Deformation band development as a function of intrinsic host-rock properties in Triassic Sherwood Sandstone

Abstract Deformation bands significantly alter the local petrophysical properties of sandstone reservoirs, although it is not known how the intrinsically variable characteristics of sandstones (e.g. grain size, sorting and mineralogy) influence the nature and distribution of deformation bands. To address this, cataclastic deformation bands within fine- and coarse-grained Triassic Sherwood Sandstone at Thurstaston, UK were analysed, for the first time, using a suite of petrographical techniques, outcrop studies, helium porosimetry and image analysis. Deformation bands are more abundant in the coarse-grained sandstone than in the underlying fine-grained sandstone. North- and south-dipping conjugate sets of cataclastic bands in the coarse-grained sandstone broadly increase in density (defined by number/m2) when approaching faults. Microstructural analysis revealed that primary grain size controls deformation band density. Deformation bands in both coarse and fine sandstones led to significantly reduced porosity, and so can represent barriers or baffles to lateral fluid flow. Microstructural data show preferential cataclasis of K-feldspar grains within the host rock and deformation band. The study is of direct relevance to the prediction of reservoir quality in several petroleum-bearing Lower Triassic reservoirs in the near offshore, as deformation band development occurred prior to Carboniferous source-rock maturation and petroleum migration.

The oil and gas industry has expressed a growing interest in deformation bands because they are subseismic, tabular zones of strain localization that can cause large changes to a reservoir's petrophysical properties . Examples of permeability alteration include the Clair Field, west of Shetland, UK, where deformation bands provide a conduit to lateral fluid flow during the early stages of deformation band formation that initially increased porosity (Baron et al. 2008). At the Anschutz Ranch East Field, Wyoming, USA, deformation bands separate clean sandstones and bitumen-stained sandstones, implying a strong impact on oil and gas movement (Solum et al. 2010). At the Arroyo Grande Field, California, USA, steam conductivity parallel to deformation bands is reported to be nine times higher than conductivity perpendicular to deformation bands, with tar deposits present on only one side of the deformation bands (Solum et al. 2010). Data presented in this study could potentially maximize near-term production targets in the Morecambe, Hamilton, Douglas and Lennox oil and gas fields within the neighbouring East Irish Sea Basin.
Millimetres to centimetres in width, with lengths of several metres or more (Schultz & Soliva 2012), deformation bands have been kinematically classified as one of three end members, namely: dilation bands (pore volume increase); shear bands (pore volume increase, decrease or no change); or compaction bands (pore volume decrease) (Aydin et al. 2006;Torabi 2014). Deformation bands in this study will be classified by the predominant deformation mechanism: disaggregation; phyllosilicate smear; cataclasis; and solution and cementation (Fossen et al. 2007).
Disaggregation bands form due to shear-induced disaggregation of grains by grain rolling, grainboundary sliding and the breakage of cements bonding grains, but show little or no evidence of grain crushing . Phyllosilicate bands form in sandstones which contain .10-15% platy minerals, and 'deformation bands with clay smearing' form in sandstones which have a clay content .40% (Fisher & Knipe 2001;Cerveny et al. 2004;Fossen et al. 2007). Mechanical grain fracturing is the dominant process in cataclastic bands, where compaction and reorganization of broken grains significantly reduces porosity (Cerveny et al. 2004;Fossen 2010). Solution bands are produced when chemical compaction, or pressure solution, is the dominant process; they commonly form at shallow depths and contain minimal cataclasis (Fossen et al. 2007). Fresh mineral surfaces exposed by grain-boundary sliding and/or grain crushing provide preferential sites for cementation, thus creating cementation bands (Fossen et al. 2007).
The primary aim of this paper is to establish how intrinsic host-rock properties (grain size, grain size distribution, porosity and mineralogy) control the nature and distribution of deformation bands. Using a range of petrographical techniques, image analysing software and field techniques, this paper will address the following specific questions, using examples of deformation bands for the first time from Thurstaston, Wirral, UK (Lower Triassic Sherwood Sandstone Group): † What types of deformation band are present at Thurstaston? † What is the spatial relationship between deformation band density and fault proximity? † What is the relationship between the nature of deformation bands and the instrinsic host-rock properties? † What is the potential impact of the presence of deformation bands in nearby reservoirs in the same lithology?

Geological setting
Early Permian rifting formed a predominantly north-south-orientated asymmetrical half-graben, deepening towards the east (Mikkelsen & Floodpage 1997). The Cheshire Basin (Fig. 1a) formed in the hanging wall of the Wem-Red Rock Fault (Knott 1994;Beach et al. 1997;Mikkelsen & Floodpage 1997), a northerly continuation of the Permo-Triassic rift system (Rowe & Burley 1997) that extended from the Wessex Basin to the Scottish Inner Hebrides. Thermal subsidence prolonged rifting until the mid-Triassic, with normal faulting during the early Triassic and Jurassic modifying the basin morphology. Tertiary contraction generated uplift of up to 1500 m (Knott 1994;Beach et al. 1997;Ware & Turner 2002), with intra-Triassic uplift resulting in 700 -900 m of erosion (Mikkelsen & Floodpage 1997;Rowley & White 1998;Ware & Turner 2002). Potential organic-rich Carboniferous sourcerock sediments (Mikkelsen & Floodpage 1997) in the Cheshire Basin are overlain unconformably by the Permian Collyhurst Sandstone, Manchester Marl and the Kinnerton Sandstone (Rowe & Burley 1997). The Sherwood Sandstone Group (the focus of this study) overlies the Permian sediments, and is a 1500 m-thick succession composed of the Cheshire Pebble Beds Formation, the Wilmslow Sandstone Formation and the Helsby Sandstone Formation (Rowe & Burley 1997). The UK migrated from approximately 108 to 308 N of the equator during the Permo-Triassic (Tellam & Barker 2006).  (1)  The depositional environment of the Sherwood Sandstone Group in the Cheshire Basin was mixed aeolian and fluvial. The Triassic river systems flowed NW from the Cheshire Basin into the East Irish Sea Basin (Meadows 2006). Post-Triassic successions have been removed across most of the Cheshire Basin following Cretaceous and Tertiary uplift; the youngest deposits (Pre-Quaternary) within the Cheshire Basin are middle Liassic (Rowe & Burley 1997). Meadows (2006) provided a synthesis of existing stratigraphic nomenclature applied to the Early and Middle Triassic Sherwood Sandstone Group in NW England, including East Irish Sea Basin equivalent units based on well correlations that contain various oil and gas discoveries (Knott 1994;Meadows 2006).

Methods
Data were collected from the Cheshire Basin within the Sherwood Sandstone Group at Thurstaston (Fig. 1), on the western side of the Wirral peninsula, 7 km SW of Birkenhead. The presence of deformation bands at Thurstaston has been documented (Knott 1994;Beach et al. 1997); however; this paper provides the first detailed analysis of deformation band distribution, as well as a petrographical description and interpretation.

Field data
The outcrop of Lower Triassic Wilmslow Sandstone Formation at Thurstaston (Beach et al. 1997), composed of aeolian dune and interdune strata, provides world-class examples of sandstone deformation bands. The Thurstaston Sandstone Member is incorporated within the Wilmslow Sandstone Formation in this study following the most recent pronouncement on the Sherwood Sandstone from the British Geological Society (see Meadows 2006 and references therein). Spatial relationships between Lower Triassic aeolian dune and interdune facies have been well documented within the Cheshire Basin, UK (Mountney & Thompson 2002;Mountney 2012). Two outcrop locations (Fig. 1b) provide a three-dimensional view of the deformation bands: † Location 1 -Telegraph Road (Figs 2a & 3a), contains conjugate sets of deformation bands within the damage zone of three slip surfaces. The orientation, density and thickness of deformation bands were recorded with proximity to three slip surfaces (over a 5 m linear scanline perpendicular to faulting), within fine-grained (mean grain size of c. 170 mm) and overlying coarse-grained sandstone (mean grain size of c. 540 mm). † Location 2 -Thurstaston Common (Figs 2b & 4), allowed for a more extensive study of the relationship between slip surface proximity and deformation band density within coarse-grained sandstone.
Two 30 m linear north -south transects (approximately perpendicular to the strike of the major slip surface) allowed for measurements of fault and deformation band density, spacing and orientation. Using a collection of field photographs covering approximately 1 m 2 of exposure subparallel to bedding, the anastomosing geometry of the deformation bands was captured in detail.
Host-rock grain size and grain-size distribution data were collected using a Beckman Coulter LS13 320 Laser Diffraction Particle Size Analyser (LPSA) for five undeformed coarse-and finegrained sandstone samples. Owing to the friable nature of both the fine-and coarse-grained sandstones, samples required only gentle disaggregation by hand, and this was analysed under an optical microscope to ensure full disaggregation. As histograms are sensitive to bin selection, sorting was defined by the gradient of cumulative frequency curves (Cheung et al. 2012). Grain-size range was calculated by D 90 -D 10 . D 90 is the grain size at the upper bound of the 90% fraction, whereas D 10 is the grain size at the upper bound of the finest 10% fraction. X-ray diffractograms generated from PANalytical X'pert Pro MPD X-ray diffractometer (XRD) quantified mineralogy of both the host rock and the deformation bands within the fine-and coarse-grained sandstones layers.

Microstructural characteristics and petrophysical properties
Orientated samples were sectioned along a northsouth plane in order to reveal depositional and diagenetic features, prior to vacuum impregnation with blue epoxy resin to reduce friability and to highlight porosity. A Meiji 9000 optical microscope fitted with an Infinity 1.5 camera with Infinity Analyser software was used to carry out an initial reconnaissance of polished thin sections. Secondary electron images (SE) were collected using a Philips XL30 SEM equipped with an Oxford Instruments Secondary X-ray detector from gold-palladium-coated deformation bands and host rock. Backscattered electron images (BSE) were collected using a Hitachi (TM3000) scanning electron microscope (SEM) and Philips XL30 SEM. Using a Philips XL30 SEM equipped with a K.E. Developments Ltd cathodoluminescence (CL) detector (D308122), SEM-CL images were obtained at 10 kV and spot size 7. SEM-CL images took up to 25 min to collect and were gathered by integrating the signal of 16 frames using a slow scanning raster.
Helium porosimetry was used to calculate porosity within the undeformed host rock of both the fine-and coarse-grained sandstone (two core plugs per sandstone). Owing to the friable nature of both sandstones, the core plugs used for the helium porosimetry were not perfectly cylindrical, resulting in a porosity error margin of 4%. Porosity heterogeneity at a deformation band scale (typically ,1 mm) cannot be captured on the scale of a core plug (c. 25 mm in diameter) and, instead, a petrographical image analysis is typically used (Antonellini et al. 1994). It should also be noted that porosity values documented within the literature using helium porosimetry are typically higher than those calculated using digital image analysis (Anselmetti et al. 1998;). In order to calculate porosity within the deformation band and host rock, BSE images (converted to an 8-bit format) have been digitized in ImageJ Analyser (Schneider et al. 2012), creating an array of pixels that are assigned a grey-level intensity. Pixel segmentation was then undertaken using a thresholding formula in which black pixels (porosity) and grey pixels (host rock) were differentiated, allowing for the quantification of total optical porosity. Fifteen images (with varying fields of view ranging from c. 0.25 to 1.00 mm 2 ) have been analysed for both fine-and coarse-grained sandstone (i.e. for both the undeformed host rock and deformation band) to ensure accurate results (Ehrlich et al. 1991).

Field data
Host-rock properties. Laser particle size analyser and X-ray diffraction data are displayed in Figure  2. Both the coarse-grained sandstone (mean grain size of c. 540 mm) and the underlying fine-grained sandstone (mean grain size of c. 170 mm) are moderately sorted, with a grain size range of approximately 325 and 340 mm, respectively (Fig.  2c). Minor fining-upward sequences (medium to coarse grained) are present within the overlying cross-stratified foresets of the coarse-grained sandstone (aeolian facies). However, the majority of any pre-existing primary depositional structures are no longer recognizable owing to intense deformation. The underlying fine-grained sandstone (interdune facies) exhibits subhorizontal, centimetre-scale, wavy sandstone laminae with negligible change in grain size both temporally and spatially. X-ray diffraction analysis of the coarse-grained sandstone identified a dominance of quartz (96%), a small quantity of K-feldspar (4-5%) and a trace of illite. X-ray diffraction analysis of the fine-grained sandstone produced a slightly lower percentage of quartz (83%), and an increase in K-feldspar (11%) and illite (6-7%). In all tested samples, there is a negligible difference in the mineralogy (and mineral abundances) of the deformation band and the host rock. The coarse-grained sandstone is classified as a quartz arenite and the fine-grained sandstone is classified as a subarkosic sandstone (Fig. 2d) according the QFR classification (Folk et al. 1970). Fault kinematics and deformation band distribution. The Wilmslow Sandstone Formation (Thurstaston Sandstone Member) at Telegraph Road ( Fig. 3a) is faulted by three WNW-trending, highangle (.808), north-dipping, normal faults (with respect to bedding), with striations suggesting a minor component of right-oblique slip. Slip surface 1 (SS 1) has an offset of 64 cm, slip surface 2 (SS 2) of 19.5 cm and slip surface 3 (SS 3) of 7 cm: all are subperpendicular to the main NE-trending Formby Point Fault (Fig. 1a) that extends many kilometres northwards, forming a bounding fault to the Lennox oilfield (Yaliz & Chapman 2003). A northand south-dipping conjugate set (an acute angle of c. 558) of deformation bands display an east-west orientation at both Telegraph Road and Thurstaston Common, parallel to faulting (Figs 3c & 4d). Deformation bands are sporadic within the underlying fine-grained subarkosic sandstone beds and form in swarms within the overlying coarse-grained quartzarenite sandstone (Fig. 3). Deformation bands are largely confined to the overlying coarse-grained sandstone, and commonly end abruptly at the finegrained sandstone boundary (Fig. 2a). Deformation bands range from 0.05 mm to 1.2 cm in width, displaying mm-scale offset. The relationship between deformation band density and fault proximity at Telegraph Road is displayed in Figure 3b. Deformation bands broadly increase in density with proximity to the faults. There is no obvious correlation between deformation band density and the magnitude of fault offset.
Slip surfaces observed at Thurstaston Common (Fig. 4) are a continuation of the high-angle WNW-trending faults that can be observed in crosssection at Telegraph Road (Figs 1b & 3). Where slip planes have initiated and offset has occurred, deformation bands tend to localize and orientate in broad zones in proximity to the slip planes. The anastomosing map pattern of the deformation bands (Fig. 4a), as well as the linkage structures (Fig. 4c), can be recognized at the mm-to cm-scale.

Microstructures and petrophysical properties
At both study sites (Telegraph Road and Thurstaston Common), deformation bands are classified as cataclastic bands, with the mechanical fracturing of grains being the predominant deformation mechanism. Secondary electron images (Fig. 5ac) highlight the friable nature of weakly quartzcemented host-rock grains and intense localized cataclasis, limited to the deformation band core, within the overlying coarse-grained sandstone. Cathodoluminescence images (Fig. 5e) reveal Hertzian grain-grain interaction, with the deformation band core composed of interlocking, fragmented quartz grains cemented by quartz.
The fine-grained host rock is moderately cemented (Fig. 6b), with pore-filling quartz reducing friability; localized comminution of grains produces a deformation band core composed of interlocking detrital clast fragments (Fig. 6c). Kfeldspar is preferentially fractured (Fig. 6d -f) within both the deformation band core and the proximal host rock, indicating a strong shear component with a K-feldspar grain being entrained into the deformation band (Fig. 6f). The undeformed host-rock porosity for coarse-and finegrained sandstone using helium porosimetry is 32 and 15%, respectively. Mean porosity data calculated using image analysis are as follows: the undeformed host-rock porosity of coarse-grained sandstone (Fig. 7a) is 26%; porosity has been reduced to 10% within the deformation band core in the coarse sandstones (Fig. 7a); undeformed host-rock porosity for the fine-grained sandstone (Fig. 7b) is 10%; porosity has been reduced to 4% within the deformation band core in the fine sandstones (Fig. 7b).

Deformation band distribution and fault proximity
Fault-zone architecture is well documented both in the field and in experimental studies (Antonellini et al. 1994;Antonellini & Pollard 1995;Caine et al. 1996;Faulkner et al. 2010). A typical fault zone comprises of a fault core surrounded by a damage zone (Faulkner et al. 2010;Schueller et al. 2013). The fault core is an area of localized strain that accommodates the majority of displacement (Faulkner et al. 2010;Schueller et al. 2013). Damage zones in porous sandstones form by growth of deformation bands prior to the initiation of a slip surface (Schueller et al. 2013). Shear strain, state of stress, rock type and microstructural deformation mechanisms are key controls in fault-zone architecture (Ngwenya et al. 2003). Hydraulic properties of faults and intrinsic properties of host rocks evolve spatially and temporarily, producing heterogeneous permeability.
At both Telegraph Road and Thurstaston Common, deformation bands broadly increase in density with fault proximity (Figs 3 & 4), consistent with 106 outcrop scanlines recording predominantly cataclastic band density in porous sandstones surrounding extensional faults documented by Schueller et al. (2013). Before the initiation of a slip surface, it is evident that deformation band density reaches a maximum of around 20 -25 bands per 30 cm section independent of fault displacement, analogous to critical microfracture density recorded within low-porosity granodiorite by Mitchell & Faulkner (2009). The trace of isolated deformation bands in outcrops at Thurstaston Common tend to be straight: however, zones of deformation have an anastomising profile showing linkage structures between neighbouring segments (Antonellini et al. 1994), similar to the duplex structure described by Cruikshank et al. (1991a, b). The presence of linkage structures suggests a sense of shear, as they resemble miniature restraining bends (Davis 1999). Deformation band lozenges (defined as the rock volumes between deformation bands) at Thurstaston Common (Figs 2b & 4b) closely compare to those documented within Goblin Valley, Utah, USA (Awdal et al. 2014). Early studies explained the development of closely spaced cataclastic bands (Aydin 1978;Aydin & Johnson 1978) in proximity to low-displacement faults (,10 m throw) by the strain-hardening model, showing an increase in deformation band density with fault displacement (Nicol et al. 2013). Cataclastic deformation bands have been suggested to strengthen during formation, thus leading to subsequent band formation within relatively weaker wall rock, adjacent to the earlier-formed bands (Nicol et al. 2013). Density counts, the positive relief of deformation bands, linkage structures and microstructural analysis (porosity reduction, interlocking quartz fragments, intense grain comminution increasing grain angularity, shear compaction and preferential quartz cementation) at Thurstaston all support a strainhardening model, resulting in an increase in deformation band density with fault proximity (Nicol et al. 2013). Anomalous results, such as the spike in deformation band density within the mapping zone at Thurstatson Common (Fig. 4e), may be explained by an alteration in host-rock cohesion by neighbouring slip surfaces. Deformation band development explained by a geometric model (see Nicol et al. 2013 and references therein) infers that deformation bands are strain weakened and form clusters at geometric complexities or irregularities on faults. Further three-dimensional analysis of the fault geometry and a better understanding of the relative timings of slip-surface formation would be required to apply a geometric model at Thurstaston.

Distribution-localization of deformation bands as a function of intrinsic host-rock properties
Mineralogy. The mineralogy of the host rock is an important controlling factor, with different minerals having varying chemical stability, shape, strength and vulnerability to cleavage fractures (Aydin et al. 2006). Mineralogically mature, coarsegrained quartzarenite samples (Fig. 2) display highly localized cataclasis within the deformation band (Fig. 5), with little host-rock fracturing in comparison to the underlying subarkosic sandstone (Fig.  6). In addition to intense cataclasis within the deformation band core, feldspathic subarkosic samples in this study show a higher degree of grain fracturing within the host rock (Fig. 6). Because feldspar fractures at lower differential stress than quartz grains (Rawling & Goodwin 2003), an increase in hostrock deformation may be a result of a selective grain-size reduction of weak grains (Fig. 6e). Preferential feldspar grain-size reduction has also been documented within conjugate sets of deformation bands within poorly consolidated arkosic sands of the Vienna Basin, Austria, by Exner & Tschegg (2012). Intense cataclasis creating angular grains and broadening the grain-size distribution considerably lowers porosity as a result of more efficient grain packing (Main et al. 2001;Tueckmantel et al. 2012).
Porosity, grain size and sorting. Cataclastic deformation bands are common in high-porosity (c. 10 -35%) sands and sandstones deformed at low confining pressures of ,40 MPa at shallow depths of ,3 km (Nicol et al. 2013). Samples with high porosity have lower rock strength than low-porosity samples as pore spaces coalesce, thus increasing the likelihood of pore collapse and so promoting volumetric reduction deformation (Aydin et al. 2006). The critical minimum porosity for deformation band development and propagation is lowered by the addition of shear to compaction (Fossen et al. 2011). In order to advance the understanding of fluid migration into subsurface reservoirs, it is important to note that the distribution of porosity (and permeability) in deformed high-porosity sandstones can be markedly anisotropic (Farrell et al. 2014). Whilst mapping of a thin section using image analysis yields a more detailed microscale (mm-scale) porosity profile, porosity values may also be dependent upon the scale of the measurement and the thin-section orientation with respect to the orientation of the pores . As expected, helium porosity values are slightly higher than those calculated using image analysis as image analysis does not include microporosity. In addition to mineralogy and porosity, factors such as grain size (Zhang et al. 1990;Yin et al. 1993;Lothe et al. 2002) and sorting (Cheung et al. 2012) significantly alter the probability of deformation band development and propagation.
It is well documented that larger grain sizes deform under lower effective stresses than finergrain material (Zhang et al. 1990;Yin et al. 1993;Lothe et al. 2002;Schultz & Siddharthan 2005;Schultz et al. 2010;Tueckmantel et al. 2012). Since sandstones at Thurstaston have a similar sorting and fall within the porosity range that allows for deformation band development, it is assumed that host-rock grain size is the principal control on deformation band density. Coarser grains have few contact points, which leads to a larger stress concentration and promotes grain-size reduction (Zhang et al. 1990;Yin et al. 1993;Lothe et al. 2002) in the form of Hertzian grain-grain interaction (Fig. 8a). Hertzian fractures are explained by a complex stress field that is set up when a spherical indenter is pressed onto the surface of an isotropic material. The stresses under and around the indenter contact are compressive; however, outside the contact circle, a radially directed tensile stress is created (Frank & Lawn 1967;Master 2012). Results are consistent with deformation band development within Navajo Sandstone sequences with varying grain size and porosity values at Buckskin Gulsch, Utah, USA . The corresponding yield envelopes for layers within the Navajo Sandstone are documented to be largest for the finegrained, less porous sandstones, and smallest for the largest values of porosity and average grain sizes . The fact that there is a higher density of deformation bands within the overlying coarse-grained sandstone compared to the fine-grained sandstone suggests strain incompatibility between the layers. However, although the density of localized deformation may be different, more strain may have been accommodated through distributed deformation via porosity loss (without fracture) in the finer-grain-sized unit.
In considering mineralogy and grain-size distribution, it is possible to surmise that an increase in K-feldspar content within sandstones will produce a wider grain-size distribution, since K-feldspar has been shown to fracture under lower differential stress than quartz (Rawling & Goodwin 2003;Exner & Tschegg 2012). Thus, a high K-feldspar content may potentially inhibit the development of deformation bands within more feldspathic sandstones, as a non-uniform grain-size distribution allows smaller grains to distribute the load over large particles, and so reduces stress concentrations between grains (Sammis & Ben-Zion 2008;Cheung et al. 2012). Petrographical evidence (Fig. 8b) within fine-grained sandstones support the 'constrained comminution' model proposed by Sammis et al. (1987), with a localized increase in grain-size distribution within the deformation band core allowing for survivor (or relict) grains.

Implication for sandstone reservoirs
During the appraisal and development of oil and gas fields, analogue studies are helpful for predicting the potential impact on subsurface fluid-flow behaviour. Unfortunately, as the classification and petrophysical measurements of cataclastic deformation bands are not systematic in the literature, it is very difficult to yield a meaningful comparison of results from different study areas . However, by combining this study with other analogue studies and experimental datasets, it is likely that reservoir quality predictions will be greatly improved. Cataclastic deformation bands in the literature commonly display lower permeability than the host rock, maximum reductions being of the order of five-six magnitudes and average reductions being around two -three orders . Clusters of cataclastic bands have been shown to be as efficient seals as fault cores, withholding up to about a 1 m column of oil and CO 2 (Torabi 2014). The porosity reduction documented in this study would (locally, at least) greatly reduce the reservoir quality, acting as a baffle to fluid flow. For deformation bands to affect well performance, bands must extend over typical well drainage areas: 0.5-1 km 2 for onshore and shallow offshore wells; and 5 km 2 in deep-water wells . Although deformation bands are commonly confined within the damage zones of faults, examples of deformation bands extending over such a large scale have been documented: for example, deformation bands extend approximately 7.5 km 2 at the Valley of Fire, Nevada, USA . In addition to vertical and horizontal continuity and intrinsic host-rock properties, the reservoir-scale impact will also depend on their permeability, orientation, connectivity and abundance (Sternlof et al. 2004;Brandenburg et al. 2012). The addition of quartz cement, lowering the porosity within the deformation band core, further increases the likelihood of reservoir compartmentalization. Unless accompanied by quartz cement, deformation bands in North Sea reservoirs have not proved to be problematic to oil and gas production . Figure 9 provides a schematic synthesis of the likely distribution of deformation bands and resulting porosity loss associated with conjugate sets of deformation bands within two sandstones with varying intrinsic host-rock properties. If encountered within core, reservoir geologists may use a combination of analogue studies in order to predict the extent of subseismic deformation bands, and the impact on petrophysical properties and reservoir performance. From another point of view, subseismic fault-development mechanisms may be understood by the intrinsic geometry of damage zones connected to the processes of fault growth (Schueller et al. 2013).

Specific importance to the East Irish Sea Basin
The Wilmslow Sandstone Formation, part of the Sherwood Sandstone Group, continues north and west into the East Irish Sea Basin, where it is locally known as the St Bees Sandstone Formation (Meadows 2006). The Sherwood Sandstone Group is Fig. 9. Schematic synthesis illustration. Deformation bands broadly increase with proximity to faulting. Deformation bands are predominantly restricted to the coarse-grained sandstone. Deformation bands in this study would lower the reservoir quality and potentially compartmentalize the sandstone reservoir. a significant petroleum reservoir within the East Irish Sea Petroleum Province (Duncan et al. 1998). A schematic burial history of the outcrop at Thurstaston, including the possible timing of deformation band development (Fig. 10), has been developed based on burial curves constructed by Rowley & White (1998) and the timing of WNWtrending faults suggested by Chadwick (1997). Apatite fission-track analysis from an outcrop 5 km NW of Thurstaston Common seemed to suggest a maximum palaeo-temperature, prior to early Tertiary uplift and cooling, of 90 -1008C (Green et al. 1997). However, in the undeformed matrix, the high intergranular volume implies limited burial and compaction, and negligible quartz cement (Fig. 6b) implies a maximum temperature much less than 808C based on depth v. host-rock quartz cement relationships (Worden et al. 2000).
Carboniferous source rocks matured during the Late Cretaceous -Early Tertiary, and then oil and gas migrated into the Lower Triassic Sherwood Sandstone (Duncan et al. 1998). The timing of the development of WNW-trending faults within both the Cheshire and East Irish Sea basins is poorly constrained. Knott (1994) suggested that WNWtrending faults within the Cheshire Basin formed under a NW-SE-trending maximum horizontal compressive stress regime, present since the Paleocene. It is unlikely that the WNW-trending faults at Thurstaston have formed under a compressive regime, since faults clearly display extensional offsets. Instead, it is possible that these faults, which have formed subperpendicular to, and cross-cut, the main northô -south-trending faults are transfer fault (Chadwick 1997). Adding to the complexity, the extensional direction may not have remained constant throughout the evolution of the Cheshire Basin (Chadwick 1997). Despite some uncertainty on the timing of deformation band development within the damage zones of the WNW-trending faults, faulting occurred prior to Tertiary uplift, source-rock maturation and hydrocarbon migration (Fig. 10). Oil and gas migration may have, therefore, been affected by deformation bands in the oil-and gas-bearing offshore equivalent outcrop. As the deformation bands are locally quartz-cemented (thus, further reducing porosity and permeability lower than achieved by simple comminution), careful analysis of cores and borehole image logs for deformation band occurrence and their stratigraphic constraints should be undertaken during field appraisal and the development of oil-and gas-bearing structures in the basin centres.
Conclusions † Deformation bands in the Triassic sandstone exposed at Thurstaston are cataclastic with a strong component of shear and porosity reduction. † Deformation bands at Thurstaston broadly increase in density (number/m 2 ) with proximity to faulting over a scale of several metres. † Deformation band distribution at Thurstaston is predominantly controlled by grain size. Deformation bands are more abundant within the overlying coarse-grained quartzarenite, and sporadic within the underlying fine-grained subarkose. K-feldspar is preferentially fractured in comparison to quartz grains. † Deformation bands in Triassic sandstones from Thurstaston have significantly altered the petrophysical properties of the intact rock. Porosity is substantially reduced relative to the matrix due to intense cataclasis and localized quartz cementation. The potential impact of deformation bands in nearby reservoirs in the same lithology could have a detrimental effect on reservoir quality and well performance.
We would like to acknowledge Oliver Hinds for his guidance on statistical grain-size analysis. We are also grateful to John Bedford for assistance during the operation of the helium porosimeter and James Utley for XRD analysis. This manuscript has benefited considerably from the comments of Dr D. Healey and an anonymous reviewer.