Abstract
Abstract: The Ketilidian orogen in SE Greenland contains a continental magmatic arc (Julianehåb batholith) and forearc which formed in the overriding plate of a Palaeoproterozoic oblique subduction boundary between c. 1854 and 1795 Ma. During uplift and deformation of the forearc between c. 1795 and 1780 Ma, its sedimentary sequences were invaded by arc-type magmas and metamorphosed at HT–LP upper amphibolite facies conditions which led to widespread anatexis and generation of abundant S-type granites. Transpression during early deformation (D1) was partitioned into arc-parallel, sinistral strike-slip in the magmatic arc rocks of the batholith and arc-normal contraction in the sedimentary and volcaniclastic sequences of the proximal forearc (Psammite Zone). Subsequently, D1 fabrics in the Psammite Zone were overprinted and transposed in an intense, flat-lying, D2 ductile shear zone with arc-parallel stretching lineations and top-NE kinematic indicators. The kinematic significance of the D2 shear zone is that it accommodated vertical differences in strain and displacement and permitted deformation to be partitioned differently at different levels in the crust. As such, it is an attachment or coupling zone that provided kinematic linkage between deformation at different crustal levels during NE transport of a forearc sliver. Overprinting of the upper crustal domain of D1 strike-slip partitioned transpression in the forearc during D2 is attributed to migration of the attachment zone to a higher crustal level. This migration was a consequence of an enhanced geothermal gradient in the proximal part of the forearc due to the emplacement of batholith-related plutonic suites during D2.
Strike-slip partitioned transpression is characteristic of upper crustal deformation at obliquely convergent plate boundaries (McCaffrey 1996; Teyssier & Tikoff 1998; St-Blanquat et al. 1998). Although partitioning of the deformation into strike-slip and thrust components may be accomplished in several different ways, strike-slip partitioning implies the existence of a low-angle ‘detachment’ horizon at the base of the partitioned system, located at the boundary between the upper crust and a region of more distributed deformation in the lower crust (Richard & Cobbold 1990; Strachan et al. 1992; Northrup & Burchfiel 1996; Teyssier & Tikoff 1998; Fig. 1 ). In a critical evaluation of the use of the term ‘detachment’ in this context, Teyssier et al. (2002) and Tikoff et al. (2002) have argued that low-angle deformation zones formed at boundaries between more- and less-rigid domains in the lithosphere are best regarded as attachment or coupling systems rather than detachments. They are accommodation systems that provide kinematic continuity or attachment between lithospheric layers in which deformation is partitioned differently, rather than discontinuities separating deformed from undeformed lithospheric layers. Inherent to this idea is that attachment may be both top-driven, by gravitational forces or bottom-driven, by mantle flow, and that an analogy can be drawn between attachment zones in the lithosphere and the clutch in an automobile that allows transfer of motion from the engine to the wheels (Tikoff et al. 2001). Here, we echo this view and use the term attachment zone to refer to structures that provide vertical mechanical coupling between layers with different rheology and modes of strain partitioning in the lithosphere. A major control on the level in the lithosphere at which an attachment zone is located is likely to be geothermal gradient, changes in which may cause vertical migration of the attachment zone and overprinting within the partitioned system.
Strike-slip partitioned transpression. Lower layer: heterogeneous transpression accommodated by distributed shear. Upper layer: partitioned transpression with the strike-slip component taken up on a vertical fault and shortening taken up on a thrust system. The differential displacements between the layers are accommodated by a ‘detachment’ or, more correctly, an attachment system (after Richard & Cobbold 1990). α is the convergence angle. ϑ is the angle between the deformation zone boundary and the thrust faults and fold axial traces in the deformed blocks.
In this paper we analyse the structure of the Ketilidian orogen in southern Greenland (Fig. 2 ) which evolved during the period c. 1850–1740 Ma. We contend that vertical changes in partitioning were accommodated by a low-angle attachment zone formed in the middle crust as a result of oblique convergence during HT/LP metamorphic conditions. In addition, we integrate the structure, kinematics, thermal history and geochronology of an internal part of the orogen in SE Greenland to demonstrate how vertical migration of an attachment zone, caused by an increase in geothermal gradient, influenced the structural evolution of the orogen.
Overview geological map of the Ketilidian orogen showing its subdivision into five zones. The inset map of Greenland shows the position of the main map (a) and generalized cross-section (b). Note that the rapakivi granites have been omitted from the generalized cross-section.
Geological setting
Our data are drawn principally from a Palaeoproterozoic continental magmatic arc and forearc exposed in an area of c.4000 km2 in the mountainous, glaciated coastal region of SE Greenland (Fig. 3 ) which was mapped in 1996 during the final year of the SUPRASYD project of the Geological Survey of Denmark and Greenland (Garde & Schønwandt 1994, 1995; Garde et al. 1997, 1998a).
Lithological and structural map of the Lindenow Fjord district of SE Greenland, showing the NE part of the Psammite Zone and adjacent parts of the Julianehåb batholith in the Ketilidian orogen (redrawn with additional structural data from Garde et al. 1998b).
Building on investigations of the Ketilidian orogen in the 1960s and 1970s (Allaart 1976; Kalsbeek et al. 1990), Chadwick & Garde (1996) showed that the orogen is divided from north to south into five domains (Fig. 2).
(1) Foreland: consisting of Archaean high-grade gneisses largely unaffected by Ketilidian processes.
(2) Border Zone: a sequence of Palaeoproterozoic Ketilidian sedimentary and volcanic rocks, with low metamorphic grade in the north, deposited unconformably on the Archaean high-grade basement rocks of the foreland. The Border Zone sequences represent deposition in the back-arc domain of the Ketilidian subduction boundary. The rocks were progressively affected by Ketilidian deformation, metamorphism and granite intrusion towards the south.
(3) Julianehåb batholith: a polyphase calc-alkaline suite dominated by granite and granodiorite with subordinate tonalite, hornblende diorite and gabbro forming a continental magmatic arc batholith. The batholith has a wedge-shaped outcrop tapering along strike from c.160 km wide in the west to c. 80 km wide in the east. The boundary between the batholith and the Border Zone is marked by steeply dipping, NE-trending ductile shear zones at Kobberminebugt in the west and at Napasorsuaq Fjord in the eastern part of the orogen (Fig. 2).
(4) Psammite Zone: a complex of variably migmatized clastic sedimentary rocks and subordinate volcanic suites interpreted as the proximal part of the Ketilidian forearc. The boundary between the forearc and the batholith is partly obscured by younger granite and syenite of the Rapakivi and Gardar suites, but data from largely inaccessible outcrops, closely observed from the air, suggest it is in part intrusive and in part a tectonically modified unconformity.
(5) Pelite Zone: a complex of heavily migmatized schists, diatexites and granites located to the SE of the Psammite Zone and believed to represent a distal part of the forearc. The boundary between the Psammite and Pelite Zones is transitional.
The geochemical and isotopic signatures of the Julianehåb batholith show that it largely comprises juvenile crustal additions (van Breemen et al. 1974; Patchett & Bridgwater 1984; Kalsbeek & Taylor 1985; Chadwick & Garde 1996; Stendal & Frei 2000). Bridgwater et al. (1973) drew attention to similarities in the large-scale crustal architecture of the Ketilidian and the Andean orogens and Windley (1991) also regarded the Julianehåb batholith as having been emplaced in the upper plate of an Andean-type subduction boundary, although he concluded that deformation in the Ketilidian orogen was due to continent–continent collision. Chadwick & Garde (1996) argued that deformation in the Ketilidian orogen was due to upper plate transpression at an ocean–continent convergent boundary and showed that there is no evidence of thrust-stacking and profound crustal shortening to support Himalayan-type continent–continent collision during Ketilidian orogeny. They interpreted the batholith as the root of a volcanic arc that developed partly on, and accreted adjacent to, a continental margin represented by the Archaean foreland. Chadwick & Garde (1996) recognized no suture within the orogen to the north of the Julianehåb batholith and consequently oceanic lithosphere was believed to have been subducted from south to north beneath the batholith and the Psammite and Pelite Zones representing the forearc (Fig. 2). No trace of oceanic lithosphere survives in the orogen, and its southern foreland is unknown.
Major sheets of rapakivi granite s.l. were emplaced in the Psammite and Pelite Zones and the southern margin of the Julianehåb batholith during late Ketilidian open folding which followed the peak of metamorphism (Gulson & Krogh 1975; Patchett & Bridgwater 1984; Brown et al. 1992; Grocott et al. 1999, Chadwick et al. 2000). The orogen forms part of the basement to sedimentary and volcanic rocks of the Gardar Province, c.1300–1120 Ma, and it hosts a number of Gardar plutons (Emeleus & Upton 1976; Kalsbeek et al. 1990, Grocott et al. 1999). Gardar dolerite dykes are common throughout the orogen and its foreland.
This paper has two principal objectives. First, following brief reviews of the Julianehåb batholith in the western part of the orogen and the depositional characteristics of the Psammite Zone, we describe the structure in the previously unknown proximal part of the forearc in SE Greenland and the adjacent part of the batholith (Fig. 3). Second, we assess the implications of the new findings for the wider issue of deformation partitioning and vertical coupling–decoupling in the lithosphere during oblique convergence.
The Julianehåb batholith in the western part of the Ketilidian orogen
Most of the granitic plutons that make up this part of the batholith (Fig. 2) are characterized by upright magmatic-state planar fabrics trending NE with shallow lineations. The fabrics are defined by the preferred orientation of feldspars, hornblende and flattened dioritic enclaves (Chadwick et al. 1994). Steeply dipping, NE-trending ductile shear zones, a few centimetres to c.1.5 km wide, with high-temperature crystal-plastic fabrics and shallow linear fabrics are also common. High-strain orthogneisses and mylonites in the shear zones include S-C fabrics and other shear criteria which indicate a predominant sinistral sense of displacement. These criteria, combined with the dominantly flattening fabrics defined by dioritic enclaves and magmatic state fabrics, led Chadwick & Garde (1996) to conclude that the bulk of the batholith was emplaced during sinistral transpression, probably in a regime of oblique convergence. The shapes of individual plutons have only been identified in the Kobberminebugt area (Fig. 2), where steep contacts and magmatic-state fabrics imply they are steeply dipping sheets elongate NE–SW (McCaffrey et al. 1998), and for some late intrusions that are circular to elliptical, gently dipping, tabular sheets.
Variably deformed plutonic complexes emplaced during the first phase of growth of the batholith have U–Pb zircon ages in the range c.1854–1805 Ma (Hamilton et al. 1996; Hamilton 1997; Garde et al. 2002; Figure 4 ). A deformed aplite dyke emplaced into one of the major sinistral shear zones near Saarloq (Fig. 2) has a zircon age of 1816 Ma, whereas anirregular body of fine-grained felsic rocks in Johan Dahl Land with a zircon age of 1799 ± 2 Ma lacks the pervasive NE-trending magmatic fabric found in the bulk of the batholith. These ages confirm that the sinistral shear zones were active during the emplacement of the western part of the batholith, and they imply that transpression and magma emplacement could have ceased there by about 1800 Ma. Between 1805 Ma and 1790 Ma, emplacement of plutonic complexes at Kobberminebugt was associated with dextral strike-slip (Garde et al. 1998c; McCaffrey et al. 1998, Garde et al. 2002). This shear sense contrasts with the consistent arc-parallel sinistral displacement found elsewhere in the western part of the orogen and in SE Greenland, which is the main concern of this paper.
Summary of age data from the Ketilidian orogen based on zircon ages obtained since 1994. Where not indicated otherwise, the mineral ages are conventional isotope dilution determinations for zircon and monazite generally within ±1–2 Ma (2 σ). A peak of titanite ages at c.1795 from the Julianehåb batholith is also shown and is interpreted as representing the uplift and unroofing of the batholith during D1 (data sources: Garde et al. 1997, 1998c, 2000, 2002; Hamilton 1997 and unpublished data; McCaffrey et al. 1998).
Depositional characteristics of the Psammite Zone
The Psammite Zone is characterized by metamorphosed feldspathic arenites with some shales, lithic arenites and rare polymict conglomerates. Where the metamorphic grade is low, the sandstones, shales and conglomerates have been interpreted as turbidite-dominated rocks deposited in fluvial and relatively shallow marine settings (Garde et al. 1997). Unitsof the Bouma sequence, with grading, cross-bedding and soft-sediment slumps, are preserved in metamorphosed turbidites at the head of Danell Fjord (Fig. 3), and in other rare low-strain areas, but intense deformation and widespread migmatization has obliterated primary depositional structures in most of the Psammite Zone. From the head of Danell Fjord southeastward, there are clear transitions from unambiguous turbidites into large areas of brown-weathered migmatized schists which are distinct from grey, variably migmatized feldspathic arenites that form the bulk of the Psammite Zone (Fig. 3). The turbidites and their migmatized schistose equivalents overlie the grey feldspathic arenites but it is unclear whether this relationship is stratigraphic or the result of tectonic stacking. Ultrabasic, basic and intermediate volcanic rocks are locally interlayered with the migmatized feldspathic arenites, principally at a low structural level. Volcanic rocks are voluminous at ‘Sorte Nunataq’ (Fig. 5 ) and at Kangerluluk along the northeastern margin of the forearc in SE Greenland (Chadwick & Garde 1996; Mueller et al. 2000; Fig. 3).
Photogrammetric diagram of ‘Sorte Nunataq’ from the south. The diagram shows a well-bedded sequence of volcanic and volcaniclastic rocks above granitic rocks of the Julianehåb batholith. (a) The western part of the diagram shows the sequence is, in part, an unconformity, possibly slightly modified by deformation and, in part, intrusive where granitic rocks of the batholith have cut the base of the supracrustal sequence. (b & c) The eastern part of the sequence is folded on the margin of a vertical, NE-trending, sinistral, strike-slip ductile shear zone of D1 age (see Fig. 7a).
Clasts of granite s.l. and U–Pb ages of detrital zircons show that the sedimentary provenance in the Psammite Zone was dominated by plutonic rocks of the Julianehåb batholith, with minor contributions from older Palaeoproterozoic and Archaean sources (Hamilton et al. 1996; Garde et al. 2002; Fig. 4). The ages of the youngest dated detrital zircons point to c.1793 Ma as the maximum age for the uplift of the forearc basin (Fig. 4). The detrital zircon data imply that the contact between the batholith and the Psammite Zone is essentially an unconformity, but original characteristics of the contact have been largely obscured because most of the proximal part of the Psammite Zone was intruded by late components of the batholith (Fig. 5) and affected by intense deformation and metamorphism.
Structure of the Ketilidian forearc in SE Greenland
Four principal episodes of deformation, D1–D4, have been recognized in the Psammite Zone. A localized fifth episode, D5, was broadly contemporaneous with D4.
D1 and D2 structures and fabrics: Danell Fjord and Paatusoq areas
D1 structures and fabrics are rarely found unaffected by D2 but near the contact between metasedimentary rocks of the forearc and the batholith, at the inner part of Danell Fjord, D1 structures and fabrics are preserved with little overprinting. In turbidites east of the batholith margin, D1 is represented by tight, upright, upward-facing minor folds (F1), with axial planar fabrics (S1) of spaced pressure solution type (Fig. 6 a). Refraction of S1 is common in graded beds. Changes in D1 minor fold and cleavage vergence indicate the axial surface trace of a major F1 syncline close to the batholith boundary (Fig. 7 a). The folds plunge gently toward the NE parallel to the plunge of the cleavage–bedding intersection (Fig. 7e). Andalusite porphyroblasts that overgrow the S1 fabric were pseudomorphed by colourless mica.
(a) F1 minor fold in bedding with a related S1 spaced cleavage. Outcrop is in turbidites on the north side of inner Danell Fjord in a domain of D1 deformation with no D2 overprint. View is to the NE. (b) West-vergent, F2 minor fold of S0 and S1, inner Danell Fjord. In the limbs of the fold, bedding and S1 are transposed completely to S2. Only in the fold hinge can bedding and S1 be distinguished from S2. View is to the NW. (c) Flat-lying structure with intense D2 fabric in south central Danell Fjord area. Pale grey migmatitic feldspathic arenites underlie darker migmatitic schists. View is to the NW. The difference in altitude between the rocks in the foreground and the summit is approximately 500 m. (d) Inverted graded bedding in the long limbs of D1 minor folds on the inverted limb of a D1 fold nappe, NW of Kangerluaraq (locality a, Fig. 3). View is to the SW. (e) Asymmetric calc-silicate boudins in psammitic rocks showing D2, dextral (top-NE) displacement on the NW-dipping limb of an F4 fold. The strong planar fabric is S2. From the central Lindenow Fjord area. (f) S-C fabric in migmatitic schists showing dextral (top-NE) displacement, inner Nørrearm area. The strong fabric is S2. View is of a vertical exposure surface looking NW. The pen in the photograph is 15 cm long. (g) Gently inclined, D3 crenulation of S2 planar fabric. From the steep limb of a large-scale D3 fold, central Danell Fjord (see Fig. 7a). View is of a vertical exposure surface looking north. The width of the field of view is approximately 1.5 m. (h) Contact between the Julianehåb batholith (pale rocks on the ridge in the distance) and metasedimentary rocks of the Psammite Zone (darker, banded rocks in the foreground and in the lower part of the ridge in the distance) NW of inner Nørrearm. The rocks of the Psammite Zone and the batholith contain intense D2 fabrics that steepen gradually from east (left side of the photograph) to west. The difference in altitude between the base and the top of the mountain ridge in the centre of the photograph is approximately 1000 m.
The intensity of D2 increases eastward away from the batholith margin. There is a narrow (< 0.5 km wide) zone of overprinting where bedding and S1 fabrics are folded by west-vergent minor folds (F2) co-planar and co-axial with F1 (Fig. 7a). Bedding and S1 are almost completely transposed to a new planar fabric, S2 (Fig. 6b) as D2 fabric intensifies eastward. The S2 planar fabric contains a marked stretching lineation (L2) plunging gently SW and NE (Fig. 7d). D2 fabric intensity increases structurally downwards away from the batholith contact together with metamorphic grade. Folded sillimanite and anatectic melt seams in S1 differentiated layering in the D2 overprinting zone show that upper amphibolite facies HT/LP conditions were reached prior to D2 deformation (Garde et al. 1998a). Seams of migmatite parallel to F2 fold axial surfaces show that these conditions persisted into D2.
Grey, migmatitic feldspathic arenites underlying heavily migmatized schistose equivalents of the turbidites north and south of central Danell Fjord (Fig. 3) contain a gently dipping, intense planar fabric (S2) with transposed bedding and S1 (Fig. 6c). Small-scale D1 folds with SE vergence, shallow westerly-dipping axial surfaces and shallow hinges coaxial with a NE-trending mineral lineation (L2) are preserved locally, but the outcrops are dominated by intense D2 deformation manifest as S2 differentiated layering and L2 mineral and grain shape elongation lineations, generally with shallow NE or SW plunge, coaxial with F2 folds (Fig. 7d). Appinite sheets and in situ anatectic granite-pegmatite sheets are common parallel to D2 fold axial surfaces: boudinage of some appinite sheets led to boudin separation of a few metres consistent with D2 arc-parallel extension. The change in orientation from upright to shallow D1–D2 structure south and north of central Danell Fjord defines a major, west-closing, recumbent D3 fold (Fig. 7a & e). Crenulations with flat-lying axial surfaces and gentle NE-plunge parallel to L1–L2 are common in the hinge and flat-limb of the structure (Fig. 6g) but are absent in the steeply dipping rocks close to the batholith boundary.
The contact between the batholith and the Psammite Zone in inner Danell Fjord is marked by a 300 m wide, NE-trending, ductile shear zone in volcanic rocks that structurally underlie the turbidites (Fig. 7a). The shear zone is vertical with a sub-horizontal stretching lineation on the planar fabric. Asymmetric boudins in the shear zone indicate a sinistral shear sense. There is no evidence of crenulation or transposition of S1 in the ground between the axial trace of the major F1 syncline in the turbidites and the shear zone (Figs 3 & 7a). In the batholith west of the shear zone, heterogeneouslydeveloped, crystal-plastic planar and linear fabrics have the same orientation and sinistral kinematics as fabrics in the shear zone. We therefore infer that D1 fabrics are continuous from the Psammite Zone into the batholith in inner Danell Fjord.
The flat-lying structure of the central Danell Fjord area is typical of the NE Psammite Zone, although the main fabrics are commonly folded by large-scale, upright folds (D4), notably in the SE (Fig. 7c). The intensity of D2 transposition makes unambiguous D1 fabrics indiscernible in most outcrops but, in addition to inner Danell Fjord, D1 structures are evident at a few widely spaced localities in the NE of the Psammite Zone where D2 strain is weak or absent. For example, widespread overturning of polymict conglomerates and feldspathic arenites NW of Kangerluaraq (Fig. 3, locality a) shows that early fold nappes are present but their geometry is unclear because of breaks in outcrop between nunataks which expose either right-way-up or overturned sequences (Fig. 6d). Gently dipping bedding is cut by a spaced cleavage associated with an intense, SW-plunging linear element of a constrictional LS tectonite fabric defined by pebble shapes in conglomerates. Changes in cleavage vergence indicate 100 m scale, SE-vergent and upward-facing folds with long inverted limbs and short right-way-up limbs on the inverted limb of a fold nappe: its gross structure is unknown because of ice cover between nunataks. The planar fabric overprints rare pre-cleavage folds, but in the absence of a second tectonic fabric we assume that the implied large-scale folds are D1 structures. Farther east, D1 differentiated layering in grey feldspathic arenites at Paatusoq (Fig. 3, locality b), is folded by gently inclined F2 minor folds. Partially melted layers within S1 and sillimanite grains folded by F2 show that here, as at inner Danell Fjord, onset of high-grade metamorphism predated D2.
D1 and D2 structures and fabrics: Nørrearm and Lindenow Fjord areas
In the cliffs of the inner reaches of Nørrearm, S2 is sub-horizontal and the effects of D3–D5 are minimal over a wide area (Figs 3 & 7b). In contrast to the Psammite Zone–batholith boundary at inner Danell Fjord, granitic rocks of the batholith overlie the forearc at Nørrearm (Fig. 7b). Shallow D2 fabrics steepen gradually as they pass westwards into comparable fabrics in the granodiorites of the batholith, which in turn steepen progressively to vertical over a distance of c. 10 km (Figs 6h & 7f). No new fabrics are associated with this steepening. It appears to represent the original orientation of the D2 strain trajectories and bears no relation to D3 or D4 folding. No overprinting of D2 on D1 fabrics like that present at Danell Fjord has been recognized. At inner Nørrearm it is clear that fabrics in the batholith (Fig. 3, locality c) are continuous with D2 fabrics in the Psammite Zone.
Kinematic indicators, including S-C fabrics, extensional structures and distorted veins found at inner Nørrearm and throughout the Psammite Zone in SE Greenland show unambiguously that D2 fabrics evolved during top-NE displacements (Figs 3, 6e & f). D2 kinematic indicators and the S2 planar fabric are folded by upright D4 folds coaxial with L2 (Fig. 7g). As a consequence, horizontal exposure surfaces and map views reveal a dextral horizontal shear sense on NW-dipping F4 limbs and sinistral sense on SE-dipping limbs (Fig. 3). When the effects of F4 folding are removed, both senses are consistent with top-NE transport related to the L2 stretching lineation during D2.
The region of sub-horizontal D2 fabrics in migmatitic metasedimentary rocks below the shallow batholith boundary at inner Nørrearm is also notable for its broadly concordant sills of granodiorite and diorite with swarms of magmatically disrupted appinite. One sill of hornblende granodiorite that extends c. 20 km NE from central Lindenow Fjord to Kangerluaraq (Figs 3 & 7c) has a zircon age of 1792 ± 1 Ma. Magmatic-state fabrics parallel to D2 planar and linear fabrics in the host rocks show that it was emplaced during D2 (Hamilton 1997; Fig. 4). On the grounds of composition and proximity to the batholith boundary, the sills are interpreted as late components of the batholith plutonic complex contemporaneous with D2 deformation at c. 1792 Ma.
D3 structures and fabrics
D3 gave rise to large-scale, close to tight folds overturned to the NW (Fig. 7a & c). F3 axial surfaces trend NE with moderate or shallow dips, and F3 hinge lines plunge gently NE or SW coaxial with L2 (Fig. 7d & e). S3 is predominantly a crenulation fabric which deforms L2, S2 and co-planar granitic veins (Fig. 6g). The fabric is present mainly in the short, steep limbs of major F3 folds. Thin sheets of garnet- and cordierite-bearing, S-type granite co-planar with S3 show that syn-tectonic, high-temperature conditions characterized D3. They are presumed to have persisted from D1–D2. Patches and irregular sheets of S-type granite are abundant in coastal exposures just south of Lindenow Fjord, where earlier structures were largely obliterated by partial melting during D3. Growth of hornblende across S3 shows that high-grade metamorphism outlasted D3. The age range of syn-D3, S-type granites is 1786–1778 Ma (Fig. 4).
D4 and D5 structures and fabrics
D4 folds are most widely developed in the outer part of the forearc (Fig. 3). They are kilometre-scale, open to tight folds with upright or steep axial surfaces trending mainly NE with shallow SW–NE plunge (Fig. 7c). S4 occurs locally as crenulation or penetrative fabrics with recrystallization of biotite and hornblende indicating amphibolite facies P–T conditions. However, D4 post-dates not only the peak of metamorphism, but also large irregular bodies of S-type granite and sheets of granitic rocks with batholith affinity. A few sheets and dykes of commingled granite s.l. and hornblende diorite crowded with cognate and xenolithic inclusions of ultrabasic to intermediate rocks and characterized locally by steep planar fabrics parallel to F4 axial surfaces, appear to have been emplaced during D4. An appinite dyke parallel to an F4 axial surface has an age of 1736 ± 2 Ma and gives a minimum age for D4 (Hamilton, unpublished data; Fig. 4). The upright to steeply inclined attitude of major D4 folds implies that they were a consequence of NW–SE shortening across the orogen.
D5 folds have a much more limited distribution. They are relatively open, large-scale warps with variable trend. One may be responsible for the basin-like fold defined by the gabbro sheet west of the outer part of Nørrearm (Figs 3 & 7c). D4–D5 fold interference was also responsible for a large-scale, open basin-like fold in outer Kangerluaraq (Fig. 3).
Partitioning model for the Ketilidian arc–forearc in SE Greenland
D1: strike-slip partitioned transpression
The bulk of the Psammite Zone of the Ketilidian forearc has a pre-D4 flat-lying structure characterized by orogen-parallel stretching with top-NE transport during the main deformation (D2). In contrast the batholith at inner Danell Fjord, NW of inner Nørrearm, and in other large areas of the western part of the orogen, is dominated by upright, NE-trending planar fabrics formed during sinistral displacements. At first sight, this contrast suggests that the kinematics of the flat-lying structure of the forearc and the steep structure of the batholith are grossly incompatible (Fig. 8 a) but we explain the apparent incompatibility as a simple consequence of different modes of deformation partitioning above and below a mid-crustal attachment system in the Psammite Zone (Fig. 8b).
(a) Cross-section A–A′ from ‘Sorte Nunataq’ to central Danell Fjord. (b) Cross-section B–B′ from inner Nørrearm to inner Kangerluaraq.(c) Cross-section C–C′–C′′ from Stendalen to central Lindenow Fjord. (d) Stretching directions (L2) from the domain of high D2 strain between inner Nørrearm and central Danell Fjord that are relatively undisturbed by F3 and F4 folds. Stretching is defined by crystallographic alignment of minerals on foliation surfaces (e.g. sillmanite, hornblende), the shape (long axis) of aggregates of grains and/or shape (long axis) of ellipsoidal objects (e.g. conglomerate pebbles, calc-silicate nodules). (e) Poles to bedding in the domain of D1 deformation in the inner Danell Fjord area and poles to bedding (transposed to S2) in the central Danell Fjord area. The pole to the best-fit great circle gives an estimate of the F3 fold axis (see cross-section a). (f) Poles to S2, inner Nørrearm. The scatter from shallow to steep dips reflects steepening of the planar fabric from shallow dips in the Psammite Zone to steep dips in the batholith. The stretching lineations (L2) on the foliation surfaces are also plotted. (g) Poles to the main planar fabric, S2 in the domain of F4 folding between north central Lindenow Fjord and Kuutseq fjord. The data define a π-girdle with an axis plunging gently to the WSW giving the approximate F4 fold axis. Note that the bedding/S1 intersection from inner Danell Fjord (b) the L1–L2 lineation (a) the F3 fold axis (b) and the F4 fold axis (d) are approximately co-axial.
(a) Schematic block diagram of the Psammite Zone–batholith boundary at Danell Fjord and inner Nørrearm showing the apparent incompatibility between the displacement patterns in the batholith and the forearc. (b) Schematic block diagram of the batholith and the Psammite Zone during D1 showing strike-slip partitioned transpression in the upper crust with sinistral strike-slip parallel to the arc in the batholith and arc-normal contraction in the forearc (Psammite Zone). The forearc is underlain by an attachment zone on which the forearc blocks were translated to the NE. This D1 attachment is below the present level of erosion.
Our model is constrained chronologically by the period of emplacement of the Julianehåb batholith c. 1854–1790 Ma, and the ages of detrital zircon grains from the Psammite Zone which mainly coincide with this range, the youngest dated being c. 1793 Ma old (Fig. 4). Thus the forearc basin was fed mainly by sediment from the arc, the age of the youngest detrital zircons marking the end of deposition and the maximum age of arc uplift and unroofing. Deformation and high-grade metamorphism of the sedimentary rocks followed almost immediately after the close of deposition in the forearc, flat-lying high grade D2 fabrics forming in the Psammite Zone by c. 1792 Ma, as indicated by the age of the syn-D2 hornblende granite (Fig. 4).
In SE Greenland pluton emplacement in the batholith was characterized by syn- and post-magmatic sinistral strike-slip. At the head of Danell Fjord, the boundary between the batholith and the Psammite Zone is marked by sinistral strike-slip shear zones and steep, ENE-trending strike-slip related fabrics in the batholith are continuous with D1 fabrics in the adjacent part of the Psammite Zone. SE-facing D1 fold nappes in the area of low D2 strain NW of Kangerluaraq and the dominant SE vergence of D1 minor folds south of central Danell Fjord are consistent with NW–SE shortening during D1. These differences between the batholith and the Psammite Zone structure imply that D1 was a partitioned transpression, with thrusting to the SE in the Psammite Zone and sinistral strike-slip in the adjacent batholith (Fig. 8b). In inner Nørrearm, shallow dip of the Psammite Zone beneath the batholith indicates a component of SE-directed thrusting during D1. However, D2 is intense at the boundary between the batholith and Psammite Zone in Nørrearm and we do not know whether the granitic sheets of the batholith overlie the forearc primarily as a result of D1 thrusting, or whether they were emplaced during D2 as sheets more or less at their present level within the sedimentary rocks of the proximal forearc.
Deformation in the western part of the orogen during the main phase of construction of the batholith is also consistent with D1 partitioned sinistral transpression in SE Greenland. Plutonic complexes emplaced at c. 1845–1805 Ma contain magmatic-state and crystal-plastic fabrics consistent with sinistral strike-slip which was contemporaneous with ENE-trending, upright and inclined folds of Palaeoproterozoic supracrustal rocks in the Border Zone (Garde et al. 2002). These chronological and kinematic relationships point to contractional deformation in the Border Zone and synplutonic sinistral strike-slip in the batholith in a regime of strike-slip partitioned transpression in this part of the orogen prior to 1805 Ma. We have shown that sinistral transpression continued during the later stages of construction of the batholith in SE Greenland although, in contrast, emplacement of late plutonic complexes in the NW of the orogen was associated with dextral strike-slip between 1805 Ma and 1790 Ma (McCaffrey et al. 1998).
D2: an attachment zone in the Ketilidian forearc
Models for partitioned transpression at transform or obliquely convergent subduction boundaries imply the presence of a flat-lying ductile attachment system below the partitioned domain (Fig. 9 a & b). The attachment marks the boundary between the partitioned zone and a zone of homogeneous or heterogeneous distributed shear at depth (St-Blanquat et al. 1998; Teyssier & Tikoff 1998). The orientation of the stretching lineation in the attachment zone and the amount of displacement will depend on the way in which transpression is partitioned in the upper crust, but the horizontal shear sense in the attachment zone must be compatible with the margin-parallel strike-slip shear sense (Fig. 9c). These principles can be applied to the Ketilidian orogen to reconcile the kinematics of D1 and D2 within a single, progressive kinematic framework.
(a) Plan view of an upper crustal transpressional domain. Deformation is partitioned into: (i) strike-slip on a vertical, sinistral fault and (ii) shortening normal to the shear plane forming folds (after Teyssier & Tikoff 1998). The original shape of the deformed block is shown with short, broken lines. (b) Plan view of a lower crustal domain, underlying the upper crustal domain of (a). Deformation is homogeneous transpression. The different response of each domain to transpression requires a sub-horizontal attachment zone to provide kinematic continuity between them. (c) Displacement vectors in the attachment zone. The vectors are determined by the difference between the displacement fields in the domains above and below the attachment.
D2 in the Psammite Zone was characterized by top-NE transport which gave rise to an orogen-parallel stretching fabric. Our evidence shows that D2 in the Psammite Zone was superimposed on D1 sinistral transpression fabrics representing the deformation regime during the emplacement of the batholith. At Nørrearm, structures in the Psammite Zone progressively steepen to the NW in a manner that does not appear to be a consequence of refolding. There are no overprinting cleavages and the structure appears to be a primary steepening of the D2 fabric at the boundary between the Psammite Zone and the Julianehåb batholith. In attachment systems, change in the orientation of planar fabrics from shallow to steep and a change from arc-parallel transport on a flat-lying shear zone to strike-slip deformation on a steeply dipping shear zone occur where a mid-crustal attachment zone meets a major arc-parallel strike-slip systems that roots into the deep crust (Figs 8b & 10 a). We propose that these changes took place in the Psammite Zone during D1–D2 in order to account for the D1 strike-slip transpression in the batholith and a D2 top-NE attachment zone in the Psammite Zone exposed at the same crustal level. The apparent conflict between D1–D2 kinematics can thus be resolved on the grounds that D1 represents a strike-slip partitioned upper crustal domain and D2 a flat-lying attachment zone (Fig. 8b).
Variation in the level of the attachment zone
Overprinting of D1 partitioned transpression fabrics by D2, top-NE fabrics in the Psammite Zone in general, and in the batholith at Nørrearm in particular, implies that the D2 attachment system moved upward in the crust during progressive deformation. Deformation of the Psammite Zone was accompanied by HT–LP metamorphism characterized by extensive migmatization. Although steep geothermal gradients may be a consequence of regional extensional deformation, we have found no extensional structures in the Psammite Zone that lend support to this idea. In their absense, the likely cause of this heat advection to relatively high crustal levels was the emplacement of thick sheets of syntectonic, batholith-related granodiorite, diorite, gabbro and abundant appinite dykes throughout the Psammite Zone during D2. Since the position of the attachment zone at the base of the partitioned upper crustal domain is likely to have been controlled by temperature (Teyssier & Tikoff 1998), upward movement of isotherms during the development of HT–LP metamorphism at high crustal levels would promote upward migration of the attachment zone causing overprinting in upper crustal domains characterized by strike-slip partitioned transpression (Fig. 10b). We contend that this migration took place during deformation and progressive metamorphism of the Psammite Zone during D1–D2. Change in level of the attachment readily accounts for the structures and fabrics in the Psammite Zone that are chronologically related to D1 strike-slip partitioned transpression but overprinted by D2 fabrics related to an attachment zone.
(a) Schematic interpretation of the Danell Fjord cross-section showing strike-slip partitioned transpression during D1. The sinistral strike-slip component is taken up on steep shear zones in the granitic rocks of the Julianehåb batholith. The arc-normal component is taken up by SE-directed thrusting in the forearc. At the base of the forearc fold-thrust belt an attachment zone separates the thrust belt from a domain of distributed shear in the forearc basement below. (b) Detail of (a) during D2. Folds of the D1 thrust belt are shown overprinted by intense D2 fabrics. D2 deformation overprinting is a consequence of the attachment zone moving to a higher level in the crust. Granitic sheet emplacement during D2 may be the source of heat that caused the attachment to migrate up the crustal section.
Conclusions
Of the five phases of deformation recognized in the Psammite Zone (forearc) of the Ketilidian orogen, D1, D2 and D3 represent a progressive deformation linked to sinistral transpression in response to oblique convergence at the Ketilidian subduction boundary. D1 was a partitioned transpression characterized by syn- and post-emplacement sinistral strike-slip in the NE part of the Julianehåb batholith adjacent to the Psammite Zone, and SE-directed thrusting in the Psammite Zone itself. D1 is regarded as a progressive development of sinistral transpressional deformation in the central and southern part of the batholith and the Border Zone in the western part of the orogen (Garde et al. 2002).
D2 is characterized by intense, flat-lying fabrics in the Psammite Zone with orogen-parallel top-NE transport. Steepening of D2 fabrics from SE to NW across the boundary between the Psammite Zone and the Julianehåb batholith in Nørrearm is interpreted as a primary change in the orientation of D2 strain trajectories. In the Psammite Zone, D2 represents a mid-crustal attachment system between partitioned transpression in the upper crust and heterogeneous or homogeneous transpression in the middle crust. Since the attachment steepens beneath the batholith, it lies at a higher structural level in the Psammite Zone. This difference in level is interpreted as an effect of HT/LP metamorphism in the Psammite Zone during D2 which resulted in upward migration of isotherms in the forearc which progressively led to the attachment system moving higher in the crust. The upward movement of the isotherms was effected by emplacement of late magmatic plutonic complexes of the arc into the forearc.
Acknowledgements
Fieldwork in SE Greenland in 1992, 1994 and 1996 was carried out as part of the SUPRASYD PROJECT under the auspices of the Geological Survey of Denmark and Greenland (GEUS). Fieldwork in 1998 and 1999 was undertaken with financial support from the Natural Environment Research Council, UK (research grant GR3/12070 to KM, JG and BC), the Danish Natural Science Research Council, the Carlsberg Foundation (research grants to AAG) and GEUS. We are deeply indebted to colleagues at GEUS for logistical and administrative support in Copenhagen and Greenland. We are grateful to our scientific colleagues on the SUPRASYD project for their part in the lively discussions that took place during the field campaigns and to C. Teyssier and M. de Saint-Blanquat for reviews that led to significant improvements of the paper. We thank the staff at Telestation Prins Christian Sund for their steadfast support, comfortable accommodation and unflagging hospitality. This paper is published with permission of the Director of the Geological Survey of Denmark and Greenland and is Geological Survey of Canada contribution no. 2001199.
- © 2002 The Geological Society of London
References
Mid-crustal partitioning and attachment during oblique convergence in an arc system, Palaeoproterozoic Ketilidian orogen, southern Greenland
