The close spatial relationship between Devonian high-pressure rocks (eclogites) and Ordovician–Silurian calc-alkaline plutonic rocks, as observed in Liverpool Land, NE Greenland, is not easily explained by existing tectonic models for the Caledonide orogen. New field studies and isotope dilution–thermal ionization mass spectrometry U-Pb geochronology demonstrate, however, that the association is just coincidental, because the two rock groups are located within distinct terranes separated by a composite structure. The major element is the Gubbedalen shear zone, a N-dipping shear zone dominated by a penetrative top-up-to-the-S ductile fabric. Superimposed brittle-ductile top-down-to-the-N shear zones are typical of the structurally uppermost part of the shear zone. The contact against the hanging wall is the N-dipping, brittle Gubbedalen extensional detachment fault. A zircon age of 399.5 ± 0.9 Ma for an eclogite body is interpreted to represent the time of high-pressure metamorphism of the footwall. The host gneiss was migmatized between ca. 388 Ma and ca. 385 Ma, as constrained by the ages of a pegmatite predating migmatization and crosscutting granites. Coeval synkinematic granites intrude along amphibolite-grade, top-to-the-S high-strain zones in the Gubbedalen shear zone.

Juxtaposition of the Ordovician–Silurian plutonic terrane (hanging wall) against the Early to mid-Devonian eclogite terrane (footwall) is best explained by a tectonic model involving early mid-Devonian buoyancy-driven exhumation followed by late mid-Devonian syncontractional extension related to thrusting on the Gubbedalen shear zone in a dextral strike-slip zone. Subsequent exhumation through the brittle-ductile transition occurred by extension on early semiductile structures and the overprinting Gubbedalen extensional detachment fault, and erosion.


The timing and kinematics of high-pressure (HP) metamorphism, exhumation, and crustal deformation are critical elements for the development of tectonic models for collisional orogens like the Caledonides. Other important factors in such reconstructions include the relationships between crustal blocks of different affinities and origins brought together by strike-slip faulting and/or thrusting or extensional faulting. The Caledonides of Scandinavia and North East Greenland represent the two flanks of the Silurian–Devonian collisional orogen between Baltica and Laurentia, where the remains now are exposed in a series of nappes and crustal blocks on the two Atlantic margins (Haller, 1971; Roberts and Gee, 1985). The study of the relationships between the various components of the Caledonides is never quite straightforward and over the years has generated many controversies and debates. In our study, we address a controversial situation in the southern part of Liverpool Land, in the East Greenland Caledonides. The fundamental problem at the outset of the study was an apparent paradox between the spatial association of Devonian eclogites (Hartz et al., 2005) and Silurian plutonic complexes (Kranck, 1935; Hansen and Steiger, 1971), the coexistence of which could not be accommodated by simple tectonic models. We solved the conundrum by discovering the importance of a major structural boundary, investigated through field work and U-Pb isotope dilution–thermal ionization mass spectrometry (ID-TIMS) geochronology. In this paper, we describe the major N-dipping composite shear zone and fault separating the Devonian eclogite terrane in the footwall from a terrane of predominately Late Ordovician to Silurian plutons in the hanging wall, report U-Pb ages that date the main steps in the evolution of the eclogite terrane and the shear zone, and describe the role of the shear zone in the exhumation of the Liverpool Land eclogite terrane.


East Greenland Caledonides

The Late Silurian to Devonian continent-continent collision between Baltica and Laurentia produced the Caledonian orogen, the remnants of which presently straddle both sides of the North Atlantic Ocean (Haller, 1971; Roberts and Gee, 1985). The main tectonic elements of the East Greenland Caledonides (Fig. 1) are: (1) an autochthonous to parautochthonous Archean to Paleoproterozoic basement with a variably preserved cover of Neoproterozoic to Silurian sediments; (2) a far-traveled thrust sheet (Niggli-Hagar thrust sheet) composed, from the base upward, of Archean and Proterozoic gneisses, high-grade Mesoproterozoic supracrustal rocks (Krummedal Sequence), and a thick package of Neoproterozoic to Ordovician sedimentary rocks (Eleonore Bay Supergroup, Tillite Group, and Kong Oscar Fjord Group); and (3) late orogenic Devonian continental deposits in fault-controlled basins (Haller, 1971; Larsen and Bengaard, 1991; Higgins et al., 2004; Andresen et al., 2007). A major unconformity separates the Middle Devonian deposits from the underlying folded and faulted Neoproterozoic to Ordovician sedimentary rocks (Larsen and Bengaard, 1991; Larsen et al., 2008). Silurian leucogranites (ca. 435–425 Ma) intrude the Krummedal Sequence and the lower part of the Neoproterozoic sequence (Watt et al., 2000; Hartz et al., 2001; Kalsbeek et al., 2001a, 2001b; White et al., 2002; Leslie and Nutman, 2003; Andresen et al., 2007). Kalsbeek et al. (2001a, 2001b, 2008) argued that most of these leucogranites are S-type granites, derived by anatexis of pelitic units within the Krummedal Sequence. All these units are crosscut by two major, late to postorogenic, low-angle, E-dipping, top-to-the-E extensional faults: the Boyd-Bastionen detachment and the Fjord Region detachment (Fig. 1), subdividing the region into three crustal-scale fault blocks (Hartz and Andresen, 1995; Andresen et al., 1998; Hartz et al., 2000, 2001). Major sinistral strike-slip shearing also accompanied continent-continent collision and postcollisional extension. This led to regions of bulk transpression and transtension, and associated extrusion and pull-apart structures (i.e., the Western fault zone and the Storstrømmen shear zone; Holdsworth and Strachan, 1991; Larsen and Bengaard, 1991; Seranne, 1992; Torsvik et al., 1996; Krabbendam and Dewey, 1998; Smith et al., 2007; Steltenpohl et al., 2009).

In a slightly modified tectonostratigraphic scheme proposed by Higgins et al. (2004), the Niggli-Hagar thrust sheet is subdivided into two tectonic units: the Niggli Spids and Hagar Bjerg thrust sheets. The nature of major faults and shear zones separating different units in the southern East Greenland Caledonides is debatable, as evidenced by conflicting interpretations of maps and cross sections presented in Higgins et al. (2008). We follow the view of Haller (1971) and Andresen et al. (2007), who interpreted structural repetitions within the Niggli-Hagar thrust sheet to be a result of large-scale recumbent folding.

Liverpool Land

Liverpool Land is an isolated area composed of pre-Carboniferous rocks in the southern part of the East Greenland Caledonides, and it is separated from the main outcrop area of Caledonian rocks to the west by a cover of Permian and Mesozoic sediments (Jameson Land Basin; Figs. 1 and 2). The link to the tectonostratigraphy established farther west is therefore not straightforward. Partly migmatized metasedimentary rocks interpreted as remnants of the Krummedal Sequence (Higgins, 1988; Johnston et al., 2009) have, however, been taken to indicate that the rocks constituting northern Liverpool Land represent an eastward continuation of the Niggli-Hagar thrust sheet.

Previous investigations indicated that most of Liverpool Land is dominated by various types of intrusions, most of which were hypothesized to be Caledonian in age (Kranck, 1935; Hansen and Steiger, 1971). The most prominent intrusion is the multiphased, Late Ordovician to Silurian Hurry Inlet composite pluton (Fig. 2; Kranck, 1935; Augland et al., 2009; Corfu and Hartz, 2011). Eclogite-bearing gneisses occur to the south of the Hurry Inlet composite pluton and its hosting paragneisses (Kranck, 1935; Sahlstein, 1935; Cheeney, 1985). Based on quartz exsolution in clinopyroxene from one of the eclogite lenses near Kap Hope (Fig. 2), Smith and Cheeney (1980) argued for high-pressure metamorphism of rocks in the area. More recently, Hartz et al. (2005) suggested that metamorphism had reached conditions of >25 kbar and 800 °C, but they presented no mineralogical information, whereas Buchanan (2008) documented minimum pressures of ∼18 kbar at average temperature of ∼870 °C for one of the eclogite lenses. Muscovite 40Ar/39Ar data from the host gneiss of the eclogite lenses gave an age of ca. 379 Ma, which was interpreted to date cooling through the muscovite closure temperature (Bowman, 2008). The Hurry Inlet composite pluton was interpreted to be intrusive into the eclogite-bearing gneisses by Coe and Cheeney (1972) and Coe (1975), thus requiring the high-pressure metamorphism to be older than intrusion of the Hurry Inlet composite pluton. Hansen and Steiger (1971) obtained an imprecise Silurian Rb/Sr biotite age from the Hurry Inlet composite pluton and considered it to be post-tectonic. These interpretations are difficult to reconcile with the Devonian age proposed by Hartz et al. (2005) for eclogitization.

To shed more light into these questions, we remapped the transition between the Hurry Inlet composite pluton and Liverpool Land eclogite terrane and show it to be a complex shear zone and fault with both subhorizontal contraction and extension. The main element of this high-strain zone is a N-dipping shear zone, a fundamental feature with a long-lived displacement history, named the Gubbedalen shear zone (Fig. 2). Superimposed on the shear zone, there is the brittle Gubbedalen extensional detachment fault (Fig. 2).

Hanging-Wall Rocks

In our study area (Fig. 2), the hanging wall is dominated by the Hurry Inlet composite pluton and its host paragneisses, but it also includes the pyroxene-bearing monzodioritic Hodal-Storefjord Pluton, the Triaselv Leucogranite, and other granitoid plutons (Coe and Cheeney, 1972; Augland et al., 2009). The hanging-wall plutonic rocks have an age range of 445–425 Ma (Augland et al., 2009; Corfu and Hartz, 2011). Except for the Triaselv Leucogranite, all these granitoids are amphibole- and partly pyroxene-bearing and have arc-granitoid geochemical signatures (Augland et al., 2009). Similarities are also noted between the plutonic rocks that intrude the high-grade paragneisses of northern Liverpool Land and similarly aged plutonic rocks on Renland farther to the west (Kalsbeek et al., 2008; Augland et al., 2009; Rehnström, 2010). The plutons and paragneisses are cut by a swarm of N-striking Late Permian lamprophyre dikes (Buchanan, 2008).

The host paragneisses underwent Caledonian high-temperature metamorphism that locally transformed them into migmatites (Johnston et al., 2009). Impure marbles occur locally, and primary sedimentary structures are preserved in some quartzites. These supracrustal rocks have been correlated with the Mesoproterozoic Krummedal Sequence (Higgins, 1988), and a study of detrital zircons from the paragneisses supports this interpretation (Johnston et al., 2009). It is important to note that there is no evidence of mid-Devonian high-grade metamorphism (Johnston et al., 2009). Titanite from the Hurry Inlet composite pluton yields U-Pb ages of ca. 420 Ma (Corfu and Hartz, 2011), and fission-track data from titanites from a high level in the batholith gave an age of ca. 413 Ma for cooling below ∼285 °C (Gleadow and Brooks, 1979; Jacobs and Thomas, 2001). Muscovite 40Ar/39Ar data from a paragneiss yields a younger cooling age of ca. 381 Ma (Bowman, 2008), but this sample was located some 20 km from the sample studied for fission tracks and probably reflects a position deeper in the crust at ca. 400 Ma.

Footwall Rocks

The structurally upper part of the footwall block consists of eclogite-bearing migmatitic granitoid (ortho-) gneisses with small and large lenses of amphibolite, eclogite, and rare ultramafites (garnet-peridotite, serpentinite, and pyroxenite; Augland et al., 2010). The mafic and ultramafic lenses largely appear as stratiform boudins parallel to the foliation, varying in size from a few centimeters to several hundred meters (Figs. 3A and 3B). Based on these field relations, the bimineralic (omphacite + garnet) nature, and the lack of hydrous phases in the least retrogressed eclogites (see following), they are interpreted as original mafic dikes and sills. The ultramafic lenses are generally serpentinized, but they locally contain relics of the original mineralogy (olivine, orthopyroxene, clinopyroxene, and rare garnet) (Augland et al., 2010).

The migmatite is homogeneous on a scale of more than 25 km2 and is interpreted as a granitoid orthogneiss. It has well-defined leucosome and melanosome domains, indicating extensive partial melting (Fig. 4A). The leucosomes are dominated by K-feldspar and quartz with only ∼10% plagioclase and minor biotite (Fig. 4F), whereas the melanosome is dominated by biotite and plagioclase in addition to K-feldspar and quartz (Fig. 4E), and in some samples amphibole. Feldspars in the migmatite are recrystallized by subgrain rotation, and quartz by grain boundary migration, indicating that the migmatites have been deformed at high temperature subsequent to migmatization (Figs. 4E and 4F). The leucosomes are locally oblate (“pancakes”) and often folded. Locally, they have a prolate (cigar-shaped) form (Figs. 4C and 4D). The long axes in the prolate leucosomes, the fold axes in the folded oblate leucosomes, and the prominent quartz aggregate lineation all trend approximately north-south (Figs. 4B–4D). In the southern part of the Liverpool Land eclogite terrane, the leucosomes appear to have been flattened only (pancakes), and the lineation is only weakly developed or absent, except in local shear zones.

Two populations of dikes intruding the footwall allow us to bracket the early stages of deformation (see geochronology section). There is an older generation of pegmatites emplaced in boudin necks and as dikes cutting the mafic boudins (Figs. 3A–3C). These pegmatites cannot be traced from the mafic boudins into the surrounding migmatite gneiss, i.e., they do not crosscut the migmatite orthogneiss foliation. The pegmatites often contain large amphibole crystals (∼3 cm). A younger, more widespread population of crosscutting sheets and some small, medium-grained plutons of granites, not affected by the foliation-forming event, intrude both the migmatite orthogneiss and the mafic boudins (Figs. 3C and 3G).

In the footwall, there is no evidence of the Late Ordovician–Silurian magmatism observed in the hanging wall.

Gubbedalen Shear Zone

The Gubbedalen shear zone is an ∼400-m-thick N- to NE-dipping, ductile shear zone deforming rocks of the footwall (Fig. 2). Based on descriptions of sheared rocks in the southeastern part of Liverpool Land (Cheeney, 1985), we interpret the shear zone to continue in a southeastern direction in this part of Liverpool Land (Fig. 2). The Gubbedalen shear zone is characterized by a foliated and lineated fine-grained quartzo-feldspathic mylonite gneiss, commonly with small feldspar augen. Less deformed lenses of amphibolite and of the original migmatitic gneiss occur locally. Amphibole porphyroblasts in the mylonite are locally common. Asymmetric feldspar porphyroclasts with core-mantle textures, S-C textures, shear bands (C′), and rotated relic foliation (of migmatite relics within the shear zone: Fig. 5A) all indicate top-up-to-the-S sense of shear.

The Gubbedalen shear zone is locally overprinted by younger brittle faults that cut and partly displace the ductile fabric. We have not found evidence, however, for large-scale rigid body rotation of the ductile fabric. We therefore interpret the fabric geometry to be representative of the original orientation and sense of shear. A pronounced downdip lineation within the mylonitic foliation, defined by parallel-oriented amphibole, quartz aggregates, and elongated feldspar porphyroclasts, trends N-S (Fig. 5F).

Granite sheets intruded the Gubbedalen shear zone contemporaneously with top-to-the-S shearing (Fig. 5A–5B). The granite sheets occur parallel to the mylonitic foliation and have an internal foliation parallel to the mylonitic foliation (Figs. 5A–5E). In one case, we also observed part of a sheet sheared into a megaporphyroclast (Fig. 5B). Except for local growth of garnet and white mica, these sheets are mineralogically almost identical to undeformed or weakly deformed granitic sheets appearing structurally lower down. The lower boundary of the Gubbedalen shear zone is gradational and marked by a systematic downward decrease in the frequency and thickness of top-up-to-the-S ductile shear zones (Figs. 5A–5D). Widely spaced top-up-to-the-S shear zones parallel to the Gubbedalen shear zone are present throughout the entire eclogite terrane. In the transition from the footwall toward the Gubbedalen shear zone, the granite sheets are swept into shear planes parallel to the mylonite foliation in the Gubbedalen shear zone (Fig. 2), but away from the shear zone, the granite is generally undeformed. The N-trending lineations also become less prominent with increasing distance downward from the shear zone.

In addition, the granite sheets in the Gubbedalen shear zone are recrystallized, having K-feldspar with core-mantle textures and myrmekite growth, both indicating that shearing took place under amphibolites-facies metamorphic conditions (Figs. 5C–5E) (cf. Vidal et al., 1980; Gapais, 1989; Gates and Glover, 1989; Simpson and Wintsch, 1989; Tribe and D'Lemos, 1996; Stipp et al., 1999; Stipp, 2002). The growth of garnets in the same rocks is consistent with recrystallization under amphibolites-facies conditions. Continued top-up-to-the-S contractional deformation under greenschist-facies conditions is evident from the presence of highly strained and locally fractured feldspar porphyroclasts, subgrain rotation recrystallization in quartz, development of grain shape fabric constituted by quartz subgrains (Figs. 5C and 5D), and buckled biotite in the mylonite rocks (Gapais, 1989; Kanaori et al., 1991; Stipp, 2002).

Extensional Structures

Superimposed on the top-up-to-the-S shear fabrics in the upper part of the Gubbedalen shear zone, there is an ∼70-m-wide zone of top-down-to-the-N extensional structures (Fig. 2). Because this zone is truncated by the brittle Gubbedalen extensional detachment fault, its original thickness is not known. The sense of shear is documented by small-scale, listric, (semi-) ductile extensional shear zones or faults overprinted by phyllonites with S-C and C′ structures and normal crenulation cleavage with associated asymmetrically folded quartz veins (Figs. 6A–6D). The phyllonite layers and textures are again overprinted by semibrittle top-down-to-the-N extensional faults (Fig. 6D). Less common are asymmetric folds and contractional faults, also indicating top-down-to-the-N movement. The Gubbedalen extensional detachment fault overprints all the other structures with a top-down-to-the-N displacement, bringing the Hurry Inlet composite pluton and its host paragneisses in contact with Gubbedalen shear zone and the Liverpool Land eclogite terrane.


Most of the mafic lenses in the footwall are amphibolites (Fig. 3C). They generally show textural and mineralogical evidence, such as diopside-plagioclase or amphibole-plagioclase symplectites (Figs. 3D and 3E), indicative of derivation via retrogression of eclogite-facies minerals (omphacite). Some garnet-pyroxenite lenses have preserved their eclogitic paragenesis (Figs. 3B and 3F) with only limited retrograde overprint. The least retrogressed eclogites are generally bimineralic (omphacite + garnet) and are often very garnet-rich (50%–70%).

Microprobe analyses of adjacent garnet and clinopyroxene pairs in eclogite and retro-eclogite are given in Table 1. Analyses from sample LEA 06-61 are from the unzoned interiors of adjacent garnets and pyroxenes. The minerals are generally compositionally homogeneous, but the clinopyroxenes show some enrichment in Ca and Fe near the margins. Symplectites of plagioclase-diopside or amphibole, up to ∼0.1 mm wide, commonly occur between garnets and omphacites. The eclogite has omphacite with an average composition of Jd44 and a maximum of Jd46 (Table 1). Clinopyroxene in sample LEA 06-59 is retrogressed, and only the interiors of garnet are unzoned. The analyses on this sample were done on omphacite inclusions in unzoned garnet domains, measuring garnet spots close to the inclusions. There are no visible reaction textures between garnet and omphacite inclusions, so mineral equilibrium is assumed. The data yield jadeite compositions up to Jd41, similar to those in LEA 06-61, corroborating the interpretation that all these rocks experienced high-pressure metamorphism. The variation in chemical composition of the analyzed omphacite inclusions is probably due to garnet growth at different pressures.

The compositions of the minerals analyzed here are very similar to those in the eclogites studied by Buchanan (2008), and his sample CP-52A is from the same eclogitic body as sample LEA 06-61. Buchanan (2008) provided quantitative pressure-temperature (P-T) estimates of minimum 18 kbar and ∼870 °C for eclogite-facies metamorphism. He also showed that the retrogressive P-T path went through granulite-facies conditions, as evidenced by secondary orthopyroxene and clinopyroxene, the appearance of symplectites with Ca-rich pyroxene and plagioclase, and Na-rich amphibole. Determining the pressure of these rocks is difficult because there are no good geobarometers applicable to bimineralic eclogites, and uncertainties in the Fe2+/Fe3+ ratios in the analyzed minerals lead to uncertainties in the Fe-Mg exchange geothermometer of Krogh Ravna (2000) used by Buchanan (2008).


We analyzed zircon, rutile, and monazite from eclogite and different intrusive rocks obtained from a traverse across the Gubbedalen shear zone and into the footwall for U-Pb isotopes in an attempt to constrain the age of metamorphism, magmatism, and deformation in the rocks of Liverpool Land eclogite terrane. All U-Pb analyses in this study were conducted by ID-TIMS at the Department of Geosciences, University of Oslo. The methodology is summarized in Appendix A.


Field Relationship and Sample Description

Sample LEA 06-59 is from a retrogressed eclogite boudin surrounded by felsic migmatite orthogneiss. The rock consists of ∼50% garnet, almost 50% symplectized omphacite (plagioclase, secondary clinopyroxene, and amphibole), and accessory rutile. Garnets contain omphacite inclusions (up to Jd42, Table 1; Fig. 3E). Interstitial rutile generally occurs together with ilmenite, which appears to pseudomorphically replace rutile (Fig. 7I). Rutile inclusions in garnet contain thin ilmenite lamellae (Fig. 7H).

Mineral Characteristics

Zircons from LEA 06-59 can be divided in three main categories based on morphology, color, size, and internal textures: (1) small (<100 μm in diameter), clear, colorless, inclusion-free, well-rounded zircons (Fig. 7G); (2) medium-sized (100–300 μm in diameter), subrounded, relatively clear, colorless zircons, with some inclusions; and (3) large (300 μm to ≥0.5 mm), metamict and inclusion-rich, slightly brownish, elongated zircons. Cathodoluminescence shows that many grains have a core-rim relationship (Fig. 7) showing dark cores and bright rims (grain A) or the opposite (grains E and F). Grain C has a weakly zoned core with a homogeneous outer rim. Grain B has more complex zoning, with inner and outer bright rims. Grain D shows an irregular diffuse zoning throughout the grain. In grains B and E, the different domains are almost homogeneous, whereas in the rims of grains A and F, the texture is somewhat more heterogeneous. The U content ranges from ∼70 to 1000 ppm, and Th/U ranges from 0.074 to 0.83 (Table 2).

Rutile grains selected for analysis were clear, brown, inclusion-free fragments of variable size. Their U contents were low, at 27–5 ppm, with initial common Pb at <0.19 ppm and Th/U <0.1 (Table 2).

Analytical Results and Interpretation: Zircon

Thirty different zircon fractions, composed of 1–15 grains, yielded mainly discordant data (Fig. 8A; Table 2). Analysis number 13 is concordant; it represents eight very small (∼50–60 μm), rounded, clear, inclusion-free zircons giving a concordia age of 399.5 ± 0.9 Ma (mean square of weighted deviates [MSWD] = 0.0084), which is interpreted to date eclogite metamorphism (Fig. 8A). One imprecise data point (analysis 11) also plots on the concordia curve, and supports the interpretation that 399.5 Ma is the crystallization age. All the discordant fractions indicate the presence of inherited components. A line calculated through all 30 analyses yields an upper-intercept age of 1601 ± 25 Ma, but with a high MSWD value of 29. Due to the scatter of the data points, and the fact that there are no points close to the upper-intercept age, it is not evident that this age is geologically meaningful. One option is that the scatter reflects different zircon populations in the protolith. A second alternative is that that the scatter is the result of recent Pb loss, pulling some of the data points down from an original mixing line. A third element affecting the interpretation is given by the five least discordant analyses (11, 12, 13, 22, and 27), which lie on a line projecting toward an intercept age of 829 ± 240 Ma, suggesting that zircons in the protolith may have been affected by a Mesoproterozoic-Neoproterozoic event (dashed, gray error ellipses in Fig. 8). Evidence for a Grenvillian-age signature has also been found in zircons from a migmatite gneiss from Liverpool Land dated by Corfu and Hartz (2011). The interpretation that the 399.5 Ma event represents eclogite metamorphism is supported by: (1) the zircon morphology (Fig. 7G); (2) the clear core-rim and two- or multidomain relationships of the zircons, with structureless or irregularly zoned and patchy texture domains (Figs. 7A–7F); and (3) the low Th/U ratio of the completely recrystallized or newly grown zircons (Table 2), which is common in eclogite-facies zircons (Corfu et al., 2003; Hoskin and Schaltegger, 2003; Bingen et al., 2004). Further support for this interpretation is given by comparable ages obtained by Corfu and Hartz (2011) on other eclogite bodies from the area.

Analytical Results and Interpretation: Rutile

Seven rutile fractions consisting of 20–50 grains or fragments yielded both discordant and concordant data, with two analyses actually being reversely discordant. The latter have 206Pb/238U ages older than those of the eclogite-facies event. There is also some correlation with an increase in 207Pb/206Pb age. This suggests that the unusual discordance could be related to disturbances of the system and fractionation of U and Pb, for example, due to only partial dissolution of some rutile. The two most precise rutile analyses, numbers 35 and 36, overlap, giving a concordia age of 382.4 ± 0.8 Ma (2σ; MSWD = 0.69; Fig. 8B).

It is unclear how the development of ilmenite inside rutile affected the isotopic behavior of rutile. Root et al. (2004) reported the existence of ilmenite plates in rutile from eclogite, interpreting them to reflect exsolution after peak pressures. In our sample, exsolution lamellae are only found in nonretrogressed rutile inclusions in garnets (Fig. 7H), together with nonretrogressed omphacite, where the host represents the preserved relicts of the high-pressure assemblage. By contrast, the interstitial rutile always occurs together with pseudomorphically growing ilmenite (Fig. 7I), which may be interpreted as a retroreaction product of decompression. Thus, rutile may exsolve ilmenite during decompression from high-pressure conditions, but ilmenite may also grow from the rutiles where there is access to fluids. Such mechanisms of exsolution and retroreaction could lead to open-system behavior and reequilibration of U and Pb in rutile. This could explain why the rutile ages postdate the time of formation of the mineral (as indicated by zircon). It could also explain why the field-based determination of rutile Pb closure temperature (400–450 °C; Mezger et al., 1989; Schmitz and Bowring, 2003) is so much lower than the experimentally determined closure temperature (∼600 °C; Cherniak, 2000). In the former case, Schmitz and Bowring (2003) noted exsolved and newly formed ilmenite in and associated with rutile, and hence their rutile ages may reflect the exsolution process, whereas the experimentally established closure temperature is related to pure diffusion of Pb.

Pegmatite in Retro-Eclogite Boudin Neck

Field Relationships and Sample Description

Sample LEA 06-18 was taken from a pegmatite within the neck of a retrogressed eclogite boudin (Fig. 3C). The pegmatite does not cut the surrounding migmatite (equivalent to the thin pegmatite in Fig. 3A). The mineralogy of the rock is simple, with K-feldspar, quartz, and biotite.

Zircon Characteristics and Analytical Results

Relatively few zircons occur in this sample. Those extracted can be divided in two groups: (1) partly metamict, brownish to reddish long prisms, mostly fragmented and dominated by {110} morphology; and (2) rounded, colorless, clear grains. The rounded grains are possibly xenocrysts from the eclogite and/or retro-eclogite and were not analyzed. The U content of the brown grains is very high, ranging from ∼12,700 ppm to ∼21,700 ppm. Th/U varies between 0.12 and 0.14 (Table 2). The four analyses are concordant or less than 3.5% discordant (Fig. 9A) and define a line with an upper-intercept age of 387.7 ± 1.8 Ma (2σ, MSWD = 0.40, anchored at 0 Ma), which is considered to represent the crystallization age of the pegmatite.

Granite Sheets and Minor Granitic Intrusive Rocks in Footwall Rocks and in the Gubbedalen Shear Zone

Field Relationships and Sample Description

Sample LEA 06-62 is from a small, undeformed granite body that clearly cuts across the SL fabric in the migmatite. Sample LEA 06-66 is from a foliated granite sheet parallel to the mylonite foliation in the Gubbedalen shear zone (see the section on Gubbedalen Shear Zone). Both samples are rich in K-feldspar and quartz, with ∼10% plagioclase and minor biotite that is partly altered to chlorite. Minor muscovite, and accessory zircon, rutile, and opaques also occur. The sample from the deformed granite dike in the shear zone (LEA 06-66) has subgrain microtextures in quartz indicating subgrain rotation recrystallization, and K-feldspar has core-mantle textures.

Zircon and Monazite Characteristics

LEA 06-62 is characterized by a heterogeneous zircon population; many grains contain visible cores and are metamict. These were not analyzed. Zircons in the analyzed fractions are weakly yellow to reddish, some are slightly metamict, and some have a rusty surface indicating partial alteration. Elongation ratios of the analyzed grains are generally high, ranging from 3.5 to 7. Most of the analyzed grains were fragmented, equant prisms with {100} dominating morphology. The high elongation ratio is indicative of rapid crystallization, probably from a water-rich melt (Corfu et al., 2003). Analyses show high U contents (1425–9366 ppm), and Th/U = 0.15–0.32. Sample LEA 06-62 also contains monazite, commonly as clear yellow fragments. The U content measured in two fractions ranges from 1616 to 6747 ppm, and Th/U is high at 16.2–18.6 (Table 2).

The zircon population of LEA 06-66 is also heterogeneous and very similar to that of LEA 06-62, but the grains are generally more fragmented, and U content (300−1300 ppm) was not as high (Table 2).

Analytical Results and Interpretation

The three most discordant zircon analyses of LEA 06-62 are collinear and define an upper-intercept age of 385.0 ± 2.7 Ma (2σ, MSWD = 0.24; Fig. 9B). The two most concordant data points, however, deviate slightly, with one to the right of the chord suggesting some minor inheritance, and the other to the left, possibly due to superimposed early Pb loss or hydrothermal effects. One reversely discordant analysis of monazite (number 5) with a 207Pb/235U age of 382.8 ± 1.7 Ma is also slightly younger than the zircon upper intercept, strengthening the possibility that some Pb loss is responsible for the apparent younger ages. Reverse discordance is common in monazite due to excess 230Th, resulting in unsupported 206Pb (Harrison et al., 2002; Oberli et al., 2004). The second monazite analysis plots, instead, right on top of the zircon chord. Two of the zircon analyses of sample LEA 06-66 are concordant, defining a concordia age of 384.5 ± 1.3 Ma (2σ, MSWD = 1.3), whereas the third analysis indicates the presence of an inherited Proterozoic component (Table 2; Fig. 9C).

The age of 385.0 ± 2.7 Ma obtained on the small, undeformed pluton far away from the Gubbedalen shear zone (LEA 06-62), and the age of 384.5 ± 1.3 Ma obtained on zircons from the deformed and foliated granite sheet within the Gubbedalen shear zone (LEA 06-66) are indistinguishable within error. This indicates emplacement during the same magmatic event in the Liverpool Land eclogite terrane, probably contemporaneous with top-up-to-the-S contractional shearing. These granite sheets cut the migmatite gneiss and are synkinematic with top-up-to-the-S displacement under amphibolites-facies conditions. The ages of the dikes thus provide a minimum age of formation of the migmatite, give important constraints on the timing, and a strong clue to the mode of exhumation of the eclogite terrane.


Chronology of Tectonometamorphic Events

The U-Pb isotope ages document that rocks constituting the footwall to the Gubbedalen shear zone represent a fragment of Precambrian crust that underwent high-pressure metamorphism during the Caledonian orogeny. Zircon ages and field observations suggest that the protoliths of the eclogite were mafic dikes or minor plutons that intruded the quartzo-feldspathic rock in the late Paleoproterozoic (Corfu and Hartz, 2011). Data reported by Augland et al. (2010) and Corfu and Hartz (2011) show that the quartzo-feldspathic rock hosting the eclogites formed at ca. 1640–1645 Ma. Corfu and Hartz (2011) also reported a ca. 1600 Ma protolith age from an eclogite lens and a late Mesoproterozoic protolith age from a migmatitic gneiss.

The most significant tectonometamorphic event in the Caledonian evolution of Liverpool Land eclogite terrane is high-pressure metamorphism at 399.5 ± 0.9 Ma, most likely related to continent-continent collision and crustal thickening. A subsequent Caledonian tectonic and metamorphic event was related to the emplacement of pegmatites, mostly in fractured eclogite lenses at ca. 388 Ma. These pegmatites do not cut the migmatites and thus provide a maximum age for the formation of the migmatite foliation. The generation of these pegmatites could have been associated with decompressional melting during the initial stage of exhumation through granulite-facies and subsequent amphibolite-facies conditions (Buchanan, 2008; Whitney et al., 2004). Further exhumation, possibly combined with introduction of fluids into the felsic host rock, is likely linked to the formation of small granitic plutons and sheeted intrusions at ca. 385 Ma (Whitney et al., 2004). These plutons and sheeted intrusions cut the foliation in the migmatite and, thus, provide a minimum age on migmatization in the Liverpool Land eclogite terrane. The age of ca. 385 Ma from the syncontractional granite in the Gubbedalen shear zone indicates that exhumation from amphibolites-facies to greenschist-facies conditions was contemporaneous with subhorizontal shortening at midcrustal levels.

Gubbedalen Shear Zone as a Terrane Boundary

The hanging wall in Liverpool Land most probably represents a magmatic arc developed on Proterozoic crust in the latest Ordovician and Silurian, with a particularly intense phase of magmatism at ca. 445–425 Ma (Augland et al., 2009; Corfu and Hartz, 2011). There is no evidence that these rocks were affected in any way by the events that exhumed, deformed, and transformed the footwall mafic rocks into eclogites, and then back to amphibolites, together with the widespread migmatization and generation of granitic magmas between 400 and 380 Ma. Conversely, there is no evidence of Silurian magmatism or metamorphism in the footwall. The time gap and the radically different types of Caledonian activity and lithologies between the hanging wall and the footwall indicate that the Gubbedalen shear zone represents an important boundary, possibly separating two terranes of highly different tectonomagmatic and metamorphic histories.

The question that is most relevant for the present paper relates to the timing and mechanism of juxtaposition of these two terranes and the role played by the Gubbedalen shear zone and the Gubbedalen extensional detachment fault in this process. The fact that the hanging wall was neither metamorphosed at ca. 400 Ma nor intruded by the 385 Ma granitic dikes that are so ubiquitous in the footwall and Gubbedalen shear zone indicates that the two terranes were separated until after 385 Ma. The Gubbedalen shear zone, with its contractional structures in the lower part and overprinting extensional structures in its upper part, must have played an important role in the exhumation of the eclogite terrane, as is evident from the decreasing metamorphic grade of the successively developed contractional and extensional structures in the shear zone. Although the observed individual top-down-to-the-N extensional structures in the upper part of the Gubbedalen shear zone are difficult to quantify, it is clear that top-to-the-N displacement must have been significant, because the transition from ductile to brittle behavior of the shear zone rocks is obvious. The amount of strain in the reactivated extensional semiductile part of the zone is, however, difficult to assess quantitatively because large parts of the structure could have been excised by the overprinting brittle Gubbedalen extensional detachment fault, and the true thickness of the zone of semiductile extension (prior to the brittle detachment) is unknown.

Final juxtaposition of the two terranes was presumably related to the extensional movements along the Gubbedalen extensional detachment fault, which brought crust exhumed from a depth of more than 50 km in contact with rocks from a middle- to upper-crustal magmatic arc (Fig. 10). Extension in the upper crust must have been important in contributing to the exhumation of the Liverpool Land eclogite terrane synchronously with ductile contractional displacements on the Gubbedalen shear zone (pre–380 Ma; see following). Muscovite 40Ar/39Ar ages at ca. 380 Ma recorded both in the hanging wall and the footwall (Bowman, 2008) indicate that the two terranes were at the same crustal level, and possibly juxtaposed, at this time.

Tectonic Model for Development of the Gubbedalen Shear Zone

Two main lines of evidence suggest that displacement on the Gubbedalen shear zone was related to oblique motions during the latest collisional stage of the Caledonian orogeny. First, the orientation (E-W) of the Gubbedalen shear zone and several related local shear zones within the Liverpool Land eclogite terrane (Fig. 2), which we interpret as the original orientation relative to the surrounding units, all are perpendicular to the main structural grain of the East Greenland Caledonides. Second, the kinematics of the Gubbedalen shear zone, with major top-up-to-the-S contraction, is consistent with N-S crustal shortening.

There is regional evidence for a Caledonian (especially mid- to late Devonian) orogenwide left-slip fault system between and within Baltica and Laurentia (Steltenpohl and Bartley, 1988; Holdsworth and Strachan, 1991; Larsen and Bengaard, 1991; Seranne, 1992; Northrup and Burchfiel, 1996; Torsvik et al., 1996; Krabbendam and Dewey, 1998; Klein et al., 1999; Titus et al., 2002; Smith et al., 2007; Steltenpohl et al., 2009). In the East Greenland Caledonides, sinistral wrench faults associated with the Western fault zone (Fig. 1) that were active until at least the mid-Devonian have been described by Larsen and Bengaard (1991). They speculatively linked the Western fault zone to the Storstrømmen shear zone farther north (Fig. 1). The Storestrømmen shear zone is a major ductile shear zone thought to have accommodated large sinistral displacements (Holdsworth and Strachan, 1991; Smith et al., 2007). These workers attributed the foreland thrusting in Dronning Louise Land (Fig. 1) to sinistral transpression on this shear zone, reflecting the oblique convergence of Baltica and Laurentia (Torsvik et al., 1996). Thrusting in Dronning Louise Land has been dated to ca. 390 Ma (Dallmeyer and Strachan, 1994), implying that, if the interpretations of Holdsworth and Strachan (1991) and Smith et al. (2007) are correct, large-scale bulk sinistral transpression in the East Greenland Caledonides was active at that time. Lower-amphibolite-facies conditions along the Storestrømmen shear zone have been dated to ca. 370 Ma (Dallmeyer and Strachan, 1994), showing that the shear zone was still active at ductile conditions in the Late Devonian.

Devonian sinistral transtension and transpression have also been reported from the Western Gneiss Region, and Krabbendam and Dewey (1998) suggested this to be the main mechanism of exhumation of ultrahigh-pressure and high-pressure rocks from amphibolites-facies conditions. It is thus clear that sinistral displacement was important across the entire Caledonide orogen during the Devonian, and this sets up a framework for our tectonic model.

In a bulk transpressional setting, shear zones that are oblique to the main direction of compression may develop as a response to structural asperities (e.g., inherited structures or lithologic boundaries). The oblique orientation of the Gubbedalen shear zone compared with the main structural grain in the East Greenland Caledonides fits with such a setting. An element of transpression in Liverpool Land could also explain the constrictional structures observed in the footwall and the strongly lineated mylonites in the Gubbedalen shear zone itself.

The initial phase of the exhumation probably occurred by nearly isothermal decompression as indicated from the presence of secondary orthopyroxenes and clinopyroxenes and Ca-rich plagioclase in the eclogites (Buchanan, 2008) and the large degree of melting in the migmatitic orthogneiss. There are different ways of achieving such nearly isothermal decompression paths (e.g., Whitney et al., 2004). In the case of the Liverpool Land eclogite terrane, the initial exhumation from eclogite-facies conditions (>50 km depth) to granulite-facies, lower-crustal conditions was probably driven by buoyancy and back thrusting of subducted crust (Chemenda et al., 1995; Augland et al., 2010). Partial melting as a consequence of nearly isothermal decompression would have further increased the buoyancy and decreased the rheology of the Liverpool Land eclogite terrane. This could have facilitated exhumation through the lower crust by mechanisms of syncontractional flow on low-angle shear zones, buckling, diapirism, or combinations of these (Hartz et al., 2001; Whitney et al., 2004; Andresen et al., 2007).

The structural and textural expressions of the exhumation from amphibolites-facies conditions to greenschist-facies conditions of the Liverpool Land eclogite terrane at the Gubbedalen shear zone are linked to contractional top-up-to-the-S displacements. If the exhumation from granulite-facies conditions to amphibolites-facies conditions was accomplished, at least in part, by flow on low-angle shear zones, a continuation of such flow at lower temperatures could have led to localizations of shear zones (i.e., the Gubbedalen shear zone). However, for continued exhumation to have occurred, thrusting must have been accompanied by extensional (and erosive) denudation in the hanging wall that exceeded the crustal thickening resulting from thrusting (Fig. 10A; Hartz et al., 2001; Andresen et al., 2007). The formation of contractional structures observed in the Gubbedalen shear zone was ongoing at ca. 385 Ma, implying that the oblique contraction in this region of the Caledonides lasted until at least the end of the mid-Devonian.

Exhumation from greenschist-facies conditions (postclosure of Ar in muscovite at ca. 380 Ma; Bowman, 2008), and juxtaposition of the footwall and hanging wall as seen today reflect movement along the Gubbedalen extensional detachment fault and other brittle faults in the area (Fig. 10B). This switch from contraction to extension is interpreted to reflect the overall change from convergence to extension in the East Greenland Caledonides at ca. 380 Ma (Fig. 10B).

An alternative exhumation model for the exhumation from amphibolites-facies to greenschist-facies conditions could be that an extrusion wedge was created above a “master fault” associated with a structural asperity comprising a restraining bend, having an extensional shear zone at the top (Dewey et al., 1998). In such a model, the structures observed in the Liverpool Land eclogite terrane would be the midcrustal ductile equivalent of asymmetric flower structures developing in the brittle crustal regime (Dewey et al., 1998).


The Liverpool Land eclogite terrane is a piece of continental crust formed during Eon 16; it has a moderate latest Mesoproterozoic overprint (Augland et al., 2010; Corfu and Hartz, 2011) and was affected by eclogite-facies metamorphism at 399.5 ± 0.9 Ma. The terrane was then exhumed and underwent extensive partial melting around 388 Ma, as high-pressure conditions changed to amphibolite-facies via granulite-facies conditions. The Gubbedalen shear zone, forming the upper boundary of the Liverpool Land eclogite terrane, is a ductile, N-dipping shear zone with a predominant top-up-to-the-S sense of motion and was active at conditions varying progressively from amphibolite-facies conditions to greenschist-facies retrogression and, together with several local shear zones within the eclogite terrane, was responsible for exhumation of Liverpool Land eclogite terrane from amphibolites-facies conditions. The structurally upper part of the Gubbedalen shear zone is characterized by top-down-to-the-N brittle-ductile structures, and the contact to the hanging wall is constituted by the brittle Gubbedalen extensional detachment fault. Exhumation in the brittle regime was provided by displacement on the Gubbedalen extensional detachment fault and by erosive denudation of the footwall block. The hanging wall is a distinct magmatic terrane of Ordovician–Silurian age. The model developed in this paper links the Gubbedalen shear zone to a large-scale sinistral transpressional system that brings together different blocks and different crustal levels of the Caledonian orogen.


After crushing and separating heavy minerals using magnetic and heavy liquid separation methods, zircon, monazite, and rutile were handpicked and discriminated on the basis of morphology, transparency, color, and internal textures. Cathodoluminescence images were obtained from a grain mount with a representative selection of 20 zircons from the eclogite. Air abrasion was carried out using the method described by Krogh (1982) to remove the marginal areas of the mineral grains most likely to have experienced Pb loss. Mineral samples were washed in dilute HNO3, ionized water, and acetone using an ultrasonic bath to remove any contamination. Each sample was then weighed on a microbalance and spiked with a mixed 202Pb-205Pb-235U tracer. Zircon and rutile were dissolved in HF and a drop of HNO3 in Teflon bombs at ∼190 °C for 5 d. Monazite was dissolved in 6 N HCl and a drop of HNO3 in Savillex vials on a hotplate at ∼125 °C for 5 d. Dissolved samples weighing more than 0.005 mg and all monazite were chemically separated using microcolumns and anion-exchange resin in order to remove cations that may inhibit ionization (Krogh, 1973). U/Pb-solutions were dried down and loaded on degassed single Re filaments with silica gel.

The samples were measured on a Finnigan MAT 262 mass spectrometer, using either Faraday cups in static mode or, for low-intensity samples, a secondary electron multiplier (SEM) in peak jumping mode. The 207Pb/204Pb ratios were measured on SEM for all samples. SEM data were corrected for nonlinearity based on measurements of the standard NBS 982-Pb + U500 (Corfu, 2004). Measurements of the standard are also used to monitor the reproducibility of the mass spectrometer.

Measurements were corrected for a 2 pg Pb and 0.1 pg U blank, with blank compositions: 206Pb/204Pb = 18.3, 207Pb/206Pb = 0.85, and 207Pb/204Pb = 15.555 (Corfu, 2004). Common Pb corrections were employed using the Pb-evolution model of Stacey and Kramers (1975) at the age in question. U source fractionation was estimated to be 0.12%/a.m.u. Pb source fractionation was corrected using the measured 205Pb/202Pb tracer ratio normalized to the certified value of 0.44050. A standard fractionation error of 0.06%/a.m.u. was incorporated in the calculations if 205Pb/202Pb was determined very precisely and the fractionation corrections became unrealistically precise. If 205Pb/202Pb was not determined, or the measured 205Pb/202Pb was far of from 0.44050, Pb fractionation was set at 0.1%/a.m.u. Pb fractionation values between 0.02% and 0.16%/a.m.u. were generated by this procedure.

The analytical errors and corrections were then incorporated and propagated using the ROMAGE 6.3 program, originally developed by T.E. Krogh. Graphic presentations and age calculations were performed using the Isoplot program of Ludwig (2003) and the decay constants referred in Steiger and Jäger (1977). All errors are reported at the 2σ confidence interval.

We thank G. Bye-Fjeld and M. Erambert for analytical assistance and P.I. Myhre and M. Steltenpohl for fruitful discussions in the field. B. Bingen and M. Steltenpohl read and made constructive suggestions on late drafts. We also appreciate critical comments by B. Hacker, W. Hames, and four anonymous reviewers.