Search

We examine the fine-scale variations in mineralogical composition, geochemical alteration, and texture of the fault-related rocks from the Phase 3 whole-rock core sampled between 3,187.4 and 3,301.4 m measured depth within the San Andreas Fault Observatory at Depth (SAFOD) borehole near Parkfield, California. This work provides insight into the physical and chemical properties, structural architecture, and fluid-rock interactions associated with the actively deforming traces of the San Andreas Fault zone at depth. Exhumed outcrops within the SAF system comprised of serpentinite-bearing protolith are examined for comparison at San Simeon, Goat Rock State Park, and Nelson Creek, California. In the Phase 3 SAFOD drillcore samples, the fault-related rocks consist of multiple juxtaposed lenses of sheared, foliated siltstone and shale with block-in-matrix fabric, black cataclasite to ultracataclasite, and sheared serpentinite-bearing, finely foliated fault gouge. Meters-wide zones of sheared rock and fault gouge correlate to the sites of active borehole casing deformation and are characterized by scaly clay fabric with multiple discrete slip surfaces or anastomosing shear zones that surround conglobulated or rounded clasts of compacted clay and/or serpentinite. The fine gouge matrix is composed of Mg-rich clays and serpentine minerals (saponite ± palygorskite, and lizardite ± chrysotile). Whole-rock geochemistry data show increases in Fe-, Mg-, Ni-, and Cr-oxides and hydroxides, Fe-sulfides, and C-rich material, with a total organic content of >1 % locally in the fault-related rocks. The faults sampled in the field are composed of meters-thick zones of cohesive to non-cohesive, serpentinite-bearing foliated clay gouge and black fine-grained fault rock derived from sheared Franciscan Formation or serpentinized Coast Range Ophiolite. X-ray diffraction of outcrop samples shows that the foliated clay gouge is composed primarily of saponite and serpentinite, with localized increases in Ni- and Cr-oxides and C-rich material over several meters. Mesoscopic and microscopic textures and deformation mechanisms interpreted from the outcrop sites are remarkably similar to those observed in the SAFOD core. Micro-scale to meso-scale fabrics observed in the SAFOD core exhibit textural characteristics that are common in deformed serpentinites and are often attributed to aseismic deformation with episodic seismic slip. The mineralogy and whole-rock geochemistry results indicate that the fault zone experienced transient fluid–rock interactions with fluids of varying chemical composition, including evidence for highly reducing, hydrocarbon-bearing fluids.

The drilling of a deep borehole across the actively creeping Parkfield segment of the San Andreas Fault Zone (SAFZ), California, and collection of core materials permit direct geological study of fault zone processes at 2–3 km depth. The three drill cores sample both host and fault rocks and pass through two currently active, narrow (1–2 m wide) shear zones enclosed within a broader (ca. 240 m wide) region of inactive foliated gouges. The host rocks preserve primary sedimentary features and are cut by numerous minor faults and small, mainly calcite-filled veins. The development of Fe-enriched smectitic phyllosilicate networks following cataclasis is prevalent in the presently inactive foliated gouges of the main fault zone and in minor faults cutting clay-rich host rocks. Calcite, anhydrite and minor smectitic phyllosilicate veins are interpreted to have formed due to local fluid overpressuring events prior to, synchronous with and after local gouge development. By contrast, the active shear zone gouges lack mineral veins (except as clasts) and contain numerous clasts of serpentinite. Markedly Mg-rich smectitic phyllosilicates are the dominant mineral phases here, suggesting that the fault zone fluids have interacted with the entrained serpentinites. We propose that weakening of the SAFZ down to depths of at least 3 km can be attributed to the pervasive development of interconnected networks of low friction smectitic phyllosilicates and to the operation of stress-induced solution-precipitation creep mechanisms.

The San Andreas Fault Observatory at Depth (SAFOD)project is yielding new insight into the San Andreas Fault (Zoback et al., 2010; Zoback et al., this issue). SAFOD drilling started in 2002 with a pilot hole, and proceeded with three phrases of drilling and coring during the summers of 2004, 2005, and 2007 (Fig. 1). One key component of theproject is curation, sampling, and documentation of SAFOD core usage at the Integrated Ocean Drilling Program’s (IODP) Gulf Coast Repository (GCR) at Texas A&M University. We present here the milestones accomplished over the past two years of sampling Phase III core at the GCR.

The microstructures, mineralogy and chemistry of representative samples collected from the cores of the San Andreas Fault drill hole (SAFOD) and the Taiwan Chelungpu-Fault Drilling project (TCDP) have been studied using optical microscopy, TEM, SEM, XRD and XRF analyses. SAFOD samples provide a transect across undeformed host rock, the fault damage zone and currently active deforming zones of the San Andreas Fault. TCDP samples are retrieved from the principal slip zone (PSZ) and from the surrounding damage zone of the Chelungpu Fault. Substantial differences exist in the clay mineralogy of SAFOD and TCDP fault gouge samples. Amorphous material has been observed in SAFOD as well as TCDP samples. In line with previous publications, we propose that melt, observed in TCDP black gouge samples, was produced by seismic slip (melt origin) whereas amorphous material in SAFOD samples was formed by comminution of grains (crush origin) rather than by melting. Dauphiné twins in quartz grains of SAFOD and TCDP samples may indicate high seismic stress. The differences in the crystallographic preferred orientation of calcite between SAFOD and TCDP samples are significant. Microstructures resulting from dissolution-precipitation processes were observed in both faults but are more frequently found in SAFOD samples than in TCDP fault rocks. As already described for many other fault zones clay-gouge fabrics are quite weak in SAFOD and TCDP samples. Clay-clast aggregates (CCAs), proposed to indicate frictional heating and thermal pressurization, occur in material taken from the PSZ of the Chelungpu Fault, as well as within and outside of the SAFOD deforming zones, indicating that these microstructures were formed over a wide range of slip rates. © 2014 Elsevier Ltd.

This thesis presents the application of 3D MT inversion to an array data set of more than 250 sites from central California to image the electrical conductivity structure of the San Andreas fault (SAF) system in the Parkfield- Cholame region where the fault changes its mechanical state. Previous extensive two-dimensional (2D) inversion revealed a high-conductivity region in the upper mantle and lower crust, which has a connection to the SAF in the northern, transitional-to-creeping section, whereas it is confined to lower- crustal to upper-mantle levels in the transitional-to-locked section. The highly conductive region and the connecting channel were interpreted as migration path for fluids rising from mantle depth into the SAF system where they are considered to contribute to the low-frictional strength of the creeping fault segment. The new 3D inversion results confirm the high- conductive zones of the 2D results and strongly support the presence of a fluid channel into the SAF system in the creeping segment. However, conductivity structures obtained with 3D inversion also showed a great non- uniqueness depending on the inversion setup. Satisfying models (in terms of data fit) could only be recovered if the inversion parameters were tuned in accordance with the particularities of the data set. Based on the Parkfield MT data and complementary synthetic 3D data sets, the model space was explored by testing the influences of a wide range of inversion settings. The results show that in presence of a pronounced regional 2D structure, 3D inversion of the complete impedance tensor still depends on the coordinate system. 2D subsurface structures can vanish if data are not aligned with the regional strike-direction. A priori models and data weighting, i.e. how strongly individual components of the impedance tensor and/or vertical magnetic field transfer functions dominate the solution, are crucial controls for the outcome of 3D inversion. If deviations from a prior model are heavily penalized, regularization is prone to result in erroneous and misleading 3D inversion models, particularly in the presence of strong conductivity contrasts. Reliable and meaningful 3D inversion models can only be recovered if data misfit is assessed systematically in the frequency-space domain. Galvanic distortion can impair 3D inversion models and result in spurious structures at depth. A new tool for automatic identification of static shift was developed and applied to synthetic and real world data, making use of the spatial coherence of MT responses. For the Parkfield data set, identification and removal of static shift improved the outcome of subsequent 3D inversion for the near-surface layers. Inversion of phase tensor and apparent resistivity & phase data was implemented into the 3D inversion scheme and proved to be a valuable asset for obtaining reliable subsurface images from galvanically distorted data, while not requiring any data preprocessing., In der vorliegenden Arbeit ist die Anwendung eines dreidimensionalen (3D) magnetotellurischen (MT) Inversionsschemas auf einen Array-Datensatz von mehr als 250 Stationen aus Zentralkalifornien (USA) dargestellt. Die Untersuchungen hatten zum Ziel, die elektrische Leitfähigkeitsstruktur der San Andreas- Verwerfung (SAF) in der Region zwischen Parkfield und Cholame, in der sich das mechanische Verhalten der Verwerfung ändert, abzubilden. Vorangegangene zweidimensionale (2D) Inversionen des Datensatzes zeigten eine Region hoher elektrischer Leitfähigkeit im Tiefenbereich des oberen Mantel und der unteren Kruste, die im nördlichen Teil des Messgebiets hin zum kriechenden Segment mit der seismogenen SAF verbunden ist. Im übergang zum blockierten Verwerfungsabschnitt Richtung Süden ist diese leitfähige Region in der Tiefe dagegen von der oberkrustalen Deformationszone abgeschnitten. Die Bereiche erhöhter Leitfähigkeit und die kanalartige Verbindung im Norden wurden als Migrationspfade für Fluide interpretiert, die aus dem Mantel in das SAF System aufsteigen. Es wird vermutet, dass Fluide wesentlich zur mechanischen Schwäche der SAF im kriechenden Segment beitragen. Die neuen, 3D Inversionsergebnisse bestätigen die elektrisch leitfähige Zone aus den 2D Modellen und unterstützen die Existenz eines Fluidkanals in das SAF-System im kriechenden Abschnitt. Jedoch zeigten die 3D Leitfähigkeitsmodelle starke Variationen der abgebildeten Strukturen, die mit der Wahl der Inversionsparameter zusammenhingen. Zufriedenstellende Modelle (im Sinne der Datenanpassung) konnten nur ermittelt werden, wenn die Inversionsparameter auf die Besonderheiten des Datensatzes abgestimmt waren. Basierend auf den MT-Daten aus Kalifornien sowie komplementären synthetischen 3D Datensätzen wurde der Einfluss der Inversionsparameter auf den Modellraum untersucht. Die Ergebnisse zeigen, dass die 3D Inversion in Gegenwart einer ausgeprägten, regionalen 2D Struktur von der Orientierung des Koordinatensystems abhängt, auch wenn alle vier Komponenten des Impedanztensors in der Inversion verwendet werden. 2D Untergrundstrukturen können gänzlich verschwinden, wenn das Koordinatensystem nicht an der regionalen geoelektrischen Streichrichtung ausgerichtet ist. A priori-Modelle und Datengewichtung, welche bestimmt, wie stark einzelne Komponenten des Impedanztensors und der vertikalen magnetischen übertragungsfunktionen die Lösung dominieren, sind wichtige Steuerungselemente für das 3D Inversionsergebnis. Werden Abweichungen von einem a priori-Modell stark regularisiert, führt dies leicht zu verfälschten und irreführenden 3D Inversionsergebnissen, insbesondere wenn starke Leitfähigkeitskontraste im Untergrund vorhanden sind. Verlässliche und aussagekräftige Modelle können nur ermittelt werden, wenn eine systematische Auswertung der Datenanpassung im Frequenz-Orts-Bereich vorgenommen wird. Galvanische Verzerrung von MT-Daten kann die Qualität von 3D Inversionsmodellen beeinträchtigen und zum Aufreten artifizieller Strukturen bis in gro\ss e Tiefe führen. Zur Identifizierung von Static Shift wurde ein neues Schema entwickelt, das auf der räumlichen Kohärenz von MT-Daten basiert, und sowohl auf synthetische als auch auf Messdaten angewendet. Für den Kalifornien-Datensatz verbesserte die Entfernung von Static Shift die nachfolgenden 3D Inversionsergebnisse. Die Inversionen von Phasentensoren und scheinbaren Widerständen & Phasen, die in das 3D Inversionspaket implementiert wurden, sind ein wertvoller Gewinn für die verlässliche Abbildung von Untergrundsstrukturen, ohne dafür verzerrte Daten vorbehandeln zu müssen.

We examine the relationships between rock properties and structure within ~ 41 m of PHASE III whole-rock core collected from ~ 3 km depth along the SAF in the San Andreas Fault Observatory at Depth (SAFOD) borehole, near Parkfield, CA.

We examine the composition and texture of whole-rock core from ~ 3 km depth in the San Andreas Fault Observatory at Depth (SAFOD) borehole, which provides a unique opportunity to characterize in situ rock properties of the near-fault environment, and how these properties vary in an area where deformation is accommodated by aseismic creep and high-rates of microseismicity. Detailed petrography and microstructural analyses coupled with X-Ray Diffraction and X-ray Fluorescence techniques are used to describe composition, alteration, and textures. All samples record multiple generations of cataclastic deformation in a complexly deformed and altered sequence of fine-grained sheared rocks. Localized shears bound multi-layered zones of medium to ultra-fine grained cataclasite. Phacoidal clasts or porphyroclasts comprised of serpentinite, quartz, and older cataclasite are embedded within the comminuted phyllosilicate-rich gouge. The intensity of damage-related features and the development of a pervasive anastomosing fabric increases towards and within the two active slip zones near ~ 3192 and 3302 m MD. Foliated fabrics alternating with discrete fractures suggest a mixed-mode style of deformation including both ductile flow and brittle deformation processes during fault zone evolution. Deformation at high-strain rates is suggested by the presence of crack-seal veins in clasts, the presence of porphyroclasts, and the development of S-C fabrics in the phyllosilicate- rich gouge. Evidence for fluid-rock interaction across the fault zone is indicated by depletion of Si and enrichment of MgO, FeO, and CaO; with significant clay alteration and/or growth of neo-mineralized vein fillings and fracture surface coatings. Shear localization may decrease porosity and inhibit fluid flow whereas fracturing may locally facilitate fluid migration and/or chemical alteration within the fault zone. These results constrain hypotheses related to fault zone behavior and broaden our understanding of the processes controlling earthquake nucleation and/or energy adsorption within the SAF. Based on the similarity of our observations to previous results from surface exposures of the SAF, we emphasize the importance of exhumed fault zone studies as proxies for understanding deformation and seismicity in the shallow crust.

The aim of the San Andreas Fault Observatory at Depth (SAFOD) project, a component of the NSF Earthscope Initiative, is to directly observe active fault processes at seismogenic depths through the drilling of a 3 km deep (true vertical depth) inclined borehole across San Andreas fault. Preliminary subsurface models based on surface mapping and geophysical data predicted different lithologies than were actually encountered. At 1920 meters measured depth (mmd), a sequence of well-indurated, interbedded arkosic conglomerates, sandstones, and siltstones was encountered. We present a detailed lithologic and structural characterization as a step toward understanding the complex geologic history of this fault-bounded block of arkosic sedimentary rocks. We divide the arkosic section into three lithologic units with different compositional, structural, and sedimentary features: the upper arkose, 1920-2530 mmd, the clay-rich zone, 2530-2680 Illtlld, and the lower arkose, 2680-3150 mmd. We interpret the section to have been deposited in a Salinian transtensional basin, in either a subaqueous or subaerial fan setting. We suggest four different possibly equivalent sedimentary units to the SAFOD arkoses, the locations of which are dependent on how the San Andreas fault system has evolved over time in the vicinity of the SAFOD site. Detailed analysis of three subsidiary faults encountered in the arkosic section at 1920 mmd, 2530 mmd, and 3060 mmd, shows that subsurface faults have similar microstructures and composition as exhumed faults at the surface, with less evidence of alteration from extensive fluid flow.

[1] Washed cuttings provide a continuous record of the rocks encountered during drilling of the main hole of the San Andreas Fault Observatory at Depth (SAFOD). Both protolith and fault rocks exhibit a wide variety of mineral assemblages that reflect variations in some combination of lithology, P-T conditions, deformation mechanisms, and fluid composition and abundance. Regions of distinct neomineralization bounded by faults may record alteration associated with fluid reservoirs confined by faults. In addition, both smectites occurring as mixed-layer phases and serpentine minerals are found in association with active strands of the San Andreas Fault that were intersected during drilling, although their rheological influence is not yet fully known. Faults containing these mineralogical phases are prime candidates for continuous coring during Phase 3 of SAFOD drilling in the summer of 2007.