American Mineralogist, Volume 85, pages 1637–1650, 2000 0003-004X/00/1112–1637$05.00 1637 INTRODUCTION The Western Gneiss Region, Norway, lies 100–300 km northwest of the exposed thrust front of the Scandinavian Caledonides (Fig. 1) and has been identified as a giant tectonic window exposing large areas of Baltica Proterozoic basement beneath a sequence of Scandian (late Silurian-early Devonian) thrust nappes (Roberts and Gee 1985; Lutro et al. 1997). In this region the basement of Baltica and its tectonic cover have been subducted beneath the overriding Laurentian plate. It is a classic area of exposure of crustal mafic rocks metamorphosed to eclogite (Eskola 1921; Krogh 1977; Griffin et al. 1985; Grif- fin 1986; Krogh and Carswell 1995), as well as garnet peridot- ites of subcontinental mantle origin (Carswell 1986; Brueckner 1998, 1999; Terry and Robinson 1999; Terry et al. 1999a). Detailed study has shown that the nappes surrounding the West- ern Gneiss Region also occur deeply infolded in narrow com- plex synclines within it, where the nappes have commonly undergone extreme tectonic thinning and complex ductile de- formation followed by development of mylonites, ultramylonites, and brittle faults associated with later phases of tectonic exhumation (Gee 1980; Krill 1985; Tucker 1986; Robinson 1995, 1997; Terry and Robinson 1996; Terry 2000). The discovery of coesite and coesite pseudomorphs in eclogites in the Nordfjord area by Smith (1984), confirmed and expanded by Wain (1997; see also Wain et al. 2000), estab- lished that part of the region had attained ultra-high-pressure (UHP) conditions of ~30 kbar, 800 °C at depths of ~100 km. Evidence of UHP conditions was expanded 120 km northeast from Nordfjord with the discovery of characteristic crustal microdiamonds in kyanite-garnet gneiss at Fjørtoft, Nordøyane (Dobrzhinetskaya et al. 1993, 1995; Larsen et al. 1998). Al- though it has been suggested that the Western Gneiss Region may be considered as a single coherent level of subducted Baltica basement, this suggestion is supported neither by struc- tural studies nor by petrologic evidence. Throughout the 50 km wide Trollheimen part of the Western Gneiss Region, Baltica basement occurs at two distinct imbricate levels separated by a nearly continuous layer of late Proterozoic quartzite uncon- formable on lower basement (Gee 1980; Krill 1980; Tveten and Lutro 1996; Robinson 1997). In the Nordfjord region, Wain (1997; see also Wain et al. 2000) described closely juxtaposed exposures of UHP (~30 kbar) and HP (~20 kbar) eclogites, strongly implying a major tectonic contact or contacts, and the results of the present study (Terry et al. 1999b) support this interpretation. Our field work in the region was begun by Robinson in 1990 (Robinson 1995), tracing infolds of nappes along the line of * E-mail: [email protected]Kyanite eclogite thermobarometry and evidence for thrusting of UHP over HP metamorphic rocks, Nordøyane, Western Gneiss Region, Norway MICHAEL P. TERRY, 1, * PETER ROBINSON, 1,2 AND ERLING J. KROGH RAVNA 3 1 Department of Geosciences, University of Massachusetts, Amherst, Massachusetts 01003, U.S.A. 2 Geological Survey of Norway, N-7491 Trondheim, Norway 3 Department of Geology, University of Tromsø, N-9037 Tromsø, Norway ABSTRACT Quantitative estimates of metamorphic conditions are combined with previously reported struc- tural analysis to develop the thermotectonic evolution of two separate lithotectonic units metamor- phosed during the Late Silurian-Early Devonian collision between Baltica and Laurentia. A structurally higher plate, regionally correlated with the Blåhø Nappe, contains kyanite eclogites associated with microdiamond-bearing kyanite-garnet-graphite gneiss on the north coast of Fjørtoft and correlated with kyanite eclogites at Nogva, Flemsøya. The kyanite eclogites, containing the assemblage kyan- ite-garnet-omphacite-coesite (now polycrystalline quartz pseudomorphs) ± zoisite ± phengite, yield conditions of 820 °C and 34–39 kbar at Fjørtoft and of 820 °C and 30–36 kbar at Nogva, best characterized by two recently recalibrated geothermobarometers. The conditions at Fjørtoft overlap the diamond-graphite phase boundary and represent the first quantitative petrologic determination of UHP diamond-forming conditions in crustal rocks outside the Dabie Mountains, China and the Kokchetav Massif, Kazahkstan. In a structurally lower plate, eclogitized mylonite with the assem- blage kyanite-garnet-omphacite-quartz-oligoclase, produced from the mid-Proterozoic Haram Gab- bro that has intruded diorite country rock, yields 780 °C and 18 kbar. This result agrees with other normal HP estimates by Mørk (1985) from partially to completely eclogitized gabbro in the same unit on Flemsøya. We propose that the UHP plate reached a maximum depth of 125 km and then experienced 65 km of exhumation during top-southeast thrusting that brought it into contact with the HP plate. Following this, both plates were exhumed together until reaching a depth of 37 km where they experienced extensive amphibolite-facies re-equilibration and top-west or left-lateral shearing. Temporal details of these histories were determined by monazite U-Th-Pb geochronology described in a companion paper.
Transcript
American Mineralogist, Volume 85, pages 1637–1650, 2000
0003-004X/00/1112–1637$05.00 1637
INTRODUCTION
The Western Gneiss Region, Norway, lies 100–300 kmnorthwest of the exposed thrust front of the ScandinavianCaledonides (Fig. 1) and has been identified as a giant tectonicwindow exposing large areas of Baltica Proterozoic basementbeneath a sequence of Scandian (late Silurian-early Devonian)thrust nappes (Roberts and Gee 1985; Lutro et al. 1997). Inthis region the basement of Baltica and its tectonic cover havebeen subducted beneath the overriding Laurentian plate. It is aclassic area of exposure of crustal mafic rocks metamorphosedto eclogite (Eskola 1921; Krogh 1977; Griffin et al. 1985; Grif-fin 1986; Krogh and Carswell 1995), as well as garnet peridot-ites of subcontinental mantle origin (Carswell 1986; Brueckner1998, 1999; Terry and Robinson 1999; Terry et al. 1999a).Detailed study has shown that the nappes surrounding the West-ern Gneiss Region also occur deeply infolded in narrow com-plex synclines within it, where the nappes have commonlyundergone extreme tectonic thinning and complex ductile de-formation followed by development of mylonites,ultramylonites, and brittle faults associated with later phasesof tectonic exhumation (Gee 1980; Krill 1985; Tucker 1986;Robinson 1995, 1997; Terry and Robinson 1996; Terry 2000).
The discovery of coesite and coesite pseudomorphs ineclogites in the Nordfjord area by Smith (1984), confirmed andexpanded by Wain (1997; see also Wain et al. 2000), estab-lished that part of the region had attained ultra-high-pressure(UHP) conditions of ~30 kbar, 800 °C at depths of ~100 km.Evidence of UHP conditions was expanded 120 km northeastfrom Nordfjord with the discovery of characteristic crustalmicrodiamonds in kyanite-garnet gneiss at Fjørtoft, Nordøyane(Dobrzhinetskaya et al. 1993, 1995; Larsen et al. 1998). Al-though it has been suggested that the Western Gneiss Regionmay be considered as a single coherent level of subductedBaltica basement, this suggestion is supported neither by struc-tural studies nor by petrologic evidence. Throughout the 50km wide Trollheimen part of the Western Gneiss Region, Balticabasement occurs at two distinct imbricate levels separated by anearly continuous layer of late Proterozoic quartzite uncon-formable on lower basement (Gee 1980; Krill 1980; Tvetenand Lutro 1996; Robinson 1997). In the Nordfjord region, Wain(1997; see also Wain et al. 2000) described closely juxtaposedexposures of UHP (~30 kbar) and HP (~20 kbar) eclogites,strongly implying a major tectonic contact or contacts, and theresults of the present study (Terry et al. 1999b) support thisinterpretation.
Our field work in the region was begun by Robinson in 1990(Robinson 1995), tracing infolds of nappes along the line of* E-mail: [email protected]
Kyanite eclogite thermobarometry and evidence for thrusting of UHP over HPmetamorphic rocks, Nordøyane, Western Gneiss Region, Norway
MICHAEL P. TERRY,1,* PETER ROBINSON,1,2 AND ERLING J. KROGH RAVNA3
1Department of Geosciences, University of Massachusetts, Amherst, Massachusetts 01003, U.S.A.2Geological Survey of Norway, N-7491 Trondheim, Norway
3Department of Geology, University of Tromsø, N-9037 Tromsø, Norway
ABSTRACT
Quantitative estimates of metamorphic conditions are combined with previously reported struc-tural analysis to develop the thermotectonic evolution of two separate lithotectonic units metamor-phosed during the Late Silurian-Early Devonian collision between Baltica and Laurentia. A structurallyhigher plate, regionally correlated with the Blåhø Nappe, contains kyanite eclogites associated withmicrodiamond-bearing kyanite-garnet-graphite gneiss on the north coast of Fjørtoft and correlatedwith kyanite eclogites at Nogva, Flemsøya. The kyanite eclogites, containing the assemblage kyan-ite-garnet-omphacite-coesite (now polycrystalline quartz pseudomorphs) ± zoisite ± phengite, yieldconditions of 820 °C and 34–39 kbar at Fjørtoft and of 820 °C and 30–36 kbar at Nogva, bestcharacterized by two recently recalibrated geothermobarometers. The conditions at Fjørtoft overlapthe diamond-graphite phase boundary and represent the first quantitative petrologic determinationof UHP diamond-forming conditions in crustal rocks outside the Dabie Mountains, China and theKokchetav Massif, Kazahkstan. In a structurally lower plate, eclogitized mylonite with the assem-blage kyanite-garnet-omphacite-quartz-oligoclase, produced from the mid-Proterozoic Haram Gab-bro that has intruded diorite country rock, yields 780 °C and 18 kbar. This result agrees with othernormal HP estimates by Mørk (1985) from partially to completely eclogitized gabbro in the sameunit on Flemsøya. We propose that the UHP plate reached a maximum depth of 125 km and thenexperienced 65 km of exhumation during top-southeast thrusting that brought it into contact with theHP plate. Following this, both plates were exhumed together until reaching a depth of 37 km wherethey experienced extensive amphibolite-facies re-equilibration and top-west or left-lateral shearing.Temporal details of these histories were determined by monazite U-Th-Pb geochronology describedin a companion paper.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1638
Moldefjord, and then, in 1992, expanded northward intoNordøyane, where intensive mapping was carried out by Terryin 1995–1997 (Terry 2000). A critical aspect of our work wasan attempt to find linkages between petrologic anddeformational features and, if possible, linkages among meta-morphism, deformation, and geochronology. This paper pre-sents geothermobarometric information focused on the earliestand deepest stages of regional metamorphism from the UHPplate and adjacent HP plate. Geochronologic information fromthe UHP plate that bears on the timing of highest-P metamor-phism and subsequent exhumation is given in a companionpaper (Terry et al. 2000).
LOCAL GEOLOGIC SETTING
Detailed geologic mapping in the Moldefjord region (Fig.1) has demonstrated the presence of four regionally extensivethrust nappes overlying Baltica basement. The basement isdominated by 1680–1650 Ma granitoid gneisses cut by ~1500Ma rapakivi granite (now augen gneiss) and by a variety of1450–950 Ma gabbros, now variably eclogitized. The nappes,from bottom to top, include: (1) Risberget, (2) Sætra, (3) Blåhø,
and (4) Støren Nappes, which correlate with Tånnås AugenGneiss, Sarv, Seve, and Köli Nappes of the SwedishCaledonides (Roberts and Gee 1985). The Støren Nappe andequivalents in the Trondheim region contain early Ordovicianfossils of Laurentian affinity and are interpreted as an oceanicand island-arc assemblage produced within Iapetus or on itsLaurentian margin, later thrust over Baltica during the LateSilurian. The nappes are exposed because they have beendownfolded into narrow synclines.
Nordøyane (literally “the North Islands”) lie north of theMoldefjord syncline (Fig. 2). They are dominated by highlydeformed Baltica basement with many narrow infolds of micaschist and amphibolite of the Blåhø Nappe, locally underlainby quartzites and amphibolites of the Sætra Nappe. Nordøyanehave been divided into three major segments, each involvingbasement and rocks of the Blåhø Nappe. In the southern seg-ment, Baltica basement contains eclogite and it is locally foundalso in the Blåhø Nappe. In the central segment, there is noevidence that any of the rocks were ever at eclogite facies con-ditions. On this basis we think the central/ southern boundaryis an early fault of large displacement, although there is no
FIGURE 1. (a) Generalized geologic map showing narrow refolded synclines of tectonic cover in Baltica crust and location of the study area.(b) Generalized tectonostratigraphic map modified from Gee et al. (1985). Large arrow shows the orientation of a time averaged (450–425 Ma)relative motion vector for Baltica with respect to Laurentia approximated from reconstructions of Torsvik (1998). Small arrows show movementvectors of allochthons.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES 1639
FIGURE 2. (a) Generalized geologic map of Nordøyane showing the major lithotectonic units. Sample localities UHP1, 929, 1066, 1294, and419, of greatest concern to this paper, are indicated. Lines A, B, and C show the location for the composite cross-section. (b) Composite sectionthrough the northern unit showing the approximate structural positions of sample locations. Heavy lines are topographic profiles along sectionlines A, B, and C.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1640
outcrop evidence for faulting. The northern segment is domi-nated by basement dioritic rocks and subordinate gabbro, withrelatively minor metamorphosed rapakivi granite, here calledthe Lower Plate, and a single overlying folded unit of schistsand mafic rocks of the Blåhø Nappe, here called the UpperPlate. All parts of the northern segment reached eclogite fa-cies, though of different intensity in the Lower and Upper Plates.Possibly the basement of the Lower Plate is actually two tec-tonic slabs juxtaposed along an early thrust fault decorated byboudins of garnet peridotite. The boundary between the cen-tral/southern and northern segments is a 50 to 200 m thick zoneof low- to middle-amphibolite-facies mylonite, the ÅkreMylonite, dominated by the subhorizontal sinistral shear fab-ric characteristic of late phases of deformation in the region.Figure 2b illustrates some of the structural relationships in thenorthern segment, dominated by an anticlinorium.
DEFORMATION HISTORY
Structural features formed during late-orogenic extensiondominate the map pattern of Nordøyane. The youngest are dis-tinct, steeply dipping mylonite and ultramylonite zones formedduring low amphibolite-facies conditions, showing consistentsinistral shear fabric (Robinson 1995; Robinson and Terry 1998)along a northeast-trending subhorizontal lineation. Beforemylonite formation, nearly all the rocks were subjected to arelated broad deformation with sinistral shear along steeplydipping surfaces and top-west shear along subhorizontal sur-faces. This deformation, which took place during upperamphibolite-facies conditions, produced the dominant east-northeast trending folds and mineral lineations, including sheathand tubular folds subparallel to the top-west or -southwest trans-port direction. Both this deformation and the subsequentmylonite formation are considered integral parts of a majorphase of upper crustal extension progressing into a sinistrallyshearing plate boundary region during late phases of continen-tal collision (Robinson and Terry 1998; Terry and Robinson1998; Terry 2000) and synchronous with deposition of Devo-nian clastic basins.
In a few places in the northern segment, in both Upper andLower Plates, an earlier northwest-trending linear shear fabricwith associated folds has been preserved, either reoriented, or(rarely) in its original orientation. After taking later deforma-tion into account, this fabric indicates top-southeast shear con-sistent with northwestward subduction of Baltica beneathLaurentia in early phases of continental collision (Robinsonand Terry 1998; Terry and Robinson 1998; Torsvik 1998). Pet-rologic studies in eclogitized gabbros (see below) indicate thatthis deformation took place under progressive eclogite-faciesmetamorphism.
Our understanding of the deformation history is summa-rized as follows. Progressive southeastward thrusting and im-brication took place as material from the Baltica margin wasprogressively transferred across the main subduction surfacefrom the underlying Baltic plate to the overlying Laurentia plate.Once brought to higher levels by thrusting, the rocks came intoa zone of gravitational extension, perhaps initially with north-west-directed transport, but rapidly turning southwestward bybroad sinistral shear along the continental collision zone. In
this way a very broad field of transtensional structural featureswas produced in the upper parts of a plate boundary region ofessentially transpressional nature, while foreland-directed con-vergence was still being expressed by continued imbricationwithin deeper and more forward locations on the Baltica mar-gin.
PETROGRAPHY AND MINERAL CHEMISTRY
Analytical methods
Mineral compositions were determined using the CamecaSX-50 electron microprobe at the University of Massachusetts,Amherst, with an operating condition of 15 kv and a 1–2 mi-crometers beam size at a beam current 15 nA with count timesof 20s. Matrix corrections were done using the PAP method(Pouchou and Pichoir 1984a, 1984b). Images showing compo-sitions of desired elements were collected using sample cur-rent (>80 nA) and rastering the stage with the electron beamfixed. Results used in thermobarometry calculations and re-ferred to below are summarized in Table 1.
Sample 419
The Haram Gabbro (Fig. 2) is a body of mid-Proterozoicgabbro with an intrusion age of 926 ± 70 Ma based on a Sm-Nd igneous mineral isochron (Mørk and Mearns 1986). Thegabbro has an exposed length of 1.3 km and a width of 0.3 kmwithin the Lower Plate. Terry (see Lutro et al. 1997; Terry andRobinson 1998) has subdivided the gabbro into coarse-grainedmafic gabbro, anorthositic gabbro, and cumulate-layered gab-bro, and there are also abundant gabbro pegmatites with pla-gioclase, pyroxene, and oxide grains 10–20 cm long. Primaryigneous minerals including plagioclase with oxide dust, aug-ite, orthopyroxene, quartz, ilmenite, and magnetite are widelypreserved, and tiny coronas of garnet are commonly the onlysign that the rocks have been through regional metamorphism.The interior of the gabbro as well as the north and south con-tacts are cut by fine-grained mylonite zones in which primaryigneous minerals are in various states of grain-size reductionand recrystallization. In these mylonite zones, there is increas-ing jadeite content of augite, partial to complete destruction oforthopyroxene, partial transformation of ilmenite to rutile, andextensive growth of fine-grained garnet particularly at the ex-pense of plagioclase. These mylonite zones characteristicallycontain a steeply plunging rodded shear fabric with north-side-up shear sense. After small circle rotation back to horizontal,the shear fabric suggests a top-southeast shear sense that weequate with subduction of Baltica beneath Laurentia undereclogite-facies conditions. Locally, both gabbro and mylonitezones are cut by younger granite pegmatite dikes along whichboth rock types are hydrated to amphibolite. Similaramphibolitization has taken place in parts of the north and southmylonite zones closest to country rock contacts. In rare loca-tions, the reaction toward eclogite has proceeded without sig-nificant shearing to the point where all primary igneous mineralsare gone, although the original gabbro texture is still preserved.
Strictly speaking, sample 419 is a HP granulite because itcontains some metamorphic plagioclase. This sample was col-lected from the eclogitized mylonite zone on the northern mar-
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES 1641
gin of the Haram Gabbro, and its mineralogy suggests it has abulk composition similar to the kyanite eclogites of the UpperPlate, but recrystallized at much lower pressure. The phase as-semblage includes garnet, omphacite, plagioclase, biotite, kyan-ite, quartz, ilmenite, and rutile. This sample has a well-developedtexture in which small euhedral garnets have formed coronasbetween extremely elongate rods of pyroxene and ilmenite andthe surrounding plagioclase. The coronas are interpreted to have
Al total 1.243 Total 99.01 100.00 100.25 100.15 Total 95.36 Total 100.66 100.01 99.183 Fe2+ 0.001
Ca 0.233Si 1.906 1.938 1.923 1.938 Si 3.493 Si 2.978 2.954 2.921 Na 0.766IVAl 0.094 0.062 0.077 0.062 IVAl 0.507 Al total 2.003 2.032 2.015 K 0.024Fe3+ 0.000 0.000 0.000 0.000 Ti 0.005 0.003 0.003
Al total 2.110 Cr3+ 0.005 0.000 0.004 Total 5.020Tet 2.000 2.000 2.000 2.000 Mg 0.682 0.992 1.510
VIAl 1.603 Fe2+ 1.375 1.116 0.620 XAn 0.227Al total 0.275 0.480 0.293 0.279 Ti 0.013 Mn 0.039 0.017 0.018 XAb 0.749
Cr3+ 0.000 Ca 0.926 0.913 0.974 XOr 0.024VIAl 0.180 0.418 0.216 0.218 Mg 0.328 Na 0.000 0.000 0.004Ti 0.008 0.003 0.004 0.003 Fe2+ 0.058 K 0.000 0.000 0.000Cr3+ 0.000 0.001 0.003 0.004 Mn 0.000Fe3+ 0.037 0.036 0.013 0.000 Ca 0.001 Total 8.013 8.027 8.069Mg 0.688 0.492 0.747 0.752 Na 0.076Fe2+ 0.087 0.050 0.017 0.024 K 0.796 XPy 0.226 0.327 0.484
FM 0.227 0.177 0.062 0.062Notes: n.d. = Not detected. n.a. = not analyzed. Structural formulae of clinopyroxene are presented twice, in terms of four cations and six oxygenatoms with inferred Fe3+ and with all FeO and six oxygen atoms. For one analysis these results are identical. Phengite analysis is presented in termsof cations per 11 oxygen atoms with all Fe as FeO, garnet analyses as cations per eight oxygen atoms with all Fe as FeO, and plagioclase as cationsper eight oxygen atoms. FM = Fe2+/(Fe2++Mg).
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1642
clasts of primary augite or orthopyroxene that are partially re-placed by fine-grained (100 µm) omphacite (Figs. 4a and 4b).These partially replaced porphyryoclasts are surrounded bygarnet that has an intervening moat of plagioclase. The garnet(250 µm), typically with composition Pyr23 Alm46 Gro30 Spe1,is euhedral and contains abundant inclusions of plagioclase(An23), omphacite [Jd14, Fe/(Fe + Mg) = 0.112], kyanite, andrare quartz. Plagioclase and kyanite (30 µm each), both show-ing a strong preferred grain-shape orientation, dominate thematrix surrounding the coronas. Plagioclase grains are zonedfrom An45 on rims to An23 in the cores, a composition similar tothe compositions of plagioclase inclusions in garnet. The cal-cic rims are considered to be a result of retrograde reactionamong garnet rims, kyanite, and quartz. Biotite is coarser-grained (500 µm) and randomly oriented with respect to thestructural fabric indicating that its growth postdates deforma-tion and may reflect retrograde re-equilibration. Rutile is re-placed locally by ilmenite that is also interpreted as a retrogradeproduct.
Sample 1066b
Locality 1066 is an exposed boudin on the north coast ofFjørtoft (Fig. 2). The exposure is a 2 m high, upward-protrud-ing relict in the center of an E-W trending basin-like structurein granitoid gneisses. The preserved eclogite and amphiboliterim extends about 5 m E-W and 8 m N-S. A “tail” of amphibo-lite embedded in gneiss extends about 3 m from the NW cornerand 4 m from the SE corner of the main mass. A coarse grani-toid pegmatite sill dips beneath the eclogite in the northeastcorner of the basin. There is another boudin consisting of weaklyto moderately foliated and lineated metamorphosed gabbro withsome garnet coronas 10 m north of the north margin of the
eclogite 1066. This rock shows no evidence of having beenthrough the same eclogite-facies conditions as the boudin 1066,and we conclude that the tectonic boundary between the UpperPlate and the Lower Plate lies in this outcrop at the base of thekyanite eclogite.
The dominant rock type in the center of the boudin, occu-pying an area 5 m E-W and 4–5 m N-S, consists of slightlyretrograded kyanite-zoisite-biotite eclogite with irregular gar-net porphyroblasts about 1 cm in diameter. Rare vertical gar-net-rich layers trend N-S, and there are numerous veins of coarsezoisite with or without kyanite In this eclogite, plagioclase-diopside symplectite has partially replaced omphacite, a veryfine sapphirine-plagioclase symplectite has partially replacedkyanite, and there are hornblende rims on garnet. In the centerof this larger area, there is a small ~1 m2 area of the freshestkyanite-zoisite-biotite eclogite from which sample 1066b wastaken. This sample contains relatively minor symplectite andvery minor hornblende. On the outer north and south marginsof the boudin there are successive layers of progressively al-tered eclogite: 0.3–1 m of kyanite-zoisite eclogite with horn-blende partially replacing plagioclase-diopside symplectite andpartially replacing garnet; 0.3–1 m of amphibolite with horn-blende-plagioclase symplectite after omphacite, sapphirine ±corundum ± spinel-plagioclase symplectite around kyanite, andhornblende pseudomorphs after garnet; and an outermost layerof foliated amphibolite with hornblende, plagioclase,clinozoisite, green biotite, and phengite with white prisms ofrelict zoisite.
Minerals interpreted to have been present at ultra-high pres-sures in sample 1066b include garnet, omphacite, phengite,zoisite, kyanite, coesite, and rutile (Fig. 4c). Garnet, typicallyPyr48 Alm20 Gro31 Spe0.6, contains inclusions (Fig. 4d) ofomphacite, kyanite, phengite, zoisite, and polycrystalline quartzpseudomorphs after coesite (Figs. 4d and 4e). Omphacite in-clusions are homogeneous and have a composition of Jd16.5 withFe/(Fe+Mg) = 0.03. In contrast, matrix grains of omphacite(Jd18.5) contain abundant rods consisting of about half quartzand half amphibole (Fig. 4c). This suggests that these may haveoriginally been high-pressure vacancy pyroxenes with the Ca-eskola component (Gasparik and Lindsley 1980), but reinte-gration of these rods into a pyroxene formula has not yet beendone. Zoisite in the matrix is commonly poikiloblastic and sur-rounds anhedral omphacite. Both omphacite and zoisite showa preferred grain-shape orientation and define a north-north-west-trending lineation. Locally the included coesite pseudo-morphs are anhedral with rims of omphacite (Fig. 4d). A singlegrain of phengite (Si = 3.5, XFe = 0.15) is present as an inclu-sion in garnet, but none is present in the matrix. Biotite is presentin the matrix and as inclusions in garnet and is explained bythe reaction:
garnet + phengite = biotite + quartz. (2)
This fluid-conservative reaction is likely responsible for thepoor preservation of phengite in this sample and other rocksfrom HP and UHP terranes. Even phengite inclusions in garnetare unlikely to survive if the rock was subjected to strong de-formation that occurred at high T during exhumation. Phengite
FIGURE 3. ACF ternary plot showing the phases present in the high-pressure granulite assemblage from the Haram Gabbro sample 419.The overall reaction 1 involved in the conversion toward eclogite isillustrated by dashed tie lines.
Fe
Fe
Fe
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES 1643
FIGURE 4. Photomicrographs of kyanite-bearing metamorphosed gabbro and kyanite eclogites. (a) and (b) Mylonitized gabbro in plane-polarized light and in cross-polarized light showing parts of a corona in the upper half the photo. The lower half of the photo shows the matrixcomposed of plagioclase and kyanite that show a strong preferred grain-shape orientation. (c), (d), and (e) Kyanite eclogite 1066b from Fjørtoft.(c) Typical assemblage showing garnet with quartz and omphacite inclusions (top), omphacite with exsolved rods of quartz (center right) andhornblende, and zoisite (bottom). (d) Inclusions in garnet including polycrystalline quartz, omphacite, and kyanite. (e) Close-up view ofpolycrystalline quartz inclusion in garnet that is interpreted to be pseudomorphous after prismatic coesite. Note how quartz is “injected” intocracks in garnet at end of prism. (f) Kyanite surrounded by polycrystalline quartz inside garnet in garnet-rich eclogite in sample 1294a.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1644
inclusions can only be preserved if garnet acts as a perfect pres-sure vessel to prevent the pure volume expansion of reaction 2.
A striking feature of this sample compared with many othereclogites is the high pyrope and grossular contents of garnet,the low jadeite content and low Fe ratio of the omphacite, andthe abundances of quartz, zoisite, and kyanite. A most likelyprotolith for this eclogite was a magmatically primitive igne-ous cumulate rock dominated by Ca-plagioclase and Mg-richpyroxene. The abundant quartz, zoisite, and kyanite are a di-rect result of plagioclase breakdown.
A convenient portrayal of the phase assemblage in the sys-tem NCFMASH is achieved by combining Fe and Mg and pro-jecting phase compositions from Si (coesite) and Al (kyanite)onto the base Na-Ca-FeMg (Fig. 5a). In this portrayal, stableparts of the garnet solid-solution series run across the base,whereas the diopside-jadeite solid solution runs from the middleof the base to the top apex. In this assemblage with zoisite, thecoexisting garnet (Gro31) is the most calcic stable garnet, andthe pyroxene (Jd16.5), lies at the low-jadeite end of the pyrox-ene solid solution. In Figure 5b, a further projection is madefrom the Ca apex to allow portrayal of phases with variable Fe/Mg ratios. This shows that garnet in 1066b has a very low XFe,and the omphacite with low jadeite content has a still moremagnesian composition.
Sample 1294a
Locality 1294 is the southern of two kyanite eclogite boudinsin biotite gneiss on the point at Nogva (Figs. 2a and 2b), whichis interpreted as part of the Upper Plate. This boudin is ex-posed over an area of about 8 × 8 m and extends an unknowndistance into the sea. Both eclogite boudins are strongly lay-ered with light-green layers rich in omphacite, zoisite, and kya-nite, and orange-red layers very rich in garnet. In the light-greenlayers, there are several veins with prisms of milky zoisite upto 12 cm long. One layering surface in the northern boudinshows very strong N-S trending lineation of zoisite. The light-green layers tend to be retrograded, with abundant plagio-clase-diopside symplectite after omphacite, abundantsapphirine-plagioclase symplectite after kyanite, clinozoisite-plagioclase symplectite after zoisite, and secondary hornblende.The garnet-rich layers are much less retrograded. Although thegarnet-rich layers lack zoisite, they contain omphacite, kyan-ite, and quartz, all of which are particularly well preserved asinclusions in garnet. Sample 1294a came from such a garnet-rich layer at the extreme southwestern edge of the outcrop.
Phases interpreted to have been present in sample 1294a atUHP include garnet, omphacite, kyanite, coesite, and rutile.Polycrystalline quartz occurs as an apparent replacement ofcoesite and locally forms a rim on anhedral kyanite inclusionsin garnet (Fig. 4f). Garnet, typically Pyr33 Alm37 Gro30 Spe0.5
also contains inclusions of omphacite [Jd41, Fe/(Fe + Mg) =0.09] that form a crude lineation that is parallel to a lineationdefined by omphacite and kyanite in the matrix.
The bulk composition of this sample is more sodic and moreferroan than sample 1066b, though still from a plagioclase-rich protolith. In the kyanite projection of Figure 5a, whichignores Fe/Mg ratios, sample 1066 contains the Ca-rich three-phase assemblage zoisite-omphacite-garnet, whereas sample
FIGURE 5. Ternary plots of UHP assemblages from kyanite eclogitesat Fjørtoft and Nogva, highlighting major compositional differencesamong phases in these two rocks. (a) Shows phases in the systemNCFMASH projected from coesite and kyanite onto the plane Na-Ca-FeMg. Sample 1066b contains the projected three-phase assemblageomphacite-garnet-zoisite; Sample 1294a has only the two-phaseassemblage omphacite-garnet. (b) The same omphacite-garnet tie linesplotted in an expanded diagram on which a projection from Ca allowsseparate portrayal of Mg and Fe. In this view omphacite-garnet tielines are compatible with the possibility of equilibration under similarconditions. The tie lines also illustrate the different bulk compositionsof these two samples. (c) Phase relations with coesite and kyaniteprojected into the tetrahedron Ca-Na-Fe-Mg (see text).
Fe
Fe
FeFe
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES 1645
1294a contains the less-calcic, two-phase assemblageomphacite-garnet. True relationships for kyanite-saturatedsamples are better shown in projection Figure 5b and in thethree-dimensional sketch in Figure 5c. These show that a moreFe-rich garnet with similar grossular content to 1066b in 1294acoexists with a more-sodic pyroxene. Figure 5c shows a trian-gular “tunnel” of zoisite-pyroxene-garnet tie planes of which1066 is one. The upper edge of this “tunnel” provides a mini-mum jadeite limit for the pyroxene field. Its front edge pro-vides a maximum grossular limit for the garnet field, whichnecessarily increases in grossular with increasing Fe/Mg ratio.In front of the tunnel lies a large volume filled with two-phasetie lines between the pyroxene and garnet planes. Sample 1294apresents such a tie line.
Samples UHP1 and 929
These samples have been the focus of monazite geochro-nology studies and they are described in detail by Terry et al.(2000). Sample UHP1, the microdiamond-bearing sample, hasa retrograde amphibolite-facies matrix assemblage that has beenused to estimate conditions of re-equilibration at about 700 °C,11 kbar (Larsen et al. 1998). Sample 929 is a kyanite-garnetmylonite from the same protolith as UHP 1, and shows partialsillimanite replacement of kyanite porphyroclasts oriented in alate subhorizontal sinstral shear fabric. It thus records contin-ued exhumation from 11 kbar toward 6–2.5 kbar.
GEOTHERMOBAROMETRY
Three approaches were used to make quantitative estimatesof metamorphic conditions for kyanite eclogites and kyanite-bearing eclogitized gabbro. The first includes use of newly cali-brated and recalibrated thermobarometers and previouslycalibrated thermobarometers that are discussed below. The sec-ond is the approach used by Berman (1991), which includesevaluation of all possible reactions for the defined system andselected end members using the TWQ program version 2.02.For this approach we use the recommendation of Berman et al.(1995) and do not correct for Fe3+ in clinopyroxene. The thirdfollows the THERMOCALC method of Powell and Holland(1994), which involves a refinement of the approach of Berman(1991) by adding a statistically rigorous least-squares methodfor determining average P and T conditions from all linearlyindependent reactions for the selected end members. The cali-brations and activity models for these three approaches are sum-marized in Table 2.
P-T estimates using coexisting garnet-clinopyroxene-kyanite-phengite-SiO2 are based on the Fe-Mg garnet-clinopyroxene exchange equilibrium (Ravna in press) and threenet-transfer equilibria:
These intersect at high angles and uniquely define the Tand P of equilibrium. We used the thermodynamic database ofHolland and Powell (1998), garnet activity model of Gangulyet al. (1996), clinopyroxene activity model of Holland (1990),and the ideal activity model for phengite and muscovite pro-posed by Holland and Powell (1998). Rigorous evaluation ofthese thermobarometers is beyond the scope of this paper andwill be presented elsewhere. In the presence of quartz, reaction3 is strongly T dependent and provides a linearly independenttest of the commonly used Fe-Mg garnet-clinopyroxene ther-mometer. This approach is particularly useful in kyanite-bear-ing eclogites below the coesite stability field or kyanite-bearing,high-pressure granulites such as sample 419. Reaction 4 withcoesite is moderately P dependent and can place important Pconstraints on UHP rocks. Use of reaction 5 involved arecalibration of the phengite barometer of Waters and Martin(1993). Phengite cation proportions are recalculated assumingCa + Na + K cations in site A = 12.000.
RESULTS
Application of the approaches described above to sample419 from the Haram Gabbro yields mixed results. Intersectionsof all end-member reactions (Table 3) calculated by TWQ (Fig.6a) cluster near the previous estimates of Mørk (1985). Se-lected reactions that have been calibrated previously asthermobarometers (Fig. 6b) are in very good agreement withestimates by Mørk (1985) from the partially to completelyeclogitized Flem Gabbro, and yield conditions of 780 °C and18 kbar. THERMOCALC indicates conditions of 593 ± 156°C and 13.1 ± 4.2 kbar (Fig. 6c). The same selected reactionsplotted in Figure 6c agree with this result; however, the Gar-Kya-Qz-Plag and Gar-Cpx-Plag-Qz barometers would yieldsimilar results to Mørk (1985) if a higher T estimate were used.The calibrated thermobarometers (Table 2 and Fig. 6d) usedfor this sample are in reasonable agreement with the results ofMørk (1985), although the reaction Albite = Jadeite + Quartzyields values of P that appear to be too low. The reason for thisresult may be the activity model of Holland (1990) foromphacite having a low jadeite content.
Each of the three approaches described above, applied tothe low-variance assemblage in sample 1066b and using themineral compositions in Table 1, indicates values of P and T inthe coesite stability field. Using the TWQ program, calculatedfluid-conservative or fluid-absent end-member reactions clus-ter near 860 °C and 45 kbar; however, fluid-dependent reac-tions show considerable scatter assuming an activity of H2O of1.00. Adjustment of the activity of H2O to 0.00036 allows con-vergence of all calculated reactions at these conditions (Fig. 7aand Table 4). This low activity of H2O may be consistent withthe proposed very low or proposed absence of H2O content influid inclusions in garnet in the microdiamond-bearing kyanitegneiss. It should be pointed out that many of these reactions
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1646
involve muscovite and paragonite end-members, and do notconsider the Fe and Mg components of phengite. Omitting micafrom the calculation does not change the estimated P and Tconditions, although the number of linearly independent reac-tions is reduced from 4 to 3. THERMOCALC yields 837 ± 146°C and 28.9 ± 5.6 kbar (2 sigma errors), barely in the coesitestability field (Fig. 7b). Application of the garnet-clinopyroxeneexchange equilibrium, and reactions 4 and 5 above, indicateconditions of 820 °C and 34–39 kbar within the coesite stabil-ity field and at the diamond-graphite phase boundary (Fig. 7c).
Sample 1294a, a kyanite eclogite from Nogva, has a higher-variance assemblage and the THERMOCALC approach couldnot be applied. Results from TWQ indicate 57 kbar and 865 °C
(Fig. 8a). The garnet-clinopyroxene exchange reaction and re-action 4 yield conditions of approximately 820 °C and 30–36kbar (Fig. 8b), which are 4 kbar lower than indicated for sample1066b (Fig. 7c) from Fjørtoft.
Interpretation and discussion of P and T estimates
The range of P and T using the approaches described aboverange from 600 °C, 10 kbar to 780 °C, 18 kbar for the LowerPlate and 700 °C, 28 kbar to 900 °C, 57 kbar for the UpperPlate. While it is possible to say that these plates experienceddifferent metamorphic histories based on a minimum P differ-ence of 10 kbar, these broad ranges of P and T conditions offeronly a little more help than our initial petrographic observa-tions for constraining the conditions of metamorphism. How-ever, we believe that consideration of petrologic and geologicobservations do allow selection of more accurate estimates ofthe metamorphic conditions.
For the Lower Plate, an estimate of 593 ± 156 °C, 13.1 ±4.2 kbar by THERMOCALC is not considered reasonable. Thisestimate is not compatible with the assemblage nor with a pre-vious P-T estimate by Mørk (1985) in the completely to par-tially eclogitized gabbro exposed at Flem. In addition,retrograde symplectite observed in the Haram gabbro mylonitezones includes orthopyroxene-diopside-plagioclase that indi-cates decompression into the granulite facies (Terry 2000). Incontrast, the THERMOCALC estimate would imply decom-pression into the amphibolite facies. Comparison of our cali-brated thermobarometers and the same end-member reactionscalculated by TWQ (Figs. 6b and 6d) with previous estimatesby Mørk (1985) indicate that the estimate of TWQ is in bestagreement, although the calibrated thermobarometers are notdramatically different. Thus we believe that the most reason-able estimate is 780 °C and 18 kbar, and extend this estimate to
TABLE 2. Summary of calibrated geothermobarometers and activity models with associated references used to estimate conditions ofmetamorphism
Approach and Reaction and Mineral and model Source of activity modeldata base calibration sourceRecalibrated and pyrope + 2 grossular + 3 celadonite = Garnet Ganguly et al. (1986)newly calibrated 6 diopside + 3 muscovite Omphacite Holland (1990)thermobarometers Waters and Martin (1993) recalibrated Plagioclase - Model 1 Holland and Powell (1992)using the data base Ravna (in prep) Phengite Holland and Powell (1990)of Holland and pyrope + grossular + 2 SiO2 =Powell (1998) 3 diopside + 2 kyanite
Ravna (in prep)Previously calibrated albite = jadeite + quartzthermobarometers Holland (1980)
quartz + pyrope + 2 grossular =3 anorthite + 3 diopsidePowell and Holland (1988)almandine + 3 diopside =pyrope + 3 HedenbergiteRavna (in press)
Thermocalc version Mica Holland and Powell (1998)2.75 (Holland and Garnet Newton and Haselton (1981)Powell 1998) Plagioclase-Model 1 Holland and Powell (1992)
Clinopyroxene Holland (1990)OmphaciteH2O Holland and Powell (1998)
TWQ version 2.02 Garnet Berman and Aranovich (1996)(Berman 1988, 1991) Mica Chatterjee and Froese (1975)
Plagioclase Fuhrman and Lindsley (1988)Clinopyroxene Berman et al. (1995)H2O Holland and Powell (1990)
TABLE 3. Reactions that are shown in Figure 6a and b
Number Reaction1 3 Hed + 4 Ky = Qz + 3 An + Alm2 Qz + 2 Ky + Gro = 3 An3 Jd + Qz = Ab4 3 Di + 4 Ky = Qz + Pyr + 3 An5 6 Ky + 3 Hed + Gro = Alm + 6 An6 2 Qz + Gro + Alm = 3 Hed + 2 Ky7 3 Qz + 2 Gro + Alm = 3 An + 3 Hed8 4 Ky + Jd + 3 Hed = Ab + Alm + 3 An9 Alm + 3 Di = Pyr + 3 Hed10 Ab + Gro + 2 Ky = Jd + 3 An11 3 Di + Gro + 6 Ky = Pyr + 6 An12 (3) 2 Qz + Pyr + Gro = 3 Di + 2 Ky13 3 Qz + Pyr + 2 Gro = 3 An + 3 Di14 4 Ky + Jd + 3 Di = Ab + 3 An + Pyr15 2 Ky + 2 Jd + 3 Hed = 2 Ab + Alm + Gro16 3 Ab + Alm + 2 Gro = 3 Jd + 3 Hed + 3 An17 2 Ky + 2 Jd + 3 Di = 2 Ab + Gro + Pyr18 3 Ab + 2 Gro + Pyr = 3 Jd + 3 Di + 3 AnNote: Assemblages on the left are on the high-pressure or high-tem-perature side of the reaction. Numbers in parentheses refer to numberedreactions in the text.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES 1647
FIGURE 6. P-T diagrams that show results from the different approaches used to estimate metamorphic conditions in kyanite-bearing eclogitizedHaram Gabbro sample 419. (a) Results from TWQ showing all possible reactions for the selected end-member components (Table 3). There arefour linearly independent reactions. (b) Selected reactions from a that represent typical thermobarometers used for these assemblages. (c)Results from Thermocalc are shown by the error ellipse (1) with selected reactions that represent typical thermobarometers used for theseassemblages. (d) Results from calibrated thermobarometers.
the surrounding gneiss based on the observation that the struc-tural fabric associated with eclogitization can be traced fromgabbro mylonite directly into the enclosing gneiss (Terry andRobinson 1998; Terry 2000).
For the Upper Plate, all three of these approaches yield re-sults compatible with values of P in the coesite stability fieldindicating UHP conditions. THERMOCALC yields the lowestvalues of P and is not in agreement with the presence ofmicrodiamonds. Values of P and T for the low-variance assem-blage in sample 1066b for the calibrated reactions and for TWQare in agreement with the assumption that microdiamonds are
stable and suggest a P range from 33 to 45 kbar. Although theestimates of 45 kbar and 860 °C are consistent with observa-tions, comparison with the P and T, 865 °C, 57 kbar for sample1294a, indicate that TWQ may overestimate P using the activ-ity model of Berman et al. (1995) for clinopyroxene (Table 2).Specifically, this overestimate is interpreted to be a result ofthe simplifying assumptions regarding the jadeite component.Note that sample 1294a has an omphacite with Jd41 in contrastto Jd16.5 for sample 1066b (Fig. 5). For this reason, we favor therange of P and T estimated by the garnet-clinopyroxene ex-change reaction and reactions 4 and 5 above, although the esti-
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1648
mate by TWQ for sample 1066b cannot be ruled out entirelybecause the jadeite component is relatively small. Applicationof reaction 4 to 1294a indicates P of about 36 kbar, which isslightly lower than reactions 4 and 5 indicate for 1066b (Fig.8b). The reason for this difference is unknown but might beexplained by the more ferroan composition of 1294a (Figs. 5a,5b, and 5c).
In summary, our preferred estimate for sample 419 fromthe Lower Plate, 780 °C and 18 kbar, is good agreement withprevious estimates by Mørk (1985) and with the newly cali-brated reactions in this study. The best estimate from the Up-per Plate is thought to be 820 °C and 34–39 kbar from sample1066b using the newly calibrated thermobarometers. This re-sult is consistent with the discovery of microdiamonds, and isvery similar and overlaps with the result from sample 1294afrom the same structural and stratigraphic unit. It is temptingconclude that the results from samples 1066b and 1294a usingTWQ are equally valid because they both fall in the diamondstability field, however, these estimates differ by 12 kbar. Thisseems unusually large for rocks interpreted to be in the samestructural and stratigraphic unit.
Evidence for a tectonic contact
The combined geologic and petrologic data provide threelines of evidence for a tectonic contact between the Upper andLower Plates. (1) The stratigraphic correlation of the UpperPlate with the Blåhø Nappe on the basis of detailed mappingsuggests this. (2) Our best estimates of P- and T-values fromkyanite-eclogites and well-equilibrated, eclogite-faciesmylonite zones indicate a difference of 21 kbar between theUpper and Lower Plates. (3) Our initial examination of eclogitesfrom the gneiss enclosing the Haram Gabbro along the north-west coast of Haramsøy does not show any evidence for poly-crystalline quartz pseudomorphs after coesite like thosepreserved in kyanite-eclogites from Fjørtoft and Nogva.
FIGURE 7. P-T diagrams that show results from the differentapproaches used to estimate metamorphic conditions in kyanite-zoisiteeclogite sample 1066b from Fjørtoft. (a) Results from TWQ showingall possible reactions for the selected end-member components (Table4). There are four linearly independent reactions in this set of reactions.Note that the activity for H2O = 0.00036. (b) Results from Thermocalcare shown by the error ellipse (1) with selected reactions that representthermobarometers used in this study. (c) Results from the thermo-barometer of Ravna (in press) and reactions 4 and 5 of the presentstudy. 1,2 refer to 1st and 2nd clinopyroxenes, Table 1.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES 1649
P-T-Deformation History
The available petrologic data indicate that the lithotectonicplates experienced different metamorphic conditions and musthave separate P-T histories (Fig. 9). The Upper Plate, whichrecords evidence for UHP metamorphism, is believed to haveexperienced a P and T of 39 kbar, 820 °C (Path 1 in Fig. 9).Detailed structural analysis of Terry and Robinson (1998; Terry2000) indicates that this occurred during top-southeast shear-ing in a constrictional strain field that is compatible with thrust-ing toward the foreland. Following subduction, the crust wasimbricated and the UHP unit was transferred to the upper sideof the subduction zone and experienced thrusting toward theforeland that is interpreted to have been synchronous with ex-humation. Note that exhumation of UHP rocks need not havebeen associated with a change in kinematics and, in fact, sucha change would be unlikely if exhumation occurred during con-tinued convergence.
Following near-isothermal decompression to 18 kbar, thisUHP plate was brought into contact with dioritic gneiss andmetamorphosed gabbro of the Lower Plate that had been expe-riencing prograde metamorphism along path 2 (Fig. 9). Thisevent was followed by further imbrication of the crust and con-tinued exhumation of the HP and UHP plates together to ap-proximately 11 kbar, which is the likely P range for thehigh-amphibolite-facies assemblage present in the matrix ofsample UHP1. This re-equilibration is interpreted to have beensynchronous with the onset of top-west and sinistral shearingthat continued during decompression through 8 kbar when thekyanite-garnet mylonite entered the sillimanite stability field.
ACKNOWLEDGMENTS
The Geological Survey of Norway (NGU) provided support for field workby Terry and Robinson. The field work of Terry was also supported by grantsfrom the Norwegian Marshall Fund, the Geological Society of America, OsloUniversity, and the University of Massachusetts. Øystein Nordgulen, Arne Solli,and Trond Torsvik of NGU and Torgeir Andersen of Oslo University providedassistance with financial and logistical aspects of the research. Michael Jercinovicand Michael L. Williams gave assistance with analytical work and extensivediscussion and review of this work. Suzanne McEnroe, Ole Lutro, and Per Terje
Osmundsen of NGU provided key links in e-mail transmission of figures. FrankSpear, Peter Dahl, and Robert Tracy made careful reviews of the manuscript,and it was also reviewed informally by Haakon Fossen and Rebecca Jamieson.To each of these persons and institutions we express our grateful acknowledg-ment.
REFERENCES CITED
Berman, R.G. (1988) Internally-consistent thermodynamic data for minerals in thesystem Na2O-K2O-CaO-MgO-FeO-Fe2O3-Al2O3-SiO2-TiO2-H2O-CO2. Journalof Petrology, 29, 445–522.
———(1991) Thermobarometry using multi-equilibrium calculations: a new tech-nique with petrological applications. Canadian Mineralogist, 29, 833–855.
Berman, R.G. and Aranovich, L.Y. (1996) Optimized standard state and solution
FIGURE 8. P-T estimates for kyanite eclogite sample 1294a from Nogva using TWQ (a) and calibrated reactions (b).
FIGURE 9. P-T deformation histories for the UHP Upper Plate (path1) and HP Lower Plate (path 2) of northern Nordøyane. Gray areas indicateP-T estimates for the Upper Plate based on the thermo-barometer of Ravna(in press) and reactions 4 and 5 of the present study (Present), onTHERMOCALC (TCALC), and on TWEEQU (TWQ). Dashed boxesrepresent previous estimates of P-T values that are discussed in the text.
TERRY ET AL.: THERMOBAROMETRY OF KYANITE ECLOGITES1650
properties of minerals; I, Model calibration for olivine, orthopyroxene, cordier-ite, garnet, and ilmenite in the system FeO-MgO-CaO-Al2O3-TiO2-SiO2. Con-tributions to Mineralogy and Petrology, 126, 1–24.
Berman, R.G., Aranovich, L.Y., and Pattison, D.R.M. (1995) Reassessment of thegarnet-clinopyroxene Fe-Mg exchange thermometer; II, Thermodynamic analy-sis. Contributions to Mineralogy and Petrology, 119, 30–42.
Brueckner, H.K. (1998) Sinking intrusion model for the emplacement of garnet-bearing peridotites into continent collision orogens. Geology, 26, 631–634.
———(1999) Sinking intrusion model for the emplacement of garnet-bearing peri-dotites into continent collision orogens: Reply. Geology, 27, 478.
Carswell, A.D. (1986) The metamorphic evolution of Cr-Mg type peridotites. Lithos,19, 279–297.
Chatterjee, N.D. and Froese, E. (1975) A thermodynamic study of the pseudo-bi-nary join muscovite-paragonite crystalline solutions in the system KAlSi3O8-NaAlSi3O8-Al2O3-SiO2-H2O. American Mineralogist, 60, 985–993.
Dobrzhinetskaya, L.F., Posukhova, T.V., Trønnes, R.G., Korneliussen, A., and Sturt,B.A. (1993) A microdiamond from eclogite-gneiss area of Norway. Proceed-ings 4th Annual Eclogite Conference, Terra Nova Abstract Supplement,Blackwell Scientific, Oxford, 5, 9.
Dobrzhinetskaya, L.F., Eide, E.A., Larsen, R.B., Sturt, B.A., Trønnes, R.G., Smith,D.C., Taylor, W.R., and Posukhova, T.V. (1995) Microdiamond in high-grademetamorphic rocks of the Western Gneiss Region, Norway. Geology, 23, 597–600.
Eskola, P. (1921): On the eclogites of Norway. Skrifter Norske Videnskaps-AkademiOslo, Kl. I, 8. 118 pp.
Fuhrman, M.L. and Lindsley, D.H. (1988) Ternary-feldspar modeling and thermom-etry. American Mineralogist, 73, 201–215.
Ganguly, J., Cheng, W., and Tirone, M. (1996) Thermodynamics of aluminosilicategarnet solid solution; new experimental data, an optimized model, and ther-mometry applications. Contributions to Mineralogy and Petrology, 126, 137–151.
Gasparik, T. and Lindsley, D.H. (1980) Phase equilibria at high pressure of py-roxenes containing monovalent and trivalent cations. In C.T. Prewitt, Ed., Py-roxenes, 7, 309–339. Mineralogical Society of America Reviews in Mineral-ogy, Washington, D.C.
Gee, D.G. (1980) Basement-cover relationships in the central ScandinavianCaledonides. Geologiska Föreningen Stockholm Förhandlingar, 102, 455–474.
Gee, D.G., Kumpulainen, R., Roberts D., Stephens, M. B., Thon, A., and Zachrisson,E. (1985) Scandinavian Caledonides-Tectonostratigraphic Map. In Gee, D.G.and Sturt, B. A., Editors. The Caledonide Orogen-Scandinavia and Related Ar-eas. Wiley, Chichester.
Griffin, W.L. (1986) On the eclogites of Norway—65 years later. MineralogicalMagazine, 51, 333–343.
Griffin, W.L., Austreim, H., Brastad, K., Brynhi, I., Krill, A.G., Krogh, E.J., Mørk,M.B.E., Qvale, H., and Tørudbakken, B. (1985) High-pressure metamorphismin the Scandinavian Caledonides. In Gee, D.G. and Sturt, B.A., Editors. TheCaledonide Orogen—Scandinavia and Related Areas. Wiley, Chichester, 783–801.
Holland, T.J.B. (1980) The reaction albite = jadeite+quartz determined experimen-tally in the range 600–1200 °C. American Mineralogist, 65, 129–134.
———(1990) Activities of components in omphacitic solid solutions; An applica-tion of Landau theory to mixtures. Contributions to Mineralogy and Petrology,105, 446–453.
Holland, T.J.B. and Powell, R. (1990) An enlarged and updated internally consistentthermodynamic data set with uncertainties and correlations: the system K2O-Na2O-CaO-MgO-FeO-Fe2O3-Al2O3-TiO2-SiO2-C-H2-O2. Journal of Metamor-phic Geology, 8, 89–124.
———(1992) Plagioclase feldspars: activity-composition relations based uponDarken’s Quadratic Formalism and Landau theory. American Mineralogist, 77,53–61.
———(1998) An internally consistent thermodynamic data set for phases of petro-logical interest. Journal of Metamorphic Geology, 16, 309–343.
Krill, A.G. (1980) Tectonics of the Oppdal area, central Norway. GeologiskaFöreningens i Stockholm Förhandlingar, 102, 523–530.
———(1985) Relationship between the western gneiss region and the Trondheimregion: Stockwerk tectonics reconsidered. In D.G. Gee and B.A. Sturt, Eds.,The Caledonide Orogen—Scandinavia and Related Areas, Wiley, Chichester,475–484.
Krogh, E.J. (1977) Evidence for a Precambrian continent-continent collision inwestern Norway. Nature, 267, 17–19.
Krogh, E.J. and Carswell, D.A. (1995) HP and UHP Eclogites and garnet peridotitesin the Scandinavian Caledonides. In Coleman, R. G. and Wang, X., Editors,Ultrahigh-pressure metamorphism. Cambridge University Press, 244–298.
Larsen, R.B., Eide, E.A., and Burke, E.A.J. (1998) Evolution of microdiamond-bearing granulites in the Western Gneiss Region, Norway. Contributions toMineralogy and Petrology, 133, 106–121.
Lutro, O., Robinson, P., and Solli, A. (1997) Proterozoic geology and Scandianoverprinting in the Western Gneiss Region. Norges Geologiske Undersøkelse,Report 97.132, 86 p.
Mørk, M.B.E. (1985) A gabbro to eclogite transition on Flemsøy, Sunnmøre, west-ern Norway. Chemical Geology, 50, 283–310.
Mørk, M.B.E. and Mearns, E.W. (1986) Sm-Nd isotopic systematics of a gabbro-eclogite transition. Lithos, 19, 255–267.
Newton, R.C. and Haselton, H.T. (1981) Thermodynamics of the garnet-plagio-clase-Al2SiO5-quartz geobarometer. In R.C. Newton, Editor, Thermodynamicsof Minerals and Melts, Springer Verlag, New York, 131–147.
Pouchou, J.L. and Pichoir, F. (1984a) A new model for quantitative X-ray microanaly-sis, Part I: application to homogeneous samples. La Recherche Aerospatiale, 3,13–38.
———(1984b) A new model for quantitative X-ray microanalysis, Part II: applica-tion to in-depth analysis of heterogeneous samples. La Recherche Aerospatiale,5, 47–65.
Powell, R. and Holland, T. (1988) An internally consistent dataset with uncertain-ties and correlations; 3, Applications to geobarometry, worked examples and acomputer program. Journal of Metamorphic Geology, 6, 173–204.
———(1994) Optimal geothermometry and geobarometry. American Mineralo-gist, 79, 120–133.
Ravna, E.J.K. (2000) The garnet-clinopyroxene Fe2+-Mg geothermometer: an up-dated calibration. Journal of Metamorphic Geology, in press.
Roberts, D. and Gee, D.G. (1985) An introduction to the structure of the Scandina-vian Caledonides. In D.G. Gee and B.A. Sturt, Eds., The Caledonide Orogen—Scandinavia and Related Areas. Wiley, Chichester, 55–68.
Robinson, P. (1995) Extension of Trollheimen tectono-stratigraphic sequence in deepsynclines near Molde and Brattvåg, Western Gneiss Region, southern Norway.Norsk Geologisk Tidsskrift, 75, 181–198.
———(1997) Quartzite of Reksdalshesten, Moldefjorden, Western Gneiss Region:Another parallel with the tectonostratigraphy of Trollheimen. Norges GeologiskeUndersøkelse, 433, 14–15.
Robinson, P. and Terry, M.P. (1998) Evolution of amphibolite-facies structural fea-tures during transition from convergence to extensional collapse, Western GneissRegion, Norway. Workshop on Ultra-High-pressure Metamorphism and Exhu-mation, Stanford University, 79–83.
Smith, D.C. (1984) Coesite in clinopyroxene in the Caledonides and its implica-tions for geodynamics: Nature, 310, 641–644.
Terry, M.P. (2000) Structural and thermal evolution of Baltica basement and in-folded cover nappes on Nordøyane and their bearing on mechanisms for pro-duction and exhumation of high-pressure rocks, Western Gneiss Region, Nor-way. Ph.D. thesis, University of Massachusetts, Amherst.
Terry, M.P. and Robinson, P. (1996) Caledonian structural fabrics related to subduc-tion of continental crust, Western Gneiss Region, Norway. Geological Societyof America Abstracts with Programs, 28, A501.
———(1998) Eclogite-facies structural features and their bearing on mechanismsfor production and exhumation of high-pressure rocks, Western Gneiss Region,Norway, including the microdiamond-bearing gneiss at Fjørtoft. InternationalWorkshop on Ultra-High-pressure Metamorphism and Exhumation, p. 74–78,Stanford University.
——— (1999) Sinking intrusion model for the emplacement of garnet-bearing peri-dotites into continent collision orogens: Comment. Geology, 27, 477–478.
Terry, M.P., Robinson, P., Carswell, D.A., and Gasparik, T. (1999a) Evidence for aProterozoic mantle plume and a thermotectonic model for exhumation of gar-net peridotites, Western Gneiss Region, Norway. EOS 80 Abstract Supplement,S359–360.
Terry, M.P., Robinson, P., and Ravna, E.K. (1999b) Evidence for Scandian coesite/diamond conditions in kyanite eclogite at Fjørtoft, Western Gneiss Region,Norway. EOS, 80 Abstract Supplement, S363.
Terry, M.P., Robinson, P., Hamilton, M.A., and Jercinovic, M.J. (2000) Monazitegeochronology of UHP and HP metamorphism, deformation, and exhumation,Nordøyane, Western Gneiss Region, Norway. American Mineralogist, 85, 1651–1664.
Torsvik, T.H. (1998) Palaeozoic palaeogeography: A North Atlantic viewpoint. GFF,120, 109–118.
Tucker, R.D. (1986) Geology of the Hemnefjord-Orkanger area, south-central Nor-way. Norges Geologiske Undersøkelse, 404, 1–24.
Tveten, E. and Lutro, O. (1996) Geologiske kart over Norge, berggrunnskart Ålesund1:250,000, foreløpig utgave. Norges geologiske undersøkelse.
Wain, A. (1997) New evidence for coesite in eclogites and gneisses: Defining anultrahigh pressure province in the western gneiss region, Norway. Geology, 25,927–930.
Wain, A., Waters, D., Jephcoat, A., and Olijynk, H. (2000) The high-pressure toultrahigh-pressure eclogite transition in the Western Gneiss Region, Norway.European Journal of Mineralogy, 12, 667–687.
Waters, D.J. and Martin, H.N. (1993) Geobarometry in phengite-bearing eclogites.Terra Abstracts, 5, 410–411.
MANUSCRIPT RECEIVED APRIL 3, 2000MANUSCRIPT ACCEPTED JULY 26, 2000PAPER HANDLED BY ROBERT J. TRACY
