9
Long-lived, cold burial of Baltica to 200 km depth Dirk Spengler a, , Hannes K. Brueckner b,c , Herman L.M. van Roermund d , Martyn R. Drury d , Paul R.D. Mason d a Department of Geology and Mineralogy, Graduate School of Science, Kyoto University, 606-8502 Kyoto, Japan b Queens College and the Graduate Center of the City University of New York, Flushing, NY 11367, USA c Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY 10964, USA d Faculteit Geowetenschappen, Universiteit Utrecht, Budapestlaan 4, 3584CD Utrecht, The Netherlands abstract article info Article history: Received 2 November 2008 Received in revised form 1 February 2009 Accepted 2 February 2009 Available online 9 March 2009 Editor: T.M. Harrison Keywords: UHP metamorphism orogenic cycle micro-diamond Caledonides majoritic garnet low Al orthopyroxene The collision of two continents causes subduction of one of the continental margins temporarily below the other into the diamond stability eld (N 3.8 GPa or N 120 km depth), coeval ultra-high pressure (UHP) metamorphism of the continental crust followed by exhumation of UHP metamorphic terrains. Recent thermo-mechanical models propose that this subductionexhumation cycle is short-lived (b 15 Ma), contradicting the range in metamorphic ages observed in several high pressure/UHP metamorphic terrains. Here we use microstructures, mineral chemistry, SmNd geochronology and NdSr isotope systematics to show that the micro-diamond bearing Western Gneiss Region in the Scandinavian Caledonides of western Norway was subjected to UHP conditions for c. 30 Ma during a long-lived cycle of subduction and exhumation related to the Scandian phase of the Caledonian orogeny. Orogenic peridotite bodies on Otrøy and Flemsøy islands are interpreted as mantle wedge fragments tectonically emplaced into Baltic continental crust during the prograde continental subduction of Baltica underneath Laurentia after c. 438 Ma. Subduction related deformation and associated strain-induced recrystallization of the mantle fragments partially to completely destroyed pyroxene exsolution microstructures in cm-scale garnet within layers of garnetpyroxenite that have a Palaeoarchaean origin (3.33±0.19 Ga, 2σ, 5 point whole rock). Millimeter scale recrystallized orthopyroxene in garnetwebsterite has low Al cores (0.10 wt.% Al 2 O 3 ) showing recrystallization conditions at UHP , 6.3 ± 0.2 GPa, and at sub-geotherm temperatures of Archaean areas, 870 ± 50 °C. Three mineral isochrons of 2 to 5 points from recrystallized mineral assemblages of garnetpyroxenite indicate overlapping, early Scandian ages (429.5±3.1 Ma, 2σ, weighted mean) showing that the Archaean mantle fragments record the prograde subduction of Baltica to 200 km depth underneath Laurentia in c. 8 Ma (25 mm a 1 vertical subduction rate). Contrasting 87 Sr/ 86 Sr in recrystallized clinopyroxene from Otrøy and Flemsøy (0.70160.7023 and 0.7131, respectively) indicates strain-induced recrystallization occurred partly at dry (uid absent) conditions. Subsequent metamorphic conditions during peridotite retrogression record exhumation through c. 120 km depth (3.8 GPa), overlapping maximum metamorphic conditions recorded in regional country-rock eclogites dated at c. 400 Ma. A slow average vertical exhumation rate of 3.6 mm a 1 is implied for the diamond phase stability. Most external eclogites crystallized or re-equilibrated, probably triggered by deformation and uids, during crustal exhumation c. 30 Ma after the peridotites crystallized early Scandian garnets. Crustal micro-diamond formed by long- lived but cold UHP metamorphism, and found almost exclusively in Phanerozoic orogens suggests a change in the nature of collisional tectonics within a cooling earth. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Ancient exhumed continentcontinent collision zones preserve evidence for the temporarily burial, down to mantle depth N 120 km, of one continental margin below the other. Evidence includes metamorphic pressure (P) and temperature (T) gradients normal to the collision front (Krogh, 1977; Grifn et al., 1985), metamorphic index minerals like coesite (Coe) and micro-diamond (Dia; other mineral abbreviations after Kretz, 1983) within the subducted continental crust (Sobolev and Shatsky, 1990; Dobrzhinetskaya et al., 1995) and orogenic (mantle and crustal) peridotites embedded in the subducted continental crust (Brueckner and Medaris, 2000; Liou et al., 2007). In addition, parts of the cycle of burial (subduction), reversal and exhumation of positively buoyant continental lithosphere have successfully been modelled in thermo-mechanical/thermo-tectonic numerical studies (Ranalli et al., 2000; Warren et al., 2008). The rst part of the cycle, the subduction of continental lithosphere, transforms Earth and Planetary Science Letters 281 (2009) 2735 Corresponding author. E-mail address: [email protected] (D. Spengler). 0012-821X/$ see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2009.02.001 Contents lists available at ScienceDirect Earth and Planetary Science Letters journal homepage: www.elsevier.com/locate/epsl

Long-lived, cold burial of Baltica to 200 km depth

Embed Size (px)

Citation preview

Page 1: Long-lived, cold burial of Baltica to 200 km depth

Earth and Planetary Science Letters 281 (2009) 27–35

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

Long-lived, cold burial of Baltica to 200 km depth

Dirk Spengler a,⁎, Hannes K. Brueckner b,c, Herman L.M. van Roermund d,Martyn R. Drury d, Paul R.D. Mason d

a Department of Geology and Mineralogy, Graduate School of Science, Kyoto University, 606-8502 Kyoto, Japanb Queens College and the Graduate Center of the City University of New York, Flushing, NY 11367, USAc Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY 10964, USAd Faculteit Geowetenschappen, Universiteit Utrecht, Budapestlaan 4, 3584CD Utrecht, The Netherlands

⁎ Corresponding author.E-mail address: [email protected] (D. Sp

0012-821X/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.epsl.2009.02.001

a b s t r a c t

a r t i c l e i n f o

Article history:

The collision of two contine Received 2 November 2008Received in revised form 1 February 2009Accepted 2 February 2009Available online 9 March 2009

Editor: T.M. Harrison

Keywords:UHP metamorphismorogenic cyclemicro-diamondCaledonidesmajoritic garnetlow Al orthopyroxene

nts causes subduction of one of the continental margins temporarily below theother into the diamond stability field (N3.8 GPa or N120 km depth), coeval ultra-high pressure (UHP)metamorphism of the continental crust followed by exhumation of UHP metamorphic terrains. Recentthermo-mechanical models propose that this subduction–exhumation cycle is short-lived (b15 Ma),contradicting the range in metamorphic ages observed in several high pressure/UHP metamorphic terrains.Here we use microstructures, mineral chemistry, Sm–Nd geochronology and Nd–Sr isotope systematics toshow that the micro-diamond bearing Western Gneiss Region in the Scandinavian Caledonides of westernNorway was subjected to UHP conditions for c. 30 Ma during a long-lived cycle of subduction andexhumation related to the Scandian phase of the Caledonian orogeny. Orogenic peridotite bodies on Otrøyand Flemsøy islands are interpreted as mantle wedge fragments tectonically emplaced into Baltic continentalcrust during the prograde continental subduction of Baltica underneath Laurentia after c. 438 Ma. Subductionrelated deformation and associated strain-induced recrystallization of the mantle fragments partially tocompletely destroyed pyroxene exsolution microstructures in cm-scale garnet within layers of garnet–pyroxenite that have a Palaeoarchaean origin (3.33±0.19 Ga, 2σ, 5 point whole rock). Millimeter scalerecrystallized orthopyroxene in garnet–websterite has low Al cores (≥0.10 wt.% Al2O3) showingrecrystallization conditions at UHP, 6.3±0.2 GPa, and at sub-geotherm temperatures of Archaean areas,870±50 °C. Three mineral isochrons of 2 to 5 points from recrystallized mineral assemblages of garnet–pyroxenite indicate overlapping, early Scandian ages (429.5±3.1 Ma, 2σ, weighted mean) showing that theArchaean mantle fragments record the prograde subduction of Baltica to 200 km depth underneath Laurentiain c. 8 Ma (25 mm a−1 vertical subduction rate). Contrasting 87Sr/86Sr in recrystallized clinopyroxene fromOtrøy and Flemsøy (0.7016–0.7023 and 0.7131, respectively) indicates strain-induced recrystallizationoccurred partly at dry (fluid absent) conditions. Subsequent metamorphic conditions during peridotiteretrogression record exhumation through c. 120 km depth (3.8 GPa), overlapping maximum metamorphicconditions recorded in regional country-rock eclogites dated at c. 400 Ma. A slow average verticalexhumation rate of 3.6 mm a−1 is implied for the diamond phase stability. Most external eclogitescrystallized or re-equilibrated, probably triggered by deformation and fluids, during crustal exhumation c.30 Ma after the peridotites crystallized early Scandian garnets. Crustal micro-diamond — formed by long-lived but cold UHPmetamorphism, and found almost exclusively in Phanerozoic orogens— suggests a changein the nature of collisional tectonics within a cooling earth.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Ancient exhumed continent–continent collision zones preserveevidence for the temporarily burial, down to mantle depth N120 km,of one continental margin below the other. Evidence includesmetamorphic pressure (P) and temperature (T) gradients normal tothe collision front (Krogh, 1977; Griffin et al., 1985), metamorphic

engler).

ll rights reserved.

index minerals like coesite (Coe) and micro-diamond (Dia; othermineral abbreviations after Kretz, 1983) within the subductedcontinental crust (Sobolev and Shatsky, 1990; Dobrzhinetskaya et al.,1995) and orogenic (mantle and crustal) peridotites embedded in thesubducted continental crust (Brueckner andMedaris, 2000; Liou et al.,2007). In addition, parts of the cycle of burial (subduction), reversaland exhumation of positively buoyant continental lithosphere havesuccessfully been modelled in thermo-mechanical/thermo-tectonicnumerical studies (Ranalli et al., 2000; Warren et al., 2008). The firstpart of the cycle, the subduction of continental lithosphere, transforms

Page 2: Long-lived, cold burial of Baltica to 200 km depth

Fig. 1. Simplified geological maps of western Otrøy (Carswell et al., 2006) and easternFlemsøy (Terry et al., 2000b) showing the position of analyzed samples and of orogenicGrt-peridotite bodies (1; M — Midsundvatnet, N — Nogvadalen, R — Raudhaugene, U —

Ugelvik) embedded in gneiss (2 — banded dioritic gneiss with abundant eclogite, 3 —

migmatitic or augen gneiss with minor eclogite, 4— granitoid gneiss and metasediment).WGR, Western Gneiss Region; VGR, Vestranden Gneiss Region.

Fig. 2.Meso- and microstructures of Grt-pyroxenite. a) Field image of a tight fold hingeof alternating peridotite and pyroxenite (pen=15 cm). b, c) Optical light micrographsshowing oriented lamellae of M2 Pyx in porphyroclastic M2 Grt and lamellae-freerecrystallized M3 Grt and Cpx in pyroxenite DS0288 (c= subset of b). d) Back-scatteredelectron image of recrystallized M3 grains with M4 symplectite of Ol+Ilm±Rt inferredto be after M3 Ti-Chu in pyroxenite 99NR6. Scale bars: 500 µm (b) and 100 µm (c, d).

28 D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

near-surface low pressure (LP) rocks to high pressure (HP) and ultra-high pressure (UHP) metamorphic crustal rocks. These includeeclogite (Eskola, 1915), Coe bearing gneiss (Chopin, 1984; Smith,1984) and micro-Dia bearing gneiss (Sobolev and Shatsky, 1990;Dobrzhinetskaya et al., 1995), which are believed to preserve themaximum depth of plate burial. The third stage, the exhumation, isgenerally assumed to overprint to varying degrees the HP/UHPmineral assemblages. The duration of both the first and third part ofthe cycle is constrained indirectly by palaeomagnetic and numericalmodels and directly by combined isotope and mineral chemicalstudies that all suggest fast rates for both continental plate subductionand crustal exhumation. Reported rates for vertical plate movementrange from 7 to more than 80 mm a−1 in a single orogen (Torsvik etal., 1996; Terry et al., 2000a; Root et al., 2004; Camacho et al., 2005;Kylander-Clark et al., 2008) that all would lead to estimates for thetotal burial time of less than 15 Ma.

Evidence for the duration of the intermediate stage two, i.e. theintegrated mantle residence time between early subduction (1) andlate exhumation (3) is sparse. The reversal of crustal rock movementis generally believed to occur rapidly as the buoyancy contrastbetween crustal and mantle lithologies increases with depth.Eclogites preserve large age ranges that suggest a larger duration ofeclogite-facies conditions of up to 25 Ma (Griffin and Brueckner,1985; Mattinson et al., 2006; Kylander-Clark et al., 2007). In addition,micro-Dia bearing UHP metamorphic terrains commonly encloseorogenic Grt-peridotite bodies that record systematically higher peakP and T conditions (Hirajima and Nakamura, 2003) than do eclogitesand thus suggest deeper subduction. It follows that chronologicaldata from crustal rocks may also systematically underestimate thetime of peak metamorphism, and therefore the residence time atmantle conditions. Herewe show that orogenic peridotite enclosed inBaltica basement gneiss at Otrøy and Flemsøy islands in westernNorway (Fig.1) records the prograde subduction and early retrogradeexhumation history of the Baltica plate margin, indicating that this

margin resided in the Dia stability field for c. 25 Ma, in the Coestability for c. 30 Ma, much longer than generally accepted for anyexhumed UHP metamorphic terrain.

2. Geological setting and sample description

The Caledonides in Scandinavia formed as a result of the Palaeozoiccontinent–continent collision between Laurentia and Baltica andconsists of a pile of tectonic nappes translated eastwards onto theBaltica continental margin (Roberts and Gee, 1985). The WesternGneiss Region (WGR) of Norway hosts a tectonic window through thisnappe pile and exposes at the lowest tectonostratigraphic position alarge segment of the former craton of Baltica that is dominated byPalaeo- to Mesoproterozoic, mainly felsic intrusives with minor maficrocks, which are often converted to high-grade gneiss and eclogite(Tucker et al., 2004). This Baltica basement gneiss encloses minororogenic peridotite bodies. Many of them are interpreted to beincorporated from the overlaying mantle wedge into the subductingcontinental crust (Brueckner andMedaris, 2000) during the final, LateSilurian to Early Devonian, stage of contraction, called the Scandianphase (or Scanidan orogeny) of the Caledonian orogeny in Scandina-via (Roberts, 2003; Brueckner and Van Roermund, 2004).

Orogenic peridotites on Otrøy and Flemsøy are melt depleted,serpentinized Grt- and Spl-harzburgite and -dunite (Fig. 1; Carswell,1968) that are compositionally banded at mm–m scale. Layers andlenses of Grt-pyroxenite and garnetite (Fig. 2a) at mm–dm scaleparallel the compositional banding (Carswell, 1973; Van Roermundand Drury, 1998) and form the focus of this study. Two types of Grt-pyroxenite occur: bi-mineralic Grt(red)-clinopyroxenite and tri-mineralic Grt(purple)-websterite. Minor components in both typesof pyroxenite are Amp, Ol, Ilm and Rt. The layering of peridotite andpyroxenite is tightly to isoclinally folded, with fold wavelengths atcm–m scale (Fig. 2a) and an axial plane foliation.

It is essential to distinguish between pre-subduction (M1 and M2)and subduction-related (M3) mineral assemblages to demonstrate thatGrt-pyroxenite records a prograde subduction history. Eight sampleswere collected (Fig. 1, Table S5) from bi- (99NR6, DS0380, DS0384) and

Page 3: Long-lived, cold burial of Baltica to 200 km depth

Fig. 3. Otrøy and Flemsøy Grt-websterite used for P–T estimates. Left: Back-scatteredelectron micrographs show recrystallized, short- to long-prismatic M3 Opx in texturalequilibrium with M3 Grt and Cpx. Right: Al2O3 content along profiles AB shown on theleft. The Al2O3 minimum content in M3 Opx cores is sensitive to the grain size.

29D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

tri-mineralic (Fl99-26, DS0246, DS0288, DS0346, DS03AO) Grt-pyrox-enite. All thin sections show a shape preferred orientation ofrecrystallized, short- to long-prismatic M3 Pyx and M3 Grt, 0.3 to 1 mminwidth (Figs. 2 and3),which locally surroundporphyroclastic grains andfragments of M2 Grt precursors, several mm inwidth (Fig. 2b, c). The M2

andM3minerals form Pyx-rich and Grt-rich stretched clusters and bands,which define an internal compositional layering atmm–cm scale. Cores ofthe porphyroclastic Grt preserve straight, crystallographically orientedlamellae ofM2 Pyxor Pyx+Ilmup to a few100µm in length and3–12µminwidth (Fig. 2b, c). Nd isotope studies indicate that the majoritic M1 Grtprecursor, out of which the M2 Pyx lamellae exsolved, formed in theArchaean (Spengler et al., 2006). Trace element analyses suggest thatexsolution of the coarse,HT (≤1380 °C) M2 Pyx lamellae occurred during

isobaric cooling of the refractory peridotite in the sub-continentallithospheric mantle. Sm–Nd isotopes record that formation of this M2

Pyx from M1 Grt occurred in the Mesoproterozoic or earlier (Spengleret al., 2006; type 1). In addition, other microstructural and trace elementstudies (Carswell and Van Roermund, 2005; Scambelluri et al., 2008; type2) have proposed that some majoritic M1 Grt exsolved their Pyxcomponent, triggered by deformation and/or fluid infiltration, duringthe Scandian at LT (800–1000 °C). Importantly, the recrystallized M3 Grtstudied here lacks Scandian M3 Pyx exsolution lamellae (Fig. 2c) in theOtrøy and Flemsøy peridotites. Thus, the strain-induced M3 recrystalliza-tion event described below postdates the exsolution of M2 Pyx from M1

Grt in the Otrøy and Flemsøy peridotites.Both typesofpyroxenite containminor symplectiteofM4Ol+Ilm±Rt

after a M3 precursor phase (Fig. 2d). These M4 break-down products areinferred to have formed fromM3 Ti-clinohumite (Ti-Chu), a phase that isstable in the mantle at comparatively high P and low T (Weiss, 1997).

3. Methods

Major element compositions ofminerals were obtained by electronmicroprobe analysis (EMPA) using a JEOL JXA8600 superprobe atUtrecht University. The superprobe operated at standard conditions(15 kV, 20 nA, 30 s counting time in WDS mode, external calibrationagainst international standards).

Laser Ablation Inductively Coupled Plasma Mass Spectrometry(LA-ICP-MS) at Utrecht University was used to measure mineral rareearth element (REE) contents in recrystallized M3 Grt and Cpx fromtwo Grt-clinopyroxenites (DS0380, −84) and one Grt-websterite(DS03AO) from Otrøy. The system comprises a Micromass PlatformICP and a GeoLas 200Q laser microsampler (193 nm excimer, beamdiameter 120 m,10 Hz, 20 J cm−2). Samples were ablated for 90–120 sper spot with signal integration over N60 s. Internal calibration wasperformed using Ca (EMPA) and external calibration was against NISTSRM 612 glass standard.

Two different techniqueswere applied on five samples formineralseparation for isotope analysis. Grt and Cpx in the porphyroclastbearing Otrøy pyroxenites, comprising two less deformed Grt-clinopyroxenites (DS0380, −84) and two intensely deformed Grt-websterites (DS0346, −AO), were separated by a new hand pickingtechnique designed to collect fractions of minerals from both M2 andM3 mineral generations.

Double polished rock-slabs (12×18×0.3–0.45 mm) were stuck tothe centre of a c. 8 cm long removable tape (Scotch Magic RemovableTape 811, 1 inch core). This stripe was then rolled back onto the taperoll, so that the rock slabwas positioned in between two layers of tape.The rock slab fractured as a result of bending the tape back onto theroll and some careful finger-nail pressing. The slab broke preferen-tially along grain boundaries and cracks, such that individual grainsbecame loose but maintained their original location in the micro-structure as a result of being stuck on the tape. Then the c. 8 cm longstripe was carefully removed from the tape roll and fixed on a glassplate with the broken rock slab facing upwards. Subsequent handpicking under an optical microscope enabled perfect separationbetween recrystallized (M3) grains and porphyroclastic (M2) grains.Only those porphyroclastic grains containing visible Pyx lamellaewere selected to represent the M2 assemblage.

Mineral separates from the porphyroclast free, intensely recrys-tallized Flemsøy Grt-websterite (Fl99-26) were obtained by classicalcrushing, sieving, magnetic separation and hand picking methods.Relict Grt with exsolved Pyx was not observed in this samplesuggesting complete M3 re-equilibration. The Grt fractionwas dividedinto two sub-fractions with distinct fragment sizes.

All mineral separates were leached in hot HNO3 and a cold solution ofHCl andHF. Sampleswere spiked anddissolved. Separation of REE fromSrand other cations were performed using Eichrom TRU-Spec resin. Nd andSm were isolated from the REE fraction using alpha-hydroxyisobutyric

Page 4: Long-lived, cold burial of Baltica to 200 km depth

Fig. 4. Average mineral REE data from Otrøy Grt-clinopyroxenite DS0380 (dashed),DS0384 (dotted) and Grt-websterite DS03AO (solid) and the analytical mean detectionlimit (dash-dotted). a) C1-chondrite normalized REE distribution patterns of M3 Grt(solid circles) and Cpx (open circles) show a normal fractionation of light and heavyREE between both phases though the REE content differs between different samples. b)Cpx-Grt REE partition coefficients of Otrøy pyroxenite (lines) compared to those of HT(1380–1260 °C) and LT (1080–870 °C) mantle xenoliths (light and dark shaded field,respectively; Schmidberger and Francis, 2001) suggest an equilibration of therecrystallized mineral assemblage M3 (T≤870 °C). Error bars are 1σ.

30 D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

acid and cation resin. Sr was isolated from the Rb–Sr fraction usingEichrom Sr-Spec resin. Sm, Nd and Sr concentrations and Nd and Srisotopes were analysed by a 9 collector VG Sector 54-30 TIMS at Lamont-Doherty Earth Observatory of Columbia University.

4. P–T estimates

Major element profiles of mineral grains N0.5 mm reveal thatrecrystallized (M3) and porphyroclastic (M2) mineral grain cores arehomogeneous in composition; compositional variations (zoning)occur at grain rims. Core compositions of small grains (b0.5 mmacross) depend on the grain size. Cores of recrystallized M3 OpxN0.5 mm preserve plateaus of low Al2O3 content, 0.10–0.24 wt.%, thatsuggest equilibration (Fig. 3, supplementary Table S1). These Al2O3

contents are lower than in any of the even larger (up to several cmacross) M2 Opx (≥0.41wt.%) found in less deformed lenses of regionalGrt-pyroxenite and garnetite (Van Roermund et al., 2002; Carswellet al., 2006; Spengler et al., 2006). Because the Al content of Opx inequilibrium with Grt decreases with increasing P (and to a minorextent with decreasing T), the M3 mineral assemblages record muchhigher metamorphic conditions than the M2 mineral assemblages.

The Al2O3 content in M3 Opx increases towards the outermost c.100 µm of the grain rims (N1 wt.%), which suggests diffusion afterdecompression.

Mg/(Mg+Fe) varies little across the same grains of M3 Opx withcore ratios being less than 0.015 higher than rim ratios (Fig. 3).Because the Mg–Fe partitioning between Opx and Grt dependsmainlyon T, flat Mg/(Mg+Fe) profiles in M3 Opx zoned in Al indicate nearisothermal decompression (Terry et al., 2000b; Carswell et al., 2003).

Core compositions of pairs of minerals larger than 200 µm fromthe Otrøy and Flemsøy Grt-websterite samples were used for Al-in-Opx/Grt geobarometry (Brey and Köhler, 1990 and Brey et al., 2008,both calibrated on natural samples at 2.8–6 GPa and 2.8–8 GPa, res-pectively). Temperature conditions were estimated using the Fe–MgOpx/Grt geothermometers of (Brey and Köhler, 1990; Harley, 1984)and using the two-pyroxene-thermometer of (Brey and Köhler,1990; TBKN90). All calibrations were performed on natural samples at900–1400 °C, except for the synthetic samples of TH84 calibrated at800–1200 °C. P and T were solved iteratively for the followingcombinations of barometers and thermometers with the 1σ precision ofeach single barometer and thermometer given in brackets (Table S1):

(I) PBBG08 (±0.3 GPa)+TH84 (±40 °C),(II) PBKN90 (±0.22 GPa)+TBK90 (±65 °C),(III) PBBG08 (±0.3 GPa)+TBK90 (±65 °C),(IV) PBBG08 (±0.3 GPa)+TBKN90 (±15 °C).

Results are shown in Fig. 7a. The most coarse-grained, yetrecrystallized M3 Opx yields the highest P–T estimates for eachthermobarometer-combination, with one exception in combinationIV. Calculated metamorphic P and T decrease with decreasing M3 Opxgrain size. Combinations I and II give overlapping P–T estimates on thecoarsest (N1 mm) and finest (b0.3 mm) M3 Opx grain fractions, 5.9–6.5 GPa/870–920 °C and 3.8–4.0 GPa/820–880 °C, respectively.Combinations III and IV differ with significant lower estimates, 5.0–5.6 GPa/810–820 °C and 2.0–4.5 GPa/470–690 °C (coarse) and3.4 GPa/800 °C and 2.0–2.7 GPa and 620–680 °C (fine), respectively.Results (1σ) from combinations I and II are within the calibrationranges for experimental P and T or exceed the calibration limits byb0.3 GPa and b20 °C. In contrast, results (1σ) from combinations IIIand IV all exceed the calibration ranges for experimental P and/or T.

5. REE

The REE content of M3 Grt and Cpx matrix grains in three samplesfrom Otrøy was measured to characterize trace element partitioningin the recrystallized mineral assemblages (Table S2). M3 Grt and Cpx

in clinopyroxenite (DS0380, −84) and websterite (DS03AO) show anormal fractionation of REE with light REE (LREE) predominantlyhosted in Cpx and heavy REE in Grt (Fig. 4a). The REE content of eachof the two minerals is higher in clinopyroxenite than in websteritereflecting distinct whole rock REE content. Cpx in clinopyroxenite hasa negative slope REE pattern. In contrast, Cpx in websterite has aconvex shape REE pattern and has Grt with LREE content belowdetection limits; both features reflect a whole rock LREE depletion.Despite these differences, the partitioning of REE between recrystal-lized Cpx and Grt in all three samples partially overlap and are steeperthan those reported from both HT (1380–1260 °C) and LT (1080–870 °C) mantle xenoliths (Fig. 4b) and thus indicates an equilibrationof the M3 mineral assemblages in the two types of pyroxenite atrelatively low T (≤870 °C).

6. Isotopes

We analysed Sm–Nd and Sr isotopes on fractions of M2 and M3

minerals to determine the timing of recrystallization (Table S3). Threeof the Otrøy samples (DS0346, DS0380, DS0384) contained relics ofM1 Grt (M2 Grt+enclosed M2 Pyx lamellae) in sufficient quantity foranalysis. Plotting these Grt with the associated whole rock on Sm–Ndisochron diagrams generates lines (‘two point isochrons’). If the slopesof these lines are taken to have age significance, they indicate variablebut clearly pre-Caledonian ages older than 500 Ma (Fig. 5a). The agedifferences between the three samples probably reflect partial re-equilibration of the porphyroclasts during strain-induced recrystalli-zation that formed M3 Grt bearing assemblages. A pre-Caledonian agecan reasonably be inferred for Grt-clinopyroxenite and Grt-websteriteprecursors on Otrøy. An ultimate Archaean origin of the Grt-pyroxenite protoliths is suggested by a whole rock ‘errorchron’ of3333±190 Ma (five points, Fig. 5a). This Palaeoarchaean agecoincides with recent Re–Os results that indicate an Archaean melt-extraction event for several orogenic peridotite bodies in the region(Brueckner et al., 2002; Beyer et al., 2004) and confirms that the Grt

Page 5: Long-lived, cold burial of Baltica to 200 km depth

Fig. 6. Sr–Nd covariance diagram. Low present day 143Nd/144Nd and high present day87Sr/86Sr in Cpx of pyroxenite from Flemsøy and the adjacent island Fjørtoft (opencircles) contrast to those from Otrøy (filled circles). Large circles, this study. Smallcircles, literature data (Brueckner, 1974; Jamtveit et al., 1991; Brueckner et al., 2002).Cross symbols, isotope composition of whole rockmantle xenoliths from South America(Conceição et al., 2005). Cross lines indicate the present day isotope composition ofBulk Silicate Earth.

Fig. 5. Sm–Nd isochron diagrams (2σ) of Otrøy and Flemsøy Grt-pyroxenite. a) Twopoint best-fit lines (dashed) between whole rock and porphyroclastic Grt (containinglamellae of Pyx) from samples DS0346, −80 and−84 indicate a pre-Caledonian originfor the mineral assemblage M2. A five point whole rock errorchron (solid) suggests aPalaeoarchaean origin for the two types of pyroxenite. b) The isotope record ofrecrystallized phases from DS0380 and −84 (2× two points) and from the completelyrecrystallized sample Fl99-26 (five points, including Amp and whole rock) indicate anearly Scandian age for the mineral assemblage M3.

31D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

bearing mantle assemblages on Otrøy and Flemsøy shared theirmantle evolution with most WGR orogenic peridotite since theArchaean (Spengler et al., 2006).

Similar plots involving lamellae-free, recrystallized M3 Grt and Cpxfrom three of the analysed samples (Fig. 5b) indicate early Scandian ages(2σ) of 422.3±7.7Ma(DS0384, twopoints), 427.0±6.0Ma(DS0380, twopoints) and 434.0±3.2 Ma (Fl99-26, five points, MSWD=0.78). Thelatter 5 point isochron includes in addition to Grt and Cpx thewhole rockand Amp. Amp is likely to be secondary in origin as the recrystallizationcondition of PN6 GPa in this sample exceeds the maximum stability forAmp of c. 3 GPa except for K-richterite stable at PN6.5 GPa (Konzett andUlmer, 1999; Hawthorne et al., 2007). Excluding Amp, the isochronchanges insignificantly to 432.9±3.5 Ma (4 points, Nd(i)=0.511802(20), MSWD=0.38), and further excluding the whole rock leads to432.9±4.2 Ma (3 points, Nd(i)=0.511802(26), MSWD=0.76).Mineral separates from the other two analysed Grt-websterites(DS0346 and DS03AO) do not form isochron relationships.

The most conservative estimate for an average recrystallizationage may include data from recrystallized phases only. All threerecrystallization ages (2× two points,1× three points) overlapwithinerror at the 95% confidence level (2σ) and give a weighted age of429.5±3.1 Ma. This age is also within error of the five point isochronage of 434.0±3.2 Ma (Fl99-26). The well defined nature of theindividual isochron from the Flemsøy peridotite and the 2σreproducibility of the recrystallization age from three differentperidotite bodies suggest that the weighted age of 429.5±3.1 Marepresents a true geological event and not an artifact reflectingmixing or disequilibrium.

Sr isotope analyses were performed in M3 Cpx to provide anindependent constraint on the history of the orogenic peridotites. The87Sr/86Sr of the crust is in average higher than that of the mantle, dueto a higher time integrated Rb content. Therefore, 87Sr/86Sr in mantlefragments that is higher than that of bulk earth (0.7052) suggestscrustal contamination. Otrøy pyroxenites have M3 Cpx with low 87Sr/86Sr, ≤0.7023 (filled circles in Fig. 6), indicating that both bi- and tri-mineralic pyroxenite did not contain abundant Rb since theirformation in the Archaean and that the M3 recrystallization onOtrøywas dry (i.e. isochemical; Brueckner et al., 2002). In contrast, M3

Cpx in the Flemsøy pyroxenite has high 87Sr/86Sr, 0.7131 (large opencircle in Fig. 6), which suggests that this sample (this orogenicperidotite body) was metasomatised by crustal fluids before or duringthe recrystallization at 429.5 Ma. Generally, dry conditions retard andwet conditions enhance mineral–chemical equilibration. The lattercondition may explain the relatively precise age given by the Flemsøyperidotite. The former conditions may explain the larger uncertaintiesin the recrystallization ages obtained from the Otrøy Grt-clinopyrox-enites. The other two Otrøy garnet–websterite samples (DS0346 andDS03AO) have very low Sr ratios in M3 Cpx, 0.716–0.717, indicatingessentially unmetasomatized conditions that apparently were too dryfor the M3 mineral assemblage in these samples to reach isotopeequilibrium during the recrystallization event.

7. Discussion

7.1. Pressure–temperature–time evolution

All the microstructural, mineral chemical and isotope data recordmajor differences between the M2 and M3 mineral assemblages in thepyroxenites on Otrøy and Flemsøy (summarized in Fig. 7). The M2

mineral assemblages, characterized by a granular, partially coronitictexture in Grt-websterite, but also in garnetite andwebsteritic garnetitein earlier studies (Carswell, 1973; Van Roermund and Drury, 1998;Spengler et al., 2006), contain exsolution lamellae of Pyx in Grt (Fig. 2b,c) and record a major element equilibration at moderate P and Tconditions in the sub-continental lithospheric mantle, 3.3–3.7 GPa/740–770 °C for samples from Otrøy (Van Roermund et al., 2000;Carswell et al., 2006; Spengler et al., 2006) and 2.6–3.1 GPa/700–750 °Cfor samples from the neighbouring island Fjørtoft (Van Roermund et al.,2002; Carswell and Van Roermund, 2005; Carswell et al., 2006) usingPBKN90+TBK90 (Fig. 7b, Table S4).

Furthermore, the coarse grained M2 mineral assemblages preservepre-Caledonian ages in different types of pyroxenite (Fig. 5a;Brueckner et al., 2002) and garnetite (Spengler et al., 2006). Our

Page 6: Long-lived, cold burial of Baltica to 200 km depth

Fig. 7. Evolution of UHPmetamorphic rocks in the WGR. a) P–T estimates from Otrøy and Flemsøy pyroxenite using different calibrations iteratively (see text for abbreviations, TableS1). 1σ error bars correspond to each single thermometer and barometer. Arrows suggest different exhumation paths depending on the thermobarometers chosen. Dashed line (II):phase change Gr/Dia (Kennedy and Kennedy, 1976). b) Thermobarometric data of Baltica basement and enclosed orogenic peridotite in the Moldefjord (filled circles) and Storfjord(open circles) areas (Fig. 1). Large circles (legend as in panel a), this study; small circles, literature data (Table S4). M2, pre-Caledonianmineral assemblages that contain Pyx lamellaein Grt. M4, conditions possible for the formation of symplectite after M3 (this study). The deduced path for continental subduction and early exhumation in the Moldefjord area(upper arrow, this study) shows that the formation of the M3 assemblages is prograde. The early retrograde evolution in the Coe stability field was accompanied with partial mineralre-equilibration of M3. M4 mantle minerals and basement rock data indicate further retrogressionwithin the Qtz stability field (lower arrow, Terry et al., 2000a; Carswell et al., 2003).Most P–T data from the Storfjord area scatter around the Moldefjord exhumation path. Curves indicate phase changes (LP/HP) of Qtz/Coe (I), Gr/Dia (II; Kennedy and Kennedy,1976), Ol+Ilm±Rt+H2O/Ti-Chu (III; Weiss, 1997), the stability of 1 mol% majorite in solid solution with Grt (IV; Gasparik, 2000) and a geotherm corresponding to 36 mW m−2

surface heat flowdensity as typical for present Archaean areas (V; Kukkonen and Peltonen,1999). c) Geochronological data versus depth of Baltica basement exposed in theWGR andenvirons. Thick black arrow, Moldefjord area (segments A–C refer to different models). Thin black arrow, Lindås nappe in the Bergen Arcs (Glodny et al., 2008); grey arrow withsquares, VGR (Fig. 1). Shaded field, time integrated period when the western edge of Baltica (Moldefjord area) was buried to UHPmetamorphic conditions. Layered mafic intrusionsin the Köli Nappe between 61 and 69°N mark the onset of the continent–continent collision (diamonds). Stars, inferred emplacement of Otrøy and Fjørtoft peridotite. White dashedline, approximate position of phase change Gr/Dia. Hatched fields and braces indicate vertical rates for subduction and early exhumation. Other symbols as in panel b. Error bars are2σ for ages and 1σ for P. Three data points (upper left) apparently record isotope disequilibrium.

32 D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

new data show that the major deformation recorded in the orogenicperidotites coincides with intense deformation and recrystallization.This recrystallization event partially to completely replaced M2 by M3

mineral assemblages, the latter are characterized by much smallergrain sizes and a strong preferred orientation (Fig. 3), lack Pyxexsolution lamellae in Grt (Fig. 2c), record major and trace elementequilibration at sub-geotherm P–T conditions deep in the Dia stabilityfield (Figs. 4b and 7a, b) and provide early Scandian ages (Fig. 5b).

Reliable estimates for metamorphic conditions during M3 recrys-tallization should fall within the experimental P–T range determinedfor the chosen thermobarometers (see section on P–T estimates,combinations I and II). Overlapping P–T estimates using fourcalibrations of two thermobarometric techniques show high max-imum metamorphic conditions of PN6 GPa and TN850 °C. We givepreference to estimates from PBKN90+TBK90 (combination II), becausethese two calibrations base on a single set of experiments. Meanvalues for P and T from samples with M3 Opx N1 mm indicate peakmetamorphic conditions of 6.34 GPa/872 °C.

The analytical procedure (acid leaching prior to acid digestion ofthe mineral separates) minimizes a crustal alteration effect to accountfor the high radiogenic Sr content in the recrystallized M3 Cpx in theFlemsøy sample (Fig. 6). 87Sr/86Sr in the Flemsøy sample (0.7131) ishigher than that of seawater (≤0.709) and arc related volcanic rocks(0.703–0.708; Notsu et al., 1987; De Astis et al., 2000; Nakamura et al.,2008) implying crustal derived fluids metasomatized this orogenicperidotite intensively. The metasomatism did not disturb thePalaeoarchaean Nd isotope signature in the whole rock of the samesample although the whole rock Nd content is low (sub-ppm; Fig. 5a).Thus, high 87Sr/86Sr and undisturbed 143Nd/144Nd preserved in theFlemsøy pyroxenite suggests a decoupling of the Sr and Nd isotope

systems in which an increase in 87Sr/86Sr during crustal metasoma-tism was not accompanied by an increase in 143Nd/144Nd. Similarchromatographic records involving high radiogenic Sr enrichment(with ratios up to more than 0.71) have been reported from mantlexenoliths of active continental margins (Conceição et al., 2005; crosssymbols in Fig. 6) and have been assigned to a subduction-relatedmetasomatism. Following this interpretation, decoupling of the Sr andNd isotope systems combined with locally high ratios of 87Sr/86Srconstrain that the studied Norwegian peridotite bodies have beenderived from an Archaean sub-continental mantle wedge. Alterna-tively, high 87Sr/86Sr in strain-induced recrystallized M3 Cpx fromFlemsøy may indicate that crustal derived fluids metasomatized theperidotite simultaneously with deformation in the crust.

The Palaeoarchaean age for the protoliths of the two types of Grt-pyroxenite within peridotite that occur embedded in Palaeo- toMesoproterozoic Baltica basement gneiss shows that the Proterozoichistory of peridotite and crust differs, i.e. the peridotites were includedin the crust not before the Phanerozoic. Integration of all results implythat the orogenic peridotite bodies are mantle fragments that weretectonically emplaced from a cold sub-continental Archaean litho-spheric mantle wedge at depths of c. 90–120 km (Gr stability field) intocontinental basement of Baltic affinity during the prograde subductionof the Baltica plate margin underneath a Laurentian sub-continentalmantle wedge (according the model of Brueckner, 1998). Ongoingsubduction transported the Baltica crust and its peridotite cargo deeper(Dia stability field). Deformation and recrystallization coincides withScandian subduction down to 200 km depth, assuming metamorphic Pequals lithostatic P and equilibration T obtained from the Fe–Mgpartitioning of M3 Grt and Opx represent peak metamorphic T(Fig. 7b). If the latter would represent post-peak re-equilibration T

Page 7: Long-lived, cold burial of Baltica to 200 km depth

33D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

thenpeakmetamorphic P equivalent to depths N200 kmare implied, i.e.within the majoritic Grt stability field (Fig. 7b). The lack of Pyxexsolution lamellae in M3 Grt does not proof such high P conditions inthe studied samples and implies lack of evidence that crustalexhumation occurred from depths greater than 200 km. This is instrong contrast to the Grt-websterite lens at Bardane (Fjørtoft island)that contains Pyx lamellae in M3 Grt (Scambelluri et al., 2008).

P–T estimates from the smallest M3 grain cores (sample DS03AO)are consistent with partial mineral re-equilibration as the orogenicperidotites where exhumed from c. 6.3 to 3.8 GPa. P–T estimates of3.8 GPa and 825 °C overlap the Gr/Dia phase boundary as well as themetamorphic P–T estimates of associated eclogite facies crustal rocks(2.8–3.9 GPa and 810–850 °C; Table S4; Fig. 7b; Dobrzhinetskayaet al., 1995; Terry et al., 2000b; Carswell et al., 2006). The occurrenceof M4 Ol+Ilm±Rt symplectite (inferred to be after M3 Ti-Chu)implies that the M3 mineral assemblages in the orogenic peridotitebodies formed by recrystallization exclusively at UHP metamorphicconditions (Figs. 2d and 7b).

Three arguments support the premise that the Sm–Nd age of429.5±3.1 Ma obtained from the M3 mineral assemblages in thisstudy, represents a true crystallization age. First, three dated samplesprovide the same age. Second, the peak metamorphic T of c. 870 °C iswell below the Dodson closure T for Nd in Grt-clinopyroxenite,N1000 °C for a crystal size of ≤1 mm (Van Orman et al., 2002),excluding a cooling age. Third, dynamic recrystallization coincidedwith REE re-partitioning between the dated mineral phases, frominitial low DCpx/Grt in M2 (Spengler et al., 2006) to high DCpx/Grt in M3

(Fig. 4b); this re-partitioning would be expected to erase olderisotope signatures. Consequently, our new constraints on theprograde and peak UHP metamorphic evolution of orogenic perido-tites combined with the post-peak, UHP and HP metamorphicevolution of eclogite and gneiss available in the literature enablesus to reconstruct the full cycle of subduction and exhumation of amicro-Dia bearing Phanerozoic continent–continent collision zone(Fig. 7b).

7.2. Baltica's long-lived subduction–exhumation cycle

The age of the youngest ophiolite/ophiolite-like rocks exposed inthe Scandinavian Caledonides (the Solund–Stavfjord ophiolite com-plex, 443±3 Ma, Dunning and Pedersen, 1988) sets a maximum agefor the closure of the Iapetus ocean. Widespread syntectonic layeredmafic intrusions and felsic plutons and pegmatites are recorded insome of the tectonic nappes (the Upper Allochthone) along the entirelength of the orogen. They range from c. 438±2 to 431±3Ma (Tuckeret al., 2004; Nilsen et al., 2007), record a shift in magmatic activityfrom mafic to felsic compositions and are interpreted to mark theclosure of the Iapetus ocean and the earliest part of the continent–continent collision before Scandian metamorphism reached its peak.

The onset of the continent–continent collision at c. 438 Matogether with the new age for the peak metamorphism at c. 430 Maconstrain the vertical rate of prograde continental plate subduction inthe Moldefjord area to 25 mm a−1 (Fig. 7c). The weightedmetamorphic age, 429.5±3.1 Ma, coincides with the time of peakmetamorphism in the Baltica basement exposed north (VestrandenGneiss Region) and south (Bergen Arcs) of theWestern Gneiss Region;i.e. 432±6 Ma (Dallmeyer et al., 1992) and 429±4 Ma (Glodny et al.,2008) respectively (Figs. 1a and 7c). The similarity in the timing ofboth collision onset and peak metamorphism along hundreds ofkilometers of the orogen suggests a large portion of the edge of Balticabehaved uniformly during collision with Laurentia. Recorded max-imum metamorphic P and T conditions vary continuously betweenhinterland and foreland (Krogh, 1977; Griffin et al., 1985), butdiscontinuously parallel to the orogen (Griffin and Brueckner, 1985),consistent with propositions that individual Baltica basement unitsare separated by exhumation related major tectonic contacts (Wenn-

berg, 1996; Labrousse et al., 2004). Eclogite-facies exhumation of theUHP metamorphic basement in the Moldefjord area from c. 200 km(depth of peak metamorphism) to c. 120 km (onset of the Dia stabilityfield) between the 429.5±3.1 Ma peak event and the 407.0±2.1 Mamean age given by micro-Dia bearing crustal rocks of the samebasement unit (Terry et al., 2000a) indicates 3.6 mm a−1 ‘Dia-facies’average vertical exhumation rate assuming exhumation beganimmediately after 429.5 Ma (path A in Fig. 7c). A faster exhumationrate would imply a significant time interval between the end ofsubduction (documented by the peridotites of the Moldefjord area)and the start of exhumation (path B in Fig. 7c). Alternatively, the lackof data between c. 425 and 410 Ma may suggest two successive UHPmetamorphic events by re-burial of temporarily exhumed basementunits (path C in Fig. 7c). The latter scenario implies continentalsubduction reached the seismic discontinuity at 410 km depth(mantle transition zone), as also suggested from high convergencerates and low geotherms (Toissaint et al., 2004). All three modelsshow the Baltica basement of the Moldefjord area resided at UHPmetamorphic conditions until c. 400 Ma.

If alternatively the 429.5 Ma age is interpreted as a post-peakmetamorphic cooling age, then there would be no correlationbetween age and metamorphic record in the Moldefjord area (blackarrow in Fig. 7c). In addition, an age mixed of pre-Caledonian andScandian isotope signatures would fall only by chance in between theonset of Caledonian subduction (c. 438 Ma) and the generallyfavoured Scandian UHP metamorphic peak at c. 400–410 Ma (Fig. 7c).

It follows that: (1) the peak UHP metamorphism in the Balticabasement gneiss and enclosed orogenic peridotite bodies occurred veryquicklyafter thebeginningof the fast, cold continent–continent collisionbetween Baltica and Laurentia during the early Scandian; (2) the time-scale for the wholeUHPmetamorphism of the edge of Balticawas long-lived, c. 30 Ma (shaded field in Fig. 7c), (3) eclogite-facies exhumationwas on average slower in theDia stabilityfield, 3.6mma−1, than in theGrstability field, 10–11 mm a−1 (Terry et al., 2000a; Carswell et al., 2003),and (4) most eclogite either formed by crystallization or re-equilibratedby re-crystallization in the Gr stability field during exhumation. Thelatter (re-equilibration) justifies earlier high metamorphic P estimateson ‘internal eclogite/pyroxenite’ of up to 6 GPa (sample U47 of Carswellet al., 1985; sample 6–20 of Vrijmoed et al., 2006; Fig. 7b, Table S4),initially regarded for decades to represent disequilibrium, and explainsthe overall absence of comparable metamorphic estimates in ‘externaleclogite’ that preserves P of b4 GPa only (Fig. 7b; Terry et al., 2000a;Carswell et al., 2006; Kylander-Clark et al., 2007). The former (crystal-lization) would require an unknown threshold of Pb4 GPa to apply forthe eclogitization process in a subduction zone during exhumation,possibly triggered by fluids (Austrheim, 1987) and/or deformation.

Most P–T–t data from the Storfjord area south of Moldefjord(Fig. 1) scatter around the Moldefjord exhumation path (open circlesin Fig. 7b, c) suggesting a similar retrograde metamorphic evolutionin both areas at P≤4 GPa. In contrast, P–T–t data from eclogite of theNordfjord area further south (Fig. 1) differ with systematically lowerT of ≤750 °C at lower P of ≤3.5 GPa (Young et al., 2007). Thepreservation of low equilibrium T in M2 and M3 mineral assemblagesin all three areas below those of cold geotherms (Fig. 7b) is consistentwith slow conductive heat transfer in continental collision zonesespecially when thick, cold and predominantly dry cratonic litho-sphere subducts rapidly (Warren et al., 2008).

8. Conclusion and implications

Orogenic peridotite bodies on Otrøy and Flemsøy are mantlefragments that preserve evidence for a prolonged, relatively cold,residence in the Dia and Coe UHP stability fields, for c. 25 and c. 30 Marespectively, together with the enclosing Baltica basement. Such longmantle residence times for continental rocks have not been recog-nized before from other UHP metamorphic areas nor has it been

Page 8: Long-lived, cold burial of Baltica to 200 km depth

34 D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

proposed in geodynamic models of collisional tectonics. Destructionof mineral exsolution microstructures by prograde deformation andhigh 87Sr/86Sr of 0.7131 in strain-induced recrystallized Cpx implyperidotite and surrounding crustal rocks shared the prograde path in asubduction zone. Grt bearing mantle mineral assemblages show thatpeak UHP metamorphic conditions of 6.3 GPa and 870 °C occurred at429.5±3.1 Ma very shortly after the onset of the continent–continentcollision between Baltica and Laurentia, not during a later Scandianstage. The entire cycle of UHP metamorphism occurs over a timeperiod of c. 30 Ma rather than during a short, punctuated event duringthe evolution of the collision zone as generally believed. Majorelement mineral chemical P–T information from crustal gneiss andexternal eclogite largely underestimate peak metamorphic conditionsin micro-Dia bearing collision zones. The correlation of isotope andmetamorphic record between orogenic peridotite and most crustalHP/UHP metamorphic rocks suggests either formation or re-equilibra-tion of crustal gneiss and external eclogite during exhumation ratherthan re-crystallization during subduction, as is generally assumed.

A cold, deep, prolonged burial of a crustal terrain is consistent withdecelerated heat transfer in a maturing planet that has enabled coldgeodynamic regimes to survive and preserve ‘Dia-facies’ UHPmetamorphic signatures in subducted continental margins since atleast the Neoproterozoic, possibly already since the Palaeoproterozoic(Cartigny et al., 2004).

This conclusion implies that the evolution of continental collisionzones changed in response to planetary cooling (Brown, 2006).

Acknowledgements

This project was initiated during the NWO-PIONEER project of M.R.Drury and subsequently funded by the Utrecht University Institute ofGeodynamic Research (GOI). Additional support provided NSF grantEAR-0000937 and grants from the Research Foundation of the CityUniversity of New York. The manuscript is an extended version of partof the senior author's Ph.D. study (http://igitur-archive.library.uu.nl/dissertations/2006-1114-200554/index.htm) and benefited fromfinancial support from the Japan Society for the Promotion of Science.D.S. acknowledges K.R. Ludwig for providing the program Isoplot v. 3.0used for the age calculations, H. Sato for additional EMP analyses andseveral reviewers notably B. Bingen for stimulating critical commentsand editorial advises.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.epsl.2009.02.001.

References

Austrheim, H.,1987. Eclogitisation of lower crustal granulites by fluidmigration throughshear zones. Earth Planet. Sci. Lett. 81, 221–232.

Beyer, E.E., Brueckner, H.K., Griffin, W.L., O'Reilly, S.Y., Graham, S., 2004. Archean mantlefragments in Proterozoic crust, Western Gneiss Region, Norway. Geology 32,609–612.

Brey, G.P., Bulatov, V.K., Girnis, A.V., 2008. Geobarometry for peridotites: experiments insimple and natural systems from 6 to 10 GPa. J. Petrol. 49, 3–24.

Brey, G.P., Köhler, T., 1990. Geothermobarometry in four-phase lherzolites II. Newthermobarometers, and practical assessment of existing thermobarometers.J. Petrol. 31, 1353–1378.

Brown, M., 2006. Duality of thermal regimes is the distinctive characteristic of platetectonics since the Neoarchean. Geology 34, 961–964.

Brueckner, H.K., 1974. ‘Mantle’ Rb/Sr and 87Sr/86Sr ratios for clinopyroxenes from theNorwegian garnet peridotites and pyroxenites. Earth Planet. Sci. Lett. 24, 26–32.

Brueckner, H.K., 1998. Sinking intrusion model for the emplacement of garnet-bearingperidotites into continent collision orogens. Geology 26, 631–634.

Brueckner, H.K., Carswell, D.A., Griffin, W.L., 2002. Paleozoic diamonds within aPrecambrian peridotite lens in UHP gneisses of the Norwegian Caledonides. EarthPlanet. Sci. Lett. 203, 805–816.

Brueckner, H.K., Medaris, L.G., 2000. A general model for the intrusion and evolution of‘mantle’ garnet peridotites in high-pressure and ultra-high-pressure terranes.J. Metamorph. Geol. 18, 123–133.

Brueckner, H.K., Van Roermund, H.L.M., 2004. Dunk tectonics: a multiple subduction/eduction model for the evolution of the Scandinavian Caledonides. Tectonics 23,1–20.

Camacho, A., Lee, J.K.W., Hensen, B.J., Braun, J., 2005. Short-lived orogenic cycles and theeclogitization of cold crust by spasmodic hot fluids. Nature 435, 1191–1196.

Carswell, D.A., 1968. Possible primary upper mantle peridotite in Norwegian basalgneiss. Lithos 1, 322–355.

Carswell, D.A., 1973. Garnet pyroxenite lens within Ugelvik layered garnet peridotite.Earth Planet. Sci. Lett. 20, 347–352.

Carswell, D.A., Brueckner, H.K., Cuthbert, S.J., Mehta, K., O'Brien, P.J., 2003. The timing ofstabilisation and the exhumation rate for ultra-high pressure rocks in the WesternGneiss Region of Norway. J. Metamorph. Geol. 21, 601–612.

Carswell, D.A., Krogh, E.J., Griffin, W.L., 1985. Norwegian orthopyroxene eclogites:calculated equilibration conditions and petrogenetic implications. In: Gee, D.G.,Sturt, B.A. (Eds.), The Caledonide Orogen — Scandinavia and Related Areas. JohnWiley & Sons, Chichester, pp. 823–841.

Carswell, D.A., Van Roermund, H.L.M., 2005. On multi-phase mineral inclusionsassociated with microdiamond formation in mantle-derived peridotite lens atBardane on Fjørtoft, west Norway. Eur. J. Mineral. 17, 31–42.

Carswell, D.A., Van Roermund, H.L.M., Wiggers de Vries, D.F., 2006. Scandian ultrahigh-pressure metamorphism of Proterozoic Basement rocks on Fjørtoft and Otrøy,Western Gneiss Region, Norway. Int. Geol. Rev. 48, 957–977.

Cartigny, P., Chinn, I., Viljoen, K.S., Robinson, D., 2004. Early Proterozoic ultrahighpressure metamorphism: evidence from microdiamonds. Science 304, 853–855.

Chopin, C., 1984. Coesite in pure pyrope in high grade blueschist of the western Alps.Contrib. Mineral. Petrol. 86, 107–118.

Conceição, R.V., Mallmann, G., Koester, E., Schilling, M., Bertotto, G.W., Rodriguez-Vargas, A., 2005. Andean subduction-related mantle xenoliths: isotopic evidence ofSr–Nd decoupling during metasomatism. Lithos 82, 273–287.

Dallmeyer, R.D., Johannson, L., Möller, C., 1992. Chronology of Caledonian high-pressuregranulite-facies metamorphism, uplift, and deformation within northern parts ofthe Western Gneiss Region, Norway. Geol. Soc. Amer. Bull. 104, 444–455.

De Astis, G., Peccerillo, A., Kempton, P.D., La Volpe, L., Wu, T.W., 2000. Transition fromcalc-alkaline to potassium-rich magmatism in subduction environments: geo-chemical and Sr, Nd, Pb isotopic constraints from the island of Vulcano (Aeolianarc). Contrib. Mineral. Petrol. 139, 684–703.

Dobrzhinetskaya, L.F., Eide, E.A., Larsen, R.B., Sturt, B.A., Trønnes, R.G., Taylor, W.R.,Posukhova, T.V., 1995. Microdiamonds in high-grade metamorphic rocks from theWestern Gneiss Region, Norway. Geology 23, 597–600.

Dunning, G.D., Pedersen, R., 1988. U/Pb ages of ophiolites and arc-related plutons of theNorwegian Caledonides: implications for the development of Iapetus. Contrib.Mineral. Petrol. 98, 13–23.

Eskola, P., 1915. On the relations between the chemical and mineralogical compositionin themetamorphic rocks of the Orijärvi region. Bull. Comm. Géol. Finl. 44,109–145.

Gasparik, T., 2000. An internally consistent thermodynamic model for the system CaO–MgO–Al2O3–SiO2 derived primarily fromphase equilibriumdata. J. Geol.108,103–119.

Glodny, J., Kühn, A., Austrheim, H., 2008. Geochronology of fluid-induced eclogite andamphibolite facies metamorphic reactions in a subduction–collision system, BergenArcs, Norway. Contrib. Mineral. Petrol. 156, 27–48.

Griffin, W.L., Austrheim, H., Brastad, K., Bryhni, I., Krill, A.G., Krogh, E.J., Mørk, M.B.E.,Qvale, H., Tørudbakken, B., 1985. High-pressuremetamorphism in the ScandinavianCaledonides. In: Gee, D.G., Sturt, B.A. (Eds.), The Caledonide Orogen — Scandinaviaand Related Areas. John Wiley & Sons, Chichester, pp. 783–801.

Griffin, W.L., Brueckner, H.K., 1985. REE, Rb–Sr and Sm–Nd studies of Norwegianeclogites. Chem. Geol. 52, 249–271.

Harley, S.L., 1984. An experimental study of the partitioning of Fe and Mg betweengarnet and orthopyroxene. Contrib. Mineral. Petrol. 86, 359–373.

Hawthorne, F.C., Oberti, R., Della Ventura, G., Mottana, A. (Eds.), 2007. Amphiboles:Crystal Chemistry, Occurrence, and Health Issues. Reviews in Mineralogy andGeochemistry, vol. 67. Mineralogical Society of America, Chantilly, Virginia.

Hirajima, T., Nakamura, D., 2003. The Dabie Shan–Sulu orogen. In: Carswell, D.A.,Compagnioni, R. (Eds.), Ultrahigh Pressure Metamorphism. Vol. 5 of EMU Notes inMineralogy. Eötvös University Press, Budapest, pp. 105–144.

Jamtveit, B., Carswell, D.A., Mearns, E.W., 1991. Chronology of the high-pressuremetamorphism of Norwegian garnet peridotites/pyroxenites. J. Metamorph. Geol.9, 125–139.

Kennedy, C.S., Kennedy, G.C., 1976. The equilibrium boundary between graphite anddiamond. J. Geophys. Res. 81, 2467–2470.

Konzett, J., Ulmer, P., 1999. The stability of hydrous potassic phases in lherzoliticmantle—anexperimental study to9.5GPa insimplifiedandnatural bulk compositions.J. Petrol. 40 (4), 629–652.

Kretz, R., 1983. Symbols for rock-forming minerals. Am. Mineral. 68, 277–279.Krogh, E.J., 1977. Evidence of Precambrian continent–continent collision in Western

Norway. Nature 267, 17–19.Kukkonen, I.T., Peltonen, P., 1999. Xenolith-controlled geotherm for the central

Fennoscandian Shield: implications for lithosphere–asthenosphere relations.Tectonophysics 304, 301–315.

Kylander-Clark, A.R.C., Hacker, B.R., Johnson, C.M., Beard, B.L., Mahlen, N.J., Lapen, T.J.,2007. Coupled Lu–Hf and Sm–Nd geochronology constrains prograde andexhumation histories of high- and ultrahigh-pressure eclogites from westernNorway. Chem. Geol. 242, 137–154.

Kylander-Clark, A.R.C., Hacker, B.R., Mattinson, J.M., 2008. Slow exhumation of UHPterranes: titanite and rutile ages of the Western Gneiss Region. Earth Planet. Sci.Lett. 272, 531–540.

Labrousse, L., Jolivet, L., Andersen, T.B., Agard, P., Hébert, R., Maluski, H., Schärer, U.,2004. Pressure–temperature–time deformation history of the exhumation of ultra-

Page 9: Long-lived, cold burial of Baltica to 200 km depth

35D. Spengler et al. / Earth and Planetary Science Letters 281 (2009) 27–35

high pressure rocks in the Western Gneiss Region, Norway. In: Whitney, D.L.,Teyssier, C., Siddoway, C.S. (Eds.), Gneiss Domes in Orogeny. Special Paper 380.Geological Society of America, Boulder, Colorado, pp. 155–183.

Liou, J.G., Zhang, R.Y., Ernst, W.G., 2007. Very high-pressure orogenic garnet peridotites.Proc. Natl. Acad. Sci. 104, 9116–9121.

Mattinson, C.G., Wooden, J.L., Liou, J.G., Bird, D.K., Wu, C.L., 2006. Age and duration ofeclogite-facies metamorphism, North Qaidam HP/UHP terrane, western China. Am.J. Sci. 306, 683–711.

Nakamura, H., Iwamori, H., Kimura, J.-I., 2008. Geochemical evidence for enhanced fluidflux due to overlapping subducting plates. Nature Geosci. 1, 380–384.

Nilsen, O., Corfu, F., Roberts, D., 2007. Silurian gabbro–diorite–trondhjemite plutons inthe Trondheim Nappe Complex, Caledonides, Norway: petrology and U–Pbgeochronology. Norw. J. Geol. 87, 329–342.

Notsu, K., Lopez-Escobar, L., Onuma, N., 1987. Along-arc variation of Sr-isotopecomposition in volcanic rocks from the Southern Andes (33° S–55° S). Geochem.J. 21, 307–313.

Ranalli, G., Pellegrini, R., D'Offizi, S., 2000. Time dependence of negative buoyancy andthe subduction of continental lithosphere. J. Geodyn. 30, 539–555.

Roberts, D., 2003. The Scandinavian Caledonides: event chronology, palaeogeographicsettings and likely modern analogues. Tectonophysics 365, 283–299.

Roberts, D., Gee, D.G., 1985. An introduction to the structure of the ScandinavianCaledonides. In: Gee, D.G., Sturt, B.A. (Eds.), The Caledonide Orogen — Scandinaviaand Related Areas. John Wiley & Sons, Chichester, pp. 55–68.

Root, D.B., Hacker, B.R., Mattinson, J.M., Wooden, J.L., 2004. Zircon geochronology andca. 400 Ma exhumation of Norwegian ultrahigh-pressure rocks: an ion microprobeand chemical abrasion study. Earth Planet. Sci. Lett. 228, 325–341.

Scambelluri, M., Pettke, T., Van Roermund, H.L.M., 2008. Majoritic garnets monitor deepsubduction fluid flow and mantle dynamics. Geology 36, 59–62.

Schmidberger, S.S., Francis, D., 2001. Constraints on the trace element composition ofthe Archean mantle root beneath Somerset Island, Arctic Canada. J. Petrol. 42,1095–1117.

Smith, D.C., 1984. Coesite in clinopyroxene in the Caledonides and implications forgeodynamics. Nature 310, 641–644.

Sobolev, N.V., Shatsky, V.S., 1990. Diamond inclusions in garnets from metamorphicrocks: a new environment for diamond formation. Nature 343, 742–745.

Spengler, D., Van Roermund, H.M.L., Drury, M.R., Ottolini, L., Mason, P.R.D., Davies, G.R.,2006. Deep origin and hot melting of an Archaean orogenic peridotite massif inNorway. Nature 440, 913–917.

Terry, M.P., Robinson, P., Hamilton, M.A., Jercinovic, M.J., 2000a. Monazite geochronol-ogy of UHP and HP metamorphism, deformation and exhumation, Nordøyane,Western Gneiss Region, Norway. Am. Mineral. 85, 1651–1664.

Terry, M.P., Robinson, P., Krogh Ravna, E.J., 2000b. Kyanite eclogite thermobarometryand evidence for thrusting of UHP over HP metamorphic rocks, Nordøyane,Western Gneiss Region, Norway. Am. Mineral. 85, 1637–1650.

Toissaint, G., Burov, E., Jolivet, L., 2004. Continental plate collision: unstable vs. stableslab dynamics. Geology 32, 33–36.

Torsvik, T.H., Smethurst, M.A., Meert, J.G., Van der Voo, R., McKerrow,W.S., Brasier, M.D.,Sturt, B.A., Walderhaug, H.J., 1996. Continental break-up and collision in theNeoproterozoic and Palaeozoic — a tale of Baltica and Laurentia. Earth-Sci. Rev. 40,229–258.

Tucker, R.D., Robinson, P., Solli, A., Gee, D.G., Thorsnes, T., Krogh, T.E., Nordgulen, O.,Bickford, M.E., 2004. Thrusting and extension in the Scandian hinterland, Norway:new U–Pb ages and tectonostratigraphic evidence. Am. J. Sci. 304, 477–532.

Van Orman, J.R., Grove, T.L., Shimizu, N., Layne, G.D., 2002. Rare earth element diffusionin a natural pyrope single crystal at 2.8 GPa. Contrib. Mineral. Petrol. 142, 416–424.

Van Roermund, H.L.M., Carswell, D.A., Drury, M.R., Heijboer, T.C., 2002. Microdiamondsin a megacrystic garnet websterite pod from Bardane on the island of Fjørtoft,western Norway: evidence for diamond formation in mantle rocks during deepcontinental subduction. Geology 30, 959–962.

Van Roermund, H.L.M., Drury, M.R., 1998. Ultra-high pressure (P N 6 GPa) garnetperidotites in western Norway: exhumation of mantle rocks from more than185 km. Terra Nova 10, 295–301.

Van Roermund, H.L.M., Drury, M.R., Barnhoorn, A., de Ronde, A., 2000. Super-silicicgarnet microstructures from an orogenic garnet peridotite, evidence for an ultra-deep origin. J. Metamorph. Geol. 18, 135–147.

Vrijmoed, J.C., Van Roermund, H.L.M., Davies, G.R., 2006. Evidence for diamond-gradeultra-high pressure metamorphism and fluid interaction in the Svartberget Fe–Tigarnet peridotite–websterite body, Western Gneiss Region, Norway. Mineral.Petrol. 88, 381–405.

Warren, C.J., Beaumont, C., Jamieson, R.A., 2008. Modelling tectonic styles and ultra-high pressure (UHP) rock exhumation during the transition from oceanicsubduction to continental collision. Earth Planet. Sci. Lett. 267, 129–145.

Weiss, M., 1997. Clinohumites: a field and experimental study. Dissertation ETH Nr.12202, Geologisches Institut der ETH, Zürich.

Wennberg, O.P., 1996. Superimposed fabrics due to reversal of shear sense: an examplefrom the Bergen Arc Shear Zone, western Norway. J. Struct. Geol. 18 (7), 871–889.

Young, D.J., Hacker, B.R., Andersen, T.B., Corfu, F., 2007. Prograde amphibolite facies toultrahigh-pressure transition along Nordfjord, western Norway: implications forexhumation tectonics. Tectonics 26 (TC1007), 1–15.