Macrofractures

The Eriboll outcrop belt is cut by faults mappable at 1:10 000 (Peach et al. 1907) and by opening-mode fractures. In addition, the outcrop belt is bounded by the ENE-striking Moine Thrust Zone to the east and is subdivided by a few large NNE- and WNW-striking faults such as those near Durness and Loch Maree. Small faults that cut the Moine Thrust Zone strike NNE and NE (020 to 065°) and NW and WNW, with wide dispersion. These late faults probably, in part, mark post-Caledonide strike-slip deformation or extension associated with opening of the Atlantic (Roberts & Holdsworth 1999). Some of the most prominent opening-mode fractures in Eriboll sandstones are associated with these late faults. Other fractures are present throughout the gently tilted sandstones, distant from faults or folds. These disseminated fractures, together with associated microfractures, represent penetrative regional deformation.

Disseminated fractures in Eriboll sandstones are planar, and, with some exceptions, opening displacements are normal to walls. Fractures have sharply defined, smooth walls, although on the millimetre scale fracture walls locally have low-amplitude sinuosity owing to cracking around, as well as across, grains. Most have steep dips (70–90°), and some are systematically aligned normal to bedding. Fractures are in parallel, systematic arrays with spacing of as much as several metres between large fractures.

Fractures include joints, or barren fractures, veins, defined as cement-filled fractures, and fractures having mixed attributes. These mixed attributes include exceedingly thin veneers of quartz cement on fracture walls, such that fractures are mostly open and thus resemble joints. Another attribute is pillar-shaped quartz-cement deposits that locally bridge fracture walls (Figs 4 and 6). These bridges commonly contain crack-seal texture, a feature generally associated with veins (Ramsay 1980). Similar isolated quartz bridges in otherwise open fractures are common in sedimentary basins (Laubach et al. 2004b). Fractures having these attributes are associated with and, as we show, are parts of the same fracture sets as arrays of millimetre- to micrometre-scale microfractures that are mostly entirely sealed with quartz cement. For simplicity we call all of these opening-mode fractures.

We interpret the veins, quartz-lined and quartz-bridged fractures and quartz-filled microfractures to have formed by some combination of remote loading and elevated pore-fluid pressure in a setting where quartz was precipitating. As we describe, differences in amount and texture of quartz fill can be accounted for by fracture age and size.

Quartz-filled fractures are long and thin; they have high aspect ratios. Owing to the small colour contrast between quartz-filled fractures and quartzite, in some outcrops even relatively large fractures are readily overlooked (Fig. 2). Nevertheless, we found a wide range of fracture sizes. Fractures range from microfractures visible only with scanned CL to macrofractures. On large bedding-plane-parallel exposures such as those near Loch an Nid (Fig. 2), fractures have trace lengths of centimetres to 100 m or more. The sum of residual porosity thickness and mineral fill in each fracture is the cumulative opening displacement, or kinematic aperture (Marrett et al. 1999). Apertures visible in outcrop range from 0.075 mm, the smallest that we could accurately measure with a hand lens and comparator (Ortega et al. 2006), to as much as 16 mm. We found significantly more small fractures than large in each fracture set.

Quartz cement present in fractures has dull luminescence mostly indistinguishable from quartz cement in the rock mass. Older fracture quartz is reddish, whereas younger quartz is blue (Fig. 4). Cement textures differ by fracture size. Fractures having apertures of c. 0.01 mm or less commonly have uniform luminescence. Fractures having apertures of c. 0.1 mm locally have varied internal textures, reflecting competition of crystals to fill space. Large fractures may have crack-seal textures marking multiple opening events. Some also contain porosity.

Some opening-mode fractures have been reactivated by slip approximately parallel to fracture walls, as shown by displaced grains, overprinting fault textures within some fracture-cement deposits and, locally, by tail cracks. They are therefore now small faults. This slip is generally inconspicuous in the field.

Microfractures

Quartz-filled micro-veins or microfractures are widespread along the entire Eriboll outcrop belt. Most are opening-mode fractures except in the vicinity of faults. Scanned CL shows that microfractures have smooth, sharply defined walls and sharp tips and are filled with quartz (Fig. 5). A few have traces of feldspar cement. Infilling quartz is in optical continuity with host grains and cement, so microfractures are inconspicuous fluid-inclusion planes in transmitted-light microscopy. Scanned CL shows that fluid inclusions are mostly sealed along medial lines by quartz deposits that grew from fracture walls.

By far the most common microfractures have straight traces and are aligned in subparallel sets. Most dip at high angles to ESE-dipping beds. These fractures have high aspect ratios; they are narrow relative to their width and are commonly in en echelon and relay patterns with straight, overlapping tip segments. The longest, transgranular microfractures cut across and around grains and across most cement. Fracture pathways are little affected by grain-scale mechanical heterogeneity, so when these fractures grew, the rock probably already had low porosity and behaved as a continuum. Inspection of fracture pathways nevertheless shows that microfractures of the oldest set have paths that locally curve around grain boundaries or hook into and end at cement patches between grains that probably mark pore space that existed when fractures grew. These patterns are consistent with fracture in rock that was less consolidated and more porous than it is now. Although differences are subtle, younger fractures have straighter traces.

Microfractures vary widely in size. Fractures grade with no discontinuities from intragranular fractures with lengths of less than 0.01 mm through transgranular fractures having lengths of 1 mm or more to macroscopically visible fractures. Kinematic aperture ranges from as little as 0.0001 mm to more than 1 mm. Microfractures are abundant even in otherwise unstructured rocks. For example, a line of observation of 138.5 mm length at 150× along the centre of one image mosaic intersected 2095 microfractures that range in aperture from less than 0.001 to 0.98 mm. The aperture population is well fitted by a power law across roughly two orders of magnitude. Cumulative apertures in this case record strain of 4.9% (Gomez & Laubach 2006), but values are mostly lower, typically ranging from 0.5 to 1%.

Not all microfractures are parts of fracture sets. Eriboll Formation sand grains contain relict microfractures that are within grains and that may be abruptly truncated at grain edges; these do not mark in situ deformation but are inherited from source areas of detrital grains. Other microfractures are products of grain-scale stress concentrations. These radiate from grain contacts and typically have triangular shapes and curved trajectories. Inherited and early compaction fractures are crosscut by all systematic fracture sets. Microfractures that can be ascribed to early compaction by crosscutting relations are sparse, a pattern consistent with compaction and cementation during gradual burial (Makowitz et al. 2006). Late microfractures having irregular shapes and radiating patterns are prevalent in fault zones, where they are associated with micro-faults.

Fracture sets

Younger fractures cut across older fractures, defining crosscutting relationships that show the relative ages of the fractures (Hancock 1985). Crossing relations occur when fracture-normal stress closes older fractures or where older fractures are sealed with cement. If a fracture in one orientation is open when a fracture in another orientation grows toward it and they intersect, the younger fracture may curve or abut the older fracture. In Eriboll Formation sandstones, prevalence of younger fractures cutting older fractures and the cement within them shows that the crosscut fractures were mostly sealed when later fractures grew (Figs 4 and 5).

Along the outcrop belt, we found five sets of opening-mode fractures having consistent crosscutting relationships and strikes. From oldest to youngest, fractures in these sets strike north, NW to WNW, NE, west, and north (Fig. 9; designated sets A–E). In most localities at least three sets are present, and two other sets can be defined in many localities (Figs 10,11,12,13,14). In some outcrops, crosscutting and abutting relations are obscure, but scanned CL indicates consistent crosscutting relations regionally. Although crosscutting relations and preferred strike and dip define sets, sets also differ in patterns of cement fill and intensity.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 9.

Fracture sets containing quartz mineralization, Eriboll Formation sandstone, west of the Moine Thrust Zone.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 10.

Set A microfracture strikes (rose diagrams) for selected samples, Loch Eriboll to Loch Maree. n, number of fractures; c, per cent outer circle. Line and circle segment show mean and 95% confidence. The reference geological map for Figures 9,10,11,12,13 is simplified from Peach et al. (1907). (See Fig. 1 for National Grid references.)

• Download figure
• Open in new tab
• Download powerpoint

Fig. 11.

Set B microfracture strikes (rose diagrams) for selected samples, Ullapool to Loch Maree. Fracture intensity increases from west to east from locality 7 to 9. Line and circle segment show mean and 95% confidence.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 12.

Set C microfracture strikes (rose diagrams) for selected samples, Ullapool to Loch Maree. Localities with NE- and NW-striking fractures: 3, 10, 11, 15 and location b in Fig. 1.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 13.

Set D microfracture strikes (rose diagrams) for selected samples, Ullapool to Loch Maree.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 14.

Set E microfracture strikes (rose diagrams) for selected samples, Ullapool to Loch Maree.

Sets differ in orientation with respect to bedding. Set A is consistently aligned normal to bedding. Other sets have steep dips (>80°) but no systematic orientation with respect to ESE-dipping beds. Rotating measured strikes to horizontal around the regional strike of bedding produces only slight shifts in strike for even the oldest fracture set, so we use rose diagrams to report strikes.

Because they are less susceptible than large fractures to reactivation as faults (tectonic overprint) or to overprint by near-surface jointing, microfractures frequently provide the clearest evidence of crosscutting relations between sets (Figs 4 and 5) and the most sensitive measure of regional changes in strike. We therefore used macrofractures to identify sets, but microfractures to quantify crosscutting relations systematically and orientation and intensity patterns regionally. Figures 10,11,12,13,14 show regional and temporal shifts in transgranular microfracture strike.

In most samples transgranular microfractures define three or more sets that match those defined by nearby macrofractures. Sets have strong preferred orientations in a given sample, but average strikes shift 5–20° over distances of tens of metres to kilometres (Figs 1, 10–14). Such shifts probably reflect regional stress perturbations during growth and are similar to orientation shifts known from joint arrays at outcrops (Cosgrove & Engelder 2004) and quartz-lined fractures in cores from sedimentary basins (Laubach 1988).

Strike dispersion obscures regional patterns. Comparison of transgranular and intragranular fractures from the same samples shows that dispersion increases with decreasing microfracture size (Gomez & Laubach 2006) owing to grain-scale heterogeneities that also cause slightly larger fractures to have sinuous traces. Sharp resolution of fracture strike and sets comes from increasing the number of transgranular fractures by measuring larger image areas. For example, strikes of intragranular and transgranular fractures measured from only one of four thin sections give a single preferred strike and a faint signal of a second set, but transgranular and intragranular microfractures from four adjoining thin sections have two preferred orientation maxima that match the strikes of nearby macroscopic fractures.

Owing to multiple fracture sets, fracture-trace patterns are highly interconnected. But the prevalence of crossing relations between cement-filled fractures shows that open-fracture networks were never as interconnected to fluid flow as trace patterns imply. Moreover, fracture-trace maps show that within a given fracture set, contemporaneous large fractures are mostly isolated.

Quartz-bridge and crack-seal textures

Whereas microfractures of all sets contain quartz, large fractures of set A and sets D and E differ markedly in degree of quartz fill: set A fractures having apertures of as much as 16 mm are entirely sealed whereas even where apertures are less than 100 μm sets D and E are mostly open. For these, residual fracture porosity is marked by faceted quartz crystals lining open pore space (Figs 4b and 6). Depth of quartz lining along open fractures is typically only a few tens of micrometres. In the field, cement lining open fractures is inconspicuous owing to small crystal sizes.

Despite this contrast in porosity preservation, we find a close resemblance between structures and textures within cement fills of set A and sets D and E. Sets D and E contain isolated quartz cement deposits or bridges displaying crack-seal texture (Figs 4b and 6). Set A fractures contain textures that we interpret to be isolated bridge deposits that have subsequently become encased or fossilized by later quartz deposits (Figs 4a, 7 and 8).

Bridges are cement deposits surrounded by porosity or later cement (Laubach et al. 2004b). They may have narrow, approximately millimetre-wide pillar or rod shapes consisting of isolated crystals or masses of crystals. Numerous fluid-inclusion planes that parallel fracture walls are trapped along the centre-lines of fractures within crack-seal texture. Scanned CL shows crack-seal texture comprising cracks of a few to hundreds of micrometres scale that formed and filled with quartz while fractures were opening. Thin layers of unfractured zoned quartz overlap crack-seal texture along bridge edges, showing that precipitation continued after fracturing ended. In sets D and E these late, unfractured quartz layers are micrometres thick (Figs 4b and 6).

Bridges in sets D and E are commonly visible in the field. However, even in thin section, fossilized bridges in set A are difficult to discern owing to optical continuity of quartz in bridges and surrounding fracture. Bridge location within fractures is marked by zones of aligned fluid-inclusion planes. These areas are separated by wider zones that lack fluid inclusions that we interpret to be former fracture porosity that filled with quartz long after bridges formed (Figs 7 and 8).

Figure 8 reconstructs the opening history of a typical set A bridge in a fracture of 176 μm aperture. Parallel to the fracture wall, the bridge dimension is slightly greater than that of crosscut grains in adjacent host rock. Multiple fracture events built up deposits in bridges (Fig. 8). The 41 gaps that opened in the bridge during fracture growth are marked by fluid-inclusion planes, contrasting CL colours and textures, and by crosscutting patterns. The average width of these gap deposits is 3.5 μm. Restoration shows that in this example fracture opening was concentrated progressively in one fracture strand, and cement accumulation became concentrated in part of the fracture. Fracture quartz, zoned to massive outside the bridge, lacks fluid inclusions or crack-seal texture. Because zoned quartz overlaps and postdates crack-seal texture, this cement must have grown after the fracture ceased opening. Zoning patterns show that quartz grew as faceted crystals into void space from fracture walls and from the top and bottom of the bridge. Most of the quartz in this fracture postdates formation of the bridge. These textures show that set A fractures contained bridge deposits surrounded by porosity. Quartz precipitation not associated with crack-seal texture later destroyed fracture porosity. All deposits in the fracture are crosscut by set B and C microfractures, showing that this set A fracture was filled prior to set B.

Quartz deposits in Eriboll sandstone fractures closely resemble those in quartz-cemented sandstone core from depths of 2–6 km (Laubach 1988; Laubach et al. 2004b). Using fluid-inclusion and isotopic data and well-constrained burial and thermal histories, Becker et al. (2008) showed that wide bridges similar to those in set A can take tens of millions of years to form, yet during the same time only small quartz deposits accumulated on fracture walls outside bridges.

Recent experiments and modelling have clarified how isolated crack-seal bridges form. According to Lander et al. (2002), such bridges reflect quartz accumulation patterns that are governed by the precipitation step rather than advection. Recent studies of quartz precipitation suggest that at a wide range of burial depths and temperatures, rock surface area and temperature fundamentally control progress of quartz accumulation (Walderhaug 1996; Lander & Walderhaug 1999). Laboratory crystal-growth experiments have shown fast growth rates on noneuhedral crystals such as those on freshly broken fracture surfaces compared with slower growth rates on euhedral crystals (Lander et al. 2008). Together with slightly greater crystal growth rates in the direction of the quartz c-axis than in the direction of the a-axis, these rates are a key to whether overall cement precipitation keeps up with incremental fracture opening. Bridges arise when increase in fracture aperture is small for single fracture events (e.g. micrometres), rate of aperture increase integrated over geological time scales is less than rate of precipitation on non-euhedral surfaces, and new non-euhedral nucleation surfaces are periodically created by fracturing.

Fluid inclusions

Our CL observations show that fluid inclusions in fractures of all sizes and sets are trapped during quartz-cement precipitation mostly along centre-lines of the largest microfractures and within bridges in macrofractures. Fluid inclusions are typically equant but range from spherical to irregular and angular. Typical inclusions in our samples are smaller than 0.1 mm. Many are monophase or too small to see a bubble (>70%), but some consist of liquid and vapor.

Using methods outlined by Goldstein & Reynolds (1988), we measured fluid-inclusion homogenization temperatures in four samples. To minimize potential resetting by beam heating, we measured fluid inclusions before imaging samples with CL. Inspection of homogenization behaviour suggests that some inclusions are naturally reset. To minimize the potential for measuring reset inclusions, we used small fluid inclusions from alignments that cross several grains, in which bubbles are the same size in both grains. These provide consistent fluid-inclusion assemblages.

Fluid-inclusion assemblages from set A collected from near Loch an Nid (Fig. 10, locality 9) have homogenization temperatures that range from 160 to 170 °C. Another possible set A sample in a rock having strong set E overprint, from near Dundonnell Bridge (locality 2), has homogenization temperatures as low as 138 °C and salinities of about 3–6 wt% NaCl eq. These homogenization values and salinities are close to the values of 132 °C and 2.6 wt% NaCl eq. reported by Blumstein et al. (2005) for vein-filling quartz (of unspecified relative fracture age) collected farther north along the outcrop belt.

Set D fractures from near Dundonnell Bridge (locality 2) have aqueous inclusions that are largely monophase at room temperature, suggesting low-temperature trapping at temperatures less than about 65–70 °C. A group of biphase aqueous inclusions from the same fracture have homogenization temperatures of 70–80 °C. A set E assemblage in the same sample is within the same temperature range. These results support late, cool conditions for formation of set D and E fractures.

Syntaxial quartz overgrowths in Eriboll sandstone homogenize at 117 °C and have salinities of about 8 wt % NaCl eq. (Blumstein et al. 2005). Our results show that set A fractures homogenize at temperatures of as much as 160–170 °C; with pressure corrections, trapping temperatures could be higher by 20–30 °C. We did not obtain data on sets B or C, but the youngest set D and E fractures formed under cool conditions, at temperatures perhaps as low as 70–80 °C.

Joints

Joints, or barren opening-mode fractures, postdate the youngest cement-lined fractures and strike NW, NE, north and west, contributing to landscape development (Geikie 1865; Gregory 1927; Auden 1954). Exceptionally large bedding surfaces reveal broadly curved arrays of joints in association with Pleistocene glacial chatter marks.

Fracture sequence

Fractures and microfractures are common even in the least-deformed Cambrian rocks of NW Scotland. Crosscutting relations, strike, and porosity preservation show that at least five sets (A–E) formed progressively in response to several deformation events. When they formed, each set of steeply dipping fractures contained the maximum horizontal stress direction. The shift of fracture set strike from north, to WNW to NW, to NE, to west, and back to north therefore constrains orientation of the least principal stress, which would have been normal to fracture strike.

Set A fractures are the only set that is systematically perpendicular to bedding. Beds now dip ESE, and we interpret set A to have been tilted along with the beds. Set A may have formed when Cambrian beds were flat lying. The timing of tilting is uncertain, but probably it was either concurrent with or after emplacement of the Moine Thrust Zone (Elliot & Johnson 1980). Set A could therefore predate emplacement of the Moine Thrust Zone. Set A is not kinematically compatible with Moine Thrust Zone deformation and is not localized near folds or faults. Set A might therefore record a stress trajectory in a platform or basin setting in Laurentia. Set A patterns, shapes and cement deposits are compatible with those of regional opening-mode fracture and joint sets identified in slightly deformed and nearly flat-lying rocks elsewhere (Cosgrove & Engelder 2004).

Deformation of the Cambro-Ordovician foreland succession occurred mainly in Silurian and Early Devonian times (McKerrow et al. 1991). The Moine Thrust Zone was emplaced west-northwestward onto the foreland c. 435–430 Ma (Strachan et al. 2002) and possibly as late as c. 408 Ma (Freeman et al. 1998). The last phases occurred at upper crustal conditions, with transient temperatures of about 250–275 °C (Johnson et al. 1985).

Kinematically compatible fractures are those whose orientations are consistent with a given tectonic loading direction (Hancock 1985). NW- to WNW-striking set B fractures are compatible with deformation associated with emplacement of NW- to WNW-directed thrusts. Set B is present in disseminated arrays up to at least several hundred metres below and west of the frontal thrust. Also, this set is locally more intensely developed near the Moine Thrust Zone or near small foreland folds that are aligned with the Moine Thrust Zone. Set B may be correlative with microfractures that have been identified to be associated with faults near the Moine Thrust Zone (Lloyd & Knipe 1992; Knipe & Lloyd 1994).

We interpret set B to be associated with Moine Thrust Zone emplacement, but set B also has the correct orientation to be associated with NW-striking pre-Moine Thrust Zone faults inferred from new mapping in the Assynt area (British Geological Survey 2007).

The structural association of set C is obscure. Fracture orientation provides ambiguous evidence because the steep dips and NE strike of set C could reflect subtle folding associated with the Moine Thrust Zone or an early phase of strike slip or extension associated with late NE-striking faults or some other event in the long interval of post-Moine Thrust Zone uplift. Fractures assigned to this set may have formed at markedly different times. Near localities 2 and 3 (Fig. 12) this set is associated with small post-Moine Thrust Zone NE–SW-trending, left-lateral faults, in some cases as tail cracks on earlier fractures, and so could record late Caledonide deformation (Treagus et al. 1999).

Sets D and E are late and crosscut all other sets; they contain only traces of quartz cement and thus preserve porosity. Because they are locally associated with or parallel to faults that cut the Moine Thrust Zone, and they are locally more plentiful near NE- and NW-striking normal faults that cut the Moine Thrust Zone, set D and E fractures probably record post-Moine Thrust Zone deformation. Post-early Ordovician deformation that tilted Cambro-Ordovician rocks 10–15° ESE may result from post-orogenic extension (Elliot & Johnson 1980), but sets D and E have no systematic attitude with respect to bedding and so might postdate tilting. Faults that cut the Moine Thrust Zone (Peach et al. 1907) are partly post-Permian and, although difficult to date precisely, may mark Mesozoic or Tertiary uplift and extension (Leedal 1951; Roberts & Holdsworth 1999).

Porosity in large set D and E fractures suggests that these fractures have not experienced protracted exposure to temperatures higher than 80 °C, which would produce rapid build-up of quartz cement (Lander & Walderhaug 1999). We found primary aqueous inclusions in set D and E fractures indicating fracture under cool (c. 80 °C) conditions, although results are insufficient to show a trapping sequence during cooling. The Highlands were a positive area throughout much of the late Palaeozoic and Mesozoic (Thomson et al. 1999). Devonian and older rocks reached temperatures of c. 110 °C or more in the Devonian to early Mesozoic (Thomson et al. 1999). Cooling may have accelerated in the early Tertiary, when landscape patterns suggest an additional 1 km of uplift (Hall 1991); at some point on the uplift and cooling path, quartz precipitation would have ceased. Sets D and E could therefore be Mesozoic to Tertiary in age.

Using cement to date set A

Dating of set A can be tested using inferred quartz accumulation rates, fluid inclusions and thermal history (Fig. 15). Although Eriboll sandstone thermal history is imprecisely known, it is sufficiently well known to constrain timing of set A when combined with recent insights into how quartz bridges form.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 15.

Fracture timing in context of thermal history. F_A to F_E indicate approximate time of sets A to E. Notes: 1–3, part of Palaeozoic time–temperature curves for two Cambrian sandstones from Laurentia (Flathead and St. Peter sandstone) Eriboll analogues; Cambro-Ordovician burial history for Durness Group from Smith & Rasmussen (2008) falls between these brackets; 2, onset of quartz cement from grain rim fluid inclusions (Blumstein et al. 2005); 3, MTZ, emplacement of Moine Thrust Zone (Freeman et al. 1998; Goodenough et al. 2006); 4, maximum temperature: filled circle, base of upper plate of Moine Thrust Zone; bar, lower plate and duration (Johnson et al. 1985); 5, maximum temperature ranges; youngest part of Durness Group near Durness has conodonts of CAI 4–5 corresponding to minimum temperatures of 350–400 °C (P. Smith, pers. comm.) but maximum temperature to the south may have been less, as Downie (1982) recorded acritarchs from beneath the sole thrust at Loch Assynt, suggesting temperatures there not above 200 °C (Brown et al. 1965; Johnson et al. 1985; Kelley 1988; Johnstone & Mykura 1989); 6, conjectural uplift history, extrapolation from maximum temperature estimate to 8; 7a, c. 470 Ma and 7b, c. 432 Ma Ar–Ar ages for authigenic K-feldspar interpreted to indicate fluid flow episodes (Mark et al. 2007); 8, uplift (Thomson et al. 1999); 9, uplift (Hall 1991). Dashed line, sets A, D, E strike and fluid inclusion homogenization temperatures (°C).

Figure 15 summarizes published constraints on Cambrian Eriboll sandstone temperature–burial history. Stratigraphical evidence for most of this history is missing, along with rocks above the Durness Group shelf carbonates. Gradual Cambro-Ordovician burial under low geothermal gradients is compatible with what stratigraphical evidence is preserved and with Eriboll sandstone intergranular volumes and compaction fracture abundances. Such an early burial history resembles that of Cambrian sandstones deposited elsewhere on the Laurentian platform (Makowitz et al. 2006, and references therein).

Rate of burial must have accelerated because Eriboll sandstones were buried to depths of about 11 km during Moine Thrust Zone emplacement in the Silurian–early Devonian (c. 435–430 Ma, possibly as late as 408 Ma) (Freeman et al. 1998; Goodenough et al. 2006). Estimated maximum temperatures for the Eriboll sandstones range from 150 to 350 °C and were probably at least 250–300 °C (Johnson et al. 1985). Organic maturation indices (Downie 1982) and lack of metamorphism in underlying Torridonian shales (Van de Camp & Leake 1997) suggest short duration of high temperatures, perhaps 20 Ma or less (Johnson et al. 1985).

The upper plate of the Moine Thrust Zone had cooled to below 300 °C by 425 Ma (Brown et al. 1965; Kelley 1988). Evidence from depositional patterns in nearby Devonian strata and apatite fission-track studies (Thomson et al. 1999) suggest that Eriboll sandstones were at temperatures of only about 100 °C (and depths of c. 3 km) by 300 Ma. These results are consistent with rapid uplift after maximum burial followed by gradual uplift to the present (Hall 1991).

Clearly, there is uncertainty in burial history, maximum temperatures achieved and their duration. We assume that foreland rocks experienced 250–270 °C at Moine Thrust Zone emplacement (Johnson et al. 1985) and uplift to high and cool conditions thereafter (Thomson et al. 1999; Trewin & Rollin 2002) (Fig. 15).

Crack-seal texture in isolated set A quartz bridges shows that these fractures formed where quartz was precipitating. Bridges probably form under conditions where quartz accumulation rate was governed by the precipitation step. Fluid inclusions having homogenization temperatures of 160–170 °C were trapped during progressive fracturing as bridges formed, although, with pressure corrections, trapping temperatures could have been higher. The shape of Eriboll temperature history suggests that set A passed through a high temperature range twice. Thus two ages for synkinematic quartz are possible, during burial prior to Moine Thrust Zone emplacement and during uplift after thrusting.

Lander et al. (2008) demonstrated that where quartz accumulation rates are governed by the precipitation step, rate of accumulation is sensitive to thermal history and to whether quartz was accumulating on euhedral or non-euhedral surfaces. Rapid cement accumulation on non-euhedral surfaces can account for thick but localized bridge deposits in fractures, because the non-euhedral surfaces are renewed with each fracture event. However, after fractures cease opening, cement accumulation between bridges progresses at the far slower, euhedral rates, although specific rates will vary with the rock's thermal history. Slow, euhedral accumulation rates accurately model quartz abundance patterns observed in sedimentary basins having well-constrained thermal histories (Lander & Walderhaug 1999; Lander et al. 2008). These rates are derived from observation of natural accumulations (Walderhaug 1996) and from experimental quartz growth (Lander et al. 2008). Quartz accumulation rates can also be inferred from bridged fractures where duration of fracturing and thermal history is known from fluid-inclusion populations that track opening history where thermal history is known independently (Becker et al. 2008). Using these rates, we can therefore explore whether set A fractures filled during burial or uplift.

The widest sealed set A fracture that we observed has an aperture of 16 mm. Synkinematic quartz in the bridge was deposited at 160 °C, but most of the fracture is filled with postkinematic quartz, similar to that in the example in Figure 8. The temperature of postkinematic cement deposition is unknown, because as a result of slow accumulation this quartz lacks fluid inclusions. Using rates from Lander et al. (2008) at 160 °C, the euhedral accumulation rate is about 3 μm Ma⁻¹, so a 16 mm fracture with accumulation from both fracture walls would not seal if the fracture formed after maximum burial while rocks were cooling, given what we know of Eriboll burial thermal history. However, for a fracture formed before maximum burial and subjected for a short time to temperatures of 275 °C or more, a 16 mm fracture could seal in 50 Ma. Using the same rates, the fracture in Figure 14 could seal in 2 Ma. This inference suggests that fluid inclusions in synkinematic quartz in set A fractures must have been trapped during burial and that fracture formation predates the Moine Thrust Zone. Postkinematic quartz accumulation may have been faster than we estimate if advection was important during part of this history; for example, during thrusting. Postkinematic quartz must have accumulated during and after Moine Thrust Zone emplacement because it is the only time that Eriboll sandstones reach a high enough temperature (>200 °C) for long enough to cement the large set A fractures.

Sealing of set A during burial is compatible with set A being overprinted by Moine Thrust Zone-related set B fractures and with the redder CL signature of set A quartz, which could arise from protracted thermal exposure. Structural and diagenetic evidence agrees that set A fractures predate the Moine Thrust Zone. North-striking set A could thus mark a Palaeozoic east–west least horizontal stress trajectory in Laurentia. Such a within-plate stress field could reflect forces exerted at plate boundaries (Heidbach et al. 2007) or a north-trending basin. Evidence of dominantly eastward palaeoflow and regional dip is compatible with a basin axis to the east (Swett 1969; McKie 1989, 1990), and elsewhere in this region of Laurentia there is evidence of north–south-trending facies belts (Higgins et al. 2001).

Controls on ephemeral fracture porosity

For Eriboll sandstone fractures, quartz precipitation accompanies fracture growth, initially without sealing large fractures. However, with time, cementation destroys fracture porosity, which is ephemeral owing to quartz precipitation. Where the precipitation step governs quartz accumulation, the potential for fractures to persist as fluid conduits depends strongly on temperature and fracture size. Our calculation for sealing set A and similar calculations for sets D and E imply that the duration of fracture sealing, for large fractures, can be of the order of millions and perhaps tens of millions of years in deep-basin settings, similar to the persistence and destruction of fracture porosity documented in young sedimentary basins by Becker et al. (2008). Persistent fracture porosity helps explain how an inactive Moine Thrust Zone could nevertheless have been a conduit for the Permian, Mesozoic, and Tertiary fluid flow evidenced by palaeomagnetic data (Blumstein et al. 2005).

In contrast, because millimetre-scale and smaller fractures have small volume relative to surface area, they will tend to seal readily. Rapid rates of microfracture sealing extrapolated from laboratory observations (Brantley 1990), inferred using rates from Lander et al. (2008) or implied by fracture porosity in natural examples in basins having well-constrained thermal and fracture sealing histories suggest that even the youngest and coolest Eriboll sandstone microfractures may have sealed in as little as a few thousand years.

All Eriboll sandstone sets have many more small fractures than large. For a given set, fracture porosity and surface area are concentrated in fractures having apertures less than 0.1 mm owing to size distributions that approximate power laws (Marrett 1996). Although all microfractures in a set are unlikely to have formed at one time, these size distributions, and crosscutting relations with other sets, imply high disseminated fracture porosity that is transient. For a specific deformation event, diagenesis rapidly converts penetrative arrays of mostly small fractures into sparse arrays of large and open fractures. Large quartz-bridged fractures may linger as potential fluid conduits, possibly even where they are misoriented with respect to evolving loading conditions that might otherwise induce mechanical closure. According to models of fluid flow in disconnected fracture systems (Philip et al. 2005), flow would tend to be in uniform fronts through penetrative arrays but might be channelized when only a few large fractures remain. Eventually, even large fractures like some of those in set A are closed. However, contrary to widely held views that emphasize that fracture systems are highly sensitive to and close in response to changing loading conditions (references given by Laubach et al. 2004a), the demise of these fractures is due to the temperature-dependent geochemical process of quartz precipitation. This hypothesis may be testable in active thrust belts if similar fracture arrays arise from episodic seismicity. Given inferred time scales of years to tens of years for filling microfractures at depth, seismic anisotropy caused by fractures might decay in a measurable and recognizable progression as time- and size-dependent fracture sealing progresses.