface structure has significant applications in resource exploration and environmental monitoring. To apply this technology to monitor the time-varying stress field associated with earthquakes, one needs highly-repeatable powerful sources that can propagate through tens of kilometers to generate similar waveform data that allows for investigating temporal changes in the crustal velocity structure at seismogenic depth. One natural source is repeating earthquakes, which are believed to occur at nearly the same location with the same source mechanism. As such, they produce virtually identical seismograms at a given station. Yet, there are subtle differences in these records that can result, in principle, from either temporal changes in the medium, or from slight differences in source parameters, or both, assuming the signal to noise ratio is high. Approaches such as using differential seismograms or making differential seismic images potentially suffer from the systematic bias introduced by changes in source parameters. For example, variations in source location can be significant when natural sources are used since changes in the medium are typically very small. In order to minimize this bias, we adopt the so-called coda wave interferometry technique. To quantify the difference between seismograms in a repeating earthquake cluster, we compute the cross correlation between the first seismogram and each subsequent seismogram within a moving time window. The lag time τ(t) is obtained when the maximum cross correlation, Cm(t), is reached, and a decorrelation index D(t) is defined as 1- Cm(t). We find that temporal changes in source location, background velocity and the scattered wave field have very different influences on the two functions, and can thus be separated on this basis. Uniform changes in the background velocity, for example, results in a monotonic increase or decrease of lag time as a function of elapsed time on the seismogram, and consequently these changes can be estimated from the slope of lag time function τ(t). Our finite difference synthetic simulations demonstrated that these various influences are distinguishable by migrating the decorrelation index D(t). We are able to image localized changes in scattering field from waveform data that substantial changes in location and the background velocity are presented. Our technique of imaging scatterer migration thus can be applied to broad regions where relatively loosely defined clusters are available.
T51C-0684
Fault Zone Structure at Seismogenic Depth from Seismic Guided Waves
The low-velocity damage zone of major faults can act as a seismic waveguide. The fault-zone guided waves provide a potential method to measure fault-zone properties in situ at depth. Recently, there has been debate over the depth extent of the observed seismic waveguide and whether fault properties can be constrained at seismogenic depth. To answer those questions, we generated elastic finite-difference synthetic seismograms for a uniform-thickness fault-zone model that includes a realistic increase in seismic velocity with depth inside and outside the fault. Two primary results have been discovered in this model: 1) Earthquakes both inside and outside of the fault zone generate strong guided waves within the near-surface part of the depth-varying fault zone, regardless of whether the waveguide is terminated at shallow depth or continues beyond the depth of the earthquakes. This result differs from previous synthetic studies for a homogeneous fault, where earthquakes off the fault do not generate guided waves. 2) Guided-wave velocity dispersion changes with depth. Near-surface fault structure traps/guides waves at low frequencies that are not efficiently trapped (travel as body waves) at greater depth. Deep fault structure can only be derived from analyses of data at higher frequencies than the guided waves that dominate at the surface. In order to determine deep fault structures, we measured the frequency-dependent group arrival times at two receivers closely spaced along the propagation path to calculate the local dispersion curve. The results closely matched the synthetic fault-zone properties between the two receivers. Similarly, a pair of closely spaced earthquakes and a single receiver can be used to derive the fault- zone dispersion between the earthquakes. This method was used to correctly reproduce the synthetic fault-zone structure at seismogenic depth. We will apply this method to existing earthquake data from the EarthScope- SAFOD borehole seismic station to attempt to constrain deep San Andreas Fault structure.
T51C-0685
San Jacinto Fault Zone Structure from Body Waveforms of Local Earthquakes
We developed a new method to determine the fine structure of fault zone (FZ) usi ng direct and FZ-reflected waves from local earthquakes. Characteristic patterns of P and S-wave arrival times across a FZ allow us to determine its boundaries. Other FZ parameters (velocity drops, strike, and dip) are determined by a least-squares inversion of the arrival times. Effects of uncertainties of event locations are also considered. This method greatly reduces the trade-off between the FZ width and velocities. In this study, we used high-frequency body waves recorded by a temporary array deployed in 1999 across the San Jacinto fault. Synthetic seismograms are computed by generalized ray theory based on the newly developed technique. Preliminary results show that the low-velocity fault zone has a width around 350 m and a 50% reduction in P and S-wave velocities compared to the host rock.
T51C-0686
Major Asperities Along The Izmit Rupture Of The North Anatolian Fault Zone Obtained From Analysis Of Aftershocks Of The August, 17th 1999 Izmit Earthquake
We investigated spatial and temporal clustering of 10066 aftershocks along the Izmit Mw=7.4 August, 17th 1999 earthquake (EQ) rupture zone with a total length of 140 km. In the time window from August, 24th 1999 to the day the Duzce Mw=7.2 EQ occurred (November, 12th 1999), we identified three spatial clusters, namely Izmit (IC), Sakarya (SC) and Karadere-Duzce (KDC). The cumulative number of EQ versus time (N(t)) significantly decreases in two of the clusters (SC and KDC) at a critical time (tc= September 29th, 1999), 45 days before the Duzce EQ occurred. We interpret this to represent seismic quiescence in aftershock activity and are able to quantify it by the change in slope of the cumulative number (dN (t)/dt) from 63 EQ/day before to 42 EQ/day after tc. Restricting our aftershock catalogue to highly precise located events with horizontal and vertical errors < 5 km, we calculated b-values (Gutenberg-Richter law) and p-values (Omori law) for 7348 events. Mean values of the aftershock sequence are determined to be b=0.75±0.01 and p=0.71±0.02, respectively. The lowest b-value (0.6) was calculated in the Izmit hypocentral region; the highest b-value (1.0) was observed east of Sapanca Lake, and at the eastern and western rim of KDC. High b-values at Sapanca lake correlate with maximum surface displacement of 5.2 m. Highest p-value (1.2) was identified again in the Izmit hypocentral area, and in the SC. Low p-values (0.6) were calculated from aftershock sequence east of Sapanca Lake and in the KDC. A pronounced (50 per cent) drop in b-value (from 0.9 to 0.45) is observed in the KDC. At SC the drop occurs 11 days earlier compared to KDC and is slightly less pronounced (0.8 to 0.6). We discriminated two zones in the depth variation of b-value. At shallower depth (1-10 km), b-values are below 0.6 (brittle crust), below 10 km the b-values monotonically increase to a value of 1.2 at a depth of 14 km, i.e. in the depth range of brittle-ductile transition where also the hypocenters of the Izmit and Duzce EQ were located. Key words: Fault asperities, Gutenberg-Richter law, Izmit earthquake, North Anatolian Fault Zone, Omori law.
T51C-0687 INVITED
Structure and mechanical properties of faults from InSAR observations of coseismic deformation due to the 1999 Izmit earthquake (Turkey)
We study the structure and mechanical properties of faults in the North Anatolian Fault system by observing near- fault deformation induced by the 1999 Mw7.4 Izmit earthquake (Turkey). We use interferometeric Synthetic Aperture Radar (InSAR) and Global Positioning System (GPS) observations to analyze the coseismic surface deformation in the near field of the Izmit rupture. The overall observed coseismic deformation is consistent with deformation predicted by a dislocation model assuming a uniform elastic crust. Previous InSAR studies revealed small-scale changes in the radar range across the nearby faults of the North Anatolian fault system. We demonstrate that these anomalous range changes are consistent with an elastic response of compliant fault zones to the stress perturbation induced by the Izmit earthquake. We examine the spatial variations and mechanical properties of fault zones around the Mudurnu Valley and Iznik faults using three-dimensional finite element models. In these models, we include compliant fault zones having various geometries and elastic properties, and apply stress changes deduced from a kinematic slip model of the Izmit earthquake. The best- fitting models suggest that the inferred fault zones have a characteristic width of a few kilometers, depth in excess of 10 km, and reductions in the effective shear modulus of about a factor of 3 compared to the surrounding rocks. The characteristic width of the best-fitting fault zone models is consistent with field observations along the North Anatolian Fault system [Ambraseys, 1970]. Our results are also in agreement with InSAR observations of small- scale deformation on faults in the Eastern California Shear Zone in response to the 1992 Landers and 1999 Hector Mine earthquakes [Fialko et al., 2002; Fialko, 2004]. The inferred compliant fault zones likely represent intense damage and may be quite commonly associated with large crustal faults.
T51C-0688
The North Anatolian Fault Zone in the broader Istanbul/Marmara region: Monitoring a ‘seismic gap'
The North Anatolian Fault Zone (NAFZ) in Turkey represents an about 1600 km long plate boundary that slips at an average rate of 20-30 mm/yr separating the westward moving Anatolian block in the South from Eurasia in the North. The 1999 M>7 Izmit and Düzce earthquakes represent the latest of a series of westward migrating mainshocks along the NAFZ starting near Erzincan, Eastern Anatolia, in 1939. The western termination of the Izmit rupture is located in the eastern Marmara Sea near the Prince Islands in direct vicinity to the Istanbul Metropolitan region with its 14 million inhabitants. There, five seismic events with M6+ occurred since 1509. The area offshore Istanbul below the Sea of Marmara today represents the eastern end of a ~140 km long seismic gap that is believed being capable of generating two M7.1 earthquakes within the next decades or could even rupture in a large single event. To monitor the microseismic activity at a presumed ‘seismic gap' along the NAFZ south of Istanbul at low- magnitude detection threshold we deployed an amphibious seismic network surrounding the Prince Islands for a period of two months. In addition for long-term monitoring of the area a permanent seismic array (PIRES) was installed on the two outermost Prince Islands consisting of two sub-arrays of five stations each. PIRES is located within only 4 km to the main branch of the NAFZ allowing to decrease the detection threshold for microseismic activity in this area by two orders of magnitude and thus to resolve the present seismotectonic setting with unprecedented detail. First results indicate that substantial microseismicity is observed along the fault trace in distinct areas off the islands at different branches of the NAFZ allowing to image the fault structure at depth and to contribute to the discussion whether this segment of the NAFZ is creeping or locked.
T51C-0689
Monitoring Seismic Velocity Variations Across the Fault System of the Gulf of Corinth Using Coda Wave Interferometry
Studying fault and fracture in-situ properties is important for gaining a better understanding of many geological and geophysical phenomena. In this work we focus on the temporal variations of fracture stiffness across an active fault system. The analysis of the temporal variations in the seismic velocity across faults can be used to estimate the in-situ stress variations. Seismic velocity of propagation depends on the fault stiffness, which is a function of stress. Traditional methods for inferring velocity variations are not sufficiently accurate because these changes are likely to be too small to be observed, a fraction of one percent. We use the Coda Wave Interferometry (CWI), technique to measure fractional velocity changes using repeating Earthquake sources or multiplets. This technique allows us to estimate with high precision, velocity changes between repeating earthquakes. CWI was applied to the CRLNET data close to the town of Aigion on the southern shore of the Gulf of Corinth. This area is the most seismically active place in Europe and is possibly the fastest extensional continental rift in the world hence is a suitable natural laboratory for studying active fault systems. Initial results show a 0.2% decrease in the seismic velocity of wave propagation across the fault system coincident or immediately after a local 4.2 magnitude event. This velocity variation indicates a probable decrease in the fracture stiffness, possibly associated with a local unclamping of the dominant fault system. In this study we also try to limit the geographical location of the change, by precise relocation of the repeating events used for the estimation of the velocity variations detected with CWI.
T51C-0690
Surface Deformation During a Magmatic Intrusion: the Example of the Dabba'hu Rift Crisis of 2005-2006 (Afar, Ethiopia)
In September 2005, a magmato-tectonic episode initiated in Western Afar (Ethiopia) when a swarm of moderate magnitude earthquakes (M<5.6) was recorded for several days. A small eruption also occurred on the eastern flank of Dabba'hu, a large silicic volcano located at the northern extremity of the Dabba'hu rift segment. The terrain is exceptionally favorable for InSAR imagery and surface fault mapping, making this event a rare opportunity to study how surface faulting and dike injection were mechanically coupled during the rifting event. Based on the combination of InSAR images shot on both ascending and descending tracks, and the correlation of SAR amplitude images, we deduced the vertical motions inside the rift during the main intrusion event. Together with the correlation of SPOT optical images, the horizontal component of opening can be retrieved. We provide evidence for an average of 6 m of opening across the 40 km northern Dabba'hu segment, during the September crisis. The inner floor of the rift subsided by ~ 2m, while the shoulders were uplifted by ~ 2m. A deficit of opening was observed in the southern segment during the main crisis. However, a second intrusive event occurred in mid-2006, leading to the further opening of this 20 km segment by an additional ~ 2m. A series of interferograms covering the post-crises periods show that significant motion also occurred between the crises and after the second crisis. Using a comparison between pre-crisis aerial photographs and post-crisis high-resolution Quickbird images, combined with SAR coherence images, we are able to map the structures that were reactivated during the crisis, and show extensive evidence of newly exposed fractures in recent basalts. The motion on a large number of en echelon faults and fissures could be observed with much greater detail than during the main rifting event. Using a DEM of the area, generated using SPOT images, the relation between faulting and rift morphology is addressed. Concentric subsidence and/or uplift occurred at various stages of the crisis on distinct volcanic edifices, pointing to a complex scenario for the possible connection between shallow and deep magmatic chambers. The estimated extension rate of 15 mm/year across the plate boundary [Vigny et al., 2006] yields a recurrence time of the order of 500 years for events of this magnitude. Surprisingly, despite the large volume of magma intruded during the September 2005 event (~ 15 km2), no basalt flows were observed.
T51C-0691
A STUDY OF ELECTRICAL STRUCTURES OF SHANCHIAO FAULT IN TAIWAN USING AUDIO-FREQUENCY MAGNETOTELLURIC (AMT) METHOD
The Shanchiao normal fault is located in the western edge of Taipei basin in an N-E to S-W direction. Since the fault crosses through the Tertiary basement of Taipei basin, it is classified as an active fault. The overburden of the fault is sediments with a thickness around few tenth meters to several hundred meters. No detailed studies related to the Shanchiao fault in the western side of Taipei Basin are reported. In addition, there are no outcrops which have been found on the surface. This part of fault seems to be a potential source of disaster for the development of western Taipei basin. The audio-frequency magnetotelluric (AMT) method is a technique used to find the vertical resistivity distribution of formation and to characterize a fault structure through the ground surface based measurement. Based on the geological investigation and lithogic information from wells, the AMT data from six soundings at Wugu site, nine soundings at XinZhuang site and eight sounding at GuanDu site were collected on a NE-SW profile, approximately perpendicular to the prospective strike of the Shanchiao fault. AMT data were then inverted for two- dimension resistivity models (sections). The features of all resistivity sections are similar; an apparent drop in resistivity was observed at the position correlates to the western edge of Taipei basin. The predicted location of Shanchiao fault matches was verified by the lithologic sections of boreholes nearby. It indicates that the Shanchiao normal fault may associate with the subsidence of Taipei basin. The basement is clearly detected as a geoelectrical unit having resistivity less than 250 . It has a trend of increasing its depth toward S-E. The uplift of layers in the east of resistivity sections may affect by the XinZhuang thrust fault from the east. As with each site, the calculated resistivity may affect by cultural interference. However, the AMT survey still successfully delineates the positions and features of the Shanchiao fault and western edge of Taipei basin. Keywords¡GCSAMT, RIP, Shanchiao fault
T51C-0692
Magnetotelluric Imaging Across the North Anatolian Fault, Erzincan Basin, Turkey
The Erzincan pull-apart basin is located on the North Anatolian Fault (NAF) which is one of the most active faults in the world. In the last century, two destructive earthquakes have occurred in this basin which have surface wave magnitudes of Ms=8.2 in 1939 and Ms=6.8 in 1992. Recent geodynamic models have shown that fluids may play an important role in the earthquake-cycle. Specifically, it has been suggested that fluids in a fault zone can raise the pore pressure, lowering the shear stress needed for rupture. These fluids, if interconnected in a fault zone will significantly decrease the resistivity of the subsurface and be detected with magnetotellurics (MT). Thus, MT can be used as an effective tool for determining the fluid content in an active fault. MT is a passive electromagnetic technique which images subsurface resistivity by recording naturally varying EM fields at the surface of the Earth. For this purpose, broad-band magnetotelluric data were acquired in the Erzincan Basin in 2005. Data were recorded at 24 stations on two parallel profiles that crossed the basin with a spacing 1-2 km. The time-series data were processed using statistically robust algorithms and remotely referenced to remove incoherent magnetic noise. All data were processed using tensor decomposition, which showed a geoelectric strike direction of N70°W. This is parallel to the strike of the NAF as expected. Two-dimensional regularized inversion models were then created and these models imaged the conductive sedimentary rocks in the Erzincan Basin to a depth of 5km. This depth is consistent with that inferred from seismic tomography. The models also contain a strong near surface conductor which underlies the surface trace of the NAF. This is similar to features observed in resistivity models of the San Andreas Fault at Parkfield where a near surface conductor in the fault zone was attributed to the presence of fluids in the fault gauge.
T51C-0693
Magmatics within the Dead Sea transform valley (insights from magnetic data)
The DST – largest continental fissure- appears to be the single shear zone consisting of several en echelon right- stepping faults, with the total sinistral displacement of about 105 km. Gravity data show obviously what the DST locates along the interface between western block with high density oceanic (transition) crust and eastern relatively light density continental crust. Assuming the interface as weakness zone the question is: Can the weakness zone to serve as magma chamber during last ten million years? To delineate and study the magmatic bodies the magnetic survey provides excellent opportunity. The water and sedimentary strata are ‘transparent' and the magnetic anomalies are caused meanly by local magmatic features. Generally the regional magnetic maps are maps of basic magmatics – outcropped and concealed. New high-resolution aeromagnetic surveys covered the southern Arava valley and the western Dead Sea area provided the valuable data to better understanding of the tectonic pattern and distribution of magmatics within the transform valley. The main findings: - presence of the rift-parallel magnetic lineaments corresponding to the active trace of the DST; - lineament system extends NNE as an almost continuous trace over the study areas and shows the same general azimuth of the magnetic lineament and of the segmented fault system as derived from the gravity, geology and topography; - the absence of any magnetic anomalies crossing the Arava Valley; all the local magnetic anomalies in the eastern and western parts of the valley appear to be completely separated; only a lineament appears to separate the local magnetic anomalies; - strong magnetic anomalies suggest hitherto unknown subsurface magmatics. Quantitative interpretation of the high accuracy detailed magnetic measurements allowed to localize the magmatic bodies and to estimate reliable theirs depth. Comparing the results with seismic reflection and geological data the belonging of magmatics located within the transform valley to corresponding magmatic event can be defined as follow: - Wide range of the prerift magmatic events (Precambrian and Phanerozoic intrusives); - Magmatic events synchronic but doesn't associated with the DST (e.g. young basalt flows); - Unrecognized magmatic bodies synchronic to the DST and associated probably with tectonic development of the DST during movement along the main shear system. Such magmatics may mark the main deep penetrated faults. Last case will be thoroughly studied in future. Magmatics are integral part of the Dead Sea transform structure and this knowledge has to be used in exploration of groundwater, oil, earthquake hazard assessment and infrastructure projects, such as the Dead Sea-Red Sea Canal This study was supported by the Earth Sciences Administration of the Ministry of National Infrastructures of Israel and the DESIRE project of the GeoForschungZentrum (Potsdam) and Kiel university.
T51C-0694
Geophysical characterization of the Muckleshoot Basin, northwestern Washington State
We characterize the Muckleshoot Basin and examine its relationship with the Tacoma and White River fault systems to the west and east, respectively. The Muckleshoot Basin, located east of Tacoma, Washington, was initially defined by seismic tomography data as a 20x25 km basin that extends to 7-9 km depth. A new analysis of gravity and vintage industry seismic-reflection data suggests that the Muckleshoot basin extends east to the base of the Cascades and is segmented into two sub-basins that exhibit post-Neogene deformation. A new 25-km high-resolution seismic-reflection transect indicates that the basin extends to within 10 km of Puget Sound. Along this seismic profile, we observe approximately 1 km of Plio-Pleistocene deposits above steeply dipping Tertiary strata. Isostatic gravity and the new seismic data also suggest that an uplift tied to the active Tacoma fault beneath Puget Sound extends east to the Muckleshoot Basin. The saddle that bisects the Muckleshoot Basin is also along strike with the White River fault that crosses the Cascades Range. We tentatively suggest that the Tacoma fault system extends to the White River fault through the Muckleshoot Basin, a total distance that exceeds 100 km. A fault of this length, if substantiated, would pose a significantly higher seismic hazard within the densely populated Puget Lowland area than current models suggest.
T51C-0695
Role of Erosion on Closely-Spaced Fault Scarps: High-resolution Laser Scanning Data and Scarp Diffusion Modelling of the Rex Hills Flower Structure, Nevada
Flower structures are often associated with strike-slip faults. Their subsurface geometry is well known but additional information about their geometry and kinematics is also contained in their surface expression, especially in surfaces exhibiting closely-spaced fault scarps. However, detailed surface descriptions of flower structures are rare. Here, we present for the first time topographic profiles extracted from high-resolution digital elevation model (DEM) data using a ground-based laser scanner (Riegl company) to examine the surface expression of a flower structure in great detail. Our study site, the Rex Hills flower structure (40 m high, ~1 km long), is located on the transpressional left-bend between the Pahrump and Amargosa segments of the dextral Stateline fault system (Guest et al., 2007, GSAB). The southern Rex Hills slope exhibits three reverse fault- scarp generations related to three reverse fault branches, respectively. The basal scarp is the youngest and most continuous one. It exhibits five fault segments with an approximately constant displacement along the scarp. The upper two fault-scarp generations are less continuous, and older. The topography of the southern Rex Hills slope also is characterized by alternating valleys and ridges (each ~100 m long) extending perpendicular to the main ridge crest. Our analysis shows that fault scarp morphology varies laterally along this slope. Scarps exposed on ridge crests are typically more numerous (up to 4-5 scarps) and smaller (on average 5 m high), whereas adjacent valleys often exhibit single large (>10 m high) scarps. The larger valley-scarps did not result from the offset across a single fault branch, but resulted from merging of the three scarps mentioned above by enhanced scarp degradation. The preservation potential of small, individual scarps is therefore better on ridge crests relative to adjacent valleys. Assuming a known age of 2 kyrs (Menges et al., 2003, AGU), we examined the spatial variability of the diffusivity parameter (κ in m2/ka) and the degradation coefficient (τ in m2, τ = 2κt) by performing linear diffusion analysis. On a plot of slope-angle vs. scarp-height, κ ranges from 2 - 4.5 m2/ka and τ from 4 - 9 m2 for the ridge crests, whereas for the valleys κ ranges from 4.5 - 9 m2/ka and τ from 9 - 18 m2. This implies that for the past 2 kyrs erosion is twice as high in valleys than on ridges. This result is consistent with the better scarp preservation on ridge crests discussed above. In order to refine the age of the basal scarp, we selected two basal scarp profiles from a ridge and a valley location, respectively, with similar scarp height but different slope angles. The graphically determined τ of the steeper ridge-scarp is smaller than the τ of the lower angle valley- scarp. Combined with a reasonable κ (1.1 m2/ka for Nevada), we obtained scarp diffusion ages of ~4.3 kyrs for the valley scarp and ~1.4 kyrs for the ridge scarp. The ridge scarp age agrees with the independent estimate of 2 kyrs, whereas the valley scarp estimate is three times too old. Therefore, ridge crest profiles provide more accurate information about closely-spaced fault branches and hence should be used to assess the evolution of flower structures or, more generally, any fault system with closely-spaced scarps. Combining high-resolution surface data analysis with subsurface data should significantly improve the understanding of flower structures in the future.
T51C-0696
Source Parameters Estimate During the 1997 Umbria Marche (Central Italy) Seismic Sequence: Fluid Diffusion or Static Stress Interaction?
The 1997 Umbria-Marche seismic sequence consisted of six moderate magnitude earthquakes (5 < Mw < 6) and thousands of aftershocks that, in 40 days, activated a ~45 km long, NW-trending, fault system. It is one of the best studied seismic sequences in Italy and most of the seismic events have been relocated using a double difference technique (Chiaraluce et al., 2003). The migration of seismicity towards the south west and the impossibility of explaining all the main events with a direct Coulomb static stress redistribution (Nostro et al. 2005), together with evidence of the presence of fluids in the area, encouraged a different approach invoking pore pressure relaxation as a possible triggering mechanism (Antonioli et al. 2005, Miller et al. 2004). In this work, we analyze seismic activity during a 2 month period of temporary seismic network recording. We compute the radiated Seismic Energy, Seismic moment, Magnitude, Static Stress Drop and Apparent Stress for any single event of the sequence both directly from the waveforms and by derivation of the analytical relationships. By examining the measured seismic parameters it is possible to infer differences between events with purely tectonic mechanisms and events with potentially different triggering mechanisms. The lower stress drop estimated south of the sequence could be an indicator of fluid diffusion, especially after comparison with travel- time and attenuation tomography.
T51C-0697
Tectonic underplating in the Nankai Trough off the Kii peninsula: Insight from Kumano 3D seismic reflection data
To reveal the evolution of the plate boundary decollement of the Nankai accretionary prism off the Kii peninsula, we use three-dimensional seismic reflection data across the accretionary prism. The characterization of the decollement in the front of the mega splay fault will support the broader goals of the Integrated Ocean Drilling Program to drill through the mega splay fault. On the seismic profiles, we observe several discontinuous reflections above the decollement seaward of the mega splay fault. Imbricated thrusts lie above the discontinuous reflections and these thrusts seem to sole down into the discontinuous reflections, but do not extend down to the basal decollement. Furthermore, the accretionary prism is thickened on the seaward side of the splay fault and the thickening cannot be explained only by increasing the thrust angle. From these observations, we interpret that the accretionary wedge between the splay fault and the trench has thickened by sediment underplating. Furthermore when we focus on the decollement geometry, we observe underplating structures. The underplating seems to be induced by elevation of the underlying oceanic crust. On the seismic profiles, we observe several thrusts within the oceanic crust that are causing the crustal thickening. These thrusts within the crust should be still active because they coincide with the location of the 2004 earthquake that occurred off Kii peninsula. Furthermore, although the polarity of the decollement in most parts of the survey area is positive (the same as the seafloor and the top of the oceanic crust), the polarity above the thrusts within the crust is negative or ambiguous. We interpret this as an indication that seismic velocity of underthrust sequence below the decollement is lower than that of the overlying accreted sequence. The fluid-supply through the thrusts (fractures) within the crust may reduce seismic velocity below the decollement. Therefore, the thrusts within the oceanic crust are important in the evolution of the decollement in our survey area.
T51C-0698
Fine fault structure in intraplate earthquake region estimated by DD method with waveform correlation analysis, using nationwide seismic network Hi-net, Japan
Determining highly resolved hypocenter distribution in intraplate earthquake region is very important for estimation of fine fault structure. The dense permanent nationwide seismic observation network (NIED Hi-net) with an average spacing of 20-30km has been developed all over the Japan region after the disasters of 1995 Kobe earthquake. The detection capability and accuracy of hypocentral location was quite improved by Hi-net. On the other hand, a double-difference (DD) earthquake location algorithm (Waldhauser and Ellsworth, 2000) was developed to improve the accuracy of relative hypocenter location. In this method, further improvement of location is expected by using waveform cross-correlation method. We applied DD algorithm using waveform cross- correlation method to the Hi-net data, to estimate the fine fault structure and confirm the availability of Hi-net routine seismic observation network. We relocated the hypocenters of aftershock sequences of two moderate crustal earthquakes, and of seismicity around the active fault system which displays one of the largest slip rates in the Japanese islands. We used the absolute travel time measurements and differential travel time measurements by waveform cross-correlation analysis. In this analysis, the velocity waveform of 0.75 sec containing the P or S wave onset was used, applying 3-20 Hz band-pass filter. Moderate crustal earthquakes occurred in the northern part of Mie Prefecture on 15 April, 2007 (Mw=5.0) and in the southwest part of Shizuoka Prefecture on 1 June, 2007 (Mw=4.4), each of which followed by aftershock sequences. In both regions, the hypocenter of aftershocks determined by routine process are scattered and we could not find any trend of main shock fault. However, the locations by DD method reveal the fine structure of seismicity. Most of events around the main shock hypocenter show a planer distribution, which is consistent with the focal mechanism of main shock. This planer distribution of hypocenters probably reflects the structure of main shock fault. In the northern part of Mie Prefecture, the locations by DD method also reveal another cluster 1 km off from the main shock hypocenter, which is parallel to the planer distribution of hypocenters around the main shock. This lineament might suggest a subsidiary fault system parallel to the fault of main shock. We relocated the hypocenter location around the Itoigawa-Shizuoka Tectonic Line active fault system (ISTL), which has large slip rate. In the northern part of ISTL, we found the planer distribution of hypocenters under the ISTL. In the central and south part of ISTL, on the other hand, the hypocenter distributions form a cloud surrounding the ISTL, which suggest that most of events did not occur on the ISTL in these areas. We concluded that the fine fault structure concerned with the moderate crustal earthquakes and the active fault system could be resolved from the hypocentral distribution by Hi-net, applying DD method. Aftershock sequences of moderate earthquakes define the planer distribution of hypocenters, which suggest that most of aftershocks concentrate on the fault of main shock.
T51C-0699
The Subsurface Images of the Chiuchiungkeng Fault in the Fold-and-Thrust Belt Area of Western Taiwan
The deformation front at the foothills area of western Taiwan is very active and showing significant crustal deformation. From the historical reports and observed seismic data, we may expect threaten of large earthquakes in that area. The Chiuchiungkeng fault is one of the young active faults situated in that fold-and-thrust belt area. Results from geological and geomorphic surveys show that the fault is an active thrust, and the fault slips should be more than 0.55mm per year. However, the subsurface features of the fault and its total length are still not clear. In this paper, we use shallow seismic reflection method to detect the Chiuchiungkeng fault and other characteristics. The results show that the Chiuchiungkeng fault strikes from north to south extends from Hobaoshan in Yunlin County, across the Niouchou River and to the Shanchulun area in Chiayi County for 18 km, and is centered at the Meishan area. According to the seismic images and the borehole data, we may know that the fault at the both ends have been truncated by another strike-slip faults. The fault dip is about 25-degree to the east, and has a sharp and narrowed fault zone at the center area. The fault dips became smaller toward the both ends, but with different styles. Close to the north end of the Chiuchiungkeng fault, where the fault zone is very wide, more than 1 km; however, at the southern section of the fault, the fault zone is narrower and the strata on both of the hanging and footwall have the similar dip angles. The total length of the Chiuchiungkeng fault is about 18 km, and the potential earthquake magnitude related to the fault is about 6.5. In 1906, an earthquake of magnitude 7.1 shocked the Meishan area and caused very severe damages. The earthquake was thought related to the left-lateral strike-slip Meishan fault only. Combining the historical map of surface ruptures, regional subsurface structures and the high-resolution shallow seismic reflection images, we may suspect that the Chiuchiungkeng fault played a more important role in that earthquake as well.
T51C-0700
Connectivity test across deformation bands: results from field pumping experiments in Tucano Basin, NE Brazil
Sandstones of the Ilhas Group in Tucano Basin, NE Brazil, commonly present outcropping fault associated deformation bands. In one of these outcrops, with approximately 1km length and 15m thick three wells were drilled and in situ permeability measurements were done across the outcrop surface. The connectivity tests were performed with one well at one side of the fault zone and the two others at the other side. Two tests were performed and at each test - whilst one well was pumped downdraws were monitored at the other two remaining wells. The permeability profiles revealed a huge variation of permeability of up to four orders of magnitude. Nonetheless, the well pumping tests revealed a moderate connectivity across the deformation band since the observed stationary downdraw at monitoring well on the opposite side of the fault zone was a considerable fraction (~1/8) of the downdraw observed in the pumped well. These results indicate that there is 3D connectivity in the field scale across the deformation band. http://www.dfte.ufrn.br
T51C-0701
A Discrete Particle Numerical Investigation of Rock Fracture Compliance in a Multi-fracture System
Fractures can greatly influence the elastic properties of a rock matrix and seismic wave propagation within it. Quantifying fracture properties from seismic data has the potential to assist in the understanding of many geophysical applications including the study of active fault zones, ground stability and underground waste storage. In this work we validate the time delay and attenuation characteristics of a single planar fracture modeled with a Discrete Particle Scheme (DPS) against analytical solutions. This validation opens the way to numerically investigate the effect of multiple fractures on wave propagation. We find that time delay and attenuation in sparsely fractured systems agree with analytical expressions of mean fracture compliance, while they depart significantly in densely fractured systems. Hence direct measurements of attenuation or time delays are generally not suitable to quantify fracture compliance in multi-fracture systems. However, we show that if the medium is sufficiently scattering the indirect measurement of time delays with Coda Wave Interferometry does allow for an accurate quantification of mean fracture compliance.
T51C-0702
Applying Pattern Informatics to Southern California to Image Fault Systems in Three Dimensions
The Pattern Informatics (PI) method (Tiampo et al., 2002) quantifies spatio-temporal variations in the seismicity of a seismogenic region, yielding a long term forecast for the locations of future earthquakes in the form of a 2-D "hotspot" map. There is an inherent link between changes in seismicity measured by the PI method and changes in the stress on a fault system. As stress accumulates on a fault, the probability of an earthquake occurring on that fault increases resulting in hotspots. Because stress tends to build preferentially on fault, the hotspots should also be preferentially located on fault, outlining its structure. It is for this reason that we propose that the PI method is not only valid for forecasting earthquakes, but can also be used to image heterogeneous stress regions on a fault surface. Here we explore applications of the 3-D PI method to multiple regions in southern California and the potential sources of error. Ultimately, our results show that adding a third dimension to the PI method is both a valid way to forecast future earthquakes at depth, as well as image the underlying fault system.
T51C-0703
Fault Zone Architecture and Deformation Processes Within Exhumed Evaporitic Rocks in The Upper Crust
Evaporitic rocks are generally considered as the ductile decollement horizon for many thrust and fold belts. Recently in the Northern Apennines geophysical data have identified the Triassic Evaporites, TE, as the source region of the major earthquakes of the area (M=6). To characterize fault zone architecture and deformation processes within the TE (anhydrites and dolomites), we have studied exhumed evaporite-bearing normal faults. The geometry and architecture of the fault zones are strongly controlled by both the inherited synorogenic ductile fabric and the different rheological behaviour of the anhydrite (brittle/ductile) and the dolomite (brittle). The interaction between these factors leads to the development of rather complex and heterogeneous fault zone geometry, where complexities within the fault core and the damage zone arise from the juxtaposition of differently deformed domains. Fault rock assemblages are very similar to those of fault zones developed within the brittle regime, and are characterized by fault breccia, gouge and cataclasites of different grain size. In particular, mesoscale damage zones consisting of fractured rocks are wide and well developed within dolomite rocks, whilst they are almost absent or weakly developed within the anhydrite rocks. The fault core of major fault zones (up to 100m displacement) appears zoned with a wider portion of ductile deformation overprinted by an inner and thinner zone of extremely localized brittle deformation. Fault parallel layers of cataclastic dolomite develop within the fault zone and they seem to localize most of the shear strain within the inner brittle zone along well developed principal slip surfaces. We present a mechanical fault evolution model to explain the development of the observed geometry as a function of the interplay between the different rheological behaviour of dolomite and anhydrite and transient fluid pressure increase and release during the fault activity. Our work is relevant to future studies aimed to understand fault controlled reseservoir properties and interpret seismogenic processes associated with other fault zones developed within evaporitic rocks around the world.
T51C-0704
Differential Geometry of Fault Surfaces and Glacial Beds: Associated Deformation Patterns
We measure glacial sliding surfaces and the striations on them and argue that these surfaces at the base of hard-bedded mountain glaciers can be compared to geologic fault surfaces albeit the obvious rheological differences. The reorientation of glacial striations around topographic anomalies provides quantifiable information about the otherwise inaccessible conditions. We use a ground-based LiDAR to measure the first high-precision orientations of these streaks at a field locality near Tenaya Lake, CA. We find that they are resolvably deflected around topographic highs. For example, bumps of about 0.3 m deflect the striations by up to 10°. Deviations from planar geometries can be quantified using the principles of differential geometry. These methods calculate the principal normal curvatures at each point on the surface and admit classification of one of eight basic shapes. Two of these shapes (synform and antiform) exhibit a zero principal normal curvature in one direction, one (plane) has no nonzero curvature, and another (perfect saddle) requires equal but opposite principal curvatures. None of these shapes are found in raw field data, so a curvature threshold and/or spectral filtering are applied to remove curvatures not distinguishable from zero. Common sliding surface shapes that can be described by this classification are domes, basins, antiformal, synformal, and perfect saddles. These non- developable shapes induce strains in the adjacent rock and ice masses as relative particle motions on either side of the fault or sliding surface are expected to be non-zero. These strains are not induced by surfaces with at most one non-zero principal normal curvature if the sliding direction is perpendicular to the direction of non-zero principal normal curvature. The additional strains may lead to nonparallel striations and off-fault deformation. Our goal is to relate the magnitudes of the two principle curvatures to changes in sliding direction. We show how the geometric shapes of the glacial sliding surfaces compare and contrast with fault surface shapes and how these shapes relate to glacial striations and slickenlines preserved on faults. Our field examples reveal the significance of geometric complexities to the mechanics of faulting and glacial sliding, and elucidate the interplay of surface geometry and slip behavior.