James P. Evans

Fault zone architecture and permeability structure

Fault zone architecture and related permeability structures form primary controls on fluid flow in upper-crustal, brittle fault zones. We develop qualitative and quantitative schemes for evaluating fault-related permeability structures by using results of field investigations, laboratory permeability measurements, and numerical models offlow within andnearfaultzones.Thequalitativeschemecomparesthepercentageofthetotalfaultzone width composed of fault core materials (e.g., anastomosing slip surfaces, clay-rich gouge, cataclasite,andfaultbreccias)tothepercentageofsubsidiarydamagezonestructures(e.g., kinematically related fracture sets, small faults, and veins). A more quantitative scheme is developed to define a set of indices that characterize fault zone architecture and spatial variability.Thefaultcoreanddamagezonearedistinctstructuralandhydrogeologicunits that reflect the material properties and deformation conditions within a fault zone. Whether a fault zone will act as a conduit, barrier, or combined conduit-barrier system is controlled by the relative percentage of fault core and damage zone structures and the inherent variability in grain scale and fracture permeability. This paper outlines a frameworkforunderstanding,comparing,andcorrelatingthefluidflowpropertiesoffaultzones in various geologic settings.

Internal structure and weakening mechanisms of the San Andreas Fault

New observations of the internal structure of the San Gabriel fault (SGF) are combined with previous characterizations of the Punchbowl fault (PF) to evaluate possible explanations for the low frictional strength and seismic characteristics of the San Andreas fault (SAF). The SGF and PF are ancient, large-displacement faults of the SAF system exhumed to depths of 2 to 5 km. These fault zones are internally zoned; the majority of slip was confined to the cores of principal faults, which typically consist of a narrow layer (less than tens of centimeters) of ultracataclasite within a zone of foliated cataclasite several meters thick. Each fault core is bounded by a zone of damaged host rock of the order of 100 m thick. Orientations of subsidiary faults and other fabric elements imply that (1) the maximum principal stress was oriented at large angles to principal fault planes, (2) strain was partitioned between simple shear in the fault cores and nearly fault-normal contraction in the damaged zones and surrounding host rock, and (3) the principal faults were weak. Microstructures and particle size distributions in the damaged zone of the SGF imply deformation was almost entirely cataclastic and can be modeled as constrained comminution. In contrast, cataclastic and fluid-assisted processes were significant in the cores of the faults as shown by pervasive syntectonic alteration of the host rock minerals to zeolites and clays and by folded, sheared, and attenuated cross-cutting veins of laumontite, albite, quartz, and calcite. Total volume of veins and neocrystallized material reaches 50% in the fault core, and vein structure implies episodic fracture and sealing with time-varying and anisotropic permeability in the fault zone. The structure of the ultracataclasite layer reflects extreme slip localization and probably repeated reworking by particulate flow at low effective stresses. The extreme slip localization reflects a mature internal fault structure resulting from a positive feedback between comminution and transformation weakening. The structural, mechanical, and hydrologic characteristics of the Punchbowl and San Gabriel faults support the model for a weak San Andreas based on inhomogeneous stress and elevated pore fluid pressures contained within the core of a seismogenic fault. Elevated fluid pressures could be repeatedly generated in the core of the fault by a combination of processes including coseismic dilatancy and creation of fracture permeability, fault-valve behavior to recharge the fault with fluid, post-seismic self-sealing of fracture networks to reduce permeability and trap fluids, and time-dependent compaction of the core to generate high pore pressure. The localized slip and fluid-saturated conditions are wholly compatible with additional dynamic weakening by thermal pressurization of fluids during large seismic slip events, which can help explain both the low average strength of the San Andreas and seismogenic characteristics such as large stress relief. In addition, such a dynamic weakening mechanism is expected only in mature fault zones and thus could help explain the apparent difference in strength of large-displacement faults from smaller-displacement, subsidiary seismogenic faults.

Permeability of Fault-Related Rocks, and Implications for Hydraulic Structure of Fault Zones

Abstract The permeability structure of a fault zone in granitic rocks has been investigated by laboratory testing of intact core samples from the unfaulted protolith and the two principal fault zone components; the fault core and the damaged zone. The results of two test series performed on rocks obtained from outcrop are reported. First, tests performed at low confining pressure on 2.54-cm-diameter cores indicate how permeability might vary within different components of a fault zone. Second, tests conducted on 5.1-cm-diameter cores at a range of confining pressures (from 2 to 50 MPa) indicate how variations in overburden or pore fluid pressures might influence the permeability structure of faults. Tests performed at low confining pressure indicate that the highest permeabilities are found in the damaged zone (10 −16 –10 −14 m 2 ), lowest permeabilities are in the fault core ( −20 –10 −17 m 2 ), with intermediate permeabilities found in the protolith (10 −17 –10 −16 m 2 ). A similar relationship between permeability and fault zone structure is obtained at progressively greater confining pressure. Although the permeability of each sample decays with increasing confining pressure, the protolith sustains a much greater decline in permeability for a given change in confining pressure than the damaged zone or fault core. This result supports the inference that protolith samples have short, poorly connected fractures that close more easily than the greater number of more throughgoing fractures found in the damaged zone and fault core. The results of these experiments show that, at the coreplug scale, the damaged zone is a region of higher permeability between the fault core and protolith. These results are consistent with previous field-based and in-situ investigations of fluid flow in faults formed in crystalline rocks. We suggest that, where present, the two-part damaged zone-fault core structure can lead to a bulk anisotropy in fault zone permeability. Thus, fault zones with well-developed damaged zones can lead to enhanced fluid flow through a relatively thin tabular region parallel to the fault plane, whereas the fault core restricts fluid flow across the fault. Although this study examined rocks collected from outcrop, correlation with insitu flow tests indicates that our results provide inexact, but useful, insights into the hydromechanical character of faults found in the shallow crust.

Fluid-rock interaction in faults of the San Andreas system: Inferences from San Gabriel fault rock geochemistry and microstructures

Optical and scanning electron microscopy and whole rock geochemical analyses are used to investigate variations in deformation mechanisms and fluid-rock interactions in rocks at three sites on the San Gabriel fault, southern California : Pacoima Canyon, Bear Creek, and North Fork. At Bear Creek, unaltered and undeformed granite, granodiorite, and diorite protolith bound a fault core several meters thick that consists of foliated cataclasite on either side of 2-20 cm thick ultracataclasite layer. The foliated cataclasite contains clays and zeolite veins which developed by alteration of protolith during slip. The ultracataclasite consists of 20-100 μm diameter feldspar and quartz fragments embedded in a clay-zeolite matrix. The matrix consists of grains <10 μm and is enriched in Fe, Mg, Mn, and Ti relative to the average composition of the protolith. In contrast, ultracataclasite at Pacoima Canyon contains little clay and zeolite and apparently evolved with little fluid-rock interaction. Whole rock geochemical analyses of the fault rock compositions at both sites are best explained as a result of mechanical mixing with local redistribution of some elements in a closed system relative to fluids. At both sites the ultracataclasite compositions can be modeled as the result of mixing of the bounding foliated cataclasites. Similarly, the foliated cataclasites were derived by mixing the protoliths on the same side of the ultracataclasite layer. Whole rock analyses for rocks from the North Fork site, which lies on a major splay of the San Gabriel fault, suggest an open system relative to fluids. The concentration of immobile elements in the fault core relative to all protoliths is best explained by fluid-assisted volume loss of 37% ± 10%. Overall, the results imply local and regional-scale variations in the hydrologic setting along the San Gabriel fault that produced contrasting styles of deformation and fluid-rock interactions.

Mesoscopic structure of the Punchbowl Fault, Southern California and the geologic and geophysical structure of active strike-slip faults

We examine the distribution, density, and orientation of outcrop-scale structures related to the Punchbowl Fault, an exhumed ancient trace of the San Andreas Fault, southern California, in order to determine the structure of the fault zone. The Punchbowl Fault has 44 km of right-lateral slip, and cuts the Cretaceous Pelona Schist in the study area. The mesoscopic structures examined include fractures, small faults, and veins; they were inventoried using scan lines at closely spaced stations along three strike-perpendicular traverses 200‐250 m long across the fault. The fault zone thickness is a function of the type of structure measured. Slip along narrow (<2 m wide) ultracataclasite cores of the faults results in foliation reorientation over a distance of 50 m from the cores: fracture and fault densities appear to increase 50‐80 m from the fault cores, and vein densities are highly variable across the fault zone. Fractures and faults in the damaged zone have a variety of orientations, but most are at high angles to the main fault zone. When coupled with previous geochemical and microstructural data, these data show that large-displacement faults of the San Andreas system, are up to 200‐250 m thick, and enclose zones of mineralogic and geochemical alteration that are 20‐30 m thick. Extreme slip localization occurs over zones 1‐5 m thick. When reconciled with geophysical imaging, our data suggest that trapped headwaves travel in the damaged zone, and that some aftershock events produce slip on faults and fractures, which often have orientations very diAerent from the principal slip surfaces. 7 2000 Elsevier Science Ltd. All rights reserved.

Structural heterogeneity and permeability in faulted eolian sandstone: Implications for subsurface modeling of faults

We determined the structure and permeability variations of a 4 km-long normal fault by integrating surface mapping with data from five boreholes drilled through the fault (borehole to tens of meters scale). The Big Hole fault outcrops in the Jurassic Navajo Sandstone, central Utah. A total of 363.2 m of oriented drill core was recovered at two sites where fault displacement is 8 and 3-5 m. The main fault core is a narrow zone of intensely comminuted grains that is a maximum of 30 cm thick and is composed of low-porosity amalgamated deformation bands that have slip surfaces on one or both sides. Probe permeameter measurements showed a permeability decline from greater than 2000 to less than 0.1 md as the fault is approached. Whole-core analyses showed that fault core permeability is less than 1 md and individual deformation band permeability is about 1 md. Using these data, we calculated the bulk permeability of the fault zone. Calculated transverse permeability over length scales of 5-10 m is 30-40 md, approximately 1-4% the value of the host rock. An inverse power mean calculation (representing a fault array with complex geometry) yielded total fault-zone permeabilities of 7-57 md. The bulk fault-zone permeability is most sensitive to variations in fault core thickness, which exhibits the greatest variability of the fault components.

Spatial Variability in Microscopic Deformation and Composition of the Punchbowl Fault, Southern California: Implications for Mechanisms, Fluid-Rock Interaction, and Fault Morphology

Abstract We examine the distribution and nature of microstructures, geochemistry, and mineralogy along two traverses across the Punchbowl fault, southern California, to determine the morphology and deformation mechanisms of the fault zone in schistose rocks. The Punchbowl fault is an exhumed fault that has two main strands of slip localization and has a total of 44 km of right-slip. Protolith in the study area consists of the Pelona Schist, which is primarily a quartz–albite–muscovite–actinolite schist with thin to medium banding, and rare metabasalts. The traverses are 1.3 km apart, and were conducted at a site with a single fault strand and a site where both principal strands of the fault are exposed. The fault-zone thickness is a function of the type of measurement that is used to define it. For the single strand site, analysis of the distribution of microfractures shows that the fault zone consists of a roughly 40-m-thick damaged zone adjacent to the fault core. The damaged zone is marked by an increase in veins, thin cataclasite bands, inter- and intragranular fracturing, and alteration relative to the country rock. Brittle grain-size reduction occurs in a zone 10 m thick as measured from the fault core, which consists of a continuous, 10-cm-thick, very fine-grained cataclasite that experienced repeated alteration, vein injection and grain-size reduction. Whole-rock geochemical analyses of the fault-related rocks suggest that the geochemically defined fault zone is less than 10 m thick. Volume loss at the site with one fault strand appears to have been small. The dominant alteration reactions associated with the fault core are the hydration of hornblende and actinolite accompanied by the alteration of muscovite to produce a quartz ± chlorite ± albite ± epidote ultracataclasite. The composition of the fault core is variable and locally influenced by one of the adjacent protoliths. The examination of the two fault strand sites shows that two damaged-zone fault-core structures are present. The region between the two strands experienced a greater degree of deformation than the protolith, but the total deformation is much less than immediately adjacent to the fault cores. The total thickness of the damaged zone around the two strands is less than 200 m. The fault core enveloped by a damaged-zone morphology, as well as the textures of the fault-core rocks are similar to rocks associated with the North Branch San Gabriel fault, which formed in crystalline rocks with a total displacement of 22 km. Fault thickness is less in the Pelona Schist than in the crystalline rocks, perhaps owing to more efficient strain localization in the schists. Thus, the faults of the San Andreas system may be thinner in regions where schists or Franciscan rocks are the protolith, but the main fault core may be a constant feature of the fault zone. Transformation-induced weakening is less important in this part of the Punchbowl fault, since the protolith has a large amount of mica. The structure of the fault zone with two principal slip surfaces is marked by both chemical changes and microstructures, and indicates that some parts of the San Andreas fault system may consist of multiple slip surfaces, each with a damaged zone, that together may create a fault zone >100 m thick, and in which slip is localized to zones meters to decimeters thick.

Thickness-Displacement Relationships for Fault Zones

Abstract Fault zone thickness and displacement data have been used in recent studies to infer a global linear relationship between the two parameters in log-log space, and based on this empirical relationship, several models of fault growth have been proposed. Examination of the original data sets, but on linear graphs, indicates that there is a wide statistical variability of many of the data, and that larger values of displacement along faults are only generally related to an increase of fault zone thickness. The earlier data do not confirm a thickness-displacement relationship for faults of similar displacement, nor for faults in similar rock types. Reasons for the lack of correlation may be the wide range of rock types, structural settings, and processes responsible for the development of the faults measured. An example of thickness variations along the Bismark fault, southwest Montana, shows the variations of fault zone thickness along strike and down plunge. Conceptual and quantitative models which rely heavily on thickness-displacement relationships should be considered with caution until further data are collected on the topic. Future studies should explicitly state the criteria used to determine faulted and unfaulted rock, present in explicit form the slip vector used to determine net offset, and present, if possible, measurements of thickness and displacement from: (1) different points along the same fault, (2) families of faults in similar rock types with different amounts of slip, and (3) faults with similar amounts of net slip in similar structural settings.

Hydrogeologic Controls on Induced Seismicity in Crystalline Basement Rocks Due to Fluid Injection into Basal Reservoirs

A series of Mb 3.8-5.5 induced seismic events in the midcontinent region, United States, resulted from injection of fluid either into a basal sedimentary reservoir with no underlying confining unit or directly into the underlying crystalline basement complex. The earthquakes probably occurred along faults that were likely critically stressed within the crystalline basement. These faults were located at a considerable distance (up to 10 km) from the injection wells and head increases at the hypocenters were likely relatively small (∼70-150 m). We present a suite of simulations that use a simple hydrogeologic-geomechanical model to assess what hydrogeologic conditions promote or deter induced seismic events within the crystalline basement across the midcontinent. The presence of a confining unit beneath the injection reservoir horizon had the single largest effect in preventing induced seismicity within the underlying crystalline basement. For a crystalline basement having a permeability of 2 × 10(-17)  m(2) and specific storage coefficient of 10(-7) /m, injection at a rate of 5455 m(3) /d into the basal aquifer with no underlying basal seal over 10 years resulted in probable brittle failure to depths of about 0.6 km below the injection reservoir. Including a permeable (kz  = 10(-13)  m(2) ) Precambrian normal fault, located 20 m from the injection well, increased the depth of the failure region below the reservoir to 3 km. For a large permeability contrast between a Precambrian thrust fault (10(-12)  m(2) ) and the surrounding crystalline basement (10(-18)  m(2) ), the failure region can extend laterally 10 km away from the injection well.

Analysis of CO 2 Leakage through 'Low-Permeability' Faults from Natural Reservoirs in the Colorado Plateau, East-Central Utah

Abstract The numerous CO2 reservoirs in the Colorado Plateau region of the United States are natural analogues for potential geological CO2 sequestration repositories. To understand better the risk of leakage from reservoirs used for long-term underground CO2 storage, we examine evidence for CO2 migration along two normal faults that cut a reservoir in east-central Utah. CO2-charged springs, geysers, and a hydrocarbon seep are localized along these faults. These include natural springs that have been active for long periods of time, and springs that were induced by recent drilling. The CO2-charged spring waters have deposited travertine mounds and carbonate veins. The faults cut siltstones, shales, and sandstones and the fault rocks are fine-grained, clay-rich gouge, generally thought to be barriers to fluid flow. The geological and geochemical data are consistent with these faults being conduits for CO2 moving to the surface. Consequently, the injection of CO2 into faulted geological reservoirs, including faults with clay gouge, must be carefully designed and monitored to avoid slow seepage or fast rupture to the biosphere.