T54A-01
Calico Fault Structure Determined Using Traveltime Data from Seismicity and Explosions
A dense array of 100 seismometers deployed near the Calico Fault, located in the eastern California shear zone, recorded explosions and seismicity between June and December, 2006. Seismic velocity reductions associated with the fault damage zone are mapped in detail along and across strike; in addition, the depth extent of the velocity reduction is investigated. Three shots exploded in and out of the fault zone reveal seismic velocities in the very shallow crust. The first shot was located within the array, a second in the fault approximately 6 km from the array, and a third outside of the fault, also about 6 km away. In addition, we use local, regional, and teleseismic earthquake body waves to measure the width of the low-velocity zone and investigate variations in that width along strike and with depth. We use the SCEC velocity model as a starting point and then perturb the velocity model in a finite zone along the Calico fault to find a match to the observed travel times. Results suggest a 1 to 2 km wide low-velocity zone, in good agreement with the InSAR-derived compliant zone. Our measurements are in accord with InSAR observations showing that the Calico fault suffered twice the strain of the surrounding bedrock during the Hector Mine and Landers earthquakes, confirming a zone of reduced shear modulus around the fault. These findings indicate that faults can affect rock properties at significant distances from the primary fault slip surfaces, a result with implications for the portion of energy expended during rupture to drive the cracking and yielding of rock.
T54A-02
Imaging the Crust of the Los Angeles Region: Forward and Inverse Velocity Models and Reflectivity Migrated from Automatic Line Drawings of Shot Gathers
In the Los Angeles Region Seismic Experiment (LARSE), conducted in the 1990's, active and passive seismic data were collected along two transects extending from offshore Los Angeles to the Mojave Desert. We discuss here imaging of the onshore, active-source segments of these transects. One segment, Line 1 (L1), extended through the Puente Hills (the vicinity of the 1987 Whittier Narrows earthquake) and San Gabriel Mts. The other segment, Line 2 (L2), extended through the Santa Monica and Santa Susana Mts (the vicinity of the 1971 San Fernando and 1994 Northridge earthquakes) and Central Transverse Ranges. We have performed forward modeling of first and secondary arrival times and also inverse modeling of first arrival times for comparison. We have also migrated automatic line drawings from shot gathers to avoid the problem of statics in stacking. In the latter method, we focus on features that can be identified in more than one migrated shot gather, 'stacking' these features by superposing the migrated lines. In addition, we focus on features within ~2 s of the first arrivals that have a slope of opposite sign from the first arrivals, in order to (a) mute first arrivals and reverberations and (b) enhance steep, fault-related reflections. For both L1 and L2, velocity features agree well between forward and inverse models (mean difference ~0 km/s; std ~0.3 km/s). On L1, steeply dipping, tabular low-velocity zones (LVZ's) correlate with the Sierra Madre fault system (SMF; ~65 deg dip) and the San Andreas fault system (SAF; steep dip). The LVZ associated with the SAF (Δv ~1 km/s) spans the ~10-km region between the Punchbowl Fault (old branch of the SAF), south of the SAF, and the Llano Fault, a 'flower-structure' reverse fault north of the SAF. Reflectivity includes a gently dipping zone that extends southward and upward from a point at 20-km depth and 7 km north of the surface trace of the SAF. This zone of coherent reflections coincides with the top of the 'San Gabriel Mts bright reflective zone' of Ryberg and Fuis (1998; seen in a trace-envelope stack) and appears to have as many as 3 southward branches: (1) a steep upward branch connecting with steep reflectivity in the hanging wall of the SMF, (2) a gently dipping branch that projects toward the hypocenter of the Whittier Narrows earthquake, and (3) a subhorizontal branch. Reflectivity is largely absent southward of the San Gabriel Mts owing to cultural noise. On L2, a narrow (4-km), shallow LVZ correlates with the SAF. In addition there are low-velocity depressions along the downward projections of the San Gabriel Fault and the Northridge Hills Fault (NHF). The depression along the NHF extends to 10-km depth. Strong reflectivity is concentrated in two oppositely dipping zones that meet ~vertically beneath the SAF in the depth range, 20-30 km. The southern zone extends upward nearly to the 13-km-deep hypocenter of the 1971 San Fernando earthquake, as demonstrated in Fuis et al. (2003, from a low- fold stack), with one upward splay toward the active San Gabriel Fault. Based on relocated aftershocks of the 1971 earthquake, Fuis et al. (2003) interpret a decollement connecting the SAF, 1971 hypocenter, and the surface trace of the NHF. Flower-structure reflective zones are seen north of the SAF.
T54A-03
Constructing a Model of a Dipping Southern San Andreas Fault, Southern California
The San Andreas Fault (SAF) in southern California is predominantly non-vertical, based on potential field observations, deep seismic imaging, and earthquake focal-mechanisms, aftershocks, and microseismicity. Dip values of the southern SAF vary systematically along its length. Between the Carrizo Plain and Lake Hughes (~180 km distance), dips are consistently to the southwest and range between 55° and 80°; between Lake Hughes and Wrightwood (~70 km), dips are essentially vertical; between Wrightwood and the Salton Sea (~220 km), dips are consistently to the northeast and vary between 37° and 75°. In the structurally complex San Gorgonio Pass area, the dip of one of the strands of the SAF in the uppermost 1-2 km may be as low as 10°. In all, 13 sites that are unequally-spaced along the southern SAF provide well-constrained observations of fault dip. Uncertainties of the fault dip observations are less than ±10°, and in some cases less than ±5°, although the different types of observations have different sensitivities to age and depth of features indicating dip. Gravity and magnetic models are sensitive to the structures that result from millions of years of strike-slip motion at shallow and intermediate crustal levels (<10-15 km); earthquake studies reflect motions at seismogenic depths since the start of the seismic catalogs (<50 years); deep seismic imaging is sensitive to the integrated history throughout the crust. Notwithstanding the different sources of observations, a self-consistent model of the fault surface may be constructed by interpolating between points with well-defined dips and assembling with three-dimensional modeling software. The three-dimensional form of this fault surface is crudely reminiscent of a propeller. The steepest sections (>75°) of the southern SAF correspond to locations where the fault trace is straightest. Where the SAF trace bends, dips are generally <75°. The variation of dip and strike of the SAF implies that strike-slip motion along the fault surface must be accompanied by changes in topography and that partial accommodation of long-term plate motion by adjacent faults may be important. Calculation of shaking potential for scenario major earthquakes on the southern SAF and calculation of geodetic deformation from buried slip on the SAF may be inaccurate unless the non-vertical dip of the fault is taken into account.
T54A-04
In-Situ Resolution of Internal Fault Zone Properties: Structure, Rock Mechanics and Rupture of the Pretorius Fault, South Africa (NELSAM Project)
We analyze the reactivation mechanism of the Pretorius fault, TauTona mine, South Africa. The analysis is based on our mapping of the fault-zone structure at a depth of 3.6 km, observations of its rupture zone during an M2.2 earthquake, and mechanical testing of the fault and host rocks in the laboratory. The slip localization mechanisms of the earthquake are modeled with the finite element method. The Pretorius fault is a 10 km long, 25-30 m wide, Archean fault with dextral displacement of 200 m and vertical displacement of 30-60 m. The fault zone consists of a network of tens of anastomosing fault segments, some of which include massive, well-cemented cataclasite. We mapped the rupture zone of the M2.2 earthquake of December 12, 2004, exposed in mining tunnels for at least 25 m horizontally and 5 m vertically. The rupture reactivated four main, quasi-planar, crosscutting segments within the complex network of the ancient Pretorius fault. The slip during this event generated 1 to 5 zones (each 0.5-1.0 mm thick) of fresh, fine-grained rock powder. This powder was located predominantly along the contacts of the quartzitic host rock and the ancient massive cataclasite and indicates slip localization during the rupture. Rock mechanics experiments were conducted on samples of the fault-rock (quartzitic cataclasite) and host rock (quartzite), collected from boreholes drilled across the fault zone. The elastic properties of the host quartzite (E = 81 GPa, v = 0.17) are similar to those of the cataclasite (E = 71 GPa, v = 0.15), but the former is twice as strong as the later (uniaxial strength of 200 MPa vs. 100 MPa). On the other hand, the host quartzite is severely damaged and shows significant strain hardening with inelastic deformation starting at ~25% of the total axial strain (brittle- plastic behavior). The cataclasite is undamaged, with only minor inelastic deformation occurring at 85% of the axial strain (brittle-elastic behavior). We developed a 2D finite element model in which a simplified fault (elliptical inclusion), composed of cataclasite- like rock, is embedded within a quartzite-like medium, using the material properties obtained from the lab tests. The model shows an abrupt increase of the shear stress at the contact between the inclusion and the host rock. A similar trend in the plastic shear strain suggests that the shear stress gradient is a result of the plastic behavior of the host rock. We propose that the mechanical contrast between the plasticity of the damaged host quartzite and the brittle cataclasite results in a steep shear stress gradient across the contact, leading to the slip localization. We will include the observed rupture complexity into our model and explore its effect on slip localization mechanisms within network fault-zones.
T54A-05 INVITED
The Structure and Fluid Flow Properties of Fault Zones: Inferences from Combined Field and Laboratory Studies
Quantifying the fluid flow properties of fault zones is important for understanding fault zone mechanics, predicting the distribution of fault-hosted economic deposits, and recovering hydrocarbons from structurally complex reservoirs. The fluid flow properties depend on the distribution of deformation within fault zones and the permeability of the individual fault components. The fault core can be a single, narrow zone of fault gouge, or consist of multiple strands with variably fractured lenses of country rock contained within them. Either structure will inhibit fluid flow across the fault zone, as typical laboratory values for the permeability of fault gouges in this direction is in the range 10-18 to 10-22 m2. The multiple strand fault core provides the greatest opportunities for fluid entrapment. Although fault gouge exhibits significant permeability anisotropy (up to 3 orders of magnitude), fluid flow rates parallel to the fault within the fault core will still be low. Field, seismological and geophysical observations suggest that faults act as significant fluid conduits, and hence the zone of fractured rock surrounding the fault core (the damage zone) must act as a high permeability pathway. The damage zone consists of microscopic and macroscopic fracturing that decreases in intensity with distance from the fault core. In order to quantify the contribution of microscopic damage to the fluid flow properties, we measured the porosity and permeability evolution of initially low porosity crystalline rocks under simulated crustal conditions during progressive deformation to failure. The data show permeability enhancement from initial values of ~10- 21 to ~10-17 immediately prior to failure. However, these values and the size of damage zones typically seen in the field are not sufficient to explain inferred fluid flow rates and imply that the macroscopic fracture network must play a significant role in fluid transport.
T54A-06
Experimental Measurements of Permeability Evolution During Brittle Deformation of Crystalline Rocks and Implications for Fluid Flow in Fault Zones
Detailed experimental studies of the development of permeability of crustal rock during deformation are essential in helping to understand fault mechanics and constrain larger scale models that predict bulk fluid flow within the crust. The strength, permeability and pore fluid volume evolution of initially intact crystalline rock (Westerly granite and Cerro Cristales granodiorite) under increasing differential load leading to macroscopic failure has been measured in a triaxial deformation apparatus. Experiments were run under pore water pressures of 50 MPa and varying effective pressures from 10 to 50 MPa. Permeability is seen to increase by up to and over two orders of magnitude prior to macroscopic failure, from 3.5 x 10-21 to 9 x 10-19 m2 with the greatest increase seen at lowest effective pressures. Post-failure permeability is shown to be over 3 orders of magnitude higher than initial intact permeabilities, as high as 4 x 10-18 m2, and approaches lower the limit of measurements of in situ bulk crustal permeabilities. Increasing amplitude cyclic loading tests show permeability-stress hysteresis, with high permeabilities maintained as differential stress is reduced. The largest permeability increases are seen between 90-99% of the failure stress. Under hydrothermal conditions without further loading, it is suggested that much of this permeability can be recovered, and pre-macroscopic failure fracture damage may heal relatively faster than post-failure macroscopic fractures. Pre-failure permeabilities are nearly seven to nine orders of magnitude lower than that predicted by some high pressure diffusive models suggesting that microfracture matrix flow cannot dominate, and agrees with inferences that bulk fluid flow and dilatancy must be dominated by larger scale structures, such as macrofractures. It is suggested that the permeability of a highly stressed fault tip process zone in low-permeability crystalline rocks could increase by more than 2 orders of magnitude, while stress drops related to fracture propagation close damage zone cracks, whereupon some permeability is maintained due to hysteresis from permanent microfracture damage.
T54A-07
Fluid Dynamic Evidence for Extremely Low Viscosity Coseismic Fault Fluids
We combine geological observations of fault rock textures with fluid mechanics to constrain the mechanics of a fault zone during a subduction earthquake. We analyze buoyant intrusive features in a fault rock that formed at 12- 14 km depth in a large-scale thrust fault embedded in a paleo-accretionary prism in Kodiak Island, AK. The fault rock can been interpreted as either a pseudotachylyte or fluidized ultracataclasite. The intrusive structures provide new, direct evidence on the coseismic rheology of the fault. The asymmetric buoyant intrusions are most readily understood as Rayleigh-Taylor instabilities with an unusually short wavelengths relative to the thickness of the layer. The geometry requires a moderately high Reynolds number flow (Re~1-10) in order to produce the observed wavelength to thickness ratio. The resulting rise velocity under these conditions is ~40 cm/s. Since the shear strain in the layer is over order 1 and the deformation is continuous, the rise velocity must be comparable to the horizontal shear velocity during emplacement. Thus, the geometry alone requires that the fault rocks were intruded coseismically. Furthermore, the Reynolds number constraint combined with the computed rise velocity provides a maximum bound on the viscosity of the fluid during emplacement. The coseismic fault fluid at this locality must have had a viscosity of \ll 10 Pa-s. This viscosity constraint is compatible with the viscosity of the silicate melt of the observed composition at 1300-1400°, which is consistent with the temperature constraints imposed by the absence of plagioclase survivor grains. In summary, both the fluid dynamical and geological evidence points to an extraordinarily low viscosity fluid in the fault zone during rupture and hence extremely low local stress in the fault during an earthquake.
T54A-08
Fluid Overpressure and Earthquakes Triggering in the Natural Laboratory of the Northern Apennines: Integration of Field and Laboratory Data
The integration of seismic reflection profiles with well-located earthquakes show that the mainshocks of the 1997-1998 Colfiorito seismic sequence (Central Italy) nucleated at a depth of ~6 km within the Triassic Evaporites (TE, anhydrites and dolostones), where CO2 at near lithostatic pressure has been encountered in two deep boreholes (4 km). In order to investigate the deformation processes operating at depth in the source region of the Colfiorito earthquakes we have characterized: 1) fault zone structure by studying exhumed outcrops of the TE: 2) rheology and permeability by performing triaxial loading tests on borehole samples of anhydrites at room temperature, 100 MPa confining pressure (Pc), and range of pore fluid pressures (Pf). Permeability and porosity development was continuously measured throughout the deformation experiments. The architecture of large fault zones within the TE is given by a distinct fault core, where most of the shear strain has been accommodated, surrounded by a geometrically complex and heterogeneous damage zone. Brittle deformation within the fault core is extremely localized along principal slip surfaces associated with dolomite rich cataclasite seams, running parallel to the fault zone. The damage zone is characterized by adjacent zones of heavily fractured rocks (dolostones) and foliated rocks displaying little fracturing (anhydrites). Static permeability measurements on anhydrite samples show increasing values of permeability for decreasing values of Pe, (k = 10E-20 - 10E-22 m2). During single cycle loading tests the permeability values immediately prior to failure are about three orders of magnitude higher than the initial values. The field data suggests that during the seismic cycle, the permeability of the dolostones, within the damage zone, is likely to be high and controlled by mesoscale fracture patterns. Conversely, the permeability of the anhydrites, due to the absence of mesoscale fracture patterns within Ca-sulphates layers, may be potentially as low as the values measured in the lab experiments (k = 10E-17 - 10E-22 m2). This suggests that fluid overpressure can be maintained in this lithology, within the damage zone, as far as the co-seismic period. Our observations and results can be applied to explain the seismicity of the Northern Apennines and other regions where fluids overpressures play a key role in triggering fault instability and earthquakes.