ABSTRACT Our field study examines two sites revealing the processes responsible for fault surface development and fault rock juxtaposition along normal faults in porous-sandstone–dominated formations. At the first site, we show that a cluster of cataclastic deformation bands made in an initially poorly consolidated sandstone localizes overprinting of a fault slip surface and brecciation during a subsequent tectonic episode, after a significant diagenesis of the formation induced by burial. Because the permeability of the clusters ranges between 6 × 100 and less than 5 × 10−1 md and because the breccia was highly dilatant, we deduce that the fault behaved as a baffle for cross-fault fluid flow at an early age of the formation and as a conduit after significant diagenetic evolution and subsequent fault surface development. At the second site, we show that the presence of clay-rich layers as thin as 80 cm (31 in.) are responsible for the initiation of a major fault slip surface in the underlying and overlying sandstone. The propagation of the fault prevents further cataclastic deformation and cluster development in these sandstones. Fault displacement juxtaposes fault surfaces, clusters of cataclastic deformation bands, clay-rich gouges, and different sedimentary units. Because both fault rocks have low permeability, their spatial juxtaposition provides a continuous baffle for cross-fault fluid flow. Our study shows that fault surface localization is related to an increase in the contrast of mechanical behavior between the cluster and the adjacent material (diagenetic hardening of the cluster or softening of the clay-rich gouge). Lithological contrasts and diagenesis are favorable conditions for localizing faulting and fault rock juxtaposition, allowing significant three-dimensional anisotropy of permeability during and/or after deformation. These processes must, therefore, be considered for fault-seal analyses in sandstone reservoirs. INTRODUCTION Faults are the economic core of structural geology because they can constitute conduits or barriers for fluids in the context of resource exploitation or storage, because of the diversity of their morphology, and because the petrophysical properties of the fault rocks are complex (e.g., Knipe, 1997; Yielding et al., 1997; Aydin, 2000; Faulkner et al., 2010). Faults generally include one or multiple core zones comprised of fault rocks, including breccias, cataclasites, or gouges, generally separated by discrete fault slip surfaces, accommodating most of shear displacement (Sibson, 1977; Caine et al., 1996). Especially in porous sandstone, despite its importance for reservoirs compartmentalization, the processes allowing fault surface formation and subsequent fault rock juxtaposition are still not clearly understood (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Fossen et al., 2018). In porous sandstone units (Figure 1A–C), the deformation is dominated by cataclasis (also referred to as “cataclastic flow”). This process is characterized by an important decrease in grain size when grains are rearranged by rolling and splitting because of stress concentrations at the grain–grain contacts (Gallagher et al., 1974; Aydin, 1978). The cataclastic deformation localizes as (1) compactional shear band (CSB, as defined by Fossen et al., 2018), within which fault rocks correspond to crush microbreccias and cataclasites, and (2) fault surfaces that are sharply striated (slickensides), along which the fault rocks generally correspond to ultracataclasites (e.g., Aydin, 1978; Blenkinsop, 1991; Fossen et al., 2007). Some models consider that fault surface initiation is mainly caused by cataclastic hardening occurring in clusters of CSB and suggest that the fault surface appears with the increase of cluster thickness (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Wibberley et al., 2007) (Figure 1A–C). This hypothesis is, however, strongly debated because fault surfaces are observed on a wide range of band cluster thicknesses and densities (see Figure 2 and Philit, 2017; Philit et al., 2018) and because other processes, such as segment linkage, which is not directly dependent on cluster thickness, might be responsible for fault slip surface development (Nicol et al., 2013). These processes are generally inspected in the host porous sandstone and rarely analyzed by looking at the surrounding lithology or its evolution through time. Sandstone-dominated formations comprise a variety of lithologies, each involving different mechanical processes of deformation (e.g., Aydin and Eyal, 2002; Schultz and Fossen, 2008) that may enhance or impede fault surface development and lead to a diversity of fault architecture and rocks (e.g., Knipe, 1997; Wibberley et al., 2008; Schmatz et al., 2010). Figure 1. Synthetic schemes of the evolution of extensional deformation to faulting in different materials: (A–C) Cataclastic deformation in porous sandstone units (poorly compacted and lithified); (D–F) brecciation in lithified (e.g., compacted, cemented) sandstone units; (G–I) shale smear in alternating shale and multiple lithified rock units. Regular bold arrows indicate the deformation kinematics (extension or compression); the thin half arrows indicate the shear direction on both sides of the fault. The red circles are zones of zoom highlighting the deformation structures in the fault core. Deformation of brittle low-porosity sandstones (Figure 1D–F) generally leads to brittle dilatant fracturing, producing breccias, protocataclasite, cataclasite, and fault surfaces as rock comminution evolves with increasing shear strain (e.g., Zhu and Wong, 1997; Woodcock and Mort, 2008). Formations bearing clay or shale units (Figure 1G–I) are likely to impede brittle failure, forming “clay gouges” deformed in a ductile manner defined as smearing (e.g., Lindsay et al., 1993; Aydin and Eyal, 2002). Few studies provide analyses of the juxtaposition of fault rocks in porous sandstones, clay units, and cemented sandstones (e.g., Foxford et al., 1998; Davatzes et al., 2005; Rawling and Goodwin, 2006), and such studies might provide interesting observations to discuss the conditions and timing of fault slip surface development. Yet, the study of the overprinting of these different mechanisms is essential to understand the complex fault core organizations resulting from the deformation and the contrasting petrophysical properties induced by the juxtaposed fault rocks (Fulljames et al., 1997; Knipe, 1997; Yielding et al., 1997; Agosta et al., 2006; Rawling and Goodwin, 2006; Childs et al., 2007). Figure 2. Graph of cluster thickness versus band density of compactional shear band (CSB) clusters containing fault slip surfaces and CSB clusters without fault slip surfaces. The cluster thickness corresponds to the sum of band thicknesses and the interband distance. The band density corresponds to the sum of band thicknesses divided by the cluster thickness. The graph presented is synthetized and modified from Philit (2017) and Philit et al. (2018). In this study, we focus on two sites in sandstone-dominated formations that have varying lithologies (variable porosity and cementation degree and presence of interbedded clay-rich materials), display fault slip surfaces adjacent to the cluster of CSBs, and allow juxtaposition of breccia and clay-rich gouge. We show that fault slip surfaces developing in fault cores can be influenced by (1) the cementation of porous sandstones and (2) the interplay of deformation between porous sandstone and clay-rich units. We discuss the implications of these processes on the heterogeneity of fault core composition and evaluate their impact on fluid flow during the evolution of the fault. GEOLOGICAL SETTINGS The Clashach Cove site is located northeast of the town of Hopeman, between the towns of Cummingston (to the west) and Lossiemouth (to the east) on the Moray Coast in Scotland (Figure 3A, B). The outcrop exposes quartz-rich sandstones from the Triassic Hopeman Sandstone Formation, buried up to 1.2–2.4 km (0.7–1.5 mi) (Quinn, 2008; Al-Hinai et al., 2008). The Hopeman Sandstone Formation is aeolian (Shotton, 1956; Glennie and Buller, 1983); the dune foresets are clearly visible in the hanging wall at the site. The sandstone is crosscut by a normal fault with a total dip-slip component of 20–50 m (66–164 ft) (Quinn, 2008) (Figure 3C). Similar to the many normal seismic-scale structures (≥20 m [≥65.6 ft] dip-slip component) in the Moray Firth, the Clashach fault is believed to have been formed with an exclusively dip-slip displacement during the pre-Jurassic and Late Jurassic rifting and possibly reactivated with an oblique displacement during the Paleocene–Eocene northeast Atlantic rifting (e.g., Roberts et al., 1990; Underhill, 1991; Al-Hinai et al., 2008). The rifting also caused the uplift of the Scottish Highlands. The fault juxtaposes gray cohesive sandstone topped by yellow sandstone of weak cohesion in the footwall to the north and the same yellow sandstone in the hanging wall to the south. Figure 3. (A) Location of the study site at the scale of the United Kingdom: Moray Coast, Scotland. (B) Geological map and local stratigraphic column of the Clashach Cove site (red arrow) in the Lossiemouth-Burghead area, along the Moray Coast, modified after Edwards et al. (1993). (C) General view of the faulted Hopeman Sandstone (Sandst.) Formation (Fm.) at Clashach Cove. (D) Stereographic projection (lower hemisphere) of the deformation structures around the fault at the study site. Blue lines represent the cataclastic structures (compactional shear band [CSB] and CSB cluster); red lines represent joints; green lines represent dilation bands; the thick black line represents the Clashach fault, with both cataclastic and dilatant compartments; and the dashed line represents the average stratigraphy at the site. The site of Goblin Valley is located in the San Rafael Desert, Utah, northeast of the Goblin Valley State Park, approximately 20 km (∼12.4 mi) north of the town of Hanksville (Figure 4A, B). The site displays the Middle Jurassic Entrada Formation of total thickness ranging from 76 to 160 m (249 to 525 ft) and buried up to 2.2 km (1.4 mi) (Rigby and Beus, 1987; Johansen and Fossen, 2008; Schultz and Fossen, 2008). The Entrada Formation is made of tabular fine-to-coarse–grained aeolian and foreshore sandstone, including thin layers of poorly lithified siltstone to clay, 10–30 cm (3.9–11.8 in.) thick each, at its base. The formation underwent extensional tectonics from the Oligocene linked to the basin and range extension. The possible burial at deformation ranges from 1100 to 2000 m (0.7 to 1.2 mi) (Doelling et al., 2015). At the site of Goblin Valley, the outcrop displays tens of clusters of CSBs oriented northwest-southeast where displacement is mostly dip-slip (Figure 4C). Even if commonly hidden by Quaternary wind deposits, these clusters are observed in an area more than 3 km (>1.9 mi) long and 800 m (>0.5 mi) wide (Figure 4B). We chose two exposures within this area to analyze the cataclastic and smear deformation processes. The first exposure (Figure 4D) consists of a fault with a dip-slip component of 6 m (19.7 ft) into a sandstone-dominated unit including thin layers of poorly lithified siltstones at the base of the Entrada Formation. The second exposure (Figure 4E) consists of a fault with a dip-slip component of 2.10 m (6.9 ft) into a sandstone-dominated formation including a 40-cm (15.7-in.)-thick shale layer. Figure 4. (A) Location of the study site at the scale of Utah: San Rafael desert. (B) Geological map (modified from Doelling et al., 2015) and local stratigraphic column (modified from Schultz and Fossen, 2008) of the two study exposures (red stars), northeast of the Goblin Valley State Park. (C) Stereographic projection (lower hemisphere) of the clusters of deformation bands at the study site. (D, E) View of the 6-m (19.7-ft) fault of exposure 1 and of the 2.1-m (6.6-ft) fault of exposure 2, respectively. Fm. = Formation; SLC = Salt Lake City. METHODOLOGY At the Clashach Cove site, to analyze the mechanical processes and to identify the relative chronology of the deformation and characterize its diagenetic setting, we studied the macroscopic deformation and the rock microstructure as well as the properties of the host material (grain size distribution, sorting [s], porosity [φ], and permeability [k]). The deformed material was sampled in the section of the cluster of cataclastic deformation bands and in the overlying oxidized gouge. The host rock was sampled a few meters away from and on both sides of the fault plane where no macroscopic deformation is observed. Microscopic observations were conducted using cold cathodoluminescence (CL) with a BX41 Olympus microscope equipped with a CL8200 MK5-1 apparatus and scanning electron microscopy of backscattered electrons (SEM-BSE) imaging mode with an FEI Quanta 200 FEG microscope. The mineralogical compositions were obtained with a combined calibration through microscopic analyses in natural light, cross-polarized light, CL, and major element composition from SEM-BSE. A quantification of the different minerals’ proportions was obtained with gray-scale contrast levels of SEM images. Because the rock is cemented by quartz, the analysis of the grain size distribution was done through CL image measurements using the Fiji software. To quantify grain size distribution, Feret diameters of the particles of the host rock and deformed material were measured from representative sections (nearly 700 mm2 [1.085 in.2] investigated) with a resolution limit of 20 μm. The s value of the host rocks was obtained from the grain size distribution data using the formula of the inclusive graphic standard deviation (Folk and Ward, 1957). Note that spectrums of the grain size distribution and the mineralogical composition of the studied sandstones of the two sites are available in the Appendix. The φ value was obtained from two-dimensional SEM and CL image analysis. A total area of 2720 mm2 [4.216 in.2] representative of the deformation band clusters and host rocks was analyzed. An error of approximately 2% is caused by variation of φ linked to the uncertainty of the upper limit of the gray-scale threshold, corresponding to the spectral limit between φ and quartz minerals. The φ value of the host rocks after cementation is given as the intergranular volume (IGV, see Ehrenberg, 1989), where any potential postdeformation cementation is subtracted; the φ of the deformation structures is given as residual φ, accounting for potential postdeformation cementation. We measured the k value at the surface of the host rocks and deformation structures with a portable air-permeameter device (TinyPerm II NER). The resolution of the device does not permit the quantification of k higher than 5 × 104 md. For each type of rock surface, k has been measured at least twice for measurements exceeding 10 min, or more for shorter measurements. This time limitation constrains the minimal k quantification to approximately 5 × 10−1 md. Also, note that the TinyPerm II NER permeameter has been calibrated by Fossen et al. (2011) for porous sandstones revealing an overestimate of permeability of 1.8 times compared with image and routine gas-based permeability measures. This suggests that the values presented in this paper might be considered overestimated by approximately a factor of 2 in addition to the error bars of the data presented for the fault and host rocks in Figure 5. Figure 5. Graph summarizing the permeability of the host rocks (HR) and their resulting fault rocks as a function of the deformation mechanisms. The square data are the measurements at the Clashach Cove site; the circle data illustrate the measurements at the Goblin Valley site. The gray scale and thickness of the frame lines reflect the grain size of the HRs: light-gray and thick line: medium-grain sandstone; medium-gray and regular line: fine-grain sandstone; dark gray without line: siltstone; black: clay. The dashed circle with the question mark illustrates that we do not know how the process of clay smear alters the permeability. ∗ = range of typical values of shale HRs are given for reference as reported by Neuzil (1994). At Goblin Valley, we focused on (1) a macroscopic description of the juxtaposition of different types of rocks in the fault cores, (2) a petrophysical and mineralogical analysis of the pristine juxtaposed host rocks and layers properties, (3) a detailed measurement of the amount of cataclastic deformation in the fault cores, and (4) k measurements using the same methodology as described before. The analysis of the grain size distribution was done through Coulter LS 13320 laboratory laser diffraction on selected samples using four types of host rocks in the Entrada Formation; the grain size of other host rocks was estimated in the field using the grain size reference chart of the American Stratigraphic Company. The φ and k values were measured with the same methodology as described before. We particularly focused on the analysis of clusters of CSBs. In the absence of an accurate definition of “cluster” in the literature, we propose to combine three typical characteristics described in the literature (a minimum band number, a typical frequency, and band thickness) to define it. A “cluster of CSBs” (or cluster) corresponds to the grouping of two or more bands (as used in Fossen and Bale, 2007) and accounts for at least 10% of the bands’ thickness (based on a usual frequency of 20–30 bands/m [6–9 bands/ft] reported by Johansen and Fossen, 2008, and a typical band thickness of approximately 4 mm [∼0.16 in.] as shown in Wibberley et al., 2007). Clusters may or may not include a fault slip surface (Figure 2). Such clusters typically have widths of 1 cm to several decimeters (0.4 in. to a couple of inches) (e.g., Wibberley et al., 2007) and accommodate centimeter to several decimeter (0.4 in. to a couple of inches) displacements (e.g., Antonellini and Aydin, 1994). From this definition, we introduce the cumulative band thickness and the cluster thickness to evaluate the deformation within clusters, and we ignore the interband deformation, which is small but not zero (Lanatà, 2014). The band thickness and the interband distance were measured with a ruler (millimeter precision). OBSERVATIONS The lithological and petrophysical parameters of the host rocks and the fault rocks presented in this section are summarized in Figure 14 and Table 1 in the Appendix; the k values and their error bars are reported in Figure 5. Deformation Band Cluster, Breccia, and Fault Surface at Clashach Cove Architecture of the Clashach Fault The fault of 20–50 m (66–164 ft) of total dip-slip component is oriented N090° 70°S on average (thick black line in Figure 3D) and is a normal extensional feature. The fault zone is made of different materials from the footwall to the hanging wall (Figure 6A, B). One to two zones of tabular cataclastic deformation showing fault slip surfaces with slickensides are located along the footwall side. An oxide-cemented crush breccia juxtaposes the cataclastic deformation on the hanging wall side of the main fault slip surface (Figure 6A–C). Figure 6. Macro- and microstructural observations at the Clashach Cove site. (A) The Clashach fault zone, with the footwall (FW) on the left and the hanging wall (HW) on the right. (B) Structural interpretation of (A). (C) Cluster on the FW edge of the fault; note the slickensides of the fault. (D) Detail of the cluster section (plane-polarized light); the dashed line indicates the orientation of the deformation bands in the cluster. (e) Oxidized breccia comprising a large amount of cataclastic material scanning electron microscope of backscattered electrons; the quartz and feldspar grains appear in gray, the oxides (Ox) wrapping the grains appear white, and the porosity (φ) appears black. CSB = compactional shear band. The zones of cataclastic deformation are composed of CSBs forming a 5–20-cm (2.0–7.9-in.) -wide cluster (Figure 6C, D). The CL microscopy reveals that clasts within the cluster are sealed by quartz cements that are up to 40 μm in thickness (also see Farrell et al., 2014). The cementation is observed along the cluster and in the host sandstone near the fault, even in the yellow generally porous low-cohesion sandstone unit (gray color along the fault plane in upper part of the footwall, Figure 3C). Although the CL imaging method does not allow precise measuring of the IGV (limit of resolution), we estimate a range from 5% to 8% in the most intensely deformed and crushed part of the cluster. In situ cluster k along the strike of the cluster is 12 (±4) md. The hanging wall edge of the cluster marks the limit with the footwall edge of the tabular breccia compartment and is made of one to two low-permeability fault slip surfaces clogged with oxides (Figure 6C, D). Slickenlines are clearly visible and pitch 71°W on average on the fault plane (red spot in Figure 3D). The k values normal to the slickenside of the fault are as low as 0.56 (±0.2) md. The tabular breccia compartment consists of a 50-cm (20-in.) -wide crush breccia. The breccia (Figure 6E) is made up of a volume of more than 40% of iron oxide coatings, 13% of quartz clasts, 10% of feldspars and other weak minerals overlain by oxide coatings, the rest of the volume being porous (35%). Among the clasts, we recognize, in particular, fragments of cataclastic sandstone material and loose sand grains. Thin sections and SEM analysis show that φ ranges between 30% and 40% (also see Table 1). These values probably overestimate the φ value because some holes visible in Figure 6E were probably quartz grains plucked during the thin section preparation. The plane separating and delineating the breccia and the yellow sandstone of the hanging wall is sharp, but no clear fault slip surface is observed. Deformation Structures in the Surrounding Exposure The host sandstone in the lower part of the footwall (grayish color) consists of a cemented and highly cohesive medium-grain sandstone (moderately sorted, IGV = 33%; k = 4.8 (± 1.8) × 102 md; see also Table 1). Microscope observation shows pervasive presence of quartz cement rims around the detrital quartz grains within the formation (Figure 7A). The quartz cements around detrital quartz grains can reach 20 μm in thickness and efficiently bond the grains to each other. The feldspars are commonly rimed with carbonate cements. A few oxides are also present in minor quantities. The main deformation structures in the cemented part of the footwall are joints (Figure 7B). Their surfaces are characterized by a sharp fracture of the grains (Figure 7C) and are clogged with very late-stage iron oxides penetrating 4–5 mm (0.16–0.20 in.) into the rock. Permeabilities along-strike of the joints are beyond the device measurement limit, corresponding to k values greater than 5 × 104 md. Scarce CSBs are observed in the cemented sands. Figure 7. (A) The host rock in the cemented part of the lower footwall (in cold cathodoluminescence [CL]). The detrital quartz (Qtz) appear in bluish hues and the Qtz cement rims (cQtz, arrows) in dark blue; note the presence of detrital feldspar (Fld) grains, light blue (Fld), surrounded by carbonate cement (yellow–green) at the bottom of the image. (B) Oxidized joint in the cemented footwall as visible at the outcrop and (C) as visible in scanning electron microscope of backscattered electrons (SEM-BSE) (the Qtz grains appear gray, the feldspar grains appear light gray, the oxides appears light gray to white, and the porosity (φ) appears black); arrows point at the sharp interface of the sandstone with the joint. (D) Host rock in the uncemented hanging wall (in CL); (E) dilation band as visible at the outcrop and (F) as visible in SEM-BSE. (G) Compactional shear bands (CSBs) in the hanging wall. (H) Zoom in the CSB section (in SEM-BSE); among the cataclastic material, we note the presence of a majority of Qtz clasts (gray), a lower Flds content (light gray), and some carbonate cement clasts (also light gray) among the finest cataclastic material; the dashed line at the bottom right corner indicates the orientation of the CSB. The yellow formation in the hanging wall (and in the upper part of the footwall) is a poorly cemented and very low-cohesive medium-grain sublitharenite sandstone (moderately sorted, φ = 31%; k = 2.3 (± 0.9) × 103 md; see also Table 1). Microscope observations show no or very thin quartz cement rims around the detrital grains (Figure 7D) and a few carbonate cements around the feldspar grains, but no significant pore filling is observed between the detrital grains. In some cases, very thin iron coatings are observed in plane-polarized light around the detrital grains, giving the sandstone its pervasive yellow color. The deformation structures in the hanging wall of the fault are scarce and include dilation bands and CSBs. Dilation bands (Figure 7E) display slightly negative relief at the exposure without visible offset. They are microscopically characterized by irregular 1–2-mm (0.04–0.08-in.)-thick high-porosity (55% ± 2%) tabular zones, where the grains are disaggregated and separated from each other without being significantly fractured (Figure 7F). CSBs (Figure 7G) are 5–10 mm (0.20–0.39 in.) thick with a positive relief and have 5–20 cm (2.0–7.9 in.) of offset. Microscope observations show that these bands are composed of a large amount of cataclastic material (Figure 7H) with significantly reduced φ (8 ± 2%). The fraction of carbonate cement is observed completely crushed in the CSB. The orientation of all the CSBs and dilatant deformation structures (joints and dilation bands) is shown in Figure 3D. The strike of CSBs (thin blue lines in Figure 3D) ranges between N075° and N170°, with 47% between N075° and N100°. The majority of them (73%) dip between 60° and 80°. A majority (55%) of the dilatant structures (joints appear as thin red lines and dilation bands as thin brown lines in Figure 3D) strike between N095° and N140°; a second significant proportion (19%) strikes between N170° and N190°; and 92% of the dilatant structures dip between 70° and 90°. Although observed interactions between these deformation structures are rare, the dilatant structures crosscut and are later than the CSBs in both the footwall and hanging wall. The distribution profile of the deformation structures is displayed in Figure 8. The profile shows a trend with increasing intensity of deformation nearer the fault; that trend is clear in the footwall and is subtler in the hanging wall. The frequency of dilatant structures is higher in the footwall where they are in the form of joints in the gray-cemented sandstone layer, commonly reaching or exceeding six joints per meter near the fault. In the hanging wall, the dilation bands in the yellow incohesive sandstone never exceed two dilation bands per meter. Figure 8. Deformation profile showing the number of deformation structures around the Clashach fault. CSBs = compactional shear bands. Deformation Band Cluster, Clay Smear, and Fault Surfaces at Goblin Valley Two exposures have been studied at Goblin Valley: (1) a normal fault zone juxtaposing porous sandstones and thin poorly lithified siltstone layers, and (2) a normal fault zone incorporating a poorly lithified shale layer into the fault core within a porous sandstone formation. Exposure 1: Deformed Sandstone Including Thin, Poorly Lithified Siltstone Layers The first fault exposure is located at the base of the Entrada Formation (Figure 4B) and displays a fault with a dip-slip component of 6 m (19.7 ft) (Figure 4D). At the outcrop, the base of the footwall is made of an orange fine-grained sandstone (c5 series as shown in Figure 4B, poorly sorted, IGV = 6%, k = 5 [± 3] × 102 md) (Figures 4D, 9A) alternating with 5–20-cm (2.0–7.9-in.)-thick layers of phyllosilicate-bearing poorly lithified siltstones (poorly sorted, IGV = 18%, k = 1.7 [± 2.3] × 102 md; Figure 9B). The top of the footwall and the whole hanging wall along the studied thickness of the outcrop is made of massive clean fine-grained sandstone (e1 series as shown in Figure 4B, well-sorted, IGV = 33%, k = 2.2 [± 1.7] × 104 md; Figure 9C). The reader is directed to Table 1 for more details on the lithological data. The damage zone is approximately 5 m (∼16.4 ft) wide on both sides of the fault core. In the damage zone, the deformation is characterized by conjugate CSBs in the clean sandstones (mean spacing of 11.1 cm [4.37 in.]) and by small fault slip surfaces in both the siltstones (mean spacing of 5.7 cm [2.24 in.]) and the orange fine-grained sandstone (mean spacing of 10.0 cm [3.94 in.]). Figure 9. Scanning electron microscopy of backscattered electrons imaging of the host rocks at Goblin Valley, at the exposure 1 (A–C), and at the exposure 2 (D–F). (A) Orange fine-grained sandstone host rock (unit c5 in Figure 4D); 25% of the rock is made of oxide-bearing oolites crushed between the stronger grains. (B) Poorly lithified siltstone layer in host rock (very top of unit c5 in Figure 4D). The matrix is phyllosilicate rich. (C) Clean fine-grained sandstone host rock (unit e1 in Figure 4D). (D) The clean fine-grained sandstone host rock at the second exposure (unit e1 in Figure 4E) has not been sampled; however, because of its field similarity in macroscopic aspect, we assume it is equivalent to the clean fine-grained sandstone at the first exposure. (E) Poorly lithified shale layer host rock (unit e2 in Figure 4E). (F) Medium-grained sandstone host rock (unit e3 in Figure 4E). φ = porosity; Fld = feldspar; ool = oolite; Ox = oxide; Qtz = quartz. At the macroscopic scale, we observe a cluster of CSBs adjacent to a major fault slip surface where the clean fine-grained sandstones are juxtaposed (Figure 10). The cumulative band thickness reaches 94 mm (3.70 in.) at the topmost end of the cluster (at 6 m [19.7 ft], Figure 10A) where the thickness of the cluster reaches 600 mm (23.6 in.) and forms the total thickness of the fault core. Both the cumulative band thickness and the cluster thickness decrease where the sandstone of the hanging wall juxtaposes against the first layers of siltstone (at 4.5 m [14.8 ft], Figure 10A) and farther downward. The cluster is still adjacent to the major fault slip surface, and the cumulative band and cluster thicknesses decrease from 50 mm (2.0 in.) and 320 mm (12.6 in.), respectively, at 4.5 m (14.8 ft), down to 15 mm (0.6 in.) and 50 mm (2.0 in.), respectively, at 2.25 m (7.38 ft). In that interval, a smeared clay-rich siltstone gouge juxtaposes against the CSB cluster separated by the major fault slip surface, and both form the fault core material. The footwall side of the fault core includes a small thickness of lenses of deformed orange fine-grained sandstone (Figure 10B). Downward (from 2.25 m [7.38 ft] and lower, Figure 10A), the cumulative band thickness decreases steadily to 4 mm (0.16 in.) (two bands) at 4 m (13.1 ft) below the siltstone juxtaposition (at 0 m, Figure 10A); here, the clay-rich siltstone gouge (smear) is almost nonexistent. Hence, the clay-rich siltstone gouge forms a fault rock along nearly 4 m (13.1 ft) below the lowest siltstone layer, after which the fault core essentially disappears, leaving the hanging wall and footwall in contact along the major fault slip surface. Below this point, the fault slip surface marks out the footwall edge of the thin cataclastic deformation structure (the end of the cluster). Note that from top to bottom of the exposure, the major throughgoing fault slip surface (striated and flat, commonly smooth; Figure 10C, D) is continuous and is observed near the footwall edge of the CSB cluster. No significant mixing between the silt derived from the poorly lithified siltstone and the cluster is observed in the core zone.--> Fault surface development and fault rock juxtaposition along deformation band clusters in porous sandstones series

Fault surface development and fault rock juxtaposition along deformation band clusters in porous sandstones series

ABSTRACT

Our field study examines two sites revealing the processes responsible for fault surface development and fault rock juxtaposition along normal faults in porous-sandstone–dominated formations. At the first site, we show that a cluster of cataclastic deformation bands made in an initially poorly consolidated sandstone localizes overprinting of a fault slip surface and brecciation during a subsequent tectonic episode, after a significant diagenesis of the formation induced by burial. Because the permeability of the clusters ranges between 6 × 100 and less than 5 × 10−1 md and because the breccia was highly dilatant, we deduce that the fault behaved as a baffle for cross-fault fluid flow at an early age of the formation and as a conduit after significant diagenetic evolution and subsequent fault surface development. At the second site, we show that the presence of clay-rich layers as thin as 80 cm (31 in.) are responsible for the initiation of a major fault slip surface in the underlying and overlying sandstone. The propagation of the fault prevents further cataclastic deformation and cluster development in these sandstones. Fault displacement juxtaposes fault surfaces, clusters of cataclastic deformation bands, clay-rich gouges, and different sedimentary units. Because both fault rocks have low permeability, their spatial juxtaposition provides a continuous baffle for cross-fault fluid flow. Our study shows that fault surface localization is related to an increase in the contrast of mechanical behavior between the cluster and the adjacent material (diagenetic hardening of the cluster or softening of the clay-rich gouge). Lithological contrasts and diagenesis are favorable conditions for localizing faulting and fault rock juxtaposition, allowing significant three-dimensional anisotropy of permeability during and/or after deformation. These processes must, therefore, be considered for fault-seal analyses in sandstone reservoirs.

INTRODUCTION

Faults are the economic core of structural geology because they can constitute conduits or barriers for fluids in the context of resource exploitation or storage, because of the diversity of their morphology, and because the petrophysical properties of the fault rocks are complex (e.g., Knipe, 1997; Yielding et al., 1997; Aydin, 2000; Faulkner et al., 2010). Faults generally include one or multiple core zones comprised of fault rocks, including breccias, cataclasites, or gouges, generally separated by discrete fault slip surfaces, accommodating most of shear displacement (Sibson, 1977; Caine et al., 1996). Especially in porous sandstone, despite its importance for reservoirs compartmentalization, the processes allowing fault surface formation and subsequent fault rock juxtaposition are still not clearly understood (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Fossen et al., 2018).

In porous sandstone units (Figure 1A–C), the deformation is dominated by cataclasis (also referred to as “cataclastic flow”). This process is characterized by an important decrease in grain size when grains are rearranged by rolling and splitting because of stress concentrations at the grain–grain contacts (Gallagher et al., 1974; Aydin, 1978). The cataclastic deformation localizes as (1) compactional shear band (CSB, as defined by Fossen et al., 2018), within which fault rocks correspond to crush microbreccias and cataclasites, and (2) fault surfaces that are sharply striated (slickensides), along which the fault rocks generally correspond to ultracataclasites (e.g., Aydin, 1978; Blenkinsop, 1991; Fossen et al., 2007). Some models consider that fault surface initiation is mainly caused by cataclastic hardening occurring in clusters of CSB and suggest that the fault surface appears with the increase of cluster thickness (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Wibberley et al., 2007) (Figure 1A–C). This hypothesis is, however, strongly debated because fault surfaces are observed on a wide range of band cluster thicknesses and densities (see Figure 2 and Philit, 2017; Philit et al., 2018) and because other processes, such as segment linkage, which is not directly dependent on cluster thickness, might be responsible for fault slip surface development (Nicol et al., 2013). These processes are generally inspected in the host porous sandstone and rarely analyzed by looking at the surrounding lithology or its evolution through time. Sandstone-dominated formations comprise a variety of lithologies, each involving different mechanical processes of deformation (e.g., Aydin and Eyal, 2002; Schultz and Fossen, 2008) that may enhance or impede fault surface development and lead to a diversity of fault architecture and rocks (e.g., Knipe, 1997; Wibberley et al., 2008; Schmatz et al., 2010).

Figure 1. Synthetic schemes of the evolution of extensional deformation to faulting in different materials: (A–C) Cataclastic deformation in porous sandstone units (poorly compacted and lithified); (D–F) brecciation in lithified (e.g., compacted, cemented) sandstone units; (G–I) shale smear in alternating shale and multiple lithified rock units. Regular bold arrows indicate the deformation kinematics (extension or compression); the thin half arrows indicate the shear direction on both sides of the fault. The red circles are zones of zoom highlighting the deformation structures in the fault core.

Deformation of brittle low-porosity sandstones (Figure 1D–F) generally leads to brittle dilatant fracturing, producing breccias, protocataclasite, cataclasite, and fault surfaces as rock comminution evolves with increasing shear strain (e.g., Zhu and Wong, 1997; Woodcock and Mort, 2008). Formations bearing clay or shale units (Figure 1G–I) are likely to impede brittle failure, forming “clay gouges” deformed in a ductile manner defined as smearing (e.g., Lindsay et al., 1993; Aydin and Eyal, 2002). Few studies provide analyses of the juxtaposition of fault rocks in porous sandstones, clay units, and cemented sandstones (e.g., Foxford et al., 1998; Davatzes et al., 2005; Rawling and Goodwin, 2006), and such studies might provide interesting observations to discuss the conditions and timing of fault slip surface development. Yet, the study of the overprinting of these different mechanisms is essential to understand the complex fault core organizations resulting from the deformation and the contrasting petrophysical properties induced by the juxtaposed fault rocks (Fulljames et al., 1997; Knipe, 1997; Yielding et al., 1997; Agosta et al., 2006; Rawling and Goodwin, 2006; Childs et al., 2007).

Figure 2. Graph of cluster thickness versus band density of compactional shear band (CSB) clusters containing fault slip surfaces and CSB clusters without fault slip surfaces. The cluster thickness corresponds to the sum of band thicknesses and the interband distance. The band density corresponds to the sum of band thicknesses divided by the cluster thickness. The graph presented is synthetized and modified from Philit (2017) and Philit et al. (2018).

In this study, we focus on two sites in sandstone-dominated formations that have varying lithologies (variable porosity and cementation degree and presence of interbedded clay-rich materials), display fault slip surfaces adjacent to the cluster of CSBs, and allow juxtaposition of breccia and clay-rich gouge. We show that fault slip surfaces developing in fault cores can be influenced by (1) the cementation of porous sandstones and (2) the interplay of deformation between porous sandstone and clay-rich units. We discuss the implications of these processes on the heterogeneity of fault core composition and evaluate their impact on fluid flow during the evolution of the fault.

GEOLOGICAL SETTINGS

The Clashach Cove site is located northeast of the town of Hopeman, between the towns of Cummingston (to the west) and Lossiemouth (to the east) on the Moray Coast in Scotland (Figure 3A, B). The outcrop exposes quartz-rich sandstones from the Triassic Hopeman Sandstone Formation, buried up to 1.2–2.4 km (0.7–1.5 mi) (Quinn, 2008; Al-Hinai et al., 2008). The Hopeman Sandstone Formation is aeolian (Shotton, 1956; Glennie and Buller, 1983); the dune foresets are clearly visible in the hanging wall at the site. The sandstone is crosscut by a normal fault with a total dip-slip component of 20–50 m (66–164 ft) (Quinn, 2008) (Figure 3C). Similar to the many normal seismic-scale structures (≥20 m [≥65.6 ft] dip-slip component) in the Moray Firth, the Clashach fault is believed to have been formed with an exclusively dip-slip displacement during the pre-Jurassic and Late Jurassic rifting and possibly reactivated with an oblique displacement during the Paleocene–Eocene northeast Atlantic rifting (e.g., Roberts et al., 1990; Underhill, 1991; Al-Hinai et al., 2008). The rifting also caused the uplift of the Scottish Highlands. The fault juxtaposes gray cohesive sandstone topped by yellow sandstone of weak cohesion in the footwall to the north and the same yellow sandstone in the hanging wall to the south.

Figure 3. (A) Location of the study site at the scale of the United Kingdom: Moray Coast, Scotland. (B) Geological map and local stratigraphic column of the Clashach Cove site (red arrow) in the Lossiemouth-Burghead area, along the Moray Coast, modified after Edwards et al. (1993). (C) General view of the faulted Hopeman Sandstone (Sandst.) Formation (Fm.) at Clashach Cove. (D) Stereographic projection (lower hemisphere) of the deformation structures around the fault at the study site. Blue lines represent the cataclastic structures (compactional shear band [CSB] and CSB cluster); red lines represent joints; green lines represent dilation bands; the thick black line represents the Clashach fault, with both cataclastic and dilatant compartments; and the dashed line represents the average stratigraphy at the site.

The site of Goblin Valley is located in the San Rafael Desert, Utah, northeast of the Goblin Valley State Park, approximately 20 km (∼12.4 mi) north of the town of Hanksville (Figure 4A, B). The site displays the Middle Jurassic Entrada Formation of total thickness ranging from 76 to 160 m (249 to 525 ft) and buried up to 2.2 km (1.4 mi) (Rigby and Beus, 1987; Johansen and Fossen, 2008; Schultz and Fossen, 2008). The Entrada Formation is made of tabular fine-to-coarse–grained aeolian and foreshore sandstone, including thin layers of poorly lithified siltstone to clay, 10–30 cm (3.9–11.8 in.) thick each, at its base. The formation underwent extensional tectonics from the Oligocene linked to the basin and range extension. The possible burial at deformation ranges from 1100 to 2000 m (0.7 to 1.2 mi) (Doelling et al., 2015). At the site of Goblin Valley, the outcrop displays tens of clusters of CSBs oriented northwest-southeast where displacement is mostly dip-slip (Figure 4C). Even if commonly hidden by Quaternary wind deposits, these clusters are observed in an area more than 3 km (>1.9 mi) long and 800 m (>0.5 mi) wide (Figure 4B). We chose two exposures within this area to analyze the cataclastic and smear deformation processes. The first exposure (Figure 4D) consists of a fault with a dip-slip component of 6 m (19.7 ft) into a sandstone-dominated unit including thin layers of poorly lithified siltstones at the base of the Entrada Formation. The second exposure (Figure 4E) consists of a fault with a dip-slip component of 2.10 m (6.9 ft) into a sandstone-dominated formation including a 40-cm (15.7-in.)-thick shale layer.

Figure 4. (A) Location of the study site at the scale of Utah: San Rafael desert. (B) Geological map (modified from Doelling et al., 2015) and local stratigraphic column (modified from Schultz and Fossen, 2008) of the two study exposures (red stars), northeast of the Goblin Valley State Park. (C) Stereographic projection (lower hemisphere) of the clusters of deformation bands at the study site. (D, E) View of the 6-m (19.7-ft) fault of exposure 1 and of the 2.1-m (6.6-ft) fault of exposure 2, respectively. Fm. = Formation; SLC = Salt Lake City.

METHODOLOGY

At the Clashach Cove site, to analyze the mechanical processes and to identify the relative chronology of the deformation and characterize its diagenetic setting, we studied the macroscopic deformation and the rock microstructure as well as the properties of the host material (grain size distribution, sorting [s], porosity [φ], and permeability [k]). The deformed material was sampled in the section of the cluster of cataclastic deformation bands and in the overlying oxidized gouge. The host rock was sampled a few meters away from and on both sides of the fault plane where no macroscopic deformation is observed. Microscopic observations were conducted using cold cathodoluminescence (CL) with a BX41 Olympus microscope equipped with a CL8200 MK5-1 apparatus and scanning electron microscopy of backscattered electrons (SEM-BSE) imaging mode with an FEI Quanta 200 FEG microscope. The mineralogical compositions were obtained with a combined calibration through microscopic analyses in natural light, cross-polarized light, CL, and major element composition from SEM-BSE. A quantification of the different minerals’ proportions was obtained with gray-scale contrast levels of SEM images. Because the rock is cemented by quartz, the analysis of the grain size distribution was done through CL image measurements using the Fiji software. To quantify grain size distribution, Feret diameters of the particles of the host rock and deformed material were measured from representative sections (nearly 700 mm2 [1.085 in.2] investigated) with a resolution limit of 20 μm. The s value of the host rocks was obtained from the grain size distribution data using the formula of the inclusive graphic standard deviation (Folk and Ward, 1957). Note that spectrums of the grain size distribution and the mineralogical composition of the studied sandstones of the two sites are available in the Appendix. The φ value was obtained from two-dimensional SEM and CL image analysis. A total area of 2720 mm2 [4.216 in.2] representative of the deformation band clusters and host rocks was analyzed. An error of approximately 2% is caused by variation of φ linked to the uncertainty of the upper limit of the gray-scale threshold, corresponding to the spectral limit between φ and quartz minerals. The φ value of the host rocks after cementation is given as the intergranular volume (IGV, see Ehrenberg, 1989), where any potential postdeformation cementation is subtracted; the φ of the deformation structures is given as residual φ, accounting for potential postdeformation cementation. We measured the k value at the surface of the host rocks and deformation structures with a portable air-permeameter device (TinyPerm II NER). The resolution of the device does not permit the quantification of k higher than 5 × 104 md. For each type of rock surface, k has been measured at least twice for measurements exceeding 10 min, or more for shorter measurements. This time limitation constrains the minimal k quantification to approximately 5 × 10−1 md. Also, note that the TinyPerm II NER permeameter has been calibrated by Fossen et al. (2011) for porous sandstones revealing an overestimate of permeability of 1.8 times compared with image and routine gas-based permeability measures. This suggests that the values presented in this paper might be considered overestimated by approximately a factor of 2 in addition to the error bars of the data presented for the fault and host rocks in Figure 5.

Figure 5. Graph summarizing the permeability of the host rocks (HR) and their resulting fault rocks as a function of the deformation mechanisms. The square data are the measurements at the Clashach Cove site; the circle data illustrate the measurements at the Goblin Valley site. The gray scale and thickness of the frame lines reflect the grain size of the HRs: light-gray and thick line: medium-grain sandstone; medium-gray and regular line: fine-grain sandstone; dark gray without line: siltstone; black: clay. The dashed circle with the question mark illustrates that we do not know how the process of clay smear alters the permeability. ∗ = range of typical values of shale HRs are given for reference as reported by Neuzil (1994).

At Goblin Valley, we focused on (1) a macroscopic description of the juxtaposition of different types of rocks in the fault cores, (2) a petrophysical and mineralogical analysis of the pristine juxtaposed host rocks and layers properties, (3) a detailed measurement of the amount of cataclastic deformation in the fault cores, and (4) k measurements using the same methodology as described before. The analysis of the grain size distribution was done through Coulter LS 13320 laboratory laser diffraction on selected samples using four types of host rocks in the Entrada Formation; the grain size of other host rocks was estimated in the field using the grain size reference chart of the American Stratigraphic Company. The φ and k values were measured with the same methodology as described before.

We particularly focused on the analysis of clusters of CSBs. In the absence of an accurate definition of “cluster” in the literature, we propose to combine three typical characteristics described in the literature (a minimum band number, a typical frequency, and band thickness) to define it. A “cluster of CSBs” (or cluster) corresponds to the grouping of two or more bands (as used in Fossen and Bale, 2007) and accounts for at least 10% of the bands’ thickness (based on a usual frequency of 20–30 bands/m [6–9 bands/ft] reported by Johansen and Fossen, 2008, and a typical band thickness of approximately 4 mm [∼0.16 in.] as shown in Wibberley et al., 2007). Clusters may or may not include a fault slip surface (Figure 2). Such clusters typically have widths of 1 cm to several decimeters (0.4 in. to a couple of inches) (e.g., Wibberley et al., 2007) and accommodate centimeter to several decimeter (0.4 in. to a couple of inches) displacements (e.g., Antonellini and Aydin, 1994). From this definition, we introduce the cumulative band thickness and the cluster thickness to evaluate the deformation within clusters, and we ignore the interband deformation, which is small but not zero (Lanatà, 2014). The band thickness and the interband distance were measured with a ruler (millimeter precision).

OBSERVATIONS

The lithological and petrophysical parameters of the host rocks and the fault rocks presented in this section are summarized in Figure 14 and Table 1 in the Appendix; the k values and their error bars are reported in Figure 5.

Deformation Band Cluster, Breccia, and Fault Surface at Clashach Cove

Architecture of the Clashach Fault

The fault of 20–50 m (66–164 ft) of total dip-slip component is oriented N090° 70°S on average (thick black line in Figure 3D) and is a normal extensional feature. The fault zone is made of different materials from the footwall to the hanging wall (Figure 6A, B). One to two zones of tabular cataclastic deformation showing fault slip surfaces with slickensides are located along the footwall side. An oxide-cemented crush breccia juxtaposes the cataclastic deformation on the hanging wall side of the main fault slip surface (Figure 6A–C).

Figure 6. Macro- and microstructural observations at the Clashach Cove site. (A) The Clashach fault zone, with the footwall (FW) on the left and the hanging wall (HW) on the right. (B) Structural interpretation of (A). (C) Cluster on the FW edge of the fault; note the slickensides of the fault. (D) Detail of the cluster section (plane-polarized light); the dashed line indicates the orientation of the deformation bands in the cluster. (e) Oxidized breccia comprising a large amount of cataclastic material scanning electron microscope of backscattered electrons; the quartz and feldspar grains appear in gray, the oxides (Ox) wrapping the grains appear white, and the porosity (φ) appears black. CSB = compactional shear band.

The zones of cataclastic deformation are composed of CSBs forming a 5–20-cm (2.0–7.9-in.) -wide cluster (Figure 6C, D). The CL microscopy reveals that clasts within the cluster are sealed by quartz cements that are up to 40 μm in thickness (also see Farrell et al., 2014). The cementation is observed along the cluster and in the host sandstone near the fault, even in the yellow generally porous low-cohesion sandstone unit (gray color along the fault plane in upper part of the footwall, Figure 3C). Although the CL imaging method does not allow precise measuring of the IGV (limit of resolution), we estimate a range from 5% to 8% in the most intensely deformed and crushed part of the cluster. In situ cluster k along the strike of the cluster is 12 (±4) md. The hanging wall edge of the cluster marks the limit with the footwall edge of the tabular breccia compartment and is made of one to two low-permeability fault slip surfaces clogged with oxides (Figure 6C, D). Slickenlines are clearly visible and pitch 71°W on average on the fault plane (red spot in Figure 3D). The k values normal to the slickenside of the fault are as low as 0.56 (±0.2) md.

The tabular breccia compartment consists of a 50-cm (20-in.) -wide crush breccia. The breccia (Figure 6E) is made up of a volume of more than 40% of iron oxide coatings, 13% of quartz clasts, 10% of feldspars and other weak minerals overlain by oxide coatings, the rest of the volume being porous (35%). Among the clasts, we recognize, in particular, fragments of cataclastic sandstone material and loose sand grains. Thin sections and SEM analysis show that φ ranges between 30% and 40% (also see Table 1). These values probably overestimate the φ value because some holes visible in Figure 6E were probably quartz grains plucked during the thin section preparation. The plane separating and delineating the breccia and the yellow sandstone of the hanging wall is sharp, but no clear fault slip surface is observed.

Deformation Structures in the Surrounding Exposure

The host sandstone in the lower part of the footwall (grayish color) consists of a cemented and highly cohesive medium-grain sandstone (moderately sorted, IGV = 33%; k = 4.8 (± 1.8) × 102 md; see also Table 1). Microscope observation shows pervasive presence of quartz cement rims around the detrital quartz grains within the formation (Figure 7A). The quartz cements around detrital quartz grains can reach 20 μm in thickness and efficiently bond the grains to each other. The feldspars are commonly rimed with carbonate cements. A few oxides are also present in minor quantities. The main deformation structures in the cemented part of the footwall are joints (Figure 7B). Their surfaces are characterized by a sharp fracture of the grains (Figure 7C) and are clogged with very late-stage iron oxides penetrating 4–5 mm (0.16–0.20 in.) into the rock. Permeabilities along-strike of the joints are beyond the device measurement limit, corresponding to k values greater than 5 × 104 md. Scarce CSBs are observed in the cemented sands.

Figure 7. (A) The host rock in the cemented part of the lower footwall (in cold cathodoluminescence [CL]). The detrital quartz (Qtz) appear in bluish hues and the Qtz cement rims (cQtz, arrows) in dark blue; note the presence of detrital feldspar (Fld) grains, light blue (Fld), surrounded by carbonate cement (yellow–green) at the bottom of the image. (B) Oxidized joint in the cemented footwall as visible at the outcrop and (C) as visible in scanning electron microscope of backscattered electrons (SEM-BSE) (the Qtz grains appear gray, the feldspar grains appear light gray, the oxides appears light gray to white, and the porosity (φ) appears black); arrows point at the sharp interface of the sandstone with the joint. (D) Host rock in the uncemented hanging wall (in CL); (E) dilation band as visible at the outcrop and (F) as visible in SEM-BSE. (G) Compactional shear bands (CSBs) in the hanging wall. (H) Zoom in the CSB section (in SEM-BSE); among the cataclastic material, we note the presence of a majority of Qtz clasts (gray), a lower Flds content (light gray), and some carbonate cement clasts (also light gray) among the finest cataclastic material; the dashed line at the bottom right corner indicates the orientation of the CSB.

The yellow formation in the hanging wall (and in the upper part of the footwall) is a poorly cemented and very low-cohesive medium-grain sublitharenite sandstone (moderately sorted, φ = 31%; k = 2.3 (± 0.9) × 103 md; see also Table 1). Microscope observations show no or very thin quartz cement rims around the detrital grains (Figure 7D) and a few carbonate cements around the feldspar grains, but no significant pore filling is observed between the detrital grains. In some cases, very thin iron coatings are observed in plane-polarized light around the detrital grains, giving the sandstone its pervasive yellow color. The deformation structures in the hanging wall of the fault are scarce and include dilation bands and CSBs. Dilation bands (Figure 7E) display slightly negative relief at the exposure without visible offset. They are microscopically characterized by irregular 1–2-mm (0.04–0.08-in.)-thick high-porosity (55% ± 2%) tabular zones, where the grains are disaggregated and separated from each other without being significantly fractured (Figure 7F). CSBs (Figure 7G) are 5–10 mm (0.20–0.39 in.) thick with a positive relief and have 5–20 cm (2.0–7.9 in.) of offset. Microscope observations show that these bands are composed of a large amount of cataclastic material (Figure 7H) with significantly reduced φ (8 ± 2%). The fraction of carbonate cement is observed completely crushed in the CSB.

The orientation of all the CSBs and dilatant deformation structures (joints and dilation bands) is shown in Figure 3D. The strike of CSBs (thin blue lines in Figure 3D) ranges between N075° and N170°, with 47% between N075° and N100°. The majority of them (73%) dip between 60° and 80°. A majority (55%) of the dilatant structures (joints appear as thin red lines and dilation bands as thin brown lines in Figure 3D) strike between N095° and N140°; a second significant proportion (19%) strikes between N170° and N190°; and 92% of the dilatant structures dip between 70° and 90°. Although observed interactions between these deformation structures are rare, the dilatant structures crosscut and are later than the CSBs in both the footwall and hanging wall. The distribution profile of the deformation structures is displayed in Figure 8. The profile shows a trend with increasing intensity of deformation nearer the fault; that trend is clear in the footwall and is subtler in the hanging wall. The frequency of dilatant structures is higher in the footwall where they are in the form of joints in the gray-cemented sandstone layer, commonly reaching or exceeding six joints per meter near the fault. In the hanging wall, the dilation bands in the yellow incohesive sandstone never exceed two dilation bands per meter.

Figure 8. Deformation profile showing the number of deformation structures around the Clashach fault. CSBs = compactional shear bands.

Deformation Band Cluster, Clay Smear, and Fault Surfaces at Goblin Valley

Two exposures have been studied at Goblin Valley: (1) a normal fault zone juxtaposing porous sandstones and thin poorly lithified siltstone layers, and (2) a normal fault zone incorporating a poorly lithified shale layer into the fault core within a porous sandstone formation.

Exposure 1: Deformed Sandstone Including Thin, Poorly Lithified Siltstone Layers

The first fault exposure is located at the base of the Entrada Formation (Figure 4B) and displays a fault with a dip-slip component of 6 m (19.7 ft) (Figure 4D). At the outcrop, the base of the footwall is made of an orange fine-grained sandstone (c5 series as shown in Figure 4B, poorly sorted, IGV = 6%, k = 5 [± 3] × 102 md) (Figures 4D, 9A) alternating with 5–20-cm (2.0–7.9-in.)-thick layers of phyllosilicate-bearing poorly lithified siltstones (poorly sorted, IGV = 18%, k = 1.7 [± 2.3] × 102 md; Figure 9B). The top of the footwall and the whole hanging wall along the studied thickness of the outcrop is made of massive clean fine-grained sandstone (e1 series as shown in Figure 4B, well-sorted, IGV = 33%, k = 2.2 [± 1.7] × 104 md; Figure 9C). The reader is directed to Table 1 for more details on the lithological data. The damage zone is approximately 5 m (∼16.4 ft) wide on both sides of the fault core. In the damage zone, the deformation is characterized by conjugate CSBs in the clean sandstones (mean spacing of 11.1 cm [4.37 in.]) and by small fault slip surfaces in both the siltstones (mean spacing of 5.7 cm [2.24 in.]) and the orange fine-grained sandstone (mean spacing of 10.0 cm [3.94 in.]).

Figure 9. Scanning electron microscopy of backscattered electrons imaging of the host rocks at Goblin Valley, at the exposure 1 (A–C), and at the exposure 2 (D–F). (A) Orange fine-grained sandstone host rock (unit c5 in Figure 4D); 25% of the rock is made of oxide-bearing oolites crushed between the stronger grains. (B) Poorly lithified siltstone layer in host rock (very top of unit c5 in Figure 4D). The matrix is phyllosilicate rich. (C) Clean fine-grained sandstone host rock (unit e1 in Figure 4D). (D) The clean fine-grained sandstone host rock at the second exposure (unit e1 in Figure 4E) has not been sampled; however, because of its field similarity in macroscopic aspect, we assume it is equivalent to the clean fine-grained sandstone at the first exposure. (E) Poorly lithified shale layer host rock (unit e2 in Figure 4E). (F) Medium-grained sandstone host rock (unit e3 in Figure 4E). φ = porosity; Fld = feldspar; ool = oolite; Ox = oxide; Qtz = quartz.

At the macroscopic scale, we observe a cluster of CSBs adjacent to a major fault slip surface where the clean fine-grained sandstones are juxtaposed (Figure 10). The cumulative band thickness reaches 94 mm (3.70 in.) at the topmost end of the cluster (at 6 m [19.7 ft], Figure 10A) where the thickness of the cluster reaches 600 mm (23.6 in.) and forms the total thickness of the fault core. Both the cumulative band thickness and the cluster thickness decrease where the sandstone of the hanging wall juxtaposes against the first layers of siltstone (at 4.5 m [14.8 ft], Figure 10A) and farther downward. The cluster is still adjacent to the major fault slip surface, and the cumulative band and cluster thicknesses decrease from 50 mm (2.0 in.) and 320 mm (12.6 in.), respectively, at 4.5 m (14.8 ft), down to 15 mm (0.6 in.) and 50 mm (2.0 in.), respectively, at 2.25 m (7.38 ft). In that interval, a smeared clay-rich siltstone gouge juxtaposes against the CSB cluster separated by the major fault slip surface, and both form the fault core material. The footwall side of the fault core includes a small thickness of lenses of deformed orange fine-grained sandstone (Figure 10B). Downward (from 2.25 m [7.38 ft] and lower, Figure 10A), the cumulative band thickness decreases steadily to 4 mm (0.16 in.) (two bands) at 4 m (13.1 ft) below the siltstone juxtaposition (at 0 m, Figure 10A); here, the clay-rich siltstone gouge (smear) is almost nonexistent. Hence, the clay-rich siltstone gouge forms a fault rock along nearly 4 m (13.1 ft) below the lowest siltstone layer, after which the fault core essentially disappears, leaving the hanging wall and footwall in contact along the major fault slip surface. Below this point, the fault slip surface marks out the footwall edge of the thin cataclastic deformation structure (the end of the cluster). Note that from top to bottom of the exposure, the major throughgoing fault slip surface (striated and flat, commonly smooth; Figure 10C, D) is continuous and is observed near the footwall edge of the CSB cluster. No significant mixing between the silt derived from the poorly lithified siltstone and the cluster is observed in the core zone.

Please log in to read the full article

ABSTRACT

Our field study examines two sites revealing the processes responsible for fault surface development and fault rock juxtaposition along normal faults in porous-sandstone–dominated formations. At the first site, we show that a cluster of cataclastic deformation bands made in an initially poorly consolidated sandstone localizes overprinting of a fault slip surface and brecciation during a subsequent tectonic episode, after a significant diagenesis of the formation induced by burial. Because the permeability of the clusters ranges between 6 × 100 and less than 5 × 10−1 md and because the breccia was highly dilatant, we deduce that the fault behaved as a baffle for cross-fault fluid flow at an early age of the formation and as a conduit after significant diagenetic evolution and subsequent fault surface development. At the second site, we show that the presence of clay-rich layers as thin as 80 cm (31 in.) are responsible for the initiation of a major fault slip surface in the underlying and overlying sandstone. The propagation of the fault prevents further cataclastic deformation and cluster development in these sandstones. Fault displacement juxtaposes fault surfaces, clusters of cataclastic deformation bands, clay-rich gouges, and different sedimentary units. Because both fault rocks have low permeability, their spatial juxtaposition provides a continuous baffle for cross-fault fluid flow. Our study shows that fault surface localization is related to an increase in the contrast of mechanical behavior between the cluster and the adjacent material (diagenetic hardening of the cluster or softening of the clay-rich gouge). Lithological contrasts and diagenesis are favorable conditions for localizing faulting and fault rock juxtaposition, allowing significant three-dimensional anisotropy of permeability during and/or after deformation. These processes must, therefore, be considered for fault-seal analyses in sandstone reservoirs.

INTRODUCTION

Faults are the economic core of structural geology because they can constitute conduits or barriers for fluids in the context of resource exploitation or storage, because of the diversity of their morphology, and because the petrophysical properties of the fault rocks are complex (e.g., Knipe, 1997; Yielding et al., 1997; Aydin, 2000; Faulkner et al., 2010). Faults generally include one or multiple core zones comprised of fault rocks, including breccias, cataclasites, or gouges, generally separated by discrete fault slip surfaces, accommodating most of shear displacement (Sibson, 1977; Caine et al., 1996). Especially in porous sandstone, despite its importance for reservoirs compartmentalization, the processes allowing fault surface formation and subsequent fault rock juxtaposition are still not clearly understood (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Fossen et al., 2018).

In porous sandstone units (Figure 1A–C), the deformation is dominated by cataclasis (also referred to as “cataclastic flow”). This process is characterized by an important decrease in grain size when grains are rearranged by rolling and splitting because of stress concentrations at the grain–grain contacts (Gallagher et al., 1974; Aydin, 1978). The cataclastic deformation localizes as (1) compactional shear band (CSB, as defined by Fossen et al., 2018), within which fault rocks correspond to crush microbreccias and cataclasites, and (2) fault surfaces that are sharply striated (slickensides), along which the fault rocks generally correspond to ultracataclasites (e.g., Aydin, 1978; Blenkinsop, 1991; Fossen et al., 2007). Some models consider that fault surface initiation is mainly caused by cataclastic hardening occurring in clusters of CSB and suggest that the fault surface appears with the increase of cluster thickness (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Wibberley et al., 2007) (Figure 1A–C). This hypothesis is, however, strongly debated because fault surfaces are observed on a wide range of band cluster thicknesses and densities (see Figure 2 and Philit, 2017; Philit et al., 2018) and because other processes, such as segment linkage, which is not directly dependent on cluster thickness, might be responsible for fault slip surface development (Nicol et al., 2013). These processes are generally inspected in the host porous sandstone and rarely analyzed by looking at the surrounding lithology or its evolution through time. Sandstone-dominated formations comprise a variety of lithologies, each involving different mechanical processes of deformation (e.g., Aydin and Eyal, 2002; Schultz and Fossen, 2008) that may enhance or impede fault surface development and lead to a diversity of fault architecture and rocks (e.g., Knipe, 1997; Wibberley et al., 2008; Schmatz et al., 2010).

Figure 1. Synthetic schemes of the evolution of extensional deformation to faulting in different materials: (A–C) Cataclastic deformation in porous sandstone units (poorly compacted and lithified); (D–F) brecciation in lithified (e.g., compacted, cemented) sandstone units; (G–I) shale smear in alternating shale and multiple lithified rock units. Regular bold arrows indicate the deformation kinematics (extension or compression); the thin half arrows indicate the shear direction on both sides of the fault. The red circles are zones of zoom highlighting the deformation structures in the fault core.

Deformation of brittle low-porosity sandstones (Figure 1D–F) generally leads to brittle dilatant fracturing, producing breccias, protocataclasite, cataclasite, and fault surfaces as rock comminution evolves with increasing shear strain (e.g., Zhu and Wong, 1997; Woodcock and Mort, 2008). Formations bearing clay or shale units (Figure 1G–I) are likely to impede brittle failure, forming “clay gouges” deformed in a ductile manner defined as smearing (e.g., Lindsay et al., 1993; Aydin and Eyal, 2002). Few studies provide analyses of the juxtaposition of fault rocks in porous sandstones, clay units, and cemented sandstones (e.g., Foxford et al., 1998; Davatzes et al., 2005; Rawling and Goodwin, 2006), and such studies might provide interesting observations to discuss the conditions and timing of fault slip surface development. Yet, the study of the overprinting of these different mechanisms is essential to understand the complex fault core organizations resulting from the deformation and the contrasting petrophysical properties induced by the juxtaposed fault rocks (Fulljames et al., 1997; Knipe, 1997; Yielding et al., 1997; Agosta et al., 2006; Rawling and Goodwin, 2006; Childs et al., 2007).

Figure 2. Graph of cluster thickness versus band density of compactional shear band (CSB) clusters containing fault slip surfaces and CSB clusters without fault slip surfaces. The cluster thickness corresponds to the sum of band thicknesses and the interband distance. The band density corresponds to the sum of band thicknesses divided by the cluster thickness. The graph presented is synthetized and modified from Philit (2017) and Philit et al. (2018).

In this study, we focus on two sites in sandstone-dominated formations that have varying lithologies (variable porosity and cementation degree and presence of interbedded clay-rich materials), display fault slip surfaces adjacent to the cluster of CSBs, and allow juxtaposition of breccia and clay-rich gouge. We show that fault slip surfaces developing in fault cores can be influenced by (1) the cementation of porous sandstones and (2) the interplay of deformation between porous sandstone and clay-rich units. We discuss the implications of these processes on the heterogeneity of fault core composition and evaluate their impact on fluid flow during the evolution of the fault.

GEOLOGICAL SETTINGS

The Clashach Cove site is located northeast of the town of Hopeman, between the towns of Cummingston (to the west) and Lossiemouth (to the east) on the Moray Coast in Scotland (Figure 3A, B). The outcrop exposes quartz-rich sandstones from the Triassic Hopeman Sandstone Formation, buried up to 1.2–2.4 km (0.7–1.5 mi) (Quinn, 2008; Al-Hinai et al., 2008). The Hopeman Sandstone Formation is aeolian (Shotton, 1956; Glennie and Buller, 1983); the dune foresets are clearly visible in the hanging wall at the site. The sandstone is crosscut by a normal fault with a total dip-slip component of 20–50 m (66–164 ft) (Quinn, 2008) (Figure 3C). Similar to the many normal seismic-scale structures (≥20 m [≥65.6 ft] dip-slip component) in the Moray Firth, the Clashach fault is believed to have been formed with an exclusively dip-slip displacement during the pre-Jurassic and Late Jurassic rifting and possibly reactivated with an oblique displacement during the Paleocene–Eocene northeast Atlantic rifting (e.g., Roberts et al., 1990; Underhill, 1991; Al-Hinai et al., 2008). The rifting also caused the uplift of the Scottish Highlands. The fault juxtaposes gray cohesive sandstone topped by yellow sandstone of weak cohesion in the footwall to the north and the same yellow sandstone in the hanging wall to the south.

Figure 3. (A) Location of the study site at the scale of the United Kingdom: Moray Coast, Scotland. (B) Geological map and local stratigraphic column of the Clashach Cove site (red arrow) in the Lossiemouth-Burghead area, along the Moray Coast, modified after Edwards et al. (1993). (C) General view of the faulted Hopeman Sandstone (Sandst.) Formation (Fm.) at Clashach Cove. (D) Stereographic projection (lower hemisphere) of the deformation structures around the fault at the study site. Blue lines represent the cataclastic structures (compactional shear band [CSB] and CSB cluster); red lines represent joints; green lines represent dilation bands; the thick black line represents the Clashach fault, with both cataclastic and dilatant compartments; and the dashed line represents the average stratigraphy at the site.

The site of Goblin Valley is located in the San Rafael Desert, Utah, northeast of the Goblin Valley State Park, approximately 20 km (∼12.4 mi) north of the town of Hanksville (Figure 4A, B). The site displays the Middle Jurassic Entrada Formation of total thickness ranging from 76 to 160 m (249 to 525 ft) and buried up to 2.2 km (1.4 mi) (Rigby and Beus, 1987; Johansen and Fossen, 2008; Schultz and Fossen, 2008). The Entrada Formation is made of tabular fine-to-coarse–grained aeolian and foreshore sandstone, including thin layers of poorly lithified siltstone to clay, 10–30 cm (3.9–11.8 in.) thick each, at its base. The formation underwent extensional tectonics from the Oligocene linked to the basin and range extension. The possible burial at deformation ranges from 1100 to 2000 m (0.7 to 1.2 mi) (Doelling et al., 2015). At the site of Goblin Valley, the outcrop displays tens of clusters of CSBs oriented northwest-southeast where displacement is mostly dip-slip (Figure 4C). Even if commonly hidden by Quaternary wind deposits, these clusters are observed in an area more than 3 km (>1.9 mi) long and 800 m (>0.5 mi) wide (Figure 4B). We chose two exposures within this area to analyze the cataclastic and smear deformation processes. The first exposure (Figure 4D) consists of a fault with a dip-slip component of 6 m (19.7 ft) into a sandstone-dominated unit including thin layers of poorly lithified siltstones at the base of the Entrada Formation. The second exposure (Figure 4E) consists of a fault with a dip-slip component of 2.10 m (6.9 ft) into a sandstone-dominated formation including a 40-cm (15.7-in.)-thick shale layer.

Figure 4. (A) Location of the study site at the scale of Utah: San Rafael desert. (B) Geological map (modified from Doelling et al., 2015) and local stratigraphic column (modified from Schultz and Fossen, 2008) of the two study exposures (red stars), northeast of the Goblin Valley State Park. (C) Stereographic projection (lower hemisphere) of the clusters of deformation bands at the study site. (D, E) View of the 6-m (19.7-ft) fault of exposure 1 and of the 2.1-m (6.6-ft) fault of exposure 2, respectively. Fm. = Formation; SLC = Salt Lake City.

METHODOLOGY

At the Clashach Cove site, to analyze the mechanical processes and to identify the relative chronology of the deformation and characterize its diagenetic setting, we studied the macroscopic deformation and the rock microstructure as well as the properties of the host material (grain size distribution, sorting [s], porosity [φ], and permeability [k]). The deformed material was sampled in the section of the cluster of cataclastic deformation bands and in the overlying oxidized gouge. The host rock was sampled a few meters away from and on both sides of the fault plane where no macroscopic deformation is observed. Microscopic observations were conducted using cold cathodoluminescence (CL) with a BX41 Olympus microscope equipped with a CL8200 MK5-1 apparatus and scanning electron microscopy of backscattered electrons (SEM-BSE) imaging mode with an FEI Quanta 200 FEG microscope. The mineralogical compositions were obtained with a combined calibration through microscopic analyses in natural light, cross-polarized light, CL, and major element composition from SEM-BSE. A quantification of the different minerals’ proportions was obtained with gray-scale contrast levels of SEM images. Because the rock is cemented by quartz, the analysis of the grain size distribution was done through CL image measurements using the Fiji software. To quantify grain size distribution, Feret diameters of the particles of the host rock and deformed material were measured from representative sections (nearly 700 mm2 [1.085 in.2] investigated) with a resolution limit of 20 μm. The s value of the host rocks was obtained from the grain size distribution data using the formula of the inclusive graphic standard deviation (Folk and Ward, 1957). Note that spectrums of the grain size distribution and the mineralogical composition of the studied sandstones of the two sites are available in the Appendix. The φ value was obtained from two-dimensional SEM and CL image analysis. A total area of 2720 mm2 [4.216 in.2] representative of the deformation band clusters and host rocks was analyzed. An error of approximately 2% is caused by variation of φ linked to the uncertainty of the upper limit of the gray-scale threshold, corresponding to the spectral limit between φ and quartz minerals. The φ value of the host rocks after cementation is given as the intergranular volume (IGV, see Ehrenberg, 1989), where any potential postdeformation cementation is subtracted; the φ of the deformation structures is given as residual φ, accounting for potential postdeformation cementation. We measured the k value at the surface of the host rocks and deformation structures with a portable air-permeameter device (TinyPerm II NER). The resolution of the device does not permit the quantification of k higher than 5 × 104 md. For each type of rock surface, k has been measured at least twice for measurements exceeding 10 min, or more for shorter measurements. This time limitation constrains the minimal k quantification to approximately 5 × 10−1 md. Also, note that the TinyPerm II NER permeameter has been calibrated by Fossen et al. (2011) for porous sandstones revealing an overestimate of permeability of 1.8 times compared with image and routine gas-based permeability measures. This suggests that the values presented in this paper might be considered overestimated by approximately a factor of 2 in addition to the error bars of the data presented for the fault and host rocks in Figure 5.

Figure 5. Graph summarizing the permeability of the host rocks (HR) and their resulting fault rocks as a function of the deformation mechanisms. The square data are the measurements at the Clashach Cove site; the circle data illustrate the measurements at the Goblin Valley site. The gray scale and thickness of the frame lines reflect the grain size of the HRs: light-gray and thick line: medium-grain sandstone; medium-gray and regular line: fine-grain sandstone; dark gray without line: siltstone; black: clay. The dashed circle with the question mark illustrates that we do not know how the process of clay smear alters the permeability. ∗ = range of typical values of shale HRs are given for reference as reported by Neuzil (1994).

At Goblin Valley, we focused on (1) a macroscopic description of the juxtaposition of different types of rocks in the fault cores, (2) a petrophysical and mineralogical analysis of the pristine juxtaposed host rocks and layers properties, (3) a detailed measurement of the amount of cataclastic deformation in the fault cores, and (4) k measurements using the same methodology as described before. The analysis of the grain size distribution was done through Coulter LS 13320 laboratory laser diffraction on selected samples using four types of host rocks in the Entrada Formation; the grain size of other host rocks was estimated in the field using the grain size reference chart of the American Stratigraphic Company. The φ and k values were measured with the same methodology as described before.

We particularly focused on the analysis of clusters of CSBs. In the absence of an accurate definition of “cluster” in the literature, we propose to combine three typical characteristics described in the literature (a minimum band number, a typical frequency, and band thickness) to define it. A “cluster of CSBs” (or cluster) corresponds to the grouping of two or more bands (as used in Fossen and Bale, 2007) and accounts for at least 10% of the bands’ thickness (based on a usual frequency of 20–30 bands/m [6–9 bands/ft] reported by Johansen and Fossen, 2008, and a typical band thickness of approximately 4 mm [∼0.16 in.] as shown in Wibberley et al., 2007). Clusters may or may not include a fault slip surface (Figure 2). Such clusters typically have widths of 1 cm to several decimeters (0.4 in. to a couple of inches) (e.g., Wibberley et al., 2007) and accommodate centimeter to several decimeter (0.4 in. to a couple of inches) displacements (e.g., Antonellini and Aydin, 1994). From this definition, we introduce the cumulative band thickness and the cluster thickness to evaluate the deformation within clusters, and we ignore the interband deformation, which is small but not zero (Lanatà, 2014). The band thickness and the interband distance were measured with a ruler (millimeter precision).

OBSERVATIONS

The lithological and petrophysical parameters of the host rocks and the fault rocks presented in this section are summarized in Figure 14 and Table 1 in the Appendix; the k values and their error bars are reported in Figure 5.

Deformation Band Cluster, Breccia, and Fault Surface at Clashach Cove

Architecture of the Clashach Fault

The fault of 20–50 m (66–164 ft) of total dip-slip component is oriented N090° 70°S on average (thick black line in Figure 3D) and is a normal extensional feature. The fault zone is made of different materials from the footwall to the hanging wall (Figure 6A, B). One to two zones of tabular cataclastic deformation showing fault slip surfaces with slickensides are located along the footwall side. An oxide-cemented crush breccia juxtaposes the cataclastic deformation on the hanging wall side of the main fault slip surface (Figure 6A–C).

Figure 6. Macro- and microstructural observations at the Clashach Cove site. (A) The Clashach fault zone, with the footwall (FW) on the left and the hanging wall (HW) on the right. (B) Structural interpretation of (A). (C) Cluster on the FW edge of the fault; note the slickensides of the fault. (D) Detail of the cluster section (plane-polarized light); the dashed line indicates the orientation of the deformation bands in the cluster. (e) Oxidized breccia comprising a large amount of cataclastic material scanning electron microscope of backscattered electrons; the quartz and feldspar grains appear in gray, the oxides (Ox) wrapping the grains appear white, and the porosity (φ) appears black. CSB = compactional shear band.

The zones of cataclastic deformation are composed of CSBs forming a 5–20-cm (2.0–7.9-in.) -wide cluster (Figure 6C, D). The CL microscopy reveals that clasts within the cluster are sealed by quartz cements that are up to 40 μm in thickness (also see Farrell et al., 2014). The cementation is observed along the cluster and in the host sandstone near the fault, even in the yellow generally porous low-cohesion sandstone unit (gray color along the fault plane in upper part of the footwall, Figure 3C). Although the CL imaging method does not allow precise measuring of the IGV (limit of resolution), we estimate a range from 5% to 8% in the most intensely deformed and crushed part of the cluster. In situ cluster k along the strike of the cluster is 12 (±4) md. The hanging wall edge of the cluster marks the limit with the footwall edge of the tabular breccia compartment and is made of one to two low-permeability fault slip surfaces clogged with oxides (Figure 6C, D). Slickenlines are clearly visible and pitch 71°W on average on the fault plane (red spot in Figure 3D). The k values normal to the slickenside of the fault are as low as 0.56 (±0.2) md.

The tabular breccia compartment consists of a 50-cm (20-in.) -wide crush breccia. The breccia (Figure 6E) is made up of a volume of more than 40% of iron oxide coatings, 13% of quartz clasts, 10% of feldspars and other weak minerals overlain by oxide coatings, the rest of the volume being porous (35%). Among the clasts, we recognize, in particular, fragments of cataclastic sandstone material and loose sand grains. Thin sections and SEM analysis show that φ ranges between 30% and 40% (also see Table 1). These values probably overestimate the φ value because some holes visible in Figure 6E were probably quartz grains plucked during the thin section preparation. The plane separating and delineating the breccia and the yellow sandstone of the hanging wall is sharp, but no clear fault slip surface is observed.

Deformation Structures in the Surrounding Exposure

The host sandstone in the lower part of the footwall (grayish color) consists of a cemented and highly cohesive medium-grain sandstone (moderately sorted, IGV = 33%; k = 4.8 (± 1.8) × 102 md; see also Table 1). Microscope observation shows pervasive presence of quartz cement rims around the detrital quartz grains within the formation (Figure 7A). The quartz cements around detrital quartz grains can reach 20 μm in thickness and efficiently bond the grains to each other. The feldspars are commonly rimed with carbonate cements. A few oxides are also present in minor quantities. The main deformation structures in the cemented part of the footwall are joints (Figure 7B). Their surfaces are characterized by a sharp fracture of the grains (Figure 7C) and are clogged with very late-stage iron oxides penetrating 4–5 mm (0.16–0.20 in.) into the rock. Permeabilities along-strike of the joints are beyond the device measurement limit, corresponding to k values greater than 5 × 104 md. Scarce CSBs are observed in the cemented sands.

Figure 7. (A) The host rock in the cemented part of the lower footwall (in cold cathodoluminescence [CL]). The detrital quartz (Qtz) appear in bluish hues and the Qtz cement rims (cQtz, arrows) in dark blue; note the presence of detrital feldspar (Fld) grains, light blue (Fld), surrounded by carbonate cement (yellow–green) at the bottom of the image. (B) Oxidized joint in the cemented footwall as visible at the outcrop and (C) as visible in scanning electron microscope of backscattered electrons (SEM-BSE) (the Qtz grains appear gray, the feldspar grains appear light gray, the oxides appears light gray to white, and the porosity (φ) appears black); arrows point at the sharp interface of the sandstone with the joint. (D) Host rock in the uncemented hanging wall (in CL); (E) dilation band as visible at the outcrop and (F) as visible in SEM-BSE. (G) Compactional shear bands (CSBs) in the hanging wall. (H) Zoom in the CSB section (in SEM-BSE); among the cataclastic material, we note the presence of a majority of Qtz clasts (gray), a lower Flds content (light gray), and some carbonate cement clasts (also light gray) among the finest cataclastic material; the dashed line at the bottom right corner indicates the orientation of the CSB.

The yellow formation in the hanging wall (and in the upper part of the footwall) is a poorly cemented and very low-cohesive medium-grain sublitharenite sandstone (moderately sorted, φ = 31%; k = 2.3 (± 0.9) × 103 md; see also Table 1). Microscope observations show no or very thin quartz cement rims around the detrital grains (Figure 7D) and a few carbonate cements around the feldspar grains, but no significant pore filling is observed between the detrital grains. In some cases, very thin iron coatings are observed in plane-polarized light around the detrital grains, giving the sandstone its pervasive yellow color. The deformation structures in the hanging wall of the fault are scarce and include dilation bands and CSBs. Dilation bands (Figure 7E) display slightly negative relief at the exposure without visible offset. They are microscopically characterized by irregular 1–2-mm (0.04–0.08-in.)-thick high-porosity (55% ± 2%) tabular zones, where the grains are disaggregated and separated from each other without being significantly fractured (Figure 7F). CSBs (Figure 7G) are 5–10 mm (0.20–0.39 in.) thick with a positive relief and have 5–20 cm (2.0–7.9 in.) of offset. Microscope observations show that these bands are composed of a large amount of cataclastic material (Figure 7H) with significantly reduced φ (8 ± 2%). The fraction of carbonate cement is observed completely crushed in the CSB.

The orientation of all the CSBs and dilatant deformation structures (joints and dilation bands) is shown in Figure 3D. The strike of CSBs (thin blue lines in Figure 3D) ranges between N075° and N170°, with 47% between N075° and N100°. The majority of them (73%) dip between 60° and 80°. A majority (55%) of the dilatant structures (joints appear as thin red lines and dilation bands as thin brown lines in Figure 3D) strike between N095° and N140°; a second significant proportion (19%) strikes between N170° and N190°; and 92% of the dilatant structures dip between 70° and 90°. Although observed interactions between these deformation structures are rare, the dilatant structures crosscut and are later than the CSBs in both the footwall and hanging wall. The distribution profile of the deformation structures is displayed in Figure 8. The profile shows a trend with increasing intensity of deformation nearer the fault; that trend is clear in the footwall and is subtler in the hanging wall. The frequency of dilatant structures is higher in the footwall where they are in the form of joints in the gray-cemented sandstone layer, commonly reaching or exceeding six joints per meter near the fault. In the hanging wall, the dilation bands in the yellow incohesive sandstone never exceed two dilation bands per meter.

Figure 8. Deformation profile showing the number of deformation structures around the Clashach fault. CSBs = compactional shear bands.

Deformation Band Cluster, Clay Smear, and Fault Surfaces at Goblin Valley

Two exposures have been studied at Goblin Valley: (1) a normal fault zone juxtaposing porous sandstones and thin poorly lithified siltstone layers, and (2) a normal fault zone incorporating a poorly lithified shale layer into the fault core within a porous sandstone formation.

Exposure 1: Deformed Sandstone Including Thin, Poorly Lithified Siltstone Layers

The first fault exposure is located at the base of the Entrada Formation (Figure 4B) and displays a fault with a dip-slip component of 6 m (19.7 ft) (Figure 4D). At the outcrop, the base of the footwall is made of an orange fine-grained sandstone (c5 series as shown in Figure 4B, poorly sorted, IGV = 6%, k = 5 [± 3] × 102 md) (Figures 4D, 9A) alternating with 5–20-cm (2.0–7.9-in.)-thick layers of phyllosilicate-bearing poorly lithified siltstones (poorly sorted, IGV = 18%, k = 1.7 [± 2.3] × 102 md; Figure 9B). The top of the footwall and the whole hanging wall along the studied thickness of the outcrop is made of massive clean fine-grained sandstone (e1 series as shown in Figure 4B, well-sorted, IGV = 33%, k = 2.2 [± 1.7] × 104 md; Figure 9C). The reader is directed to Table 1 for more details on the lithological data. The damage zone is approximately 5 m (∼16.4 ft) wide on both sides of the fault core. In the damage zone, the deformation is characterized by conjugate CSBs in the clean sandstones (mean spacing of 11.1 cm [4.37 in.]) and by small fault slip surfaces in both the siltstones (mean spacing of 5.7 cm [2.24 in.]) and the orange fine-grained sandstone (mean spacing of 10.0 cm [3.94 in.]).

Figure 9. Scanning electron microscopy of backscattered electrons imaging of the host rocks at Goblin Valley, at the exposure 1 (A–C), and at the exposure 2 (D–F). (A) Orange fine-grained sandstone host rock (unit c5 in Figure 4D); 25% of the rock is made of oxide-bearing oolites crushed between the stronger grains. (B) Poorly lithified siltstone layer in host rock (very top of unit c5 in Figure 4D). The matrix is phyllosilicate rich. (C) Clean fine-grained sandstone host rock (unit e1 in Figure 4D). (D) The clean fine-grained sandstone host rock at the second exposure (unit e1 in Figure 4E) has not been sampled; however, because of its field similarity in macroscopic aspect, we assume it is equivalent to the clean fine-grained sandstone at the first exposure. (E) Poorly lithified shale layer host rock (unit e2 in Figure 4E). (F) Medium-grained sandstone host rock (unit e3 in Figure 4E). φ = porosity; Fld = feldspar; ool = oolite; Ox = oxide; Qtz = quartz.

At the macroscopic scale, we observe a cluster of CSBs adjacent to a major fault slip surface where the clean fine-grained sandstones are juxtaposed (Figure 10). The cumulative band thickness reaches 94 mm (3.70 in.) at the topmost end of the cluster (at 6 m [19.7 ft], Figure 10A) where the thickness of the cluster reaches 600 mm (23.6 in.) and forms the total thickness of the fault core. Both the cumulative band thickness and the cluster thickness decrease where the sandstone of the hanging wall juxtaposes against the first layers of siltstone (at 4.5 m [14.8 ft], Figure 10A) and farther downward. The cluster is still adjacent to the major fault slip surface, and the cumulative band and cluster thicknesses decrease from 50 mm (2.0 in.) and 320 mm (12.6 in.), respectively, at 4.5 m (14.8 ft), down to 15 mm (0.6 in.) and 50 mm (2.0 in.), respectively, at 2.25 m (7.38 ft). In that interval, a smeared clay-rich siltstone gouge juxtaposes against the CSB cluster separated by the major fault slip surface, and both form the fault core material. The footwall side of the fault core includes a small thickness of lenses of deformed orange fine-grained sandstone (Figure 10B). Downward (from 2.25 m [7.38 ft] and lower, Figure 10A), the cumulative band thickness decreases steadily to 4 mm (0.16 in.) (two bands) at 4 m (13.1 ft) below the siltstone juxtaposition (at 0 m, Figure 10A); here, the clay-rich siltstone gouge (smear) is almost nonexistent. Hence, the clay-rich siltstone gouge forms a fault rock along nearly 4 m (13.1 ft) below the lowest siltstone layer, after which the fault core essentially disappears, leaving the hanging wall and footwall in contact along the major fault slip surface. Below this point, the fault slip surface marks out the footwall edge of the thin cataclastic deformation structure (the end of the cluster). Note that from top to bottom of the exposure, the major throughgoing fault slip surface (striated and flat, commonly smooth; Figure 10C, D) is continuous and is observed near the footwall edge of the CSB cluster. No significant mixing between the silt derived from the poorly lithified siltstone and the cluster is observed in the core zone.

Figure 10. Micro- and macrostructural observations and quantification of the deformation at the first exposure – Goblin Valley. (A) Graph displaying the quantification of the deformation along the length of the cluster of the 6-m (19.7-ft) fault exposure including alternating thin layers of poorly lithified siltstone (Figure 4D). The total core thickness (given as an approximation for reference) accounts for the sum of the cataclastic fault core, the clay-rich gouge (smear), and lenses of deformed host rock. The diagram below the graph represents the type of host rock in the hanging wall (HW) and footwall (FW), and the type of fault material between (the thickness of the fault material is not to scale). (B) Detail from Figure 4D showing the organization of the deformation along the fault plane. (C) Detail from Figure 3D showing a branch of the cluster and the major fault slip surface. (D) Scanning electron microscopy of backscattered electrons (SEM-BSE) image of the major fault slip surface at the edge of the cluster (right hand side); the edge of the fault slip surface (20-μm thickness to the right of the dashed line) is a well-developed cataclasite. Med. = Medium; sst. = sandstone; v. = very.

Exposure 2: Deformed Sandstone Including a Shale Layer

The second exposure is located in the Entrada Formation (Figure 4B) and displays a fault zone with a dip-slip component of 2.1 m (6.6 ft) (Figure 4E), allowing the smearing of shales (mainly illite and kaolinite in a lesser proportion) in a continuous zone of clay gouge between two parallel and overlapping segments of CSB clusters (Figure 11). The exposure displays a clean fine-grained sandstone at the base of the outcrop (considered as analog to the clean fine-grained sandstone described in the first exposure – e1 series as shown in Figure 4B), overlain by an 80-cm (31.5-in.)-thick poorly lithified shale layer (e2 series as shown in Figure 4B; Figure 9E), overlain by a medium-grained sandstone (e3 series as shown in Figure 4B, moderately sorted, φ = 19%; Figure 9F). All lithological data are summarized in Table 1 in the Appendix. The damage zone is approximately 5 m (∼16.4 ft) wide on both sides of the fault core. In the damage zone, the deformation is characterized by conjugate CSBs in the sandstones (mean spacing of 13.1 cm [5.16 in.]) and by small shear surfaces in the clay (mean spacing of 3.1 cm [1.22 in.]).

Figure 11. Micro- and macrostructural observations and quantification of the deformation at the second exposure – Goblin Valley. (A) Graph displaying the quantification of the deformation along the length of the cluster of the 2.1-m (6.9-ft) fault offset including a layer of clay (Figure 4E). The total core thickness (given as an approximation for reference) includes the sum of the cataclastic fault core and the smear. The diagram below the graph represents the type of host rock in the hanging wall (HW) and footwall (FW) and the type of fault material between them (the thickness of the fault material is not to scale); see Figure 10A for legend. (B) Detail from Figure 4E showing the organization of the deformation along the fault slip surface around the clay-rich gouge. (C) Scanning electron microscopy of backscattered electrons image of the deformation of the clay layer within the gouge; the dashed lines encompass a zone where sheared particles are particularly visible.

From the top of the outcrop to the limit of the shale layer in the footwall (from 5–3.50 m [16.4–11.5 ft]; Figure 11A) the fault core is exclusively made of a CSB cluster adjacent to a major fault slip surface. In this interval, the cumulative band thickness is quite stable, ranging between 122 and 165 mm (4.80 and 6.50 in.); the cluster thickness ranges between 290 and 340 mm (11.4 and 13.4 in.). The major fault slip surface is located at the footwall side of the cataclastic cluster (red line in Figure 11A). We note that the position of the fault slip surface beyond 5 m (16.4 ft) (not represented in Figure 11A) is not clearly identified as it varies within the thickness of the cluster. From 3.50 to 1.30 m (11.5 to 4.3 ft), the cluster localizes in the hanging wall side of the fault core against a smeared clay gouge on the footwall side (Figure 11B). In this interval, the cumulative band thickness drops from 112 to 34 mm (4.41 to 1.34 in.), and the cluster thickness decreases from 350 to 105 mm (13.8 to 4.1 in.). Below that, the cluster thickness on the hanging wall side decreases to 0. The fault slip surface marks out the limit with the clay gouge; it is major (flat, smooth, and striated) at 3.3 m (10.8 ft); however, it becomes minor (corrugated and barely striated) at 1.3 m (4.3 ft) and then disappears. No fault slip surface is visible at the macroscopic scale in the small thickness of the clay gouge. However, microscopic scale imaging reveals distributed shear in the shales (Figure 11C). At 1.50 m (4.9 ft), a thickness of 1 mm (0.04 in.) of CSB (one band) is measured against the clay gouge on the footwall side of the fault core. From this location down to the bottom of the outcrop (at 0 m), the cumulative band thickness on the footwall increases up to 67 mm (2.64 in.), corresponding to 155 mm (6.10 in.) of cluster thickness. From 1.5 m (4.9 ft), a minor fault slip surface marks out the limit between the cluster in the footwall and the clay gouge. Below that, the clay gouge disappears, and the fault slip surface is continuous and evolves to a major fault slip surface near the bottom of the outcrop. Apart from a few fragments of 1–5 cm (0.4–2.0 in.) in thickness made of poorly consolidated cataclastic structures included in the clay gouge, no significant mixing between the shale and the cluster is observed in the core zone.

The measured k values of the different fault rocks in the core zone as well as of the different types of host rocks are reported in Figure 5 and Table 1, which also includes the data from the Clashach Cove site for comparison. The host sandstone in Goblin Valley has a k value as high as 2.2 × 104 md for both the clean fine sandstone at the clay-rich siltstone gouge exposure and the clean fine sandstone at the clay gouge exposures. We measured a k value of 5 × 102 md for the orange fine sandstone of the bottom of the footwall at the clay-rich siltstone gouge exposure and down to 1.7 × 102 md for the siltstone layers. The k value of the medium sandstone at the clay gouge exposure was not measured. The k value of the shale was not measured either because the shale at the outcrop was very poorly lithified and incohesive; we give instead the range of typical values of shales as reported by Neuzil (1994) as a reference, spanning from 1 × 100 to 1 × 10−8 md. Cataclastic clusters have k values of 6.1 md parallel to the fault plane and lower than 5 × 10−1 md normal to a major fault slip surface. The sheared surface of the siltstone at the first outcrop in Goblin Valley has a k value of 8.3 × 101 md. The clay gouge k in the second outcrop was not measured but is probably within the range for shales mentioned above.

DISCUSSION

Chronology of the Deformation and Diagenesis at Clashach Cove

The Clashach Cove site records at least two tectonic events, each corresponding to the development of a specific fault rock: (1) an episode of north-south extension characterized by cataclastic deformation and followed by a quartz cementation, and (2) a north-northeast–south-southwest extensional episode characterized by brittle reactivation, including fault slip surface development, brecciation, and jointing, followed by an oxide mineralization. Because joints and dilation bands (dilatant deformation structures) crosscut the CSB both in the footwall and in the hanging wall, they postdate cataclastic deformation. Given the close orientation of individual CSBs and the clusters at the edge of the fault zone, we assume that this cataclastic deformation (shear and compactive deformation) occurred during the same extensional tectonic event (Figure 12A). The N090° strike of the fault surface and its normal-sense kinematics suggest a north-south extension consistent with the Late Jurassic rifting tectonics of the Moray Firth (Roberts et al. 1990; Underhill, 1991). The N100-120° main orientation of the subsequent dilatant deformation structures and the 71°W pitch of the lineation on the fault slip surface are kinematically consistent and suggest a later extension oriented north-northeast–south-southwest, with an oblique reactivation of the fault (Figure 12B). Brittle faulting hence developed during this later stage, as proved by the azimuth of the slickenlines and the oxidized breccia observed at the hanging wall side of the fault surface. This second extensional episode matches with the oblique extension during the Paleocene–Eocene during the Scottish Highlands uplift as suggested by Roberts et al. (1990). The orientation of the dilatant structures from N140° to N190° could mark a gradual east-west reorientation of this latter extensional episode or potential stress perturbation because of fault interactions (e.g., Kattenhorn et al., 2000). Both the close orientation and the formation time of dilation bands and joints suggest a genetic similarity of these structures. They form under extensional conditions as a function of the cohesion and confining pressure. When cohesion is small (no cementation), we observe dilation bands, and when it is higher (because of cementation), we observe joints instead of dilation bands.

Figure 12. Synthetic diagrams summarizing the chronology of the different tectonic events and the associated fault rock generation at the Clashach Cove site. (A) Late Jurassic extension setting. (B) During the Paleocene–Eocene extension. The orange rims around the white grains in the zooms represent the quartz cements sealing the grains; the dark-gray background represents the porosity. CSB = compactional shear band; Dv = vertical displacement.

The quartz cementation of the gray sandstone at the lower part of the footwall probably occurred during the burial of the Moray Firth sandstone during the Late Jurassic. Underhill (1991) estimates that the Hopeman Sandstone Formation near the Moray Coast was buried from a maximum of 1.7 km (1.1 mi) at the end of the Late Jurassic to a maximum of 1.9 km (1.2 mi) at the end of the Cretaceous. Haszeldine et al. (1984) describes a quartz overgrowth development in Lower Jurassic clay-poor sandstone from the Beatrice oil field (middle Moray Firth) from 68°C (no depth mentioned). However, Pollington et al. (2011) showed that quartz cementation in sandstone can form at temperatures as low as 40°C. Assuming a mean gradient of at least 30°C/km (which is a minimum given the interrifting tectonic of the region at the time), this range of temperature seems compatible with the estimated burial depth of the formation. Because the quartz cements along the cluster are sealing the cataclastic textures of the CSB and are crosscut by the joints, the quartz diagenesis may therefore be coeval or postdate the cluster and predate the brittle structures observed. Hence this diagenesis could also have happened with the maximum burial of the formation. However, Fisher and Knipe (1998), Lander et al. (2008), and Williams et al. (2015) have documented the enhancement of cementation in cataclastic structures. In particular, Milliken et al. (2005) and Philit et al. (2015) have shown that high quartz cementation of CSB clusters is possible even at shallow burial (depth <2 km [<1.2 mi] and <1 km [<0.6 mi], respectively). Ngwenya et al. (2000) revealed that this preferential cementation in the cataclastic material can occur through high silica concentration release during cataclastic faulting and subsequent self-sealing. Consequently, and because the sandstone formation in the footwall edge of the cluster is not significantly cemented, it is possible that the quartz cementation in the cluster along the fault slip surface was favored more by the presence of cataclasis than by significant burial. Differential diagenesis in the sandstone series probably occurred during burial because of the presence of cataclasis in the cluster, contributing to the hardening of the cluster and allowing faulting. Fault slip surface formation during the first stage of deformation is unlikely because (1) the fault slip surface slickenline is not kinematically consistent with this first stage of deformation and (2) cementation is an efficient process of rock hardening and subsequent brittle behavior.

The relative chronology between the fault slip surface and the breccia remains undetermined. In brittle rocks, brecciation might occur first because of fracture coalescence (see Figure 1B, e.g., Davatzes et al., 2005; Woodcock and Mort, 2008). However, in this case of differential diagenesis, hardening of the cluster because of its cementation might be sufficient to allow fault slip surface initiation and slip localization in zones where the cluster is adjacent to porous sandstone, and brecciation might initiate first in zones where the cemented sandstone is damaged against the fault. Unfortunately, the distribution of the breccia, which seems present along the whole fault, with respect to the stratigraphy, does not give clear field evidence of this deformation sequence (too large displacement [20–50 m (66–164 ft)] compared to layer thicknesses [<5 m (<16 ft)]).

Oxides seal the fault breccia (grain and clast coating) and are clogging the joints. They have precipitated after the reactivation of the fault, leading to brecciation and fault slip surface formation during the Paleocene–Eocene. We cannot exclude that some oxides may also be contemporaneous with the brecciation. Given that the Paleocene–Eocene rifting caused the uplift of the Scottish Highlands (continental conditions), the massive oxide clogging of the breccia and of the surface of the joints most likely originates from meteoric fluid, drained downward into the fault core and the joints. The dilatant characteristic of brecciation is evidenced by the presence of these massive iron oxide coatings filling the dilatant fault core. Both the composition of the fault core (i.e., the breccia) and the slickenside lineations indicate that the reactivation of the fault involved hardened and therefore more lithified sandstone. The occurrence of the breccia parallel to and against the hardened cluster (preferential cementation evoked before) indicates that the former mechanical process of deformation determined the localization of the brecciation. Joints, which are kinematically consistent with fault reactivation, suggest that the brecciation of the cemented material occurred at relatively low effective stress. The example of Clashach shows that juxtaposition of fault rocks can provide very high variation of permeability through time. Between the two deformation events, fluid flow has probably been retarded across the cluster after the cataclastic deformation and diagenesis because the cluster permeability is approximately four orders of magnitude lower than in the porous sandstone (Figure 5, also see the synthesis in Ballas et al., 2015), whereas flow was likely highly enhanced along dilatant breccia when juxtaposed against the cluster. Brittle deformation factors, such as fracture opening, coalescence, and brecciation with large IGV, provide a significant increase of permeability, and they are macroscopic conduits for meteoric fluid flow downward, as suggested by the iron oxide cements. We remark that the permeability increase was transient if we consider that the oxides now seal the breccia. To a lesser extent, a permeability anisotropy of one order of magnitude can be obtained in the host rock walls near the fault core because of pervasive microscale deformation and grain reorientation (Farrell et al., 2014). Consistent with the results of Williams et al. (2017), we document here that the behavior of a fault through time is highly variable depending on the diagenetic evolution of the host rock.

Spatial Evolution of Faults in a Formation of Alternating Sandstone and Clay-Rich Layers at Goblin Valley

The Goblin Valley site shows deformation structures formed during the same tectonic episode through a porous-sandstone–dominated formation including layers of poorly lithified clay-rich siltstone or clay. Our study evidences (1) the location of major fault slip surfaces at the interface between clusters and clay-rich gouges and (2) the juxtaposition of distinct fault cores made of two fault materials: clusters of CSBs and clay-rich gouges. Both exposures at Goblin Valley display major fault slip surfaces along the same edge of the cluster, at the interface between the clay-rich gouge, which is a major lithological contrast of stiffness and relief at the outcrop. Siltstone and shale layers therefore appear determinant in the initiation and localization of a discrete fault slip surface in the cluster. This gives examples of siltstone and clay gouge juxtaposition along cataclastic clusters in porous sandstone units and suggests that the process of clay smearing prevents continuation of the process of cataclasis and leads to fault surface localization and slip adjacent to clusters at early stages of displacement (≥2.1 m [≥6.9 ft]).

Even if the Entrada Formation was buried to at least 1100 m (0.68 mi) at the time of deformation, both exposures record cataclasis as the dominant mechanical process of deformation in the porous sandstone. Although it is true that the thickness of a cluster may vary considerably along its length at the scale of 1 m (3.3 ft) (Fossen and Bale, 2007), we observed as a general trend that the cumulative band thickness is always maximum between directly juxtaposed sandstones (without clay-rich gouge in between) and decreases down to zero where the sandstones are displaced against siltstones or shales. Thin clusters are observed at all the cluster-smear transitions identified in the field (two more outcrops of lesser quality also expose this thinning), and this large gradient of thickness decrease (∼0.3 m/1.5 m [∼1.0 ft/4.9 ft]) is exactly located at places where the clay-rich gouge is juxtaposed (see Figure 13A), probably marking a genetic relation between the two. In addition, conventional gradients of decrease in cluster thickness in homogeneous sandstones reported by Philit (2017) and Philit et al. (2018) show values two orders of magnitude below (∼0.3 m /100 m [∼1.0 ft /328 ft]). This observation suggests that clusters grow by increasing the number of cataclastic bands where the sandstones are still directly in contact (Figure 13A, B). They can form early as strain is accommodated by self-juxtaposition of the sand units and can be simply shifted against the clay-rich gouge when displacement increases. In other words, it is highly probable that CSBs do not form and develop into clusters where they juxtapose against the clay-rich gouge. Smearing results in the reduction of the shear stresses because clay (or phyllosilicate-bearing siltstone) represents a weaker material than sandstone or cataclastic structures (Byerlee, 1978; Schmatz et al., 2010; Kettermann et al., 2016). At Goblin Valley, although distributed in the gouge thickness (clay smear; Figure 11C), shear tends to localize as a fault surface at the cluster contact. Displacement as small as 2.1 m (6.9 ft) is sufficient to propagate this fault surface as a major fault slip surface (faulting) in the underlying and overlying sandstones (Figure 13C), consequently stopping further cluster growth and deformation in the damage zone. This behavior deserves to be studied on more field examples and particularly in contractional regimes, as this behavior may have major implications for the deformation in sandstones, including the evolution of faulting or the extent of growth of the deformation zone (as studied, for instance, by Harris et al., 2003; Nicol et al., 2013; Schueller et al., 2013; Ballas et al., 2014; Soliva et al., 2016). We note that the initiation of clustering may occur simultaneously in both the sandstone units above and below the siltstones and shale layers because the clusters in the two units are aligned in the same plane at the second exposure. The different thickness of the cataclastic structures on both sides of the smear (hanging versus footwall) may be related to the clusters not forming in the same sandstone unit and may therefore have been generated under different conditions of fault propagation.

Figure 13. Synthetic diagrams summarizing the evolution of the deformation and its associated processes and the generated fault rocks in the sandstone-dominated formations at Goblin Valley: (A) at the incipient stage of deformation (displacement [D] of ∼10 cm [∼3.9 in.]), (B) when both smearing and cataclasis are active (D of ∼1 m [∼3.3 ft]), and (C) when localized faulting and slip occur and cataclasis is inhibited (D > ∼2 m [>∼6.6 ft]). k = permeability.

At Goblin Valley, any sort of sandstone or siltstone juxtaposed to sandstone with φ values lower than 6% or including weak material, such as phyllosilicates or oxide-bearing oolites, between the quartz grains, or both, develop no cataclastic deformation. Knipe (1997) suggests that an amount of 10%–15% of phyllosilicate in siliciclastic material is sufficient to promote grain boundary sliding rather than grain fracturing. We assume similar behavior would happen with other weak materials.

The example of Goblin Valley shows that highly heterogeneous host rocks can generate heterogeneous mechanical behavior through space and time during deformation and affect the permeability of the fault core. The juxtaposition of cataclastic clusters and clay-rich gouges leads to a localized permeability reduction. Although the case of mixing between sands and clay layers is commonly reported (Gibson, 1998; Rawling et al., 2001; Bense et al., 2003; Kettermann et al., 2016), we note that this process is not significant in the present cases of clusters juxtaposed against clay-rich gouges. Yet, in the case of thin alternating sand and clay layers, the latter authors showed that the mixing of the smeared clays and sand grains as displacement increases leads to a sealing material as long as the clay volume exceeds the pore volume of the sand. Linked to the porosity reduction, the permeability reduction of the host rock by cataclasis (at Clashach and Goblin Valley) is between three to four orders of magnitude in the CSB clusters and is up to four to five orders of magnitude when a major fault slip surface exists. These results are consistent with the measurements of Fossen and Bale (2007) and Ballas et al. (2015), for instance. On the one hand, Eichhubl et al. (2004) and Tueckmantel et al. (2012) report that permeability reduction of two orders of magnitude is enough to baffle meteoric fluids. However, many studies (e.g., Yielding, 2002; Bretan et al., 2003) have shown that sandstone–sandstone juxtapositions commonly cannot support significant hydrocarbon column heights where clay smear is limited. Hence, the juxtaposition of clusters with low-permeability clay-rich gouges may at least make the whole fault core a transverse fault baffle for fluid flow early in the evolution of the deformation. The baffle effect can be efficient because clusters can be up to hundreds of meters long and form networks on distances of several kilometers (Figure 4B).

Implications for Deformation Band Cluster Growth and Fault Surface Development

We have shown that important fault slip surfaces can develop and persist adjacent to CSB clusters as the consequence of (1) cementation of porous sandstone and (2) smearing of clay-rich units. These examples therefore provide new field observations suggesting that fault slip surface initiation and stabilization might be independent of the thickening of CSB clusters previously invoked (e.g., Aydin, 1978; Schultz and Siddharthan, 2005; Wibberley et al., 2007; Figure 1C). Our study reveals that fault slip surface localization can be related to an increase in the contrast of mechanical behavior between the cluster and the adjacent material (i.e., to a diagenetic hardening of the band cluster adjacent to porous sandstone at Clashach and to a softening deformation inherent to clay-rich layer juxtaposed [smeared] against the cluster). The fact that (1) both diagenesis and clay smearing do not directly depend on cluster thickness and density and the fact that (2) there is no clear correlation between the presence of a major fault slip surface and cluster thickness and density (Philit, 2017; Philit et al., 2018) suggest that these processes might be responsible for the presence of fault slip surfaces in the clusters, as shown in Figure 2. It is worth considering that significant hardening caused by thin and discrete quartz cementation at grain contacts is possible even for shallow burial conditions (Ngwenya et al., 2000; Milliken et al., 2005; Philit et al., 2015; Philit, 2017). Also, note that clay-rich layers above or below porous sandstones and responsible for fault surface development might not crop out exactly where a cluster is observed. These processes, as well as cluster segment linkage at relay zones, should therefore be inspected with care when trying to understand fault surface development adjacent to clusters in porous sandstone units.

CONCLUSION

Our field study focuses on the development of fault surfaces and the related juxtaposition of fault rocks in normal fault cores during deformation of porous-sandstone–dominated formations at two sites.

In agreement with the previous works related to the influence of the lithology on the mechanical process of deformation, we confirm that, in sandstone-dominated formations, cataclasis organized as clusters of CSBs occurs in poorly consolidated sandstone, whereas faulting and brecciation occurs if the rock is significantly lithified and cohesive. Hence, we evidence at the site of Clashach that the evolution of a sandstone through time can lead to evolving fault fluid flow behavior. The formation was first deformed while poorly lithified, which led to cataclastic deformation and the formation of a cluster of CSBs. Because of the permeability reduction in this fault rock, the fluid flow may have been at least baffled across the fault. A burial of 1.9 km (1.2 mi) appears sufficient to allow the cementation, hardening, and permeability reduction of the formation and the favored cementation of the cluster. The presence of the former hardened cluster determined the localization of a fault slip surface and brecciation during later normal tectonic loading (fault reactivation). As a result, the dilatant behavior of this mechanical process of deformation allowed an efficient fluid circulation along the fault, providing an intense oxide clogging of the breccia.

At Goblin Valley, we confirm the smearing of the formation when the sandstone-dominated formation includes thin clay-rich layer(s). We note that the cataclasis was possible at a burial depth of at least 1.1 km (0.7 mi). Our study suggests that the mechanical process of clay smearing inhibits the process of cataclasis and enhances the localization of future slip on the fault surface at the interface between the cluster and the clay-rich gouge and adjacent to (and sometimes in) the clusters in places of sand–sand juxtaposition. Consequently, mixing would not necessarily occur during smearing when clusters are overlapping the clay-rich gouges. Because the permeability of cataclastic fault rocks is three to five orders of magnitude lower than that of the host rock and because of the juxtaposition with the low-permeability clay-rich gouge, the clusters of CSB are likely to form continuous and efficient barriers to fluid flow across the fault.

As a general conclusion, our contribution shows that fault slip surface localization and stabilization are related to an increase in the contrast of mechanical behavior between the cluster and the adjacent material (diagenetic hardening of the cluster or softening of the clay-rich gouge). Such evolution of the deformation in a sandstone-dominated formation through time and space may promote fault slip surface stabilization along the entire cluster, fault rock juxtaposition, and strong permeability anisotropy. This will result in contrasted fluid flow behaviors through the reservoir depending on the lithology of the faulted units and the tectonic history of the formation. For instance, subseismic structures (displacement < 20 m [<66 ft]) can evolve into seismic-scale faults (displacement ≥ 20 m [≥66 ft]) with juxtaposing fault rocks and drastically different fluid flow along and across the fault. However, subseismic displacements are sufficient to induce continued low-permeability structures through the juxtaposition of clusters and clay-rich gouge likely to act as barriers for fluid flow.

APPENDIX

excxcqcbaasssfqxvvcft Figure 14. (A, B) Grain size distribution of the host rock sandstones at Clashach Cove. The grain populations were obtained by measuring the Feret diameter of a number (n)of host rock grains in thin section. A statistical factor of 21/2 was applied to correct for the underestimation of the diameter in two dimensions because of the truncation effect (sampling effect is neglected), as suggested by Pelto (1952). (C, D) Grain size distribution of the host rock sandstones at Goblin Valley. The grain size distribution is obtained through laser diffraction analysis. c Sst. = coarse sandstone; f Sst. = fine sandstone; FW = footwall; HW = hanging wall; m Sst. = medium sandstone; Siltst. = siltstone; vc Sst. = very coarse sandstone; vf Sst. = very fine sandstone.

REFERENCES CITED

Agosta, F., M. Prasad, and A. Aydin, 2006, Physical properties of carbonate fault rocks, fucino basin (Central Italy): Implications for fault seal in platform carbonates: Geofluids, v. 7, no. 1, p. 19–32, doi:10.1111/j.1468-8123.2006.00158.x.

Al-Hinai, S., Q. J. Fisher, B. Al-Busafi, P. Guise, and C. A. Grattoni, 2008, Laboratory measurements of the relative permeability of cataclastic fault rocks: An important consideration for production simulation modelling: Marine and Petroleum Geology, v. 25, no. 6, p. 473–485, doi:10.1016/j.marpetgeo.2007.07.005.

Antonellini, M., and A. Aydin, 1994, Effect of faulting on fluid flow in porous sandstones: Petrophysical properties: AAPG Bulletin, v. 78, no. 3, p. 355–377.

Aydin, A., 1978, Small faults formed as deformation bands in sandstone: Pure and Applied Geophysics, v. 116, p. 913–930, doi:10.1007/BF00876546.

Aydin, A., 2000, Fractures, faults, and hydrocarbon entrapment, migration and flow: Marine and Petroleum Geology, v. 17, p. 797–814.

Aydin, A., and Y. Eyal, 2002, Anatomy of a normal fault with shale smear: Implications for fault seal: AAPG Bulletin, v. 86, no. 8, p. 1367–1381, doi:10.1306/61EEDC9C-173E-11D7-8645000102C1865D.

Ballas, G., H. Fossen, and R. Soliva, 2015, Factors controlling permeability of cataclastic deformation bands and faults in porous sandstone reservoirs: Journal of Structural Geology, v. 76, p. 1–21, doi:10.1016/j.jsg.2015.03.013.

Ballas, G., R. Soliva, A. Benedicto, and J.-P. Sizun, 2014, Control of tectonic settings and large-scale faults on the basin-scale distribution of deformation bands in porous sandstone (Provence, France): Marine and Petroleum Geology, v. 55, p. 142–159, doi:10.1016/j.marpetgeo.2013.12.020.

Bense, V. F., E. H. Van den Berg, and R. T. Van Balen, 2003, Deformation mechanisms and hydraulic properties of fault zones in unconsolidated sediments; the Roer Valley Rift System, The Netherlands: Hydrogeology Journal, v. 11, no. 3, p. 319–332, doi:10.1007/s10040-003-0262-8.

Blenkinsop, T. G., 1991, Cataclasis and processes of particle size reduction: Pure and Applied Geophysics, v. 136, p. 59–86, doi:10.1007/BF00878888.

Bretan, P., G. Yielding, and H. Jones, 2003, Using calibrated shale gouge ratio to estimate hydrocarbon column heights: AAPG Bulletin, v. 87, no. 3, p. 397–413, doi:10.1306/08010201128.

Byerlee, J., 1978, Friction of rocks: Pure and Applied Geophysics, v. 116, p. 615–626, doi:10.1007/BF00876528.

Caine, J. S., J. P. Evans, and C. B. Forster, 1996, Fault zone architecture and permeability structure: Geology, v. 24, no. 11, p. 1025–1028, doi:10.1130/0091-7613(1996)024<1025:FZAAPS>2.3.CO;2.

Childs, C., J. J. Walsh, T. Manzocchi, J. Strand, A. Nicol, M. Tomasso, M. P. J. Schöpfer, and A. C. Aplin, 2007, Definition of a fault permeability predictor from outcrop studies of a faulted turbidite sequence, Taranaki, New Zealand: Geological Society, London, Special Publications 2007, v. 292, p. 235–258, doi:10.1144/SP292.14.

Davatzes, N. C., P. Eichhubl, and A. Aydin, 2005, Structural evolution of fault zones in sandstone by multiple deformation mechanisms: Moab fault, southeast Utah: Geological Society of America Bulletin, v. 117, no. 1–2, p. 135–148, doi:10.1130/B25473.1.

Doelling, H. H., P. A. Kuehne, G. C. Willis, and J. Buck Ehler, 2015, Geologic map of the San Rafael Desert 30 x 60-minute quadrangle, Emery and Grand Counties, Utah: Salt Lake City, Utah, Utah Geological Survey, scale 1:62,500, 3 sheets.

Edwards, H. E., A. D. Becker, and J. A. Howell, 1993, Compartmentalization of an aeolian sandstone by structural heterogeneities: Permo-Triassic Hopeman Sandstone, Moray Firth, Scotland: Geological Society, London, Special Publications 1993, v. 73, p. 339–365, doi:10.1144/GSL.SP.1993.073.01.20.

Ehrenberg, S. N., 1989, Assessing the relative importance of compaction processes and cementation to reduction of porosity in sandstones: Discussion; Compaction and porosity evolution of Pliocene sandstones, Ventura Basin, California; Discussion: AAPG Bulletin, v. 73, no. 10, p. 1274–1276.

Eichhubl, P., W. L. Taylor, D. D. Pollard, and A. Aydin, 2004, Paleo-fluid flow and deformation in the Aztec Sandstone at the Valley of Fire, Nevada—Evidence for the coupling of hydrogeologic, diagenetic, and tectonic processes: Geological Society of America Bulletin, v. 116, no. 9–10, p. 1120–1136, doi:10.1130/B25446.1.

Farrell, N. J. C., D. Healy, and C. W. Taylor, 2014, Anisotropy of permeability in faulted porous sandstones: Journal of Structural Geology, v. 63, p. 50–67, doi:10.1016/j.jsg.2014.02.008

.

Faulkner, D. R., C. A. L. Jackson, R. J. Lunn, R. W. Schlische, Z. K. Shipton, C. A. J. Wibberley, and M. O. Withjack, 2010, A review of recent developments concerning the structure, mechanics and fluid flow properties of fault zones: Journal of Structural Geology, v. 32, no. 11, p. 1557–1575, doi:10.1016/j.jsg.2010.06.009.

Fisher, Q. J., and R. J. Knipe, 1998, Fault sealing processes in siliciclastic sediments: Geological Society, London, Special Publications 1998, v. 147, p. 117–134, doi:10.1144/GSL.SP.1998.147.01.08.

Folk, R. L., and W. C. Ward, 1957, Brazos River Bar [Texas]: A study in the significance of grain size parameters: Journal of Sedimentary Research, v. 27, no. 1, p. 3–26, doi:10.1306/74D70646-2B21-11D7-8648000102C1865D.

Fossen, H., and A. Bale, 2007, Deformation bands and their influence on fluid flow: AAPG Bulletin, v. 91, no. 12, p. 1685–1700, doi:10.1306/07300706146.

Fossen, H., R. A. Schultz, Z. K. Shipton, and K. Mair, 2007, Deformation bands in sandstone: A review: Journal of the Geological Society, v. 164, no. 4, p. 755–769, doi:10.1144/0016-76492006-036.

Fossen, H., R. A. Schultz, and A. Torabi, 2011, Conditions and implications for compaction band formation in the Navajo Sandstone, Utah: Journal of Structural Geology, v. 33, p. 1477–1490, doi:10.1016/j.jsg.2011.08.001.

Fossen, H., R. Soliva, G. Ballas, B. Trzaskos, C. Cavalcante, and R. A. Schultz, 2018, A review of deformation bands in reservoir sandstones: Geometries, mechanisms and distribution: Geological Society, London, Special Publications 2018, v. 459, p. 9–33, doi:10.1144/SP459.4.

Foxford, K. A., J. J. Walsh, J. Watterson, I. R. Garden, S. C. Guscott, and S. D. Burley, 1998, Structure and content of the Moab Fault Zone, Utah, USA, and its implications for fault seal prediction: Geological Society, London, Special Publications 1998, v. 147, p. 87–103, doi:10.1144/GSL.SP.1998.147.01.06.

Fulljames, J. R., L. J. J. Zijerveld, and R. C. M. W. Franssen, 1997, Fault seal processes: Systematic analysis of fault seals over geological and production time scales, in P. Møller-Pedersen and A. G Koestler, eds., Hydrocarbon seals: Importance for exploration and production: Oslo, Norway, Norwegian Petroleum Society Special Publication7, p. 51–59, doi:10.1016/S0928-8937(97)80006-9.

Gallagher, J. J., Jr., M. Friedman, J. Handin, and G. M. Sowers, 1974, Experimental studies relating to microfracture in sandstone: Tectonophysics, v. 21, no. 3, p. 203–247, doi:10.1016/0040-1951(74)90053-5.

Gibson, R. G., 1998, Physical character and fluid-flow properties of sandstone-derived fault zones: Geological Society, London, Special Publications 1998, v. 127, p. 83–97, doi:10.1144/GSL.SP.1998.127.01.07.

Glennie, K. W., and A. T. Buller, 1983, The Permian Weissliegend of NW Europe: The partial deformation of aeolian dune sands caused by the Zechstein transgression: Sedimentary Geology, v. 35, no. 1, p. 43–81, doi:10.1016/0037-0738(83)90069-6.

Harris, S. D., E. McAllister, R. J. Knipe, and N. E. Odling, 2003, Predicting the three-dimensional population characteristics of fault zones: A study using stochastic models: Journal of Structural Geology, v. 25, no. 8, p. 1281–1299, doi:10.1016/S0191-8141(02)00158-X.

Haszeldine, R. S., I. M. Samson, and C. Cornford, 1984, Quartz diagenesis and convective fluid movement: Beatrice Oilfield, UK North Sea: Clay Minerals, v. 19, no. 3, p. 391–402, doi:10.1180/claymin.1984.019.3.10.

Johansen, T. E. S., and H. Fossen, 2008, Internal geometry of fault damage zones in interbedded siliciclastic sediments: Geological Society, London, Special Publications 2008, v. 299, p. 35–56, doi:10.1144/SP299.3.

Kattenhorn, S. A., A. Aydin, and D. D. Pollard, 2000, Joints at high angles to normal fault strike: An explanation using 3-D numerical models of fault-perturbed stress fields: Journal of Structural Geology, v. 22, no. 1, p. 1–23, doi:10.1016/S0191-8141(99)00130-3.

Kettermann, M., S. Thronberens, O. Juarez, J. Lajos Urai, M. Ziegler, S. Asmus, and U. Kruger, 2016, Mechanisms of clay smear formation in unconsolidated sediments—Insights from 3-D observations of excavated normal faults: Solid Earth, v. 7, no. 3, p. 789–815, doi:10.5194/se-7-789-2016.

Knipe, R. J., 1997, Juxtaposition and seal diagrams to help analyze fault seals in hydrocarbon reservoirs: AAPG Bulletin, v. 81, no. 2, p. 187–195.

Lanatà, P., 2014, Full-field experimental characterization of mechanical behaviour and failure in a porous rock in plane strain compression: Homogeneous deformation and strain localization, Ph.D. thesis, Université de Grenoble, Grenoble, France, 387 p.

Lander, R. H., R. E. Larese, and L. M. Bonnell, 2008, Toward more accurate quartz cement models: The importance of euhedral versus noneuhedral growth rates: AAPG Bulletin, v. 92, no. 11, p. 1537–1563, doi:10.1306/07160808037.

Lindsay, N. G., F. C. Murphy, J. J. Walsh, and J. Waterson, 1993, Outcrop studies of shale smear on fault surfaces, in I. D. Bryant and S. S. Flint, eds., The geological modelling of hydrocarbon reservoirs and outcrop analogues: Ghent, Belgium, International Association of Sedimentologists Special Publication 15, p. 113–123.

Milliken, K. L., R. M. Reed, and S. E. Laubach, 2005, Quantifying compaction and cementation in deformation bands in porous sandstones, in R. Sorkhabi and Y. Tsuji, eds., Faults, fluid flow, and petroleum traps: AAPG Memoir 85, p. 237–249, doi:10.1306/1033726M85252.

Neuzil, C. E., 1994, How permeable are clays and shales?: Water Resources Research, v. 30, no. 2, p. 145–150, doi:10.1029/93WR02930.

Ngwenya, B. T., S. C. Elphick, I. G. Main, and G. B. Shimmield, 2000, Experimental constraints on the diagenetic self-sealing capacity of faults in high porosity rocks: Earth and Planetary Science Letters, v. 183, no. 1–2, p. 187–199, doi:10.1016/S0012-821X(00)00261-2.

Nicol, A., C. Childs, J. J. Walsh, and K. W. Schafer, 2013, A geometric model for the formation of deformation band clusters: Journal of Structural Geology, v. 55, p. 21–33, doi:10.1016/j.jsg.2013.07.004.

Pelto, C. R., 1952, The mechanical analysis of sediments from thin-section data: A discussion: Journal of Geology, v. 60, no. 4, p. 402–406, doi:10.1086/625988.

Philit, S., 2017, Elaboration of a structural, petrophysical and mechanical model of faults in porous sandstones; Implication for migration and fluid entrapment, Ph.D. thesis, University of Montpellier, Montpellier, France, 298 p.

Philit, S., R. Soliva, R. Castilla, G. Ballas, and A. Taillefer, 2018, Clusters of cataclastic deformation bands in porous sandstones: Journal of Structural Geology, v. 114, p. 235–250, doi:10.1016/j.jsg.2018.04.013.

Philit, S., R. Soliva, P. Labaume, C. Gout, and C. Wibberley, 2015, Relations between shallow cataclastic faulting and cementation in porous sandstones: First insight from a groundwater environmental context: Journal of Structural Geology, v. 81, p. 89–105, doi:10.1016/j.jsg.2015.10.001.

Pollington, A. D., R. Kozdon, and J. W. Valley, 2011, Evolution of quartz cementation during burial of the Cambrian Mount Simon Sandstone, Illinois Basin: In situ microanalysis of δ18O: Geology, v. 39, no. 12, p. 1119–1122, doi:10.1130/G32195.1.

Quinn, O. F., 2008, Fault-controlled fluid flow and quartz cementation in porous sandstones, Ph.D. thesis, the University of Edinburgh, Edinburgh, Scotland, 270 p.

Rawling, G. C., and L. B. Goodwin, 2006, Structural record of the mechanical evolution of mixed zones in faulted poorly lithified sediments, Rio Grande rift, New Mexico, USA: Journal of Structural Geology, v. 28, no. 9, p. 1623–1639, doi:10.1016/j.jsg.2006.06.008.

Rawling, G. C., L. B. Goodwin, and J. L. Wilson, 2001, Internal architecture, permeability structure, and hydrologic significance of contrasting fault-zone types: Geology, v. 29, no. 1, p. 43–46, doi:10.1130/0091-7613(2001)029<0043:IAPSAH>2.0.CO;2.

Rigby, J. K., and S. S. Beus, 1987, Stratigraphy and structure of the San Rafael Reef, Utah: A major monocline of the Colorado Plateau, in Decade of North American geology, : Boulder, Colorado, Rocky Mountain Section of the Geological Society of America, Centennial Field Guide, v. 2, p. 269–273.

Roberts, A. M., M. E. Badley, J. D. Price, and I. W. Huck, 1990, The structural history of a transtensional basin: Inner Moray Firth, NE Scotland: Journal of the Geological Society, v. 147, no. 1, p. 87–103, doi:10.1144/gsjgs.147.1.0087.

Schmatz, J., P. J. Vrolijk, and J. L. Urai, 2010, Clay smear in normal fault zones—The effect of multilayers and clay cementation in water-saturated model experiments: Journal of Structural Geology, v. 32, no. 11, p. 1834–1849, doi:10.1016/j.jsg.2009.12.006.

Schueller, S., A. Braathen, H. Fossen, and J. Tveranger, 2013, Spatial distribution of deformation bands in damage zones of extensional faults in porous sandstones: Statistical analysis of field data: Journal of Structural Geology, v. 52, p. 148–162, doi:10.1016/j.jsg.2013.03.013.

Schultz, R. A., and H. Fossen, 2008, Terminology for structural discontinuities: AAPG Bulletin, v. 92, no. 7, p. 853–867, doi:10.1306/02200807065.

Schultz, R. A., and R. Siddharthan, 2005, A general framework for the occurrence and faulting of deformation bands in porous granular rocks.: Tectonophysics, v. 411, no. 1–4, p. 1–18, doi:10.1016/j.tecto.2005.07.008.

Shotton, F. W., 1956, Some aspects of the new red desert in Britain: Geological Journal, v. 1, no. 5, p. 450–465, doi:10.1002/gj.3350010505.

Sibson, R. H., 1977, Fault rocks and fault mechanisms: Journal of the Geological Society, v. 133, no. 3, p. 191–213, doi:10.1144/gsjgs.133.3.0191.

Soliva, R., G. Ballas, H. Fossen, and S. Philit, 2016, Tectonic regime controls clustering of deformation bands in porous sandstone: Geology, v. 44, no. 6, p. 423–426, doi:10.1130/G37585.1.

Tueckmantel, C., Q. J. Fisher, T. Manzocchi, S. Skachkov, and C. A. Grattoni, 2012, Two-phase fluid flow properties of cataclastic fault rocks: Implications for CO2 storage in saline aquifers: Geology, v. 40, no. 1, p. 39–42, doi:10.1130/G32508.1.

Underhill, J. R., 1991, Implications of Mesozoic—Recent basin development in the western Inner Moray Firth, UK: Marine and Petroleum Geology, v. 8, no. 3, p. 359–369, doi:10.1016/0264-8172(91)90089-J.

Wibberley, C. A. J., J.-P. Petit, and T. Rives, 2007, The mechanics of fault distribution and localization in high-porosity sands, Provence, France: Geological Society, London, Special Publications 2007, v. 289, p. 19–46, doi:10.1144/SP289.3.

Wibberley, C. A. J., G. Yielding, and G. D. Toro, 2008, Recent advances in the understanding of fault zone internal structure: A review: Geological Society, London, Special Publications 2008, v. 299, p. 5–33, doi:10.1144/SP299.2.

Williams, R. T., J. R. Farver, C. M. Onasch, and D. F. Winslow, 2015, An experimental investigation of the role of microfracture surfaces in controlling quartz precipitation rate: Applications to fault zone diagenesis: Journal of Structural Geology, v. 74, p. 24–30, doi:10.1016/j.jsg.2015.02.011.

Williams, R. T., L. B. Goodwin, and P. S. Mozley, 2017, Diagenetic controls on the evolution of fault-zone architecture and permeability structure: Implications for episodicity of fault-zone fluid transport in extensional basins: Geological Society of America Bulletin, v. 129, no. 3–4, p. 464–478, doi:10.1130/B31443.1.

Woodcock, N. H., and K. Mort, 2008, Classification of fault breccias and related fault rocks: Geological Magazine, v. 145, no. 3, p. 435–440, doi:10.1017/S0016756808004883.

Yielding, G., 2002, Shale gouge ratio—Calibration by geohistory: Norwegian Petroleum Society Special Publications, v. 11, p. 1–15, doi:10.1016/S0928-8937(02)80003-0.

Yielding, G., B. Freeman, and D. T. Needham, 1997, Quantitative fault seal prediction: AAPG Bulletin, v. 81, no. 6, p. 897–917.

Zhu, W., and T. Wong, 1997, The transition from brittle faulting to cataclastic flow: Permeability evolution: Journal of Geophysical Research: Solid Earth, v. 102, p. 3027–3041, doi:10.1029/96JB03282.

ACKNOWLEDGMENTS

We wish to thank the Laboratory of Géosciences Montpellier, the Institut Universitaire de France, and Total Exploration and Production for their support in this work. We are also grateful to Doriane Delmas and Christophe Nevado for the laborious work they did by providing the numerous and quality thin sections necessary to perform this work. We thank Jean-Jacques Cornée for the detailed insight he gave us about the depositional environment of some studied rocks. Finally, we thank Rob Knipe and Marco Lommatsch for their constructive comments on the manuscript.

excxcqcbaasssfqxvvcft

You may also be interested in ...