T11A-0333
High velocity frictional properties of clay-bearing fault gouges : experiments and modelling
The present study focuses on the experimental measurement and theoretical understanding of the high-velocity frictional properties of the MTL (Median Tectonic Line at the Tsukide outcrop, Japan) fault gouge and on its possible consequences on the large scale behaviour of clay bearing fault gouges during co-seismic slip. Several experiments were conducted on the MTL fault gouge using a rotary-shear apparatus at high velocities (up to 1.03 m.s-1), low normal stresses (up to 1.4 MPa), for displacements up to 60 m. During these experiments, we observed systematically a slip-weakening behaviour, i.e. a dramatic decrease in the coefficient of friction, from a value of ~1.2 to a value of ~0.3. In addition, the slip-weakening distance Dc also decreased with increasing normal stress. Optical and SEM observations show the presence of a very thin slipping zone, with important grainsize reduction. TEM analysis using a FIB section shows that this zone is partially amorpheous, with only a few remaining oxides grains of the order of a few nanometers in grainsize. XRD analysis show that the initial kaolinite content disapeared after shearing, probably due to co-seismic dehydration of kaoliniti into metakaolinite (an amorpheous mineral) and bounded water exsolution. For comparison, additional experiments were conducted under similar conditions on pure kaolinite gouge samples. Monitoring humidity, we did observe a release of water vapor during these experiments. Further, we present a numerical model in which thermal pressurization of pore fluid is coupled to dehydration reactions. Our modellling shows that thermal dehydration of hydrous clay minerals may be a non-negligible phenomenon during co-seismic slip. In well documented fault gouges such as San Andreas, Aegion, or the MTL where the water content can reach up to 10% in weight, pore pressure raise due to frictional exsolution of bonded water may : 1) be at least comparable to the thermal pressurization term, 2) help overcome normal stress and induce damage by hydrofracturation of the fault walls, 3) limit the temperature rise and thus prevent the formation of pseudotachylites.
T11A-0334
Clinker formation in basaltic lava flows
Basaltic lava flows are classified according their surface morphology. They can be either aa, displaying a rough clinkery surface or pahoehoe, displaying a smooth clinkerless surface. These two surface types differ also in their emplacement and rheology, and can be differentiated in a shear-strain rate vs. apparent viscosity diagram (Hon & al., 2003). To understand clinker formation, one way is to see how a pahoehoe lava converts to an aa through shear-viscosity changes. Two possibilities occur: 1) the viscosity can increase (e.g. by levee formation) and clinkers will be formed by torque on the flow edges, or 2) the shear-strain will increase (lava influx increasing, topographic obstacles, slope change) and clinker will be formed by crust-breakage. These two clinker formatting processes are called magmatic fragmentation. Clinker will be formed on the summit and to the edges of the flow and they will appear at the base of its according the caterpillar motion usually associated with flows. However, in the Chaîne des Puys (French), basal clinker appears without top clinker (i.e. a pahoehoe lava flow with basal clinker) and thus another explanation is needed to explain them. Clinker samples were collected in different emplacement contexts and different part of flows. The SEM analysis of these samples and comparisons with ash samples from the literature show classical magmatic fragmentation textures (stepped fractures, non- synchronic fractures) in three aa lava flows. However, in one pahoehoe flow there are typical phreatomagmatic textures (blocky shapes, adhering fine particles). There are also shearing structures, such as microfaults in an intermediate flow. Thus, there are at least three different ways to form clinker: 1) classically by fragmentation at the flow base and the edges; 2) by phreatomagmatism at the base flow; 3) by shearing at the flow base and the edges. Basal shearing structures include fault gauges and welded clasts, indicating possible shear heating at the base of some flows. Heated water in breccia may also aid to lubricate the base, aiding flow advancement by slippage.
T11A-0335
The Development of Vein and Pressure Solution Arrays in a Compressive Orogenic Belt, County Clare, Ireland.
The Carboniferous Ross Sandstone, exposed on the Loop Head Peninsula, County Clare, Ireland is part of the deep-water deposits of the Shannon Basin fill that was subject to compressive stresses at the end of the Carboniferous during the Variscan Orogeny. This deformation produced broad regional east-west trending folds and also tight low-amplitude folds that are concentrated around thrust faults. Near these faults, orthogonal sets of pressure solutions and veins form a pervasive structural assemblage within the sandstone. Abutting relationships of the veins against the pressure solution seams indicate contemporaneous formation of the two structure sets. The east-west orientation of the pressure solution seams are sub-parallel to local thrust fault traces and fold axes, and reflects the regional Variscan structural fabric. A subsequent stress or material rotation during the Variscan Orogeny (or perhaps a major second stage of deformation either in late phase of the orogeny or following the orogeny) has lead to left-lateral shear of the pressure solution seams evidenced by pressure solution splays and pull-aparts between their sheared segments, and right-lateral shear on the veins evidenced by splay crack features. With increased shear, extensive splay cracks and pull-apart networks form damage zones through which strike-slip fault systems develop. Ultimately, this study documents the formation of damage zones through the formation and evolution of veins and pressure solution seams and the role of these fundamental structures in palaeo- and present-day fluid flow.
T11A-0336
Grain Fragmentation: Dynamic Implications
Granular flow has long been recognised as the dominant process in many geological phenomena, from bedload sediment transport through debris flows to rock avalanches and fault motion. The high confining stresses occurring in large-scale geological phenomena cause intact grains in a grain-flow to fragment as they shear. Grain flow with fragmentation thus affects the dynamics of many large-scale geophysical phenomena. Brittle grain fragmentation has been much studied in the context of shear band formation and fault motion, and recent 3D micromechanical simulations have succeeded in representing the kinematics of comminution, leading to successful reproduction of catacalastically deformed grains, shear localisation and fracture sets. In these simulations, however, the dynamics of fragmentation were investigated only in the context of acoustic emissions and induced seismicity. The kinetic energy arising from fragmentation was artificially and arbitrarily damped, decreasing the energy available for further deformation and obscuring the dynamic effects. We demonstrate that grain fragmentation under sufficiently high confining stress causes instantaneous local dispersive pressures in the GPa range. At sufficiently high strain rate the spatial concentration of fragmentation events can generate stresses sufficient to dominate the dynamics of grain flow; this leads to low intergranular effective stresses and correspondingly low frictional resistance at normal (Byerlee) values of friction coefficient. We also show that the spatial concentration of fragmentation is self-regulating to a calculable value. During coseismic rupture, high-velocity sliding generates strain rates high enough to produce dynamically-fragmenting grain flow in a localised principal slip zone; kinetic energy release within this zone quantitatively explains the observed low frictional resistance to shearing motion on the San Andreas and other large faults. It also satisfactorily explains the motion of the Heart Mountain and Waikaremoana blockslides and the Socompa volcanic debris avalanche.
T11A-0337
No Energy Loss to Surface Energy in Rock Fragmentation
Rock fragmentation is of industrial importance, and has been researched for some centuries, even before modern energy concepts were formalised. It takes energy to break rock, and all methods for doing it use the same process — a mass is deformed by application of mechanical energy until it breaks. In breaking, new surfaces are created. All material surfaces exhibit an ability to influence the behaviour of matter in their immediate surroundings; it is described in physical chemistry by the concept of surface energy. In fragmentation, the total chemical surface energy of a grain mass increases; but its internal chemical energy decreases by the same amount, because chemical energy is associated with molecules, irrespective of their location. Confusion between the surface energy, and the mechanical energy (termed fracture surface energy) used to create new surfaces through breakage, has led to the notion that fragmentation is an energy sink, consuming energy that might otherwise remain available to affect motion. The notion is incorrect; the increase in chemical surface energy does not arise through an increase in the total energy of the mass; it arises solely because more of the mass is exposed at surfaces. When mass remains unchanged, and surface area is increased, chemical surface energy always increases, but there is no over-all increase in chemical energy associated with the mass. The increase in grain surface area is not where energy is lost in grain fragmentation; the energy loss in fragmentation arises from conventional friction between grains and with imperfect elasticity associated with the deformation needed to cause fragmentation. The mechanical potential energy stored in elastic deformation is all initially converted to kinetic energy when grains break, none goes to chemical surface energy. This kinetic energy is available to further grain motion and significantly affects the dynamics of high-stress granular flows.
T11A-0338
Gouge Powder from Earthquakes Rupture-zones and Laboratory Rupture Experiments: Sub- microscopic Observations and Particle size Distribution
We examine grain features and estimate the particle size distribution (PSD) in gouge collected from three sources: rupture zones of two recent earthquakes in South African mines (m3.7 event in 1997 and m2.2 event in 2004); unstable faulting experiments of quartzite cylinders; and the San Andreas fault-zone at Tejon Pass, California. Studies in nanotechnology indicate that PSD measurements of fine powders are susceptible to major errors due to aggregation and agglomeration of the fine particles that cause a systematic bias toward coarser PSD. Thus, our central analytical objective is to determine the true grain size of the gouge that forms during an earthquake and we have employed several techniques. In the TEM (Jeol-2010F) we use bright-field and dark-field modes for magnifications smaller than 200,000, and FFT-filtering mode for high-resolution (HREM) magnifications of 200,000-500,000. We also test several methods to disperse the agglomerates (with ethanol, silanes and toluene) and measure the PDS in the Dynamic Light Scattering method (using the Zeta particle analyzer ZetaPALS-90Plus of Brookhaven Instruments). The grains in the TEM analyses of eight gouge samples display similar shapes: aspect ratios range up to 1:3 with small amounts of very elongated grains, and most grains are angular to very angular while grains of the rock mechanics experiment are the most angular. The samples display wide ranges of grain sizes (5 nm to 5 microns), however, quantitative PSD cannot be determined as most (if not all) grains appear in clusters and aggregates that are hardly separable in the bright-field mode. We thus use the dark-field and FFT-filter methods to map the internal structure of tens of grains ranging in size from a few nanometers to about 2 microns. With a few exceptions, all examined grains are composites of 3 to more than 15 (and probably more) secondary grains that are as small as a few nanometers in size. The PSD of these samples is measured with the Zeta analyzer with resolution range of 1 nanometer to 3 micron. For half of the samples we use ‘as is' gouge powder and for the other half we use the gouge fraction that passes 63 micron sieve. Several dispersion methods are used. The observed PSD is multimodal, and with the exception of one sample, the grain sizes are 0.1-1.0 micron. The TEM direct observations of the agglomerate internal structure indicate that the true grain size of gouge powder is in the sub-micron range as confirmed by the Zeta analyzer. We currently attempt to quantify the abundance of particles in the 10-100 nanometer range in gouge powder by testing nanotechnology dispersion techniques. We discuss the significant implications of such fine grains to earthquake energy balance and to slip weakening mechanisms during earthquakes. This study was supported by a SCEC 2007 grant and by NSF Continental Dynamic grant 0409605.
T11A-0340
Sublinear Displacement-Length Scaling of Igneous Dikes in Basalt
Displacement-length ( D-L) scaling relations of faults are well understood, with an exponent of n = 1, whereas dilatant structures such as veins and dikes may scale as n = 0.5. We investigate the scaling relations from a new dataset obtained from 39 igneous dikes from northwestern Ethiopia that intrude 30 Ma Precambrian mylonites and gneisses capped by a 100-m-thick layer of altered basalt. Dike lengths were measured using Spot and ASTER satellite images with measurements of dike thicknesses (maximum opening displacements) obtained in the field. The data demonstrate sublinear displacement-length scaling, with Dmax = 0.088 L0.48 consistent with those of other opening-mode fractures available in the literature, including basaltic dikes intruding shales at Shiprock, New Mexico and two populations of veins. All four datasets are characterized by a power-law slope of about n = 0.5, indicating propagation under conditions of constant rock fracture toughness, rather than constant driving stress, as in the case of faults. The intercept for the population can be solved for fracture toughness to obtain KIc = 303-909 MPa m1/2 for the basalt. Several factors may contribute to the large fracture toughness value. First, the country rock is not intact basalt, but a jointed and altered basaltic rock mass. The small tensile strength of the basaltic rock mass was likely to have been exceeded within a volume surrounding the tips of propagating igneous dikes, thus increasing the intrinsic resistance to propagation. Second, given field evidence for syn- or post-emplacement shearing along several dike margins, the orientations of the dikes were likely misaligned relative to the remote least compressive principal stress direction. Propagation of the dikes was probably controlled by the mixed-mode fracture toughness of the rock mass, which is invariably larger than that associated with pure opening of a dike. Third, fracture toughness of basalt increases nonlinearly with temperature, making dike propagation substantially more difficult for country-rock temperatures adjacent to the dike above °sim700 °C. Similarity in ages of dikes and intruded basalt suggest that the temperature effect may also have contributed to some degree to an increased resistance to dike propagation. http://mines.unr.edu/geo-eng/schultz
T11A-0341
Dynamics of grains induced by the competition of coarsening and fragmentation
We propose a dynamical model for the grain evolution in systems dominated by grain coarsening and fragmentation, or rotation recrystallization. Specifically, we formulate a generic rate equation for the size distribution of grains. As an example of the application of our equation, we employ data on the evolution of crystal assemblies from the North Greenland Icecore Project (NorthGRIP). The size distributions of ice crystals in the upper 880m of the ice core covers a time-span of approximately 5300 years and provide an excellent record of the grain evolution. Slowly, the size distributions evolve towards a universal curve and a steady state is reached where the fragmentation balances the coarsening. The model provides an excellent one-parameter fit of the steady state ice crystal distribution in terms of a Bessel function. The full dynamics is described by two parameters controlling the rate of coarsening and fragmentation, respectively.
T11A-0342
Crack Coalescence in Molded Gypsum and Carrara Marble
This research investigates the fracturing and coalescence behavior in prismatic laboratory-molded gypsum and Carrara marble specimens, which consist of either one or two pre-existing open flaws, under uniaxial compression. The tests are monitored by a high speed video system with a frame rate up to 24,000 frames/second. It allows one to precisely observe the cracking mechanisms, in particular if shear or tensile fracturing takes place. Seven crack types and nine crack coalescence categories are identified. The flaw inclination angle, the ligament length and the bridging angle between two flaws have different extents of influence on the coalescence patterns. For coplanar flaws, as the flaw inclination angle increases, there is a general trend of variation from shear coalescence to tensile coalescence. For stepped flaws, as the bridging angle changes from negative to small positive, and further up to large positive values, the coalescence generally progresses from categories of no coalescence, indirect coalescence to direct coalescence. For direct coalescence, it generally progresses from shear, mixed shear-tensile to tensile as the bridging angle increases. Some differences in fracturing and coalescence processes are observed in gypsum and marble, particularly the crack initiation in marble is preceded by the development of macroscopic white patches, but not in gypsum. Scanning Electron Microprobe (SEM) study reveals that the white patches consist of zones of microcracks (process zones).
T11A-0343
Analysis of Fragmentation Mechanisms in Fault-Core and Damage Zone: Application of Fractal Fragmentation Theory
The study of physical mechanisms triggering fragmentation processes in fault zones is of paramount importance for understanding the mechanics of faulting. In the last twenty years a plethora of works indicated that a suitable characterization of fragmentation processes can be attained by studying the size distribution of rock fragments around fault zones. In particular, fractal geometry techniques have been proven to unravel information that would be otherwise inaccessible by conventional particle size analysis. Although these studies allowed to understand some of the basic mechanisms acting on a rock formation and producing the observed fractal fragment size distributions, a few works focussed on the continuous characterization of the size distribution from the fault core to the damage zone. The aim of this contribution is to study the evolution of fragmentation processes within cataclastic fault rocks from the fault core into the damage zone by using fractal analysis. From a small displacement (slip <10m), dolomite-bearing fault, a large rock sample (40x20x20 cm) containing the fault core and a portion of the damage zone has been collected. The sample has been cut along both fault orthogonal and fault parallel cross sections. For each section the size distribution of fragments has been obtained by image analysis. Results indicate that all distributions are fractal as they can be approximated by a power-law relationship whose exponent (D) is the fractal dimension of fragmentation. An unusual result from our analysis is the switch of D values by passing from the damage zone (D=2.84±0.03) to the fault core (D=2.61±0.03); this variation is consistently observed for both parallel and orthogonal cross sections. Results from our analysis are interpreted in the light of two fractal fragmentation models: the so-called "pillar-of- strength" and "comminution" model. The first model is typically invoked to account for the catastrophic fragmentation occurring during the initiation of faulting, i.e. when the fault plane develops, and generates fragment size distributions with D=2.84. The second model, is commonly claimed to account for the grinding of fragments occurring along the fault core as the result of localised deformation, and produces fragment size distributions with D=2.58. The difference in the D values from the fault core to the damage zone suggests that the latter experienced catastrophic events, D=2.84±0.03, whilst the fault core was affected by a progressive grinding of fragments generating a fragment size distributions with D=2.61±0.03.
T11A-0344
Fragmentation, Gouge Production and Surface Roughness Evolution on Experimentally Simulated Faults
To investigate the physical processes operating in active fault zones, we conduct analogue laboratory experiments where we track the morphological, mechanical and thermal evolution of an interface during slip. Our laboratory friction experiments consist of a halite (NaCl) slider held under constant normal load that is dragged across a coarse sandpaper substrate. This set-up is a surrogate for a fault surface, where brittle and plastic deformation mechanisms operate simultaneously during sliding. Surface morphology evolution, frictional resistance and heat emission are recorded with accumulated slip. After each experiment, we characterize the roughness developed on slid surfaces, to nanometer resolution, using white light interferometry. We directly observe the formation of deformation features, such as slip parallel linear striations, as well as deformation products or gouge. The striations are often associated with marginal ridges of positive relief suggesting sideways transport of gouge products in a snow-plough-like fashion. Deeper striations are commonly bounded by triangular brittle fractures that fragment the salt surface and efficiently generate a breccia or gouge. An abundance of gouge at the sliding interface reduces the shear resistance, demonstrating that accumulated slip may reduce the friction coefficient. The relative importance of these deformation mechanisms may influence gouge production rate, fault surface roughness evolution and mechanical behavior. Finally, our experimental results are scaled to nature by comparing the experimental surfaces to an actual fault surface, whose striated morphology has been characterized (to cm resolution) using a laser scanner. For both natural and experimental faults, it is demonstrated that the stress field is heterogeneous at all scales during the maturation of the interface with accumulated slip.
T11A-0345
Breaking up: Fault zone evolution during shear
To better understand fault zone dynamics we need an improved understanding of the underlying micro- mechanics of the fault evolution process. Our basic data is gleaned from quantitative observations of structural fabrics associated with natural fault systems. However, such datasets generally record the final state of evolution, from which it is often difficult to discern the dynamic micro-scale processes involved, such as preferential fragmentation and strain partitioning, or the macro-mechanical behaviour of a fault. Laboratory experiments give valuable insights into links between microscale processes and macro-mechanical behaviour. Similarly, numerical simulations are very useful tools for visualising dynamic grain-scale interactions not readily visible from nature. Together these tools can help us identify and isolate first order parameters that are relevant for the faulting process. We present recent results from 3D simulations that implement realistic gouge evolution during shear. Our particle based simulation includes breakable elastic bonds between individual particles allowing fracture of aggregate grains that are composed of many bonded particles. With accumulated strain, aggregate grains fragment in different ways, gradually evolve in size and shape to produce a textural signature reminiscent of natural faults. We use a new image analysis tool to characterise grain shape and size distributions from thin sections obtained from natural and experimental fault rocks and are now applying this tool to 3D model results. This approach allows us to build on existing observations and permits closer investigation of dynamic bumping, grinding and fragmentation processes that may be operating in evolving fault zones.
T11A-0346
An Empirical Wear Law for Rocks during High-velocity Fault Motion
An empirical wear law of rocks is examined by using a rotary-shear high-speed frictional testing apparatus, in order to reveal the gouge-generation processes during coseismic sliding of faults. Hollow cylindrical specimens of gabbro, granite and sandstone with inner and outer diameters of 15 and 25 mm, respectively, were slid at velocities of 0.02 to 0.3 m/s and normal stresses of 0.3 to 4.2 MPa under unconfined and dry conditions. Powdered rock (gouge) was continuously produced by abrasive wear of initially bare fault surfaces during sliding. Because the fault was not confined in our experiments, the gouge was extruded from the sliding surfaces, resulting in shortening of axial length of specimen. In this study the wear rate was defined that an axial shortening rate of the specimen was divided by velocity, although it is commonly described as the ratio of thickness of gouge zone to total slip. The experimental results on gabbro and granite indicate that the wear rate increases drastically with velocity and normal stress. The relationship between the wear rate and normal stress can be fit well with an exponential equation. The increase in wear rate must be caused by a subtle thermal cracking owing to the frictional heating on the sliding fault surfaces. When the velocity is of more than 0.08 m/s, the wear rate of sandstone decreases with increasing in normal stress. It might be associated with the formation of a consolidated or sintered layer on sliding surface. The wear rate of gabbro and granite is an order of 10-5 at normal stress of 2 MPa and velocity of 0.2 m/s. If this wear rate is extrapolated to high normal-stress conditions in nature (i.e., 50 MPa), the estimated value is higher by several orders of magnitude than the wear rates of 10-3 to 100 reported from natural fault (e.g., Scholz, 1987). This large gap between laboratory and nature might be associated with the extrusion of gouge from sliding surface in the experiment. We will try to prevent the leak of gouge from sliding surfaces and modify the wear law of rocks during coseismic sliding.
T11A-0347
Pulverized Tejon Lookout Granite: Attempts at Placing Constraints on the Processes
We have described and analyzed pulverized Tejon Lookout granite recovered from several transects of the western segment of the Garlock fault on Tejon Ranch in southern California. Observations and data collected at this location are compared to a sampled transect of the San Andreas fault at Tejon Pass previously studied by Wilson et al. (2005), also exposing the Tejon Lookout granite. The purpose of this study is to characterize the physical and chemical properties of the pervasively pulverized leucocratic rocks at multiple locations and to hopefully place constraints on the processes producing them. To accomplish this we performed particle size analysis with the use of both laser particle analyzer and pipette methodology; major and trace chemistry analyses determined by XRF; clay mineralogy determined by XRD; and we evaluated fabric and texture through the study of thin sections. Recovered samples met the field criteria of pulverization developed by Dor et al., 2006 - that is, the individual 1-2 mm-sized crystals can be recognized in the field but the granite (including quartz and feldspar) can be mashed with ones fingers and exhibits the texture of toothpaste. All samples were analyzed on a Horiba LA930 Laser Particle Analyzer in an attempt to reproduce the earlier results of Wilson et al. (2005) with similar methodology. We also utilized the classic pipette methodology to ensure complete discrimination of particle sizes. Our PSD analysis shows that the dominant particle size falls in the 31-125 micron range, much coarser than previously reported by Wilson et al. (2005), with >90% of the total sample falling in the >31 micron size range. We can reproduce the previously documented results by allowing the samples to circulate for long periods of time at slow circulation speeds in the laser particle size analyzer, during which time the coarse fraction settles out, thereby leaving only the fine fraction for detection. However, subsequent increase in the circulation speed leads to a complete recovery of the original PSD. Our XRF and XRD analyses provide evidence of the lack of major weathering products and their inability to skew the PSD results in a significant way. Dor et al. (2007) and Stillings et al. (2007) document evidence that support theoretical predictions and previous inferences of pulverization occurring in the upper few kilometers, especially along faults of the southern San Andreas system. Geophysical observations of Lewis et al. (2005, 2007) provide evidence that low velocity fault- parallel layers, which are likely made of pulverized or highly damaged material, are dominant in the upper few kilometers of the crust. Their asymmetric position with respect to the slipping zone, in agreement with asymmetric patterns of small scale mapped rock damage (Dor et al., 2006), suggest that pulverized rocks are likely the product of a preferred rupture direction during dynamic slip. Our results combined with the above mentioned works imply that pulverized fault zone rocks at multiple locations are much less damaged than suggested in previous studies.
T11A-0348
Effect of Magma Degassing on Dike Propagation at Mid-Ocean Ridges
Dike propagation from a subaxial magma body at a mid-ocean ridge may initiate either near the tips or at the center of the magma body. Previous research showed that for magma bodies in the shallow crust, tip dikes could propagate to the seafloor, whereas central dikes, or dikes propagating from deep seated magma lenses may not. Previous work neglected the effect of volatile exsolution on dike propagation, however. Volatile exsolution increases magma compressibility and hence the exsolution of volatiles either prior to or during dike propagation may increase the driving force for dike propagation and increase the propagation distance. Expressing the magma compressibility as a function of the volume fraction of gas, we investigated this scenario through a range of parameters related to the magma lens characteristics. Our results show that tip dikes emerging from magma bodies at the base of the crust may reach the upper crustal magma lens provided the volatile content is greater than 3.5%. Central dikes emerging from the shallow magma chamber can reach the seafloor if the volatile content is 1.75% and the average crystal content of the lens is approximately 40-50%.
T11A-0349
Fracture Controlled Dyke Intrusion Patterns And Inferred Stress Conditions In Geo-je Island, Southeast Korea
Tertiary basic dykes systematically intruded Cretaceous sedimentary rocks in Geo-je Island, southeast Korea, where a variety of non-planar intrusive patterns are observed. The characteristics of the intrusive patterns are analyzed and classified, and their associated stress conditions are inferred. Based on this kinematic analysis, the relationship between dyke intrusion patterns and fractures development is interpreted. The predominant orientations of the dykes are NNE-SSW and NNW-SSE, which are also consistent in vein and fracture orientations. It means that the dykes intruded along the pre-existing weakness indicating passive intrusion controlled by pre-existing fractures. This relationship is also supported by the fracture analysis based on cross-cutting relationship, restoration of the dyke intrusion gaps, and inferred deformation history in this area. Therefore, dyke intrusion patterns in this study area are mainly controlled by pre-existing fracture patterns as well as stress condition during magma intrusion. This study might be very useful to interpret local tectonic stress changes and origins. Keywords: dyke intrusion patterns, pre-existing fractures, stress condition, passive intrusion, deformation history
T11A-0350
Fragmentation Wave in Viscoelastic Medium Containing Bubbles and Crystals
We conducted fragmentation experiment using viscoelastic silicone compound with various pressure, vesicularity, crystallinity and permeability to understand the magma fragmentation in an explosive volcanic eruption. We used a vertical shock tube to generate rapid decompression. The specimen was pressurized with nitrogen very slowly so that the pores are filled with the high-pressure gas. Then the membranes separating the high- pressure part from the atmospheric pressure part are artificially ruptured, and the specimen is rapidly decompressed. The fragmentation behavior of the specimen is photographed by a high-speed video camera. The fragmentation wave velocity is measured from the video images. After each experiment, the fragments are taken out of the chamber on top of the shock tube carefully and the structures are observed. We obtained the following results from the observation of the fragmentation speed. The fragmentation speed is in the range of 20-40 m/s. Its dependence on the void fraction is not clear in the present experimental conditions with void fraction ranging from 0.33 to 0.44. It tends to be decreased by existence of crystals and increase of permeability. The larger it is, the larger is the acceleration of the fragments. We have noticed significant cavitation in the viscoelastic compound after rapid decompression. The bubbles are generated homogeneously within the compound. The cavitation occurred regardless of the initial gas-saturation condition of the compound. It was not observed with slow decompression, though the decompression amplitude is the same. The bubble nucleation depending on the decompression rate might be significant also in the volcanic processes with rapid decompression and magma fragmentation.
T11A-0351
Ground Subsidence in the Granada City and Surrounding Area (Spain) using DInSAR Monitoring
Differential SAR interferometry (DInSAR) is a remote sensing technique which has been successfully used, since the final of the eighties, for different applications such as coseismic deformation mapping, volcano deformation monitoring, landslides monitoring and subsidence detection among others. This is an alternative technique to obtain measurements of the surface displacement providing better spatial resolution and comparable accuracy while being less time consuming than conventional surveying methods. However, spatial and temporal decorrelation and atmospheric signal contributions in repeat-pass SAR interferometry often hamper the accurate measurement of surface displacements in SAR interferograms. In 1999, the POLIMI SAR group implemented a different process allowing overcoming these difficulties by interpreting time-series of interferometric phases at coherent point scatterers. This technique was called the Permanent Scatterer techniques, which allow us to measure deformation with accuracies of millimetres per year. Since then, other techniques have been suggested following similar processing lines, such as the Persistent Scatterer Interferometry (PSI) developed at Delft Technical University (The Netherlands). In this study, we apply PSI technique using two time-series of 32 ERS-1/2 and 22 ENVISAT ASAR acquisitions of the Granada Basin, located at the central sector of the Betic Cordillera (southern Spain), covering the period from 1992 to 2005. This is the first time that the PSI technique is applied to derive displacement information in this southern Iberia region. This technique is very useful for the analysis of subsidence in urban areas, where angular structures produce efficient reflectors that dominate the background scattering. However, man-made structures are absent from most of the Earth's surface, such is the case in part of the studied area. After this first data processing, several subsidence areas have been detected in the southern part of Granada city and nearby villages. At the moment, some investigations are being carried out in order to find the relationship between the detected deformation and the tectonic deformation present in the area.
T11A-0352
Flash Heating and Weakening of Crustal Rocks During Coseismic Fault Slip
During fault slip, rocks become heated by the dissipation of friction at transient, microscopic, highly stressed asperity contacts. For sufficiently high slip rates, this 'flash' heating may yield high temperatures and even melting of contacts, causing severe degradation of contact shear stress and dramatic reductions in fault strength. We have reported the results of high-speed friction experiments on a variety of crustal rocks, including quartz rocks, feldspar rocks, granite, gabbro, and calcite marble, for slip rates up to 0.36 m/s and slips of <0.045 m, conditions conducive to extreme heating at asperity contacts but insignificant heating of the entire fault surface. Above ~0.1 m/s, such tests reveal an inverse relationship between friction and slip velocity, and slip weakening distances are on the order of the presumed contact size, in agreement with theory, for all rocks tested except calcite marble. Fits of constitutive relations for flash heating (Rice, 1999, 2006; Beeler et al., 2007) to experimental data yield extrapolated values of the friction coefficient of ~0.2 or less at a seismic slip rate of 1 m/s. Here we report results of high-speed friction experiments on quartz rocks that have larger initial surface roughness, resulting from grinding with #24 grit alumina powder, than smoother quartz rocks tested previously, roughened with #60 and #120 grit alumina. Despite the expectation that flash heating of the rougher samples should be greater - due to larger asperity contacts - experiments on the rougher samples reveal no weakening at slip rates up to 0.36 m/s. The stark contrast in behavior of rough and smooth samples may result from increased interlocking of asperities for the rough samples, resulting in a thicker generated gouge zone. If the gouge layer undergoes distributed shearing, then the contact-scale slip velocity is less than the far-field velocity by a factor N, the number of gouge particles across the gouge layer; equivalently, the weakening velocity for flash weakening is increased by a factor N. Estimation of an apparent weakening velocity for distributed shear of modest gouge thicknesses (<1 mm) using lab-like contact dimensions suggests values of the weakening velocity greater than seismic slip rates, i.e., >1 m/s. Our results emphasize the critical roles of contact size and the degree of slip localization in determining whether dynamic fault weakening due to flash heating occurs during earthquake slip.
T11A-0353
A High-Precision, 3-Dimensional Fracture Characterization Study in the Bandelier Tuff, Pajarito Plateau, Northern New Mexico
For certain proposed new facilities at Los Alamos National Laboratory, located on the Pajarito Plateau of northern New Mexico, completion of seismic hazards studies must precede any construction to ensure appropriate engineering for damaging earthquakes is included in building design. For one such study, a two-phase excavation exposed approximately 4000 sq. m (42,000 sq. ft) of the two uppermost subunits of the 1.22 Ma Tshirege Member of the Bandelier Tuff (Qbt3 and Qbt4) and interbedded pyroclastic surge in excavation walls for detailed geologic investigations. Using traditional paleoseismic trenching techniques in an unconventionally large setting, we obtained detailed data on exposed geologic features, including contacts, fractures, and faults. Geologic mapping at an approximate scale of 1:36 on digital photographs, coupled with high-precision geodetic surveys of identified features, place the data within a geospatial reference frame and provide an unparalleled 3- dimensional composite dataset of localized contact relationships and structure within Qbt3 and Qbt4. The majority of fracture orientations trend east-northeast, with an overwhelming predominance of vertical or near- vertical dips. The pyroclastic surge exhibits radical thickness changes over short distances. Qbt3 and Qbt4 are intensely fractured in places, with the fractures exhibiting variable amounts of vertical continuity and displacement across the pyroclastic surge in particular. Fractures are commonly curvilinear and terminate at the contact with the pyroclastic surge. Minor faults with less than 0.6 m (2 ft) of vertical displacement across the pyroclastic surge are present. Many of these small faults exhibit displacement only on the upper or lower contact of the surge deposit. In several locations, these small faults bound funnel-shaped, fines-depleted zones in Qbt3 with significant overlying shattering in Qbt4. Thus, very little of the observed deformation is apparently tectonic in origin; rather, most deformation probably occurred during cooling and compaction of the thick tuff deposit, and some pronounced deformation is associated with zones of localized paleo-fumarolic activity.
T11A-0354
Amorphous Material Formed by the Mechanochemical Effect in Natural Pseudotachylyte of Crushing Origin: A Case Study of the Iida-Matsukawa Fault, Nagano Prefecture, Central Japan
Glass or amorphous material in pseudotachylyte are interpreted as melt texture; however, naturally occurring amorphous material can form via the rapid cooling of melt, alteration, and mechanochemical effects induced by mechanical energy during the comminution process. Recent experimental studies suggested that amorphization and the formation of gel via shearing and comminution under wet conditions is one of the weakening mechanisms of faults during seismic slip (Goldsby and Tullis, 2002; Di Toro et al., 2004). These studies show that the presence of amorphous material formed by comminution is an essential factor in understanding fault strength and faulting processes during earthquakes; however, the presence of amorphous material formed by comminution has yet been reported from natural fault rocks. In the present study, the crush-origin pseudotachylytes from the Iida-Matsukawa Fault (Iida pseudotachylytes) are described down to the nanometer scale using transmission electron microscopy (TEM), scanning electron microscopy (SEM), analyses of chemical compositions, and mercury intrusion porosimetry. The matrix of the Iida pseudotachylyte chiefly consists of nanoscale particles and elongated submicron fragments with biotite compositions. Amorphous materials of several tens of nanometers in size are scattered randomly among the submicron crystalline fragments within the matrix, despite the absence of melt textures. Lattice fringe images reveal that the amorphous phase coexists with lattice distortion in deformed biotite fragments that are several hundreds of nanometers in size. These submicrostructures indicates that the amorphous material formed by mechanical stress during the comminution process. The fragments in the Iida pseudotachylyte have a low degree of roundness (< 0.4), suggesting that the pseudotachylyte formed by processes other than melting. Most of the pores in the pseudotachylyte are smaller than 100 nm, and that the host rock is more porous than the pseudotachylyte, thereby suggesting that fluid is unlikely to have passed through the pseudotachylyte subsequent to its formation. It is therefore inferred that the Iida pseudotachylyte is negligibly affected by chemical dissolution in groundwater and hydrothermal alteration subsequent to fault movement. We conclude that amorphous material does not always provide evidence of the rapid cooling of melt in pseudotachylyte. The amorphous materials present in the Iida pseudotachylyte resulted from the mechanochemical effect due to both shear stress and normal stress during the comminution process that accompanied fault motion.
T11A-0355
Scanning ESR microscopy revealing multiplex frictional heating events in the Nojima fault rocks, Japan
Whether or not the temperature rise due to frictional heating in a fault zone universally occurs is a significant problem in connection to the San Andreas fault heat flow paradox or the earthquake energy budget. Since paramagnetic iron hydroxides (γ-FeOOH or Fe(OH)3) inside the fault gouge change into ferrimagnetic iron oxides, maghemite (gamma-Fe2O3), by frictional heating [Fukuchi, 2003; Fukuchi et al., 2005; 2007], ferrimagnetic minerals in fault rocks are available as indicators of ancient frictional heating events. I thus started detecting FMR (ferrimagnetic resonance) signals derived from ferrimagnetic minerals in fault rocks using a scanning ESR (electron spin resonance) microscopy technique. In the scanning ESR microscopy, microwaves leaking out of a pinhole bored on a cavity resonator are directly absorbed by a flat slab sample, and continuous ESR data or ESR images are obtained by one- or two-dimensionally moving the slab sample with an X-Y stage. At the present stage, the resolution for detection is estimated as 0.25mm in using a 2.6mmφ pinhole. This resolution is enough to detect ancient frictional heating events recorded in natural fault rocks although frictional heat temperature changes at a unit of 1 mm or less with the distance from a fault plane. As a result of scanning ESR microscopy of the Nojima fault rocks in Japan, I have succeeded in detecting multiplex frictional heating events recorded in the fault gouge and pseudotachylyte. I now attempt to reconstruct the temperature of ancient frictional heat by inversion using each FMR signal peak showing the multiplex frictional heating events. A preliminary computer simulation indicates that the temperature rise due to frictional heating has universally occurred in the Nojima fault zone and the maximum temperature during ancient seismic fault slips may have momentarily risen above 1000°C. References Fukuchi, T. (2003) J. Geophys. Res., 108, No.B6, 2312, doi:10.1029/2002JB002007. Fukuchi, T., Mizoguchi, K., and Shimamoto, T. (2005) J. Geophys. Res., 110, B12404, doi:10.1029/2004JB003485, 2005. Fukuchi, T., Yurugi, J., and Imai, N. (2007) Tectonophysics, doi:10.1016/j.tecto.2007.01.020. http://web.cc.yamaguchi- u.ac.jp/~fukuchi/
T11A-0356
Extreme Ground-Motion Rockfall Deposits on the Nevada Test Site
In order to detect the evidence of extreme ground motion in the past, we have begun to catalog geomorphic characteristics that distinguish slope deposits strongly influenced by extreme ground motion from deposits primarily influenced by climate processes. Underground nuclear explosions (UNEs) of yields between 200 kilotons and 1.3 megatons were conducted under Pahute Mesa at the Nevada Test site from 1962 to 1992. The primary surface effects from these tests were surface cracks, triggered earthquakes, offsets on pre-existing faults, and changes in land surface topography. Rockfall and rock spall were observed along cliffs after a few nuclear tests; however, few observations of accumulations of shattered rock were documented. A large volume of rockfall located along a 1.5–km¬–long cliff of welded ash-flow tuff resulted from extreme ground motions from two nearby UNEs. In 1968 UNE Rickey released maximum ground motions of 500 cm/s peak ground velocity (PGV) at the closest cliff face and PGV decreased to about 300 cm/s at the north end of the cliff. Large boulders with 1–3–m average diameters were shaken loose from fracture planes and cooling joints to form a stack of jumbled boulders at the base of the cliff. Very few large boulders rolled to the base of the hillslope. Subsequently, in 1976, UNE Pool induced 300-350 cm/s PGV along the same cliff. A significant volume of rock, also released along fractures and joints, was added to the coarse boulder colluvium shaken loose in 1968. Ground motion from Pool also rearranged the hillslope boulders from UNE Rickey, but did not cause many boulders to roll downslope. Extreme ground motions from these two UNEs resulted in 1.5–3.0 m of physical erosion to the cliff face. Rockfall from less welded ash-flow tuff units situated above and below the cliff produced significantly less boulder colluvium. Our observations indicate that boulder size and rockfall volume from a cliff or ridge crest due to extreme ground motion are strongly influenced by physical properties of the rock, by unit thickness of the source rock, by pre-existing fracture patterns, and by weathering history. Boulder distribution on volcanic hillslopes governed by climate processes is significantly different from the concentrated piles of large boulders generated by extreme ground motion. Unshaken cliffs of welded ash-flow tuff are undercut by accelerated erosion of underlying non-welded and partially to moderately welded tuffs. As the undercut cliff recedes boulders are gradually released along weathered joints, which results in a pattern of large boulders distributed more or less evenly downslope except where debris flows and hillslope channels concentrate clasts. Be-10 and Cl-36 exposure ages of welded tuff boulders indicate that large boulders on southern Nevada volcanic hillslopes have been preserved for 200,000 to greater than 500,000 years. Thus, coarse rockfall deposits that resulted from past earthquakes characterized by extreme ground motion may persist on this volcanic landscape for several hundreds of thousands of years and may be distinguishable from climate- generated hillslope colluvium.
T11A-0357
Changes in permeability of fractured rocks by principal stress axes rotation
The use of underground facilities such as waste reservoirs and power station caverns is increasing. Excavation of such spaces results in a change in stress distribution by principal stress axes rotation. Such changes alter the mechanical properties of rock mass such as strength and deformability, and the hydraulic conductivities and hence influence contaminant pathways. The stress redistribution is caused by excavation disturbed zone (EDZ) and rotation of tectonic stress. When the high-level nuclaer waste isolate long term, we have to consider changes in permeability by principal stress axes rotation. In this study, true tri-axial compresshon test was carried out under various stress conditions. The tested rock is damaged sedimentary rock, and has many fractures. Samples are loaded up to following conditions. Case1) isotropic stress condition, case 2) extension stress condition (maximum principal stress is equal to intermediate Principal Stress), and case 3) compression stress condition (intermediate Principal Stress is equal to minimum principal stress). After reached setting stress condition, permeability was measured under 0.05MPa pore pressure. As a result, permeability is the twice different by the change of the direction of principal stress even if mean stress is the same. The direction of fracture and principal stress axes are important factor to estimate long term stability of hydro-mechanical properties of fractured rock mass. This research project has been conducted under the research contract with the Japan Nuclear Energy Safety Organization (JNES).