Abstract
A palaeomagnetic study of different lithologies exposed along the Moine Thrust Zone from Skye to Durness, Scotland, has identified three chemical remanent magnetizations (CRMs), a potential fourth CRM, and one possible primary component that all reside in hematite. Local and regional fold tests suggest that the CRMs are post-folding and post-thrusting. A contact test in the Torridon Group sandstone indicates that a Permian CRM is localized in and near the fault zone. The Permian CRM, which is also found in the Durness Group, and associated alteration are direct evidence for post-orogenic activity, in which the thrusts vented excess heat during regional crustal extension. On the Isle of Skye, sandstones (Sleat, Torridon and Eriboll Groups) in and around the fault zone contain a Tertiary CRM interpreted to be related to hydrothermal fluids associated with Tertiary intrusions. A possible Mesozoic CRM is also present but it could represent the vector addition of the Permian and Tertiary CRMs. A Devonian CRM is interpreted to be related to hydrothermal fluids associated with Devonian volcanic rocks or to fluids triggered by late stages of the Caledonian orogeny. Geochemical and petrographical studies provide supporting evidence for multiple fluid flow events along the Moine Thrust Zone. These results show that palaeomagnetism can be used to date multiple fluid migration events and that dormant fault zones can act as conduits for flow.
Chemical remanent magnetizations (CRMs) residing in hematite are common in a variety of lithologies over a large geographical area in Scotland (e.g. Torsvik & Sturt 1988; Potts 1990; Parnell et al. 2000; Elmore et al. 2002). Most of these CRMs post-date the Caledonian orogeny and many are Late Palaeozoic in age. Because of their different ages and their occurrence in different geological terranes, all of these CRMs cannot be explained with one model, such as the orogenic fluid hypothesis, in which fluids are forced into the basin by thrust sheets (Oliver 1992). Other mechanisms, therefore, must be sought to explain their origin. Several studies have reported localized CRMs around faults (Torsvik & Sturt 1988; Potts 1990; Elmore et al. 2002) and veins (Parnell et al. 2000) in Scotland. These CRMs may be related to a regional-scale Late Palaeozoic remagnetization in Scotland proposed by Torsvik et al. (1989).
The objective of this study is to test if the Moine Thrust Zone in Scotland was a conduit for remagnetizing fluids. Palaeomagnetic and geochemical methods were used to test for evidence of such fluids, and to date and determine the origin of the fluids. Samples were collected from in and around the fault zone in Cambrian–Ordovician limestones or dolomites (Durness Group), Cambrian sandstones (Eriboll Group), and Precambrian sandstones (Torridon Group and Sleat Group of the Torridonian Supergroup). Palaeomagnetic analysis was used to identify remagnetizations in those rocks and fold tests were conducted to constrain the timing of remagnetization. A large-scale contact (fault) test was performed to determine whether the Moine Thrust Zone was a conduit for fluids or a barrier to flow. Petrographical, rock magnetism and geochemical studies provide evidence for the origin(s) of the magnetizations.
Geological setting of study area
The study area is located in NW Scotland along the Moine Thrust Zone (Fig. 1). Scotland lies on ancient continental crust known as the Hebridean craton (Harris 1991). The Archaean to Proterozoic Lewisian complex, exposed along and in the Moine Thrust Zone, represents the last stages of the growth and reworking of the Hebridean craton.
(a) Map of study area showing the major geological units along the Moine Thrust Zone from Durness south to the Isle of Skye. Inset map shows the location of the study area in Scotland. (b) Detail map showing sample locations on mainland across faulted recumbent syncline in the Torridon Group from site K18 and sites BB1–BB8. (c) Detail map showing sample sites on the Isle of Skye and the mainland near the fault contact. Symbols for the sample sites correspond to the lithology from which the samples were collected: □, Sleat Group (sst); ○ Torridon Group (sst); ▵, Eriboll Group (sst); ⋄, Durness Group (lst). (Modified from Johnstone & Mykura 1989, p. 33.)
The Stoer Group overlies the Lewisian rocks and was deposited during initial stages of rifting into a closed basin c. 80 km wide (e.g. van de Kamp & Leake 1997). The Stoer Group, composed of conglomerate, sandstone and shale (Johnstone & Mykura 1989), was derived from the local Lewisian rocks to the east and west based on palaeocurrent indicators (Nicholson 1993; van de Kamp & Leake 1997; Rainbird et al. 2001) and was deposited in alluvial, fluvial and lacustrine environments. Turnbull et al. (1996) gave an age of 1199 ± 70 Ma for the Stoer Group based on a calcite Pb/Pb age from a thin limestone bed.
The Torridon Group overlies Lewisian rocks or the Stoer Group (Fig. 2). On Skye, the Sleat Group is older than and conformable with the Torridon Group. It was deposited to the east and moved to its current position during the Caledonian orogeny (Stewart 1991; Nicholson 1993). The Sleat Group consists of grey or buff distal fluvial and lacustrine sandstones and shales (Johnstone & Mykura 1989).
Stratigraphic section showing the units sampled and the approximate positions of the sample sites. Radiometric dates on left side of diagram are from Turnbull et al. (1996).
The basal Torridon Group unit, the Diabaig Formation, was deposited in a closed basin composed of feldspathic synrift sediments similar to those found in the Stoer Group. The unit has a diagenetic age of 994 ± 48 Ma from a Rb–Sr whole-rock study of shales (Turnbull et al. 1996). The middle to upper Torridon Group was deposited in an open basin or passive margin with most of the sediment derived from the Laurentian craton to the NW and SE and not from a local source, as evinced by the abundance of subrounded to rounded quartz grains and the lack of feldspars (Nicholson 1993; van de Kamp & Leake 1997; Rainbird et al. 2001). Turnbull et al. (1996) reported an age of 977 ± 39 Ma from a Rb–Sr whole-rock study of shales in the Applecross Formation of the middle Torridon Group.
After a c. 400 Ma depositional hiatus and erosion of the Torridon Group, the Lower Cambrian Eriboll Group was deposited. This group consists of clastic sediments at the base representing a transgressive sequence of barrier island to tidal shelf deposits, overlain by storm-dominated shelf deposits (McKie 1990). The transgressive Cambrian deposits are dominantly quartz-cemented quartzarenites and are overlain by the Durness Group, a thick Lower Cambrian to Lower Ordovician carbonate sequence.
In the mid- to late Ordovician, Northern Scotland, a promontory on the Laurentian craton, experienced deformation and metamorphism associated with early Caledonian collision of an island arc system (Dalziel & Soper 2001). Baltica collided with Avalonia and they subsequently collided with Laurentia to complete the Laurussian continent during Silurian and earliest Devonian times (Torsvik & Rehnstrom 2003). This orogenesis resulted in the development of the Moine Thrust Zone, in which Proterozoic metasedimentary rocks of the Caledonian belt were carried over the foreland sequence of the Lewisian complex, Torridonian sandstones and the passive margin Cambro- Ordovician shelf quartzites and limestones (Coward 1988). The Moine Thrust Zone or ‘Zone of Complication’ (Peach et al. 1907) consists of centimetre- to kilometre-scale thrusts and folds. The rocks in the thrust zone have been subjected to low-grade metamorphism (Johnson et al. 1985). This zone forms a narrow band that trends from Loch Eriboll south to the Isle of Skye, and is bounded on the east by the Moine thrust fault. The Moine Thrust Zone varies in outcrop width from a few hundred metres to 19 km (Johnstone & Mykura 1989). The exact timing of movement along the Moine thrust and associated thrusts as well as the associated folding and tilting is difficult to ascertain. Dates of 430 ± 4 and 439 ± 4 Ma were determined from U–Pb age dating of zircons found in the Loch Borralan and Loch Ailsh intrusive complexes within the Assynt window, respectively (van Breeman et al. 1979; Halliday et al. 1987). The Loch Ailsh complex has been interpreted to have intruded prior to the Ben More thrust, one of the thrusts in the Moine Thrust Zone, and the Loch Borralan complex was intruded after the main movements along the Ben More thrust (Brown 1991). Johnson et al. (1985) concluded that major thrust movement in the Moine Thrust Zone ceased about 425 Ma ago.
In the Devonian, extensional faulting along with magmatic and volcanic activity occurred in Scotland (Butler & Coward 1984; Serranne 1992). The early Permian rifting in the North Atlantic involved siliciclastic sedimentation in extensional basins and widespread extrusive and intrusive magmatic activity (Francis 1991; Hitchen et al. 1995). Coward et al. (1989) stated that normal faults developed in the hanging wall of Caledonian thrusts, suggesting that the thrust zones were reactivated during this extension (Hitchen et al. 1995). The normal fault bounded basins were part of the initial stages of the breakup of Pangaea and the opening of the Atlantic (Hitchen et al. 1995). The major rift eventually developed west of Scotland (present orientation).
In the Tertiary, a central igneous complex was intruded on the Isle of Skye in the SW of the study area, part of the much wider North Atlantic magmatic province. This intrusion has been dated to 55 Ma (Dagley et al. 1990). Associated with this Tertiary intrusion is a dyke swarm, part of which cuts through the study area on Skye to the south (Johnstone & Mykura 1989).
Palaeomagnetic and other methods
Samples (5–8 for most sites) from the Durness Group (Cambrian–Ordovician), Eriboll Group (Cambrian), Torridon Group (Precambrian) and Sleat Group (Precambrian) (Figs 1 and 2) were collected using a hand-held drill and oriented with an inclinometer and Brunton compass. Samples were collected from 15 sites on the Isle of Skye, four between Kyleakin and Balmacara, 11 sites from Bealach na Ba, seven sites at the head of Loch Kishorn, 12 sites near Ullapool, eight sites near Knockan, and five sites south of Durness (Fig. 1). At Bealach na Ba, specimens were collected near the Kishorn Thrust and progressively further away from the thrust as part of a contact (fault) test. The specimens were cut to standard lengths (2.2 cm) in the laboratory.
The natural remanent magnetization (NRM) was measured using a 2G Enterprises cryogenic magnetometer with d.c. squids in a magnetically shielded room. Most specimens were subjected to stepwise thermal demagnetization in 19 steps by heating them in an ASC Thermal Demagnetizer for 60 min. Some specimens were subjected to alternating field (a.f.) demagnetization to 120 mT in a 2G Automated Degaussing System before undergoing thermal demagnetization. The thermal and a.f. demagnetization data were plotted in Zijderveld (1967) diagrams using the Super-IAPD99 software. Magnetic directions were determined using principal component analysis (Kirschvink 1980) with mean angular deviations (MAD) of less than 15°. Fisher (1953) statistics were used to compute the mean directions. The Watson & Enkin (1993) parameter fold test was conducted on site means with n ≥3 and α95 <21°.
Isothermal remanent magnetization (IRM) acquisition and tri-axial IRM decay was performed on one specimen from each site where a stable remanent direction was found. An impulse magnetizer was used to apply a stepwise IRM. Subsequently, three perpendicular IRMs (120 mT, 500 mT and 2500 mT) were applied and the specimens were thermally demagnetized to give the tri-axial IRM decay curves.
Petrographical studies were conducted by examining thin sections in transmitted and reflected light to identify magnetic mineralogy and diagenetic phases. Also, the carbonate thin sections were classified according to Dunham (1962) and the sandstone sections according to Folk (1974). Samples of representative lithologies were also examined by SEM at the University of Oklahoma. Samples were examined in backscatter mode and elemental abundances were determined by energy-dispersive X-ray analysis (EDAX).
Stable isotope analyses were performed on whole-rock dolomite samples from 14 sites from the Durness Group using the procedure of Rosenbaum & Sheppard (1986). Each sample was ground to a fine powder and digested in 100% phosphoric acid for 1.5 h at 100 °C. The resultant CO2 gases were isolated cryogenically and analysed using a Finnigan Delta E isotope ratio mass spectrometer. The δ18O values were corrected to 25 °C using the fractionation factors of Rosenbaum & Sheppard (1986). The δ13C and δ18O values are reported in per mil relative to PDB carbonate. The standard error in our laboratory for replicate analyses of a pure carbonate is ±0.03‰ (δ13C) and ±0.2‰ (δ18O).
Fluid inclusion analyses were conducted at the University of Aberdeen following the methods described by Elmore et al. (2002).
Results and interpretations
Palaeomagnetism
The thermal demagnetization of specimens resulted in the identification of five magnetic components (Table 1). One or more components were found in each lithology sampled along the length of the Moine Thrust Zone. In some cases, different specimens in the same site contain different components (Table 1). The demagnetization vector in some specimens has curved trajectories, suggesting overlapping components, and individual components could not be resolved. Because of these issues, the numbers of specimens that contain a specific component in some sites are low (Table 1). Some specimens also contain multiple components (Fig. 3).
Paleomagnetic results
Orthogonal projection diagrams (Zijderveld 1967) of representative stepwise thermal demagnetizations of specimens. Open symbols, vertical projection; filled symbols, horizontal projection. (a) Torridon Group (sst) within the fault zone (between Kishorn and Moine Thrusts) showing Component 1 removed by thermal demagnetization. (b) Torridon Group (sst) west of the fault zone showing Component 1 at moderate temperatures and Component 4 at high temperatures. (c) Sleat Group (sst) from Skye showing removal of Component 2. (d) Eriboll Group (sst) in the Assynt area showing removal of Component 3 at moderate temperatures and a potential Component 1 at higher temperatures. (e) Durness Group near Kishorn thrust showing removal of Component 3. (f) Torridon Group (sst) outside thrust zone showing removal of antipode of Component 4 at high temperatures and Component 1 at low temperatures. Component 5 with northerly declinations and shallow inclinations may also be present at intermediate temperatures. (g) Torridon Group (sst) outside thrust zone showing removal of Component 5. Components are delineated by grey lines along the apparent inclinations for each diagram.
Component 1
Component 1, with southerly declinations and shallow negative inclinations (Fig. 3a and b), is identified at 19 sites from the Sleat Group, Torridon Group, Eriboll Group and Durness Group (Table 1), and was removed from 300 to 680 °C. It is found along the entire length of the Moine Thrust Zone, but is not common on Skye.
Two fold tests were conducted to constrain the timing of remanence acquisition. One fold test was performed across the faulted recumbent syncline. Four (BB1, BB2, BB7, K18) of the six sites with n ≥4 and α95 (the cone of 95% confidence about the mean direction) <21° that contain this component were used in the fold test (Fig. 1a and b, Table 1). The K18 site is overturned and across the fault in Loch Kishorn (Barber 1965) from the other sites in the fold test. Using the Watson & Enkin (1993) fold test, the best grouping for these sites is at 19.3 ± 25.2% unfolding (Fig. 4, Table 2). Although the error bar is large, the post-tilting result (0% unfolded) is included within the 95% confidence limit for the fold test, suggesting that the magnetization was acquired after thrusting and folding, which occurred in the late Silurian–early Devonian (Elliott & Johnson 1980). Therefore, the mean geographical direction (declination (D) 188.9°, inclination (I) −11.2°, k (precision parameter, after Fisher 1953) 35.5, α95 15.6°) was used to calculate the pole of 38.0°N, 163.1°E (Table 2).
Summary of palaeomagnetic results (mean component directions, fold test means, and pole positions)
(a) Equal area projections of site means for Component 1 from the faulted recumbent syncline in geographical and stratigraphic coordinates. Open symbols, negative inclinations; filled symbols, positive inclinations. The α95 confidence circles are shown around the site means. (b) Graph of percent unfolding v. grouping (k) for the fold test (Watson & Enkin 1993) of the faulted recumbent syncline. The continuous grey line marks the maximum and the dashed grey line represents the significance for the fold test (Watson & Enkin 1993).
A regional fold test was also conducted using all sites containing Component 1 with n ≥4 and α95 <21° (Table 1). The best grouping for the regional fold test is at 8.0 ± 25.1% unfolding, which is not statistically different from a post-tilting result (Table 2). This result is similar to that for the faulted recumbent syncline and Component 1 is interpreted to have been acquired after thrusting and folding. The mean geographical direction was used to calculate the pole of 40.0°N, 162.4°E (Table 2). Comparing the poles from both fold tests with the European apparent polar wander path (APWP) from Van der Voo (1993) indicates that Component 1 is of Permian age (Fig. 5).
Apparent polar wander curve (Van der Voo 1993) with the poles and calculated β95 for Components 1–5. □, Recumbent syncline and regional fold test result of Component 1 at 0% unfolding; ▵, regional fold test result on Skye of Component 2 at 0% unfolding; ○, regional fold test result of Component 3 at 0% unfolding; open star, mean for Component 4 at 100% unfolding; ⋄, regional fold test result for Component 5 at 0% unfolding. The dashed circle represents the α95 for the mean Torridon pole from Buchan et al. (2000).
Component 2
Component 2, with southerly declinations and steep negative inclinations (Fig. 3c), is found at seven sites from the Sleat, Torridon and Eriboll Groups on the Isle of Skye (Table 1). This component was removed from 300 to 680 °C. Component 2 is well developed at sites LM1 and LM2, which are at the Moine thrust contact (Fig. 1a and c), but is not present c. 50 m away from the fault in samples from sites LM3, LM4 and LM5. This suggests that Component 2 could be associated with the Moine thrust. The other sites on Skye at which Component 2 is found (SI1, SI2, SI6, SO1 and S03) are also in close proximity to faults (Fig. 1c). A regional fold test with the five sites (LM1, LM2, SI1, SI2, SI6) on Skye indicates that the best grouping is −6.3 ± 5.1% untilting (Fig. 6, Table 2). This result suggests that the magnetization is younger than the late Silurian–early Devonian tilting. The mean geographical direction was used to calculate the pole of 73.8°N, 168.3°E (Table 2), which suggests a Tertiary age for Component 2 (Fig. 5).
(a) Equal area projections of site means for Component 2 in geographical and stratigraphic coordinates. (b) Graph of percent unfolding v. grouping (k) for the regional fold test (Watson & Enkin 1993) of Component 2. Conventions for stereonets and fold test are as in Figure 4.
Component 3
Component 3 has southerly declinations and moderate negative inclinations (Fig. 3d and e) and is found at six sites along the length of the Moine Thrust Zone in the Sleat Group, Torridon Group, Eriboll Group and Durness Group (Table 1). In most specimens Component 3 was removed from 350 to 680 °C. In a few specimens (e.g. Fig. 3d), Component 3 is present at moderate demagnetization temperatures whereas at higher temperatures the inclination becomes shallower and Component 1 may be removed. The best grouping for the regional fold test was −3.7 ± 12.6% unfolded (Fig. 7, Table 2). Component 3 is interpreted as a post-thrusting and post-folding (late Silurian–early Devonian) magnetization with a mean geographical direction of declination 185.6° and an inclination of −46.0° (k 63.7, α95 11.6°; Table 2). The pole (59.6°N, 164.6°E; Table 2) position suggests a Triassic or Cretaceous age for Component 3 (Fig. 5).
(a) Equal area projections of site means for Component 3 in geographical and stratigraphic coordinates. (b) Graph of percent unfolding v. grouping (k) for the regional fold test for Component 3. Conventions for stereonets and fold test are as in Figure 4.
Component 4
Component 4, removed from 550 to 680 °C, has northwesterly declinations and steep negative inclinations or southeasterly declinations with steep positive inclinations (Fig. 3b and f), and is present in the Torridon and Sleat Groups at eight sites. The sites occur at a range of distances from the Moine Thrust Zone (Fig. 1, Table 1). A reversal test (McFadden & McElhinny 1990) was performed on the four site means with n >2 and α95 <20. The deviation from antipodality is 17.8, which is less than 38.9 (the 95% critical value), showing the two polarities to be insignificantly different. The results pass the reversal test at classification Rc.
Although a fold test was not possible, the Component 4 mean is better grouped (Table 2) in the stratigraphic (k 18.2, α95 18.4°) than in geographical coordinates (k 4.9, α95 38.6°). The pole of the stratigraphic mean for Component 4 (9.9°S, 205.3°E; Table 2) is similar to Torridon age poles (Smith et al. 1983; Torsvik & Sturt 1988; Potts 1990; Buchan et al. 2000), suggesting a Torridon age for Component 4 (Fig. 5).
Component 5
Component 5 has northeasterly declinations and shallow negative inclinations (Fig. 3g), and is found at seven sites along the length of the Moine Thrust Zone in the Sleat, Torridon, Eriboll and Durness Groups (Table 1). Maximum unblocking temperatures for this component are above 580 °C, although in many specimens the magnetization passes through the origin, indicating that another component may be present. This additional component, however, could not be resolved.
A regional fold test was conducted on three (n ≥4 and α95 <20°) sites that contain Component 5. Some sites were eliminated because they only had one specimen with the component (AP1 and SI1) or they had high α95 values (BB5 and BB6, Table 1). The best grouping for the fold test was −1.5 ± 39.75% unfolding (Fig. 8a and b, Table 2), consistent with remanence acquisition after the late Silurian–early Devonian thrusting and folding. The mean geographical direction (Table 2) yields a pole (18.8°N, 143.9°E) that suggests a Devonian age (Fig. 5).
(a) Equal area projections of site means for Components 4 and 5 in geographical and stratigraphic coordinates. Component 4 has both NW and up and the antipodal equivalent direction in stratigraphic coordinates (squares). Component 5 has a NE and shallow up direction (circles). (b) Graph of percent unfolding v. grouping (k) for the regional fold test of Component 5. Conventions for stereonets and fold test are as in Figure 4.
Other sites
Some sites such as BB4 (Torridon Group) and LM4 (Sleat Group) have southeasterly declinations in stratigraphic coordinates similar to the Component 4 direction but the inclinations are too shallow and they are not included with Component 4 (Table 1). Specimens at site K11 (Durness Group) have a southeasterly declination and moderate down inclination that does not correspond to any of the Phanerozoic components (Table 1). The age and origin of the magnetizations in these sites are unknown and they will not be considered further.
Other sites (B1, B2, E1 and IN1) are not listed in Table 1 because the specimens have weak magnetic intensities or contain only a modern viscous remanent magnetization (VRM). Other sites (D2, D3, D4, D5, K13, K14, UL4 and UL7) are not shown in Table 1 because the specimens do not display linear demagnetization trajectories, perhaps because of vector addition of unresolved components. Data for one site (GL) are not listed because the magnetic intensity is high and the magnetization is interpreted to have been caused by a lightning strike.
Contact (fault) test
A large-scale contact (fault) test was conducted using sites from the faulted recumbent syncline in a traverse from within the fault zone (site K18) to away from the fault zone (site BB8, Fig. 1a and b). Site K18 and sites BB1–BB8 were all collected from Torridon Group sandstones. Component 1 is found both in and away from the fault zone, although the intensity of the component decreases away from the fault zone. Site K18 is in the fault zone, and contains only Component 1 (300–680 °C). Away from the fault zone Component 1 (425–525 °C) makes up a lower percentage of the NRM and has lower absolute intensities (compare Component 1 in Fig. 3a with that in Fig. 3b and f). Component 4 (550–680 °C) also emerges away from the fault zone and dominates the NRM (Fig. 3b and f). This suggests that the contact (fault) test is positive and that Component 1 is related to the Moine Thrust Zone.
Rock magnetism
The a.f. demagnetization of specimens from the different lithologies showed no decay, suggesting that the magnetic phase(s) is of high coercivity. The IRM acquisition curves are not saturated by 300–500 mT, suggesting that a high-coercivity magnetic phase dominates (Fig. 9a). Based on the triaxial IRM decay (Fig. 9b–e), most of the magnetic intensity in the specimens resides in the 2500 mT curves, indicating that a high-coercivity mineral dominates. The 2500 mT curve decays at temperatures above 580 °C, indicating the presence of hematite. There is little or no magnetic intensity in the 120 mT curves, signifying that a low-coercivity phase is not a significant contributor to the magnetizations. These results indicate that hematite is the dominant magnetic mineral for all components. The Torridon and Sleat Group specimens, however, have higher medium-coercivity (500 mT) contributions (Fig. 9d and e) as compared with the other two lithologies (Fig. 9b and c). This could be due to a very fine-grained magnetic phase.
(a) Isothermal remanent magnetization acquisition curves for representative specimens from each lithology sampled. Magnetic saturation does not occur in any of the lithologies, indicating that the magnetic material has high coercivity. (b)– (e) Representative triaxial IRM decay curves for four lithologies: (b) Durness Group; (c) Eriboll Group; (d) Torridon Group; (e) Sleat Group. The high-coercivity curve dominates and decays at temperatures above 580 °C. This suggests that hematite dominates the magnetization.
Petrography
The sandstones sampled range from subarkoses to quartzarenites (Folk 1974). The limestones include dolomitic mudstones, wackestones, packstones and grainstones (Dunham 1962). The lithologies with the magnetic components contain a variety of types of authigenic hematite and associated diagenetic features. Sandstones with Component 1 commonly have authigenic hematite cement intergrown with clay cement (Fig. 10a and b). The clay types are kaolinite and illite based on elemental EDAX analyses. Specular hematite is found in muscovite grains (Fig. 10c) and is interpreted to be authigenic because it is squeezed between and folded around detrital grains. Some samples have hematite in the calcite veins (Fig. 10d). Hematite is also found in the sandstones replacing feldspars and rimming some detrital magnetite grains. Authigenic hematite is found in Component 1-bearing dolomites partially filling veins, associated with dedolomite in veins (Fig. 10e and f), and as cement between dolomite crystals (Fig. 10g). Also, calcite veins with some barite cut through most of the diagenetic features in the dolomites (Fig. 10h).
Representative photomicrographs of rocks from the study area. (a) Authigenic hematite cement surrounds grains in Torridon sandstone; cross-polarized reflected light (CPRL), sample BB1-4. (b) Kaolinite and illite cement identified by morphology and EDAX intergrown with authigenic hematite near Moine thrust in sandstone; backscattered SEM image, sample SI1-4. (c) Specular authigenic hematite replacing muscovite grains in sandstone; SEM, sample SI1-4. (d) Authigenic hematite in calcite vein and between grains in sandstone; CPRL, sample K18-1. (e) Large vein filled with dolomite and authigenic hematite associated with dedolomite in vein; thermoluminescence (TL), sample K15-7. (f) Authigenic hematite filling voids from dedolomitization; SEM, sample K15-7. (g) Authigenic hematite filling voids in dolomite cement; TL, sample DUR4. (h) Authigenic hematite filling voids in late-stage dolomite cement and with calcite vein cutting hematite cement and large dolomite crystal; TL, sample DUR4. (i) Authigenic hematite intergrown with clay and muscovite replacement of authigenic hematite on Skye at the Moine thrust in sandstones; TL, sample LM2-2. (j) Authigenic hematite (specularite) surrounding grains in dolomite; SEM, sample K17-1. (k) Large magnetite pseudomorphs replaced by authigenic hematite in dolomite; SEM, sample K17-1. (l) Detrital hematite grains and hematite cement in sandstones; TL, sample BB7-3.
Rocks with Component 2 appear to have more authigenic hematite cement than samples with Component 1 but do not contain higher magnetic (e.g. NRM) intensities. Sandstones with this component contain authigenic hematite intergrown with kaolinite and illite and replacing muscovite (Fig. 10i).
At one site (K17), Component 3-bearing dolomites contain authigenic hematite as specularite surrounding grains (Fig. 10j), as pseudomorphs of authigenic magnetite (Fig. 10k), in large authigenic concentrations, and in veins. Late calcite veins are present in these dolomites, similar to those in Component 1-bearing dolomites. Torridon sandstones with this component contain the same types of hematite as samples with Component 1. The CON1 site is a quartzarenite that contains abundant hematite cement.
Abundant detrital hematite grains (Fig. 10l) and hematite–clay cement characterize Torridon sandstones with Component 4. Component 5 sandstones contain abundant authigenic hematite–clay cement, and the muscovite grains are almost completely replaced by authigenic hematite. The dolomites have dolomite–dedolomite veins with authigenic hematite in the matrix similar to Component 1 samples.
In summary, authigenic hematite is found in a variety of forms but none appear unique to any of the magnetic components. The presence of different types of authigenic hematite suggests that a number of different processes caused hematite authigenesis.
The paragenesis of these carbonates and sandstones is complex, partly because it was necessary to integrate results from several areas along the Moine Thrust Zone (Figs 1 and 11). The carbonates provide more definitive information about relative timing of diagenetic events than do the sandstones. The carbonates were originally deposited as limestone and experienced several phases of dolomitization. Fine-grained dolomite probably formed relatively early and could be supratidal in origin. The larger dolomite crystals fill voids in the fine-grained dolomite matrix. Dolomite veining subsequently cut through the previous stages of dolomite and some of the dolomite in the veins was dedolomitized. Late-stage calcite cement with associated authigenic hematite fills vugs in the dolomites. Calcite veins, many containing barite, cut across all of the previously formed dolomite types (Fig. 10h) and could be related to dedolomitization and precipitation of the vug-filling calcite.
Authigenesis of coarse magnetite presumably occurred relatively early, perhaps with the void-filling dolomite. The replacement of the authigenic magnetite by hematite probably occurred at the same time as the specular hematite that surrounds the dolomite grains. The hematite associated with dedolomite and vug-filling calcite cements formed late in the sequence. The relative timing for the other hematite types is difficult to ascertain although they probably formed in the middle to late part of the sequence.
In sandstones, some of the feldspar alteration may have occurred prior to deposition although most alteration probably occurred post-depositionally. Hematite and clay (kaolinite and illite) cements are found around detrital grains. The hematite and the clays appear intergrown, suggesting that they co-precipitated. The authigenic specularite in the muscovite grains is interpreted as a diagenetic feature because it conforms to post-compaction deformation of the muscovite, which could have occurred any time after lithification. Calcite veins containing hematite have cut through the sandstones and hematite was precipitated in the matrix around the veins.
Geochemistry
Stable isotope analyses were performed on whole-rock samples from 14 sites of the Durness Group (Fig. 12a). The δ13C values ranged from −3.02 to −0.07‰. The δ18O values ranged from −9.22 to −3.86‰. The isotopic compositions of rocks with Components 1 and 3 are indistinguishable. The one sample of the host rock for Component 5 is distinct from the data for Components 1 and 3, suggesting a different diagenetic history.
Analysis of fluid inclusions from the Durness Group near Durness, Scotland, found that the inclusions were associated with three phases of calcite (early-stage white calcite, pink calcite and late-stage clear calcite; Figs 11c and 12b). The fluid inclusions had moderate homogenization temperatures between 60 and 175 °C and salinities between 4 and 18 wt.% NaCl equivalent (Fig. 12b). Petrographical analysis indicates that diagenetic phases that contain low to moderate homogenization temperatures bracket authigenic hematite (Fig. 11). The diagenetic relations depicted in Figure 11 come from rocks that are dominated by the Component 1 magnetization.
Diagrams showing the paragenetic relationships in three lithologies (vein-filling in Torridon Group sandstone, cement and veining in Eriboll Formation sandstone, mineralized Durness Group limestone) that are dominated by the Component 1 magnetization. The petrographical relationships indicate that diagenetic phases that contain low to moderate fluid inclusion homogenization temperatures (T) bracket authigenic hematite, suggesting that the hematite was precipitated at low to moderate temperatures.
(a) Stable carbon (PDB) and oxygen (PDB) isotope values for the 14 sites from the Durness Group. (b) Fluid inclusion data from dolomites of the Durness Group near Durness, Scotland.
Discussion
Origin of magnetic components
All five of the components reside in hematite, based on the maximum unblocking temperatures (600–680 °C) as well as on the rock magnetic and petrographic analyses. Johnson et al. (1985) concluded that rocks below the Moine thrust reached temperatures of 300–350 °C in the mid-Silurian and those in the foreland reached a maximum burial temperature of 275 ± 50 °C during the late Silurian. The four secondary components were all younger and probably experienced lower temperatures. The only constraints on the temperatures during formation of the hematite are from fluid inclusions in carbonates, which indicate a maximum homogenization temperature of 175 °C. Using the temperatures experienced in the Silurian as a maximum, and comparing these with the curves of Pullaiah et al. (1975), indicates that the unblocking temperatures for Components 1, 2, 3 and 5 are too high for a thermal viscous remanent magnetization (TVRM). Therefore, the components cannot be TVRMs and are interpreted as CRMs.
Permian CRM
Component 1 is post-folding and post-thrusting, and the pole position suggests that remanence acquisition was in the Permian. The contact test is positive, which suggests that the component is associated with the thrust zone. Results from other studies of Torridonian rocks (e.g. Stewart & Irving 1974; Smith et al. 1983; Robinson & McClelland 1987; Torsvik & Sturt 1987; Piper & Poppleton 1991; Williams & Schmidt 1997) in which sample locations were far from the Moine Thrust Zone also support this interpretation, as these studies found normal and reversed Torridon age, but not Permian age magnetizations. These results and the contact test suggest that a fluid flow event occurred along the Moine Thrust Zone in the Permian.
Component 1 is found in rocks that also have kaolin mineralization that formed from hydrothermal fluids (Parnell et al. 2003) that migrated along the length of the Moine Thrust Zone. Kaolin is abundant on fracture surfaces in Lewisian, Moinian, Torridonian, Cambrian and Durness Group rocks and increases in the vicinity of the Moine Thrust Zone. The high-temperature polytype dickite is also present in the immediate vicinity of the thrust planes (Parnell et al. 2003). The kaolin and dickite post-date all deformation fabrics in the thrust zone. Petrographical studies indicate coeval precipitation of kaolin and quartz at Bealach na Ba, where coeval quartz contains primary fluid inclusions with temperatures between 60 and 80 °C (Parnell et al. 2003). Just east of Loch Kishorn, kaolin is intergrown with coeval hematite filling fractures in the footwall of the Kishorn thrust. The dickite is confined to the thrust planes, which is consistent with a hydrothermal model, in which fluids were channelized along thrust planes.
Torsvik & Sturt (1988) reported a magnetization similar to the Permian CRM in hematite from the Torridon Group in the Lochalsh Syncline, which they suggested could be attributed to localized shearing in the Permian. A study along the Highland Boundary Fault of dolomitized Cambro-Ordovician serpentinite in the Highland Border Complex reported a Late Palaeozoic CRM in hematite that was interpreted to be related to a fluid migration event (Elmore et al. 2002). This is similar to the interpretation made in the current study for the Permian CRM in hematite in the Moine Thrust Zone. The Elmore et al. (2002) study, in conjunction with the present study, indicates that a regional crustal extension event occurred in the Permian in which excess heat (as fluids) was vented along major faults (e.g. Moine Thrust Zone and Highland Boundary Fault). The hydrothermal fluids responsible for kaolinization and hematite authigenesis could have been associated with a dense swarm of late Carboniferous–early Permian dykes (Rock 1983) and associated volcanic vents in Scotland (Rock 1983). This event could be linked to the earliest stages of rifting in the North Atlantic region and related magmatic activity (Smythe et al. 1995). It is also possible that the Permian CRM could be related to fluids triggered during the late stages of the Variscan orogeny.
Tertiary CRM
Component 2 is found only on the Isle of Skye and a regional fold test indicates a post-folding CRM that is Tertiary in age. The Tertiary age for Component 2 is similar to the results of other studies conducted on Skye by Potts (1990) and Woods et al. (2002). Potts (1990) found a Tertiary remagnetization in the Torridon Group in magnetite and hematite that was interpreted as a CRM associated with the Tertiary intrusion. Woods et al. (2002) identified a dual polarity Tertiary CRM residing in magnetite from the Jurassic rocks on Skye that was interpreted to be related to clay diagenesis caused by hydrothermal fluids associated with the Tertiary igneous intrusions. Component 2 is interpreted as a CRM that was acquired when hydrothermal fluids utilized the Moine and Kishorn thrusts on Skye as conduits. The fluids also caused kaolin mineralization along faults and fractures on Skye. Component 1 is not common on Skye, possibly because the Tertiary hydrothermal fluids caused recrystallization of previously formed hematite.
Component 3
Component 3 was found along the entire length of the Moine Thrust Zone and the regional fold test results indicate that it was acquired after folding and thrusting. The pole for this component falls between the Triassic and Cretaceous poles on the European APWP. Two possibilities exist for the origin of this component: a CRM or a vector addition of Components 1 and 2. In terms of the latter, Component 3 has a direction intermediate between Components 1 and 2. A few specimens contain components 1 and 3 although the inclinations are curved between the components (e.g. Fig. 3d). This would support the interpretation that Component 3 is a vector addition. On the other hand, the Tertiary CRM is found only on Skye but Component 3 is found from Skye north to Durness. This suggests that Component 3 is not a vector addition but could be a single component. The evidence for the origin of Component 3 is contradictory and therefore inconclusive. We acknowledge the possibility that it could be a vector addition.
If Component 3 is a CRM, there is no independent evidence to suggest a Triassic or Cretaceous age. Other studies report that some units in Northern Scotland also contain Mesozoic magnetizations that reside in hematite (e.g. Turner 1977; Turner et al. 1978; Tarling 1985; Piper & Poppleton 1991; Parnell et al. 2000). If it is a CRM, it could be related to rifting in the Atlantic.
Torridon component
Component 4 and its antipodal equivalent are found in the Torridon and Sleat Groups from Skye north to Ullapool on the mainland. These antipodal directions passed the reversal test and are statistically better grouped after the structural correction (stratigraphic coordinates), suggesting acquisition prior to thrusting and folding. The pole for the stratigraphic mean is interpreted to be a Torridon age pole because it is similar to other Torridon poles (e.g. Smith et al. 1983; Torsvik & Sturt 1987; Potts 1990; Buchan et al. 2000). The Torridon age component is interpreted as a detrital remanent magnetization (DRM) or early CRM residing in hematite.
Devonian CRM
The Component 5 CRM is post-folding and is interpreted to be Devonian in age. This CRM is similar to the Devonian remagnetization reported by Torsvik & Sturt (1988) that apparently resides in magnetite, which was found in Torridon rocks near the Moine Thrust Zone at Balmacara. Torsvik & Sturt (1988) concluded that the remagnetization was related to early Devonian uplift along the Moine Thrust Zone with related localized shearing and brittle thrusting and was near coincident with the emplacement of lamprophyre and felsic dykes. In the present study, Component 5 is interpreted as a CRM because it is found in different lithologies along the length of the Moine Thrust Zone and is therefore not likely to be the result of a localized structural event. This CRM could be the result of hydrothermal fluids associated with Devonian igneous activity or fluids triggered during the late stages of the Caledonian orogeny. Although there are no outcrops of Devonian extrusive volcanic rocks in NW Scotland, they can be found in southern and central Scotland and the Orkney Islands (Armstrong & Paterson 1970; Storetvedt & Petersen 1972; Gandy 1975; Thirlwall 1981; Robinson 1985). Hydrothermal fluids associated with the dykes and volcanic rocks could have intersected the Moine Thrust Zone at a depth of several kilometres and migrated along this conduit to the surface.
Synthesis
Multiple CRMs and multiple types of authigenic hematite are present in the rocks along the Moine Thrust Zone. Because of the paragenetic complexity, it is not possible to associate hematite types with specific CRMs. Hematite in the Moine Thrust Zone rocks (1) co-precipitated with kaolinite and illitic clays, (2) precipitated during dedolomitization, (3) replaced authigenic magnetite, (4) co-precipitated with calcite in veins, (5) replaced mica, and (6) precipitated as a late vug-filling cement. Although some of these hematite types are probably related genetically, it is clear that there were several episodes of hematite authigenesis. The fluid inclusion study, although limited in scope, also provides evidence for several different fluid events. The fluids that precipitated the veins were a likely source for some of the iron that formed the hematite. Iron released during dedolomitization of iron-rich dolomites and the iron in replaced magnetites indicates that some iron was locally derived.
Faults can act as barriers to flow or they can be conduits for flow, thereby controlling the distribution of diagenetic alteration, mineralization and remagnetization (e.g. Knipe 1993). Multiple CRMs, occurring from the Devonian to the Tertiary, have been identified in the rocks along the Moine Thrust Zone in NW Scotland. This suggests that the rocks have experienced several different events that precipitated hematite or altered the magnetic mineralogy of the rocks. These CRMs, along with the multiple types of hematite present, suggest that the Moine Thrust Zone has been a conduit for fluid migration for much of its post-orogenic existence. As noted above, the Devonian and Permian CRMs could be related to local events or to the migration of ‘orogenic fluids’ associated with the late stages of the Caledonian and Variscan orogenies, respectively. The possible Mesozoic component and Tertiary CRM are younger than any major orogenic event and cannot be related to the orogenic fluid hypothesis (Oliver 1992) that is invoked to explain many CRMs (McCabe & Elmore 1989). These CRMs are therefore significant because they show that dormant faults or fault systems can contain localized CRMs that are not directly related to major orogenic events. Instead, these localized CRMs in fault zones can be caused by fluids related to volcanic and igneous activity as well as other tectonic events.
Conclusions
Remagnetization along the Moine Thrust Zone occurred three and perhaps four times after the major movements on the faults. The three CRMs are Devonian, Permian and Tertiary in age whereas another potential CRM is Mesozoic in age. These localized CRMs coincide with events such as the migration of hydrothermal fluids related to Devonian igneous activity, regional crustal extension of Scotland and NW Europe in the Permian, migration of orogenic fluids, possible Atlantic rifting in the Mesozoic, and Tertiary intrusions. This study indicates that localized CRMs along dormant faults are not necessarily the result of fluids related to orogenesis (i.e. continent–continent collision). The remagnetizations along the Moine Thrust Zone provide evidence that remagnetizing fluids can use faults as conduits and that fluid migration events can be dated by palaeomagnetism.
Acknowledgements
This work was supported by DOE grant DE-FGO3-96ER14643 to R.D.E. and M.H.E. The authors thank C. Aubourg and C. Niocaill for helpful reviews.
- © 2005 The Geological Society of London
