Géochronologie Par Méthode Conventionnelle

Géochronologie par méthode conventionnelle

[1] Bruguier O, Dada SS & Lancelot JR (1994) Early Archean crust (>3.56 Ga) within a 3.05 Ga orthogneiss from Northern Nigeria. U-Pb zircon evidence. Earth and Planetary Science Letters 125: 89-103.

[2] Bruguier O (1996) U-Pb ages on single detrital zircon grains from the Tasmiyele Group: Implications for the evolution of the Olekma Block (Aldan Shield, Siberia). Precambrian Research 78: 197-210.

[3] Bruguier O, Lancelot JR & Malavieille J (1997). U-Pb dating on single detrital zircon grains from the Triassic Songpan-Ganze flysch (Central China): evidence for a Sino-Korean provenance and tectonic correlations. Earth and Planetary Science Letters 152: 217-231.

[4] Bosch D & Bruguier O (1998) An Early Miocene age for a high temperature event in gneisses from Zabargad Island (Red Sea, Egypt): mantle diapirism. Terra Nova 10: 274-279.

[5] Bruguier O, Becq-Giraudon JF, Bosch D & Lancelot JR (1998) Late Visean hidden basins in the Internal Zones of the Variscan belt: U-Pb zircon evidence from the French Massif Central. Geology 26: 627-630.

[6] Bruguier O, Bosch D, Pidgeon RT, Byrne D & Harris LB (1999) U-Pb chronology of the Northampton Complex, Western Australia - evidence for Grenvillian sedimentation, metamorphism and deformation and geodynamic implications. Contributions to Mineralogy and Petrology 136: 258-272.

[7] Dhuime B, Bosch D, Bruguier O, Caby R & Archanjo C (2003) An early Cambrian U-Pb apatite cooling age for the high-temperature regional metamorphism in the Pianco area, Borborema province (NE Brazil). Compte-Rendus Geosciences, 335: 1081-1089.

[8] Bruguier O, Becq-Giraudon J.F, Clauer N & Maluski H (2003) From late Visean to Stephanian: Pinpointing a two-stage basinal evolution in the Variscan Belt. A case study from the Bosmoreau basin (French Massif Central) and its geodynamic implications. International Journal of Earth Sciences 92: 338-347.

[9] Bosch D, Bruguier O. Kranshobaiev A & Efimov A (2006) A Middle Silurian age for the Uralian Platinum-bearing Belt (Central Urals, Russia): U-Pb zircon evidence and geodynamic implication. In "European Lithosphere Dynamics", Gee, D.G. and Stephenson, R.A. (eds), Memoirs of the Geological Society of London 32: 443-448.

Pre[nmbrinn Resenrth ELSEVIER PrecambfianResearch 78 (1996) 197-210

U-Pb ages on single detrital zircon grains from the Tasmiyele Group: implications for the evolution of the Olekma Block (Aldan Shield, Siberia)

O. Bruguier Laboratoire de G£ochronologie-Gdochimie-P£trologie, CNRS-URA 1763, Case courrier 066, Universit£ de Montpellier 11, PI. Eugene Bataillon, 34,095 Montpellier Cedex 5, France

Received 22 December 1994; revised version accepted 19 September 1995

Abstract

The Aldan Shield of Siberia is one of the largest exposures of the Siberian Craton and has been divided into different units according to their geological characteristics. In the central part of the shield, the main divisions are the Olekma and West Aldan Blocks. The former contains supracrustal rocks and typical greenstone belts. We report U-Pb isotopic analyses on 51 single detrital zircon grains from 5 samples of quartzite collected at different stratigraphical levels from clastic metasediments of the Tasmiyele Group situated in the Olekma Block. The youngest sub-concordant grain (2963 ___ 5 Ma) provides an older limit to the deposition. Combined with other information on the geological evolution of this part of the Aldan Shield, the results show that sediments were deposited between 2500 and ~ 2960 Ma, and that detritus was derived from the neighbouring basement of the Olekma Block. The age spectrum presented by detrital zircons implies the creation of large amounts of differentiated material during the period 2900-3000 Ma which represents an important crustal event for this part of the Aldan Shield. Moreover, it appears from these results and previous works that the Tasmiyele Group and the Tungurcha Group, initially grouped together to form the Tungurcha Greenstone Belt, are two distinct and unrelated units. The Tasmiyele Group, whose affiliation with greenstone belts is uncertain, was deposited at the same time or after the formation of the Olondo Greenstone Belt.

1. Introduction carry important information on the composition, tectonic setting and evolution of the source(s) region(s) Despite covering up to 80% of the Earth surface, from which they come from (e.g. Maas and McCul- sediments have not been intensively studied using loch, 1991). Furthermore, they may represent the conventional geochemical techniques. This is mainly only remnants of source rocks which have since due to mixing between components of various origin disappeared (Froude et al., 1983; Compston and and also to the complexity of the processes occurring Pidgeon, 1986). In sedimentary environments, detri- during diagenesis and/or alteration. However, clas- tal zircons constitute a mixture of grains from source tic sedimentary rocks and their detrital minerals, can rocks whose origin, age and evolution may be totally

different. Therefore, even two grains which appear which (such as the Kurulta Group) may have been identical in shape and colour, cannot be considered deposited and subsequently metamorphosed in the to have been derived from the same source rock. Age Early to Mid-Archaean (> 3.3 Ga) (Bibikova et al., determinations of single zircon grains are therefore 1989; Glebovitsky and Drugova, 1993). Geochrono- essential to identify different age populations and logical results (Bibikova et al., 1989; Morozova et this paper presents U-Pb ages on 51 detrital zircon al., 1989a; Nutman et al., 1992) indicate that the grains from the Tasmiyele Group (Olekma Block, basement of the West Aldan Block is made up of Aldan Shield). The study aims to resolve the age Archaean and even Early Archaean rocks (> 3.5 Ga) spectrum presented by the zircon populations and to as demonstrated by the 3500 Ma minimum 2°Tpb/ obtain information on the provenance of the detritus 2°6pb age of a discordant zircon fraction from a and on the age of deposition of the Tasmiyele Group. tonalitic gneiss (Morozova et al., 1989b). Furthermore, the age patterns reflect the evolution of The basement of the Olekma Block consists the source regions and allow a direct comparison mainly of granitoids and tonalitic to granodioritic between the Tasmiyele Group and other supracrustal gneisses. Supracrustal belts (metavolcanic and units of the Olekma Block. Another purpose of this metasedimentary rocks) are also preserved. All these study was to search for evidence of very old (Early rocks have undergone amphibolite-facies metamor- Archaean) zircon grains. With the discovery of 3.4 phism. Eastwards, the grade of metamorphism and Ga old Archaean rocks within the Omolon massif deformation increases and relict eclogites are found (Bibikova, 1984), it was hoped to find, in the Olekma in the easternmost part of the Olekma Block (Smelov, Block, the most ancient core of the Aldan Shield. 1989), assigned to be of Mesoproterozoic age (Nut- man et al., 1992). Published geochronological results (Jahn et al., 1991; Nutman et al., 1992; Glebovitsky 2. Geological setting and Drugova, 1993; Neymark et al., 1993; Ve- likoslavinsky et al., 1993), emphasize that the base- The Tasmiyele Group outcrops on the Olekma ment of the Olekma Block is mainly constituted of Block (Fig. 1) in the middle part of the Archaean 2.9-3.0 Ga old rocks. So far, the oldest rocks identi- Aldan Shield. The Aldan Shield, which constitutes fied are 3.25 Ga old orthogneisses (Nutman et al., the largest exposure of the Siberian Craton, has been 1992), although Neymark et al. (1993) proposed for traditionally subdivided into various geological units the 2.98 Ga Amnunnakta granitoid massif a Nd according to structural and metamorphic characteris- model age of 3700 Ma, that reflects the occurrence tics (Dook et al., 1989). In the middle part of the of much older sialic material. Greenstones belts out- shield, the main divisions are constituted by the cropping on the Olekma Block have not been dated Olekma granite-greenstone terrain and the West AI- yet, except for the Olondo Greenstone Belt. Ages dan granulite-gneiss Block (Fig. 1). These two from volcanics of this typical greenstone belt are blocks are separated by the Amga fault which ap- indistinguishable from ages of the basement rocks pears to represent a ductile shear zone, related to ( ~ 3.0 Ga) (Baadsgaard et al., 1990). The Tungurcha thrusting of the West Aldan Block westward over the Greenstone Belt, more than 170 km long and 25-30 edge of the Olekma Block (Smelov, 1989; Smelov km wide, is constituted by the Tungurcha Group and Beryozkin, 1993). The age of this event has been (lower part) and the Tasmiyele Group (upper part). clearly shown to be Proterozoic (1.9-2.0 Ga) (Nut- The Tungurcha Group outcrops as isolated tectonic man et al., 1992). The West Aldan Block consists fragments which may have constituted a single se- mainly of granulites and amphibolite-facies migma- quence (Bogomolova and Smelov, 1989). The slabs titic gneisses and it is generally thought that high- differ from each other in their constituent rocks grade events occurred several times during Archaean which comprise volcanics, mafic plutonic rocks, and Proterozoic time (Bibikova et al., 1989; Gle- schists, carbonates and clastic sediments. A mini- bovitsky and Drugova, 1993). The block also pre- mum age for the deposition and subsequent thrusting sents supracrustal sequences of mature sediments of the Tungurcha Group is given by a tonalitic gneiss (quartzites, mafic schists and calc-silicate), some of (3016_+8 Ma) intruding gneisses and ultramafic O. Bruguier / Precambrian Research 78 (1996) 197-210 199 rocks of the group on the west side of the Olekma mineral assemblages in the diabases indicate that the ri,/er (Nutman et al., 1992). The metasediments of rocks have equilibrated under metamorphic condi- the Tasmiyele Group outcrop in a north-south tions of about 500-530°C and 1-2.5 kbar graben-like structure whose boundaries are defined (Bogomolova and Smelov, 1989). The sedimentary by tectonic contacts (Fig. lc). On its western flank a sequence is composed of mica-quartz schists, blastomylonite zone is developed in rocks of both quartzites and micaceous quartzites. Primary sedi- the adjacent gneissic basement and the graben. The mentary textures and beddings are well preserved eastern boundary is a two-mica granite massif clearly and six bedded units (units 1 to 6 in Fig. lc) can be cross-cutting the sedimentary sequence. Sills of distinguished which consist of basal coarse-grained metadiabases occur in the lower part of the series. sediments which progressively fine upward (Bogo- Sediments and diabases have been deformed and molova and Smelov, 1989). The total thickness of metamorphosed under metamorphic conditions that the group does not exceed 1000 m. Though no do not exceed the middle amphibolite facies. The geochronological data are available, the Tasmiyele

Aldan Shield, Siberia "" .."..".-".." .. "')1~i~:~l/iV ~ [ II I rest AIdan Granulite- ::: Gneiss Block

Olekma Granite- Greenstone Block

~;:rea

L-% ~::::::::::::::::::: .. ".. ".,-'.. ".. ,... Tc~alite ,~

Units 5-- and 6 =am Unit4 ~

Unit 3 ~,~~ ¢1 I 1 unit2 ~ • Phanerozolc I LateArchaean mallc- covers, intrusions ultramafic intr., TTG gn. 0 60 Unit 1 SM0 ~ Proterozoic ~ arly Arch. intrusives, Krn coverseries less-rrG gnelsses Diabase Pmterozolc ~ Arch. Supr. seq., Archaean Two-mica Gra~te ]•1 enderbitic gneiss Intrusive complex, mainly sedim., volc. My!or,it~ Late Arch. intrus., I Archaean Archaean TTG grey gnelsses r lesslEG gn. greenstone belts o 2 Km

Fig. 1. Composite map. (a) The geographical location of the Archaean Aldan Shield in the former Soviet Union. (b) A geological map of a portion of the central part of the Aldan Shield, showing the location of the various units and blocks mentioned in the text: Ol = Olondo Greenstone Belt; Tn = Tungurcha Greenstone Belt; Ts = Tasmiyele Group. Geology is from Dook et al. (1989), as clarified by B.M. Jahn. The rectangle labelled c encloses the sampling site. (c) A more detailed map of the sampling area showing sample location. Geology is from Bogomolova and Smelov (1989), 200 O. Bruguier / Precambrian Research 78 (1996) 197-210

Group has been considered as representing the upper ever, the occurrence of rounded grains also indicates part of the Tungurcha Greenstone Belt and its depo- that some grains have probably been transported sition has been proposed to occur before 2.5 Ga over a longer distance than the main population. No (Bogomolova and Smelov, 1989). metamorphic multifaceted grains, commonly found in high-grade metamorphic rocks have been observed and there is no evidence that the zircons are 3. Samples other than detrital in origin.

Five samples of quartzites and micaceous quartzites (Si-10, Si-11, Si-12, Si-13 and Si-14) have 4. Analytical technique been collected during the IGCP field trip in 1989. Sampling has been done at different stratigraphical Zircons were separated from 3 to 5 kg of rock levels of the sedimentary pile in order to determine using standard heavy liquids. Grains used for U-Pb an age spectrum for the whole series, from base analyses were hand picked under a binocular micro- (sample Si-10) to top (sample Si-14), on the basis of scope according to colour and morphology and air primary sedimentary features. The samples will also abraded during a few hours following the technique allow detection of rocks of different ages in the of Krogh (1982). Grains were then carefully washed source areas of the Tasmiyele Group. The samples with highly purified reagents (4 N HNO3, tridistilled typically consist of angular fragments of quartz water and 2 N HNO 3) and weighted on a Cahn showing undulose extinction, microcline, plagioclase electronic micro-balance. Zircons were dissolved in with myrmekite structure, tourmaline, garnet and 48 h at 195°C in a Teflon microbomb with 5 p.l of opaques. The matrix (groundmass) is composed of tridistilled 48% HF. After dissolution, the solution chlorite, quartz, biotite and muscovite. Zircon is an was evaporated to dryness and the microbomb was accessory mineral. The occurrence of feldspar, to- then filled with tridistilled 6 N HC1 and heated for a gether with the angular shape of quartz, implies few hours. Following Lancelot et al. (1976) and relatively short sedimentary transport. The samples Bruguier et al. (1994), an aliquot was then spiked yielded abundant zircon grains which are relatively with a 2°8pb-235U tracer. The unspiked part was homogeneous with respect to external morphology used for measurement of the Pb isotopic composi- but show a wide range in size. Most samples are tion. Both solutions, evaporated with 0.25 N H3PO 4 dominated by light pink to colourless translucent (2 /xl), were loaded with 1 mg/ml silicagel (6 /xl) crystals, the morphology of which can be divided onto a single Re filament without previous chemical into two broad categories. (1) A major component of separation of the elements. Isotopic measurements the total population is constituted by prismatic euhe- were carried out on a VG Sector mass spectrometer dral crystals with undamaged or almost undamaged using a Daly detector. As Pb and U were loaded faces and corners. These grains clearly show no together on the same filament, Pb was first measured signs of metamorphic or erosion-related rounding. before running U. While heating the samples, signals (2) A minor component is represented by rounded on masses 201, 203 and 205 were commonly ob- zircons showing signs of abrasion of external sur- served, corresponding to TI + and BaPO~- (com- faces. Some grains present only slight rounding of pounds of 137 Ba, 138 Ba, 16 O, 17O and 18 O) interfer- the corners whereas others are well rounded suggest- ences. Pb isotopic ratios were therefore only mea- ing a rather long transport and/or that they may sured at high temperatures (1400-1500°C) after have passed through more than one cycle of erosion complete vaporization of T1 and Ba compounds. and deposition. Total Pb blanks over the period of the analyses range The preponderance of well-preserved crystal forms from 16 to 25 pg. The calculation of common Pb is attributed to short sedimentary transport of the was made by subtracting blanks and then assuming grains and to a provenance from a source area close the remaining common Pb has been incorporated to to the basin as previously pointed out by the angular the crystal and has a composition determined from shape of quartz and the occurrence of feldspar. How- the model of Stacey and Kramers (1975). A correc- O. Bruguier / Precambrian Research 78 (1996) 197-210 201 tion of 0.24 _+ 0.05% a.m.u, for mass fractionation 2°7pb/2°6pb age determined for each grain provides was applied. Corrected isotopic ratios were calcu- reliable minimum ages for the source rocks from lated according to Briqueu and De La Boisse (1990), which the zircons are derived. Of the thirteen grains and regression lines and intercepts according to Lud- analyzed, nine provide 2°7pb/2°6pb ages ranging wig (1987). Analytical uncertainties are listed as 2o- from 2900 to 2990 Ma. Four grains (3, 4, 8 and 13, and uncertainties in ages as 95% confidence levels. in black in Fig. 2) have slightly older 2°7pb/2°6pb Decay constants are those recommended by the IUGS ages ranging from 3040 to 3090 Ma and, as we will Subcommission on Geochronology (Steiger and see below, exceeding those of the other grains ana- J~iger, 1977). lyzed in this study.

5. Results 5.2. Quartzite Si-ll

5.1. Quartzite Si-lO Seven grains have been analysed on this sample. U contents range from 169 to 725 ppm (Table 1). Thirteen crystals selected among the various grain Surprisingly, for a set of detrital zircons, and unlike types present in this sample have been analysed. U the previous sample, experimental points can be contents range from 76 to 434 ppm (Table 1) and fitted to a chord (Fig. 3) intersecting the concordia there is no simple relationship between U concentra- curve at 2997 _ 28 Ma and 242 + 98 Ma (MSWD = tion and degree of discordance. However, grains 6 37). The scattering of experimental points, expressed and 7, among the richest in U (280 and 434 ppm, by the high value of the MSWD, is essentially due to respectively), are the most discordant and grain 8, grain 20. Removing this data point from the regres- with a low U concentration (76 ppm) is the least sion, for example on the assumption that this grain discordant. Reported on the concordia diagram (Fig. underwent zero-age Pb losses, leads to values of 2), experimental points do not form a linear array 2997 + 4 Ma for the upper intercept and 280 + 13 indicating that the zircons analyzed do not form a Ma for the lower intercept (MSWD = 0.9). As these single, homogeneous population. This is commonly grains are detrital in origin, and because two grains, observed for detrital zircons which originated from even when identical in shape and colour criteria, various source rocks (Davis et al., 1990; Krogh and cannot be assumed to be cogenetic, we prefer not to Keppie, 1990; Ross et al., 1992). Considering the delete any of the results and favour the 2997 + 28 discordance of experimental points, the apparent Ma value. The observed alignment could then correspond to U-Pb discordancy patterns for zircon grains of identical or nearly identical ages having suffered a similar evolution (Schgtrer and All~gre, 1982). There- i ~ Tasmiyele Group 50 ~"~I - fore, this could be regarded as the age of rocks of the 54 ~ ~ Olekma Block 2~"~//~I/ - [ ~" Quartzite Si-lO ~ /~//~.~'~ ] source area feeding the Tasmiyele Group which is consistent with previously published radiometric data i from basement rocks of the Olekma Block (e.g. Nutman et al., 1992). The lower intercept is significantly different from zero and suggests disturbance of the U-Pb systems in the past but this does not correlate with any known geological event. More- over, the apparent 2°7pb/2°6pb ages for all grains

4 6 8 10 12 14 16 18 but one (grain 17, the most discordant and the richest in U) span a narrow range of time (from 2900 to Fig. 2. Concordia diagram showing U-Pb results on single zircon grains from the quartzite sample Si-10. Dots show the four oldest 2980 Ma), which encompasses the range of varia- grains (> 3040 Ma) analysed in this study. Heavy lines are tions observed for sample Si-10. This small range in 2°7pb/2°6pb age reference lines. age, together with the good alignment of experimen- 202 O. Bruguier / Precambrian Research 78 (1996) 197-210

Table 1 U-Pb isotopic data for single zircon grains from the Tasmiyele Group of the Olekma Block (Aldan Shield, Siberia) Sample Weight U Pb 2°6pb/2°4pb Z°aPb/Z°6pb 2°6pbr/238U 2°7pbr/235U 2°7pbr/2°6pb r Apparent Disc.(%) (rag) (ppm) (ppm) age (Ma) Si-lO 1Zr, C, Eu, L 0.005 178 97 524 0.1208 0.4664_+20 14.062_+062 0.2187+_4 2971 17 2 Zr, P, Eu, S 0.005 194 122 533 0.2099 0.5009 +_ 23 14.528 _+ 064 0.2104_+ 3 2909 10 3 Zr, C, Rd, L 0.005 237 113 1351 0.0835 0.4148 _+ 17 13.102 _+ 053 0.2291 _+ 3 3046 26 4Zr, P, Eu, L 0.006 318 192 1386 0.1136 0.5222+-21 16.874+_067 0.2343_+2 3082 12 5 Zr, C, Eu, L 0.007 127 61 1273 0.1072 0.4158 _+ 27 12.560 _+ 095 0.2191 _+ 8 2974 25 6Zr, P, Rd, L 0.008 280 108 2402 0.1251 0.3298_+20 09.508_+066 0.2091_+6 2899 37 7Zr, P, Eu, L 0.009 434 169 3359 0.1165 0.3362_+ 18 10.011_+055 0.2159+_2 2951 37 8Zr, C, Rd, S 0.010 76 48 1029 0.1187 0.5321_+30 16.869+-098 0.2299_+5 3052 10 9Zr, P, Rd, L 0.010 117 58 1122 0.0842 0.4344_+30 13.079+-088 0.2184+_3 2969 22 10Zr, C, Eu, L 0.013 114 71 1382 0.2314 0.4912+_32 14.983+_106 0.2212_+5 2990 14 I1Zr, P, Eu, S 0.013 185 105 2831 0.1291 0.4896_+32 14.906+_095 0.2208_+2 2987 14 12Zr, C, Eu, S 0.013 166 85 607 0.0918 0.4502_+29 13.178_+089 0.21234-4 2923 18 13 Zr, C, Eu, L 0.015 193 107 2186 0.0983 0.4785 _+ 29 15.095 _+ 095 0.2288 +_ 4 3044 17

Si-l l 14Zr, C, Eu, L 0.005 435 163 1372 0.0480 0.3401 _+ 14 09.923_+042 0.2116_+2 2918 35 15Zr, C, Eu, S 0.005 169 93 408 0.1000 0.4714_+24 14.234+-071 0.2190_+5 2973 16 16 Zr, C, Eu, L 0.005 575 191 903 0.0360 0.3047 +_ 23 08.804 _+ 066 0.2096 _+ 2 2903 41 17 Zr, P, Eu, L 0.006 725 170 1612 0.0700 0.2073 +- 21 05.635 ± 057 0.1972 _+ 2 2803 57 18 Zr, P, Rd, L 0.006 200 112 1826 0.0620 0.4974 _+ 22 15.067 _+ 065 0.2197 _+ 3 2979 13 19Zr, C, Eu, S 0.006 331 196 1780 0.1220 0.5055_+27 15.335+_079 0.2200_+3 2981 12 20Zr, C, Eu, L 0.008 314 113 2307 0.1070 0.3144_+12 09.309_+036 0.2147_+3 2942 40

Si-12 21Zr, P, Eu, L 0.002 242 138 714 0.1315 0.4613+-41 14.156_+133 0.2226_+6 2999 18 22 Zr, P, Eu, S 0.002 560 228 719 0.0863 0.3556 _+ 16 10.324_+ 043 0.2105 _+ 3 2910 33 23Zr, P, Eu, L 0.002 221 130 480 0.0865 0.4887_+49 14.717_+ 142 0.2184_+4 2969 14 24Zr, C, Rd, L 0.002 175 114 432 0.1388 0.5218+60 15.905_+ 177 0.2211 4-4 2989 9 25Zr, C, Eu, L 0.003 330 82 438 0.1112 0.1896_+ t7 05.652+_044 0.2161:24 2952 62 26 Zr, C, Rd, S 0.003 570 112 963 0.0832 0.1689 +_ 08 05.044 _+ 026 0.2166 -+_ 5 2955 66 27Zr, P, Eu, L 0.005 309 179 1574 0.1094 0.5018_+20 15.192_+061 0.2196_+3 2978 12 28Zr, C, Eu, L 0.005 191 110 942 0.1325 0.4853_+29 14.604_+089 0.2183_+6 2968 14 29Zr, C, Eu, S 0.005 48 30 426 0.1526 0.5212_+29 15.873+_085 0.2209_+4 2987 9 30 Zr, P, Rd, L 0.006 168 103 t 102 0.0674 0.5447 _+ 25 16.533 _+ 075 0.2202 _+ 3 2982 6 31Zr, P, Rd, L 0.006 211 1t7 1684 0.0527 0.4929_+19 14.923_+056 0.2196_+3 2978 13 32Zr, C, Eu, S 0.006 309 135 1814 0.055l 0.3984_+49 11.667_+ 143 0.2124_+2 2924 26 33 Zr, P, Rd, S 0.007 1205 386 3842 0.0763 0.2880 _+ 11 08.209 _+ 032 0.2068 _+ 1 2881 43 34Zr, C, Eu, L 0.007 146 86 2232 0.0638 0.5211 _+36 15.840_+ 107 0.2205_+2 2984 9 35 Zr, P, Rd, S 0.007 209 108 1614 0.0743 0.4582 _+ 24 13.628 _+ 069 0.2157 ± 3 2949 18 36Zr, C, Eu, L 0.008 172 94 1051 0.1045 0.4778_+22 14.352_+064 0.2179_+3 2965 15

Si- 13 37 Zr, C, Eu, L 0.003 235 133 792 0.1042 0.4982 + 13 15.009 +_ 038 0.2185 _+ 2 2970 12 38Zr, P, Eu, L 0.004 300 152 1939 0.0746 0.4547+_ 16 13.671 _+046 0.2180+_2 2966 19 39Zr, P, Eu, L 0.007 188 105 921 0.0870 0.4968_+21 14.953_+067 0.2183_+4 2968 12 40Zr, C, Eu, S 0.009 221 119 723 0.0771 0.4813_+24 14.418_+071 0.2173_+3 2961 14 41Zr, C, Eu, L 0.010 229 135 1816 0.1200 0.5113__+20 15.636_+061 0.2218_+2 2994 11 42Zr, P, Eu, S 0.014 377 193 2521 0.0887 0.4583_+60 13.313_+ 175 0.2107_+6 291l 16

Si-14 43 Zr, P, Eu, L 0.004 233 131 573 0.0995 0.5170 ± 22 15.721 _+ 068 0.2205 _+ 4 2984 10 44Zr, P, Eu, S 0.005 320 152 1020 0.0391 0.4407_+59 12.801 ± 168 0.2106_+3 2910 19 O. Bruguier /Precambrian Research 78 (1996) 197-210 203

Table 1 (continued) Sample Weight U Pb 2°6pb/2°4pb 2°spb/:°6pb 2°6pbr/238U 2°7pbr/235U 2°7pbr/2°6pbr Apparent Disc.(%) (mg) (ppm) (ppm) age (Ma) Si-14 45 Zr, P, Eu, L 0.006 140 80 653 0.1504 0.4832 + 21 14.724 + 060 0.2210 + 4 2988 15 46Zr, P, Eu, L 0.007 190 112 1184 0.1339 0.5068+45 15.322_+ 144 0.2192_+3 2975 11 47 Zr, P, Eu, S 0.007 170 89 734 0.1081 0.4606 _+ 26 13.729 _+ 077 0.2161 4- 4 2952 17 48 Zr, C, Rd, L 0.008 91 49 407 0.0995 0.4777 -+ 26 14.042 _+ 068 0.2132 _+ 5 2930 14 49Zr, C, Eu, L 0.009 205 119 2219 0.1019 0.5089_+46 15.481 _+ 137 0.2206_+3 2985 11 50Zr, C, Rd, S 0.009 226 145 1435 0.1172 0.5556_+42 16.865_+130 0.2202_+6 2982 4 51Zr, P, Eu, L 0.011 152 97 1563 0.0880 0.5666_+63 17.003_+194 0.2176_+7 2963 2 P = pink to purple; C = colourless; Eu = euhedral; Rd = rounded; L = elongated; S = squat; r = radiogenic lead corrected from blank, fractionation and initial Pb (after Stacey and Kramers, 1975). The right-hand column is percentage discordance assuming recent lead losses.

tal points, strongly suggests that the source region is 281 + 344 Ma for the upper and lower intercept, homogeneous chronologically and likely consists es- respectively. The upper intercept is identical to that sentially of ~ 3000 Ma old rocks or that the grains determined for sample Si-ll, and, in a same way, were derived from only one type of source rock. Due could be interpreted as reflecting an average age for to the observed variety of morphological types,(shape rocks of the source area. The lower intercept, within and degree of rounding) we favour the hypothesis of error margins, is not significantly different from a uniform source area. zero. Of the sixteen grains analysed, thirteen plot on or close to a line (dashed line in Fig. 4) connecting 5.3. Quartzite Si-12 the origin and the 2996 Ma intercept. The two most discordant analyses (grain 25 and 26) lie on this line Sixteen grains have been analysed for this sample, and testify to a rather simple history of the U-Pb collected near the middle of the sedimentary pile. U systems of the grains, controlled by recent Pb losses. contents range from 48 to 570 ppm, excepted for This also indicates that the crystals have not suffered grain 33 which presents a higher U concentration significant ancient Pb losses. Three grains, however (1205 ppm). Reported on the concordia diagram (22, 32 and 33), are markedly displaced to the left of (Fig. 4), experimental points show variable degrees this line and lie on a chord that has a non-zero lower of discordance. The least discordant grains (circle on Fig. 4) can be fitted to a discordia line (MSWD-- 10.6), intersecting concordia at 2996 + 27 Ma and Tasmiyele Group ' ~I 0.50 .~ Olekma Block 2soo./j/ ~ t

0.40 - ~ Tasmiyele Group ~i 2000 / / " ~ { 0.50 OlekmaBlock 2500 ~ ~ 0.30 040 ! ~ 2000 ! 0.20 0.30 0.10 020 t ajectory 207pb/235U i1 oo~~ 17 £ 2 4 6 8 10 12 14 16 18 0.10 ~ [ 2425:98 Ma i 2°7pb/235U i Fig. 4. Concordia diagram showing U-Pb results on single zircon 2 6 10 I4 18 grains from the quartzite sample Si-12. Regression of the least discordant grains (circles) provides an average age of 2996 + 27 Fig. 3. Concordia diagram showing U-Pb results on single zircon Ma (heavy line). The dashed line traces a chord between the grains from the quartzite sample Si- 11. "origin and 2996 Ma. 204 O. Bruguier / Precambrian Research 78 (1996) 197-210

loss chord calculated from these analyses provides o 57 ~ Tasmiyele Group 2~oo.¢J/ ' Olekma Block j~ ,~, 51 an upper intercept of 3014 + 59 Ma, the lower inter- - .800 / ~l/ cept (559 + 532 Ma) being close to zero within error margins (MSWD = 39). Analysis 51 (Fig. 5) is only 2% discordant and the 2°7pb/2°6pb age of this grain

2501) c • (2963 + 5 Ma), assuming inheritance is not a factor o 45 24 Quartzites in the interpretation, represents a good estimate of ~oo/ /~ / si-13 :. the age of crystallization of the rock from which it originated. This age places a maximum age constraint on the deposition of the Tasmiyele Group. 8 10 12 14 16 18 Analysis 50 is only 4% discordant and yields a Fig, 5. Concordia diagram showing U-Pb results on single zircon 2°7pb/2°6pb age of 2982 _+ 3 Ma. The 2°7pb/2°6pb grains from the quartzite samples Si-13 and Si-14. Heavy lines are ages determined for the nine grains analysed range 2°Tpb/2°6pb age reference lines. from 2910 to 2988 Ma.

intercept. These grains are among the richest in 6. Discussion uranium (309-1205 ppm) and they may have experienced a more complex Pb loss history compared to This study on the metasediments of the Tasmiyele the main population. They may also have derived group has been made on samples collected at differ- from parent rocks significantly younger than the bulk ent levels of the sedimentary pile in order to obtain grains but older than 2881 Ma (2°7pb/2°6pb age of chronological information on the whole series. Zir- grain 33). Moreover, the observed trend again sug- cons have been selected from different morphologi- gests that the source area is uniform chronologically. cal types allowing access to a great variety of grains The sixteen grains yield apparent 2°7pb/Z°rpb ages and thus to the age spectrum of rocks from the ranging from 2881 to 2999 Ma. source areas. This age spectrum reflects the evolution of the eroded crustal segments and makes it 5.4. Quartzite Si-13 possible to determine the origin of the sediments. Moreover, the results enable the age of deposition of Six zircon grains have been analysed in this sam- the sediments to be constrained and allow a direct ple. They belong essentially to the euhedral grain comparison with other supracrustal rocks from the type (see Table 1). U contents range from 188 to 377 Olekma Block. ppm. Reported on the concordia diagram (Fig. 5), the six grains do not define a simple alignment but scatter along a line (MSWD = 35) whose intersec- 6.1. Age spectrum tions with the concordia curve are 3006 _+ 72 Ma and 429_+ 661 Ma for the upper and lower intercept, Average 2°7pb/2°6pb ages for the bulk of the respectively. The upper intercept value is identical to zircons extracted are consistent along the sedimen- the average age obtained for samples Si-11 and Si-12 tary pile and range from 2996 to 3014 Ma. The and the lower intersection is not significantly differ- distribution of data from quartzites Si-10, Si-ll, ent from zero. The 2°7pb/Z°6pb apparent ages for the Si-13 and Si-14 allows us to infer that the dominant six grains analyzed range from 2911 to 2994 Ma. time for Pb loss from the zircons was post 500 Ma. Zircons from quartzite Si-12 clearly show evidence 5.5. Quartzite Si-14 of strong episode of Pb loss in recent times. More- over, the similarity in 2°Tpb/Z°6pb ages for almost This sample has been collected close to the top of all the grains (see Table 1) indicates a simple Pb loss the sedimentary pile. The nine crystals analysed have history. These observations suggest that for most U concentrations ranging from 90 to 320 ppm. A Pb grains a great proportion of the radiogenic lead loss O. Bruguier / Precambrian Research 78 (1996) 197-210 205

occurred recently, in which case the 2°7pb/2°6pb N=45 [] Si-14 apparent ages could be regarded as a reasonably Mean: 2961_+56 Ma (2 a) ~0ood approximation to the age of crystallization. The Median: 2969 Ma [] Si-13 20 ¸ D [] Si-12 7pb/2°6pb age distribution of detrital zircons from [] Si-ll the Tasmiyele Group is remarkably homogeneous • Si-10 and contrasts with the age spectrum commonly observed for sedimentary formations (Froude et al., e. 1983; Compston and Pidgeon, 1986; Krogh and Kep- pie, 1990; Rainbird et al., 1992; Ross et al., 1992). 10 This distribution has important implications for the tectonic and magmatic evolution of the source areas. Indeed, the 2°Tpb/2°6pb apparent ages of the 51 grains analysed are restricted to a range of 2800- 3100 Ma. The lower limit is due to grains 17 and 33 which present 2°Tpb/z°6pb ages of 2803 Ma and 2881 Ma, respectively, and high U contents (725 and 2815 2845 2875 2905 2935 2965 2995 3025 3055 3085 1205 ppm). Moreover, these analyses are among the (207Pb/206Pb) age in Ma most discordant and it is likely that they may have been prone to loose significant proportions of radio- Fig. 6. Frequency histogram showing the distribution of 2°7pb/ 2°6pb ages of single zircon grain analyses from the Tasmiyele genic lead under relatively mild conditions or by Group quartzite samples. diffusion from radiation-damaged domains. This phenomenon may have occurred sometime in the past in addition to the recent lead losses. We therefore consider that for these grains, the 2°7pb/2°6pb uncertainties and more than 95% within 20- uncer- age is probably not a reasonably good approximation tainties (+ 56 Ma). As pointed out before, because to the age of crystallisation. With the exception of these grains present a simple evolution, controlled by these two experimental points, the age spectrum recent lead losses, the apparent 2°7pb/2°6pb age of spans only 200 Ma and ages range from 2900 to the grains should not be really different from the true 3100 Ma. It is also noteworthy that the age distribu- age of the rock from which they derived. Then, tion is similar in the four samples Si-ll, Si-12, Si-13 assuming that the zircons analysed reflect the evolu- and Si-14, and, to some degree, in sample Si-10. The tion of the eroded source areas, the age spectrum latter, collected at the bottom of the sedimentary indicates that a large proportion of these areas con- pile, exhibits a slightly different age distribution with sists of rocks formed during the period (2961 + 56 four zircon grains showing apparent ages older than Ma). Because zircon grains are relatively scarce in 3040 Ma. However, because of the degree of discor- mafic lithologies, it is very likely that most of the dance shown by experimental points, it is not possi- grains derived from acidic to intermediate rocks. The ble to decide unequivocally whether these four grains period identified above is then likely to represent a derived from source material emplaced during a time of accretion of important volumes of crustal distinct, older event (i.e. > 3100 Ma) or originated material in the source area. Moreover, these results from rocks belonging to one, single protracted event do not show the imprint of the Proterozoic (1900- broadly occurring between 2900 and 3100 Ma. A 2000 Ma) granulite-facies metamorphism that affects 2°7pb//2°6pb age frequency distribution is shown in the West Aldan Block and the eastern part of the Fig. 6. Excluding grains 17 and 33, which exhibit Olekma Block (Nutman et al., 1992). This is sup- younger apparent ages, as well as grains 3, 4, 8 and ported by a simple evolution of the U-Pb systems of 13 whose 2°Tpb/2°6pb ages range from 3040 to 3090 the detrital zircons from the Tasmiyele Group which Ma, the remaining 45 grains exhibit a normal-like provide no evidence of ancient isotopic disturbances. distribution with a geometric mean of 2961 Ma. In This observation is in agreement with the conclu- this distribution, 2/3 of the analyses are within 1 0- sions of Nutman et al. (1992) showing that the 206 O. Bruguier / Precambrian Research 78 (1996) 197-210

1400 tuted by 2900-3000 Ma old rocks. The age spectrum Tasmiyele Group 12(X} • and the inferred characteristics of the analyzed grains are consistent with an origin of the clastic metasedi- 1000 Tonalites ] ~ S'~I'/ [ Granodiorites ments from the neighbouring basement of both the r- 80o Olekma and West Aldan Block. However, the great ; o proportion of rocks with ages ranging from 2900 to 600 o • 3100 Ma on the Olekma Block (Bibikova, 1989; u o Volcanic rocks 400 o Baadsgaard et al., 1990; Nutman et al., 1992; Gle-

20(} o bovitsky and Drugova, 1993; Neymark et al., 1993; Velikostavinsky et al., 1993) and the well-preserved I) i ~ : 0. l (1 2 shapes of most grains suggest that the source rocks 208Pb/206Pb for the metasediments may be located in the adjacent Fig. 7. 2°sPb/2°6pb versus U (ppm) diagram showing data points crystalline basement of the Olekma Block. Whether of single zircon grains from the Tasmiyele Group quartzite sam- the West Aldan Block constitutes a plausible source ples. Fields shown after Bibikova (1984). depends on its position relative to the Olekma Block at the time of deposition. However, several authors (Nutman et al., 1992; Smelov and Beryozkin, 1993) have proposed that these two blocks underwent a effects of this event are restricted to the easternmost separate Archaean evolution and were juxtaposed part of the Olekma Block. These results also agree only during Proterozoic time (1900-2000 Ma). with the proposition that the Aldan Shield is made Therefore, an origin of the sediments from the up of undisturbed Archaean crustal segments that Olekma Block alone, without noticeable contribution have evolved separately and that were welded to- of materials from the West Aldan Block, is more gether during Proterozoic time. likely. The four older grains detected in sample Si-10, assuming that they are truly older than the 6.2. Sediment sources main population, may have originated from ~ 3.2 Ga old rocks recognized on the Olekma Block Bibikova (1984) proposed that the isotopic char- (Putchel et al., 1989b; Nutman et al., 1992) which acteristics of Archaean zircon grains (U contents and have suffered strong lead losses during the ~ 3000 2°sPb/Z°6pb ratio) could be correlated with the na- Ma event. At least, the likelihood that some of the ture of the rock from which they derived. Indeed, the zircon grains analyzed derived from volcanic materi- 2°spb/2°6pb ratio is linearly correlated with the als and their similarity in age with grains derived 23ZTh//Z38u ratio and reflects the Th/U ratio of the from plutonic rocks is consistent with the proposition rock, providing no differential movement of Th and that some greenstone belts and the bulk of the base- U or of uranogenic and thorogenic Pb had occurred. ment on the Olekma Block developed at the same In such a diagram (Fig. 7) most data-points fall in time, as suggested by previous studies (Baadsgaard the field of granodioritic and tonalitic gneisses of et al., 1990; Nutman et al., 1992). magmatic origin. A few analyses, however, are located in the field of volcanic rocks. This distribution suggests that zircons from the Tasmiyele Group orig- 6.3. Constraints on the age of deposition inated mainly from gneissic rocks. The lack of metamorphic grains (high 2°SPb/2°6pb ratio and low U Analyses of detrital zircons from sedimentary content) is also in agreement with the fact that no rocks make it possible to constrain the age of deposi- typical, multifaceted, metamorphic grains have been tion of a sedimentary pile. Indeed, the youngest detected during examination of the zircon concen- detrital zircon in a sediment provides an older limit trate. As seen above, the age distribution for detrital to the deposition (Armstrong et al., 1990; Davis et zircons of the Tasmiyele Group indicates an origin al., 1990; Robb et al., 1989; Krogh and Keppie, of the sediments from a source area mainly consti- 1990). The accuracy of this constraint is dependant O. Bruguier / Precambrian Research 78 (1996) 197-210 207 on a number of factors among which are the fact that material, has been previously associated with the the youngest zircon present in the rock has been Tungurcha Group and interpreted as representing the effectively analyzed, the time interval between the upper part of the Tungurcha Greenstone Belt last magmatic or metamorphic event occurring in the (Bogomolova and Smelov, 1989). However, notice- source areas and the deposition, the velocity of the able differences have already been pointed out by uplift in these areas and the time spent before ero- these authors. The Tungurcha Group, which repre- sion and transport of the sediments in the basin with sents the lower part of the Tungurcha Greenstone the possibility for detrital zircons to pass through Belt, consists of numerous tectonic units. Metasedi- more than one cycle of erosion and deposition. All ments of this group have been deposited and subse- these factors can combine together to precisely con- quently carried onto the basement of the Olekma strain the deposition age. In this study, only few Block before 3016 _ 8 Ma as indicated by the age of concordant to sub-concordant points (less than 5% a tonalitic gneiss intruding the sediments (Nutman et discordance) have been obtained and thus, the con- al., 1992). The Tasmiyele Group, on the contrary, straint on the age deposition is weak. Nevertheless, outcrops in a graben-like structure and it is not grain 45 from sample Si-14 is only 2% discordant certain whether this sedimentary complex is really and presents a 2°7pb/2°6pb age of 2963 ___ 5 Ma part of a greenstone belt (Bogomolova and Smelov, which constitutes a maximum age for the deposition 1989). From this study, U-Pb single zircon grain of the Tasmiyele Group. Moreover, Bogomolova and results indicate that most of the analyzed crystals Smelov (1989) have indicated a 2500 Ma age for yield 2°7pb/2°6pb ages ranging from 2900 to 3000 diabases intruding the metasediments. This age con- Ma and grain 51 provides a maximum age of ~ 2960 stitutes a minimum value for deposition of the sedi- Ma for the deposition of the sediments. This clearly ments and is in good agreement with a 2420 Ma shows that deposition of the Tasmiyele Group took minimum U-Pb zircon age obtained from the two- place at least 50 Ma after that of the Tungurcha mica granite intruding the east flank of the Tas- Group. These two formations are therefore likely to miyele Group (Bruguier, 1993). These values com- be two distinct units. This study also allows a direct bined suggests that deposition of the Tasmiyele comparison with the Olondo Greenstone Belt from Group took place during Archaean time, between which geochronological results have been published. 2500 and 2960 Ma. Previous geochronological data According to these results, the formation of this from the Olekma Block indicate three main mag- typical greenstone belt occurred between 2966_ 16 matic periods at 2.70-2.75 Ga (Nutman et al., 1992), Ma (Putchel et al., 1989a) and 3006+ 11 Ma 2.9-3.0 Ga (Bibikova, 1989; Putchel et al., 1989a; (Baadsgaard et al., 1990). Therefore, sedimentary Baadsgaard et al., 1990) and before 3.2 Ga (Putchel units of the Tasmiyele Group were deposited at the et al., 1989b; Nutman et al., 1992). The two last same time or after those of the Olondo Greenstone magmatic periods (2.9-3.0 Ga and > 3.2 Ga) obvi- Belt. ously took place before deposition of the sediments, but the lack of zircon grains with minimum ages younger than 2800 Ma suggests that the Tasmiyele Group was already deposited before the third 2.70- 7. Conclusions 2.75 Ga magmatic period.

Conclusions from this study on single detrital zircon grains from five samples of the Tasmiyele 6.4. Correlation with different units of the Olekma Group are as follows: Block (1) The age of deposition of the Tasmiyele Group is bracketed by the age of the youngest concordant to The Tasmiyele Group outcrops in the western part sub-concordant grain analysed (2963 + 5 Ma, this of the Tungurcha Greenstone Belt and, due to its study) and that of intrusive material (~ 2500 Ma, geographical location and the scarcity of volcanic Bogomolova and Smelov, 1989). Geochronological 208 O. Bruguier /Precambrian Research 78 (1996) 197-210

data from the Olekma Block also suggest that depo- Acknowledgements sition probably occurred before 2.70-2.75 Ga. (2) The age spectrum is coherent with a local This work was carried out while the author was in origin for the sediments. No 'exotic' origin is re- receipt of a M.R.T. grant. I would like to thank D. quired to explain the age spectrum presented by the Bosch and J.R. Lancelot for constructive criticisms zircon populations. It is proposed that most of the and helpful comments on the paper. The paper bene- detrital zircons found in the metasediments origi- fited greatly from reviews of R.T. Pidgeon and two nated from gneissic source rocks which form the anonymous reviewers. This research is a contribution bulk of the Olekma Block. Some grains possibly to the IGCP Project 280 'The Oldest Rocks on derived from volcanic material from adjacent green- Earth'. The samples have been collected by J.P. stone belts, Respaut during the 1989 IGCP field trip in Siberia. (3) This part of the Aldan Shield has experienced an important crustal event ~ 3.0 Ga ago. This period is now well identified in many Archaean cratons References (Sino-Korean, Kaapvaal and Yilgarn cratons; see for

example Zhang et al., 1984; KrSner et al., 1989; Armstrong, R.A., Compston, W., De Witt, M.J. and Williams, Pidgeon and Wilde, 1990) and is believed to repre- I.S., 1990. The stratigraphy of the 3.5-3.2 Ga Barbenon sent a worldwide event of crust formation (Jahn et Greenstone Belt revisited: a single zircon ion microprobe al., 1991). Since then, this Archaean segment has not study. Earth Planet. Sci. Lett., 101: 90-106. undergone major geological perturbations. The ~ 2.0 Baadsgaard, H., Nutman, A.P. and Samsonov, A.V., 1990. Geochronology of the Olondo Greenstone Belt. 7th Int. Conf. Ga high-grade event, found in the west Aldan Block, Geochronology, Cosmochronology and Isotope Geology. Geol. is not recorded by our data indicating that its effects Soc. Aust. Abstr., 27: 6. do not extend to this part of the Olekma Block, as Bibikova, E.V., 1984. The most ancient rocks in the U.S.S.R. pointed out by previous authors (Nutman et al., Territory by U-Pb data on accessory zircons. In: A. Kr~Sner, 1992). G.N. Hanson and A.M. Goodwin (Editors), Archaean Geochronology. Springer Verlag, Berlin, pp. 235-250. (4) This study also failed in the search for ancient Bibikova, E.V., 1989. U-Pb ages of the metavolcanics from the (Early Archaean) material in the Aldan Shield. The Olondo Greenstone Belt. In: V.L. Dook, L.A. Neymark and oldest grains identified present 2°7pb/2°6pb ages V.A. Rudnick (Editors), The oldest Rocks of the Aldan- ranging from 3040 to 3090 Ma and may be attributed Stanovik Shield, Eastern Siberia, USSR. Soviet Committee for to ~ 3.2 Ga old source material having suffered 1GCP Project 280, pp. 12-13. Bibikova, E.V., Morozova, 1.M., Gracheva, T.V. and Makarov, strong lead losses during the ~ 3.0 Ga event. As V.A., 1989. U-Pb age of granulites from the Kurulta Com- target materials for a search for ancient witnesses, plex. In: V.L. Dook, L.A. Neymark and V.A. Rudnick (Edi- quartzites from the Tasmiyele Group have not the tors), The Oldest Rocks of the Aldan-Stanovik Shield, Eastern required qualities. Indeed, their local derivation re- Siberia, USSR. Soviet Committee for IGCP Project 280, pp. duces the likelihood that the source areas may con- 89-91 Bogomolova, L.M. and Smelov, A.P., 1989. The Tungurcha tain very old zircons. A more mature sediment, greenstone belt. In: V.L. Dook, L.A. Neymark and V.A. sampling wider areas of Archaean surfaces would Rudnick (Editors), The Oldest Rocks of the Aldan-Stanovik have been more promising. Moreover, the impor- Shield, Eastern Siberia, USSR. Soviet Committee for IGCP tance of the ~ 3.0 Ga event in the Olekma Block Project 280, pp. 35-42. also suggests that older zircons may have been de- Briqueu, L. and De La Boisse, H., 1990. U Pb geochronology: systematic development of mixing equations and application stroyed. of Montc Carlo numerical simulation to the error propagation (5) The Tasmiyele Group and the Tungurcha in the Concordia diagram. Chem. Geol., 88: 69-83. Group, initially grouped together to form the Tun- Bruguier, O., 1993. Applications de la gEochronologie U Pb sur gurcha Greenstone Belt, are more probably two dis- monocristal de zircon abras6 en domaine sEdimentaire ct tinct units deposited at different times and without magmatique: source des matrriaux drtritiques, tdmoins ArchEens cmstaux et g~odynamique globale. Thesis. Univ. relationship to one another. The sediments of the Montpellier I1, 330 pp. Tasmiyele Group were deposited at the same time or Brnguier, O., Dada, S.S. and Lancelot, J.R., 1994. Early Archaean after those of the Olondo Greenstone Belt. component (> 3.5 Ga) within a 3.05 Ga orthogneiss from O. Bruguier / Precambrian Research 78 (1996) 197-210 209

northern Nigeria: U-Pb zircon evidence. Earth Planet. Sci. tonalite-trondhjemitic gneisses from the probable basement. Lett., 125: 89-103. In: V.L. Dook, L.A. Neymark and K.A. Rudnick (Editors), Compston, W. and Pidgeon, R.T., 1986. Jack Hills, evidence of The Oldest Rocks of The Aldan-Stanovik Shield, Eastern more old detrital zircons in Western Australia. Nature, 321: Siberia, U.S.S.R. Soviet Committee for the IGCP Project 280, 766-769. pp. 98-102. Davis, D.W., Pezzutto, F. and Ojakangas, R.W., 1990. The age Neymark, L.A., Kovach, V.P., Nemchin, A.A., Morozova, I.M., and provenance of metasedimentary rocks in the Quetico Kotov, A.B., Vinogradov, D.P., Gorokhovsky, B.M., Ovchin- Subprovince, Ontario, from single zircon analyses: implica- nikova, G.V., Bogomolova, L.M. and Smelov, A.P., 1993. tions for Archaean sedimentation and tectonics in the Superior Late Archaean intrusive complexes in the Olekma granite- Province. Earth Planet. Sci. Lett., 99: 195-205. greenstone terrain (eastern Siberia): geochemical and isotopic Dook, V.L., Neymark, L.A. and Rudnick, K.A., 1989. The oldest study. Precambrian Res., 62: 453-472. rocks of the Aldan-Stanovik Shield, Eastern Siberia, USSR. Nutman, A.P., Chernyshev, I.V., Baadsgaard, H. and Smelov, Soviet Committee for IGCP Project 280. Excursion Guide for A.P., 1992. The Aldan Shield of Siberia, USSR: the age of its Geological Field Trip to the Aldan-Stanovik Shield. Archaean components and evidence for widespread reworking Leningrad-Mainz, 121 pp. in the mid-Proterozoic. Precambrian Res., 54: 195-210. Froude, D.O., Ireland, T.R., Kinny, P.D., Williams, I.S., Comp- Pidgeon, R.T. and Wilde, S.A., 1990. The distribution of 3.0 Ga ston, W., Williams, I.R. and Myers, J.S., 1983. Ion micro and 2.7 Ga volcanic episodes in the Yilgam Craton of Western probe identification of 4100-4200 Myr-old terrestrial zircons. Australia. Precambrian Res., 48:309-325 Nature, 304: 616-618. Putchel, I.S., Samsonov, A.V., Simon, A.K. and Zhuravlev, D.Z., Glebovitsky, V.A. and Drugova, G.M., 1993. Tectonothermal 1989a. Petrology, geochemistry and Sm-Nd age of metavol- evolution of the Western Aldan Shield, Siberia. Precambrian canics from the Olondo Greenstone Belt, the Tungurcha Res., 62: 493-505. Greenstone Belt. In: V.L. Dook, L.A. Neymark and K.A. Jahn, B.M., Capdevila, R., Gruau, G., Bemard-Griffiths, J. and Rudnick (Editors), The Oldest Rocks of The Aldan-Stanovik Cornichet, J., 1991. The granite-greenstones terranes of the Shield, Eastern Siberia, U.S.S.R. Soviet Committee for the Aldan Shield, Siberia: geochemical characterisation, ages and IGCP Project 280, pp. 27-34. petrogenesis. Terra Abstracts, EUG V1, 3: 196-197. Putchel, I.S., Samsonov, A.V., Simon, A.K. and Zhuravlev, D.Z., Krogh, T.E., 1982. Improved accuracy of U-Pb zircon ages by 1989b. Geology, geochemistry and Sm-Nd geochronology of the creation of more concordant systems using an air abrasion the oldest rocks from the Olekma Granite Greenstone Terrain. technique. Geochim. Cosmochim. Acta, 46: 637-649. In: V.L. Dook, L.A. Neymark and K.A. Rudnick (Editors), Krogh, T.E. and Keppie, J.D., 1990. Age of detrital zircon and The Oldest Rocks of The Aldan-Stanovik Shield, Eastern titanite in the Meguma Group, Southern Nova Scotia, Canada: Siberia, U.S.S.R. Soviet Committee for the IGCP Project 280, clue to the origin of the Meguma Terrain. Tectonophysics, pp. 57-62. 177: 307-323. Rainbird, R.H., Heaman, L.M. and Young, G., 1992. Sampling KrOner, A., Compston, W. and Williams, I.S., 1989. Growth of Laurentia: detrital zircon geochronology offers evidence for an early Archaean crust in the Ancient Gneiss Complex of Swazi- extensive Neoproterozoic river system originating from the land as revealed by single zircon dating. Tectonophysics, 161: Grenville orogen. Geology, 20: 351-354. 271-298. Robb, L.J., Davis, D.W. and Kamo, S.L., 1989. U-Pb ages on Lancelot, J.R., Vitrac, A. and Allegre, C.J., 1976. Uranium and single detrital zircon grains from the Witwatersrand basin, lead isotopic dating with grain-by-grain zircon analysis: a South Africa: constraints on the age of sedimentation and on study of complex geological history with a single rock. Earth the evolution of granites adjacent to the basin. J. Geol., 98: Planet. Sci. Lett., 29: 357-366. 311-328. Ludwig, K.R., 1987. lsoplot 200, a plotting an regression program Ross, G.M., Parrish, R.R. and Winston, D., 1992. Provenance and for isotope geochemists, for use with HP series 200 comput- U-Pb geochronology of the Mesoproterozoic Belt Supergroup ers. U.S. Geol. Surv. Open-File Rep., 105: 85-513. (Northwestern United States): implications for age of deposi- Maas, R. and McCulloch, M.T., 1991. The provenance of Ar- tion and pre-Panthalassa reconstruction. Earth Planet. Sci. chaean clastic metasediments in the Narryer Gneiss Complex, Lett., 113: 57-76. Western Australia: Trace element geochemistry, Nd isotopes Schiirer, U. and Allbgre, C.J., 1982. Investigation of the Archaean and U-Pb ages for detrital zircons. Geochim. Cosmochim. crust by single-grain dating of detrital zircons: a graywacke of Acta, 55: 1915-1932. the Slave Province, Canada. Can. J. Earth Sci., 19: 1910-1918. Morozova, I.M., Bogomolov, E.S., Jakovleva, S.Z., Beliatsky, Smelov, A.P., 1989. 'Exotic' mafic rocks of the Olekma Com- B.V. and Berezhnaya, N.G., 1989a. U-Pb dating of granite- plex. In: V.L. Dook, L.A. Neymark and K.A. Rudnick (Edi- gneisses and tonalite-trondhjemitic gneisses. In: V.L. Dook, tors), The Oldest Rocks of The Aldan-Stanovik Shield, East- L.A. Neymark and K.A. Rudnick (Editors), The Oldest Rocks ern Siberia, U.S.S.R. Soviet Committee for the IGCP Project of The Aldan-Stanovik Shield, Eastern Siberia, U.S.S.R. So- 280, pp. 55-56. viet Committee for the IGCP Project 280, pp. 105-107. Smelov, A.P. and Beryozkin, V.I., 1993. Retrograded eclogites in Morozova, I.M., Ovchinnikova, G.V., Neymark, L.A. and Levsky, the Olekma granite-greenstone region, Aldan Shield, Siberia. L.K., 1989b. U-Pb and Pb-Pb dating of hypersthene-biotite Precambrian Res., 62: 419-430. 210 O. Bruguier / Precambrian Research 78 (1996) 197-210

Stacey, J.S. and Kramers, J.D., 1975. Approximation of terrestrial R.I. and Rudnik, V.A., 1993. Geochemical mapping of basic lead isotope evolution by a two stage model. Earth Planet. Sci. complexes in the early Precambrian Aldan-Stanovoy Shield Lett., 6: 15-25. of Siberia. Precambrian Res., 62: 507-525. Steiger, R.H. and J~ger, E., 1977. Subcommission on geochronol- Zhang, Z.M., Liou, J.G. and Coleman, R.G., 1984. An outline of ogy: convention on the use of decay constants in geo- and the plate tectonics of China. Geol. Soc. Am. Bull., 95: 295- cosmochronology. Earth Planet. Sci. Lett., 36: 359-362. 312. Velikoslavinsky, S.D., Tolmacheva, E.V., Dook, V.L., Milkevich, Earth and Planetary Science Letters 152Ž. 1997 217±231

U±Pb dating on single detrital zircon grains from the Triassic Songpan±Ganze flyschž/ Central China : provenance and tectonic correlations

O. Bruguier a,), J.R. Lancelot a, J. Malavieille b a Laboratoire de GeochimieÂÂÁ Isotopique, CNRS-UMR 5567, UniÕersite de Montpellier II, case 066, Place Eugene Bataillon, Montpellier, Cedex 5 34095 France b Laboratoire de GeophysiqueÂÂÁ et Tectonique, CNRS-UMR 5573, UniÕersite de Montpellier II, case 060, Place Eugene Bataillon, Montpellier, Cedex 5 34095 France Received 1 May 1997; revised 4 August 1997; accepted 4 August 1997

Abstract

The Songpan±Ganze flysch beltŽ. Central China covers a huge triangular area of more than 200,000 km2 and is bounded by the continental blocks of South China, North China and the Tibetan plateau. Detrital zircons extracted from three flysch samples collected in the central part of the belt were analyzed grain by grain using the U±Pb method. Two samples of Middle Triassic sandstones, collected at different locations in the belt, provide identical results, which suggests similar source regions. The detrital zircons yield a wide range of ages and indicate their principal derivation from Mid-Proterozoic Ž.1.8±2.0 Ga source rocks with minor contribution from late Archean Ž ca. 2.5±2.6 Ga . material. The discordance and Pb loss patterns from low-U zircons indicate disturbances during a subsequent event which may be of Caledonian age, as suggested by concordant zircon grains at ca. 420 and 450 Ma. One sample collected within the Palang Shan Pass zone provides concordant zircon grains at around 230 MaŽ. 231"1 Ma and 233"1 Ma . These Triassic ages are synchronous to flysch deposition and suggest intense geological activityŽ. calc-alkaline volcanism? at that time in the area close to the basin. The data support an origin of the clastic material mainly from a northeastern landmass, corresponding to the southern margin of the Sino±Korean craton. To a lesser degree, inputs from the Yangtze craton and possibly from the northern margin of the basinŽ. Kunlun arc are also detected. The age spectrum from the Upper Triassic sandstone is significantly different and shows predominance of SinianŽ. ca. 760 Ma grains, probably derived from the Yangtze craton. This change in the source region is interpreted as reflecting the tectonic evolution of this area and in particular as being linked to the late Triassic collision between South China and North China. In the Middle Triassic, while subduction of the Songpan sea northward beneath the North China plate was still taking place, continental subduction of South China in the Dabie region was responsible for uplift of the overriding plateŽ. i.e. the Sino±Korean craton which supplied large volumes of sediments. During the Late Triassic, clockwise rotation of the South China block uplifted the Indo-Sinian part of the Qinling belt and closed the basin. As the accretionary wedge was thickening along the southern margin of North China, detritus derived from this source region were unable to reach the flysch basin. The age spectrum presented by detrital zircons indicates

) Corresponding author. E-mail: [email protected]

0012-821Xr97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S0012-821XŽ. 97 00138-6 218 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 predominance of Sinian material derived from source area located on the northern margin of the Yangtze craton; a source region which was until this period swamped by Luliang material from the Sino±Korean craton. q 1997 Elsevier Science B.V.

Keywords: zircon; UrPb; PbrPb; provenance; flysch; tectonics

1. Introduction corresponding to a minimum volume of detrital material of about 2.0=1063 kmŽ. Fig. 1a . Sediments in During the last decade much attention has been the basin are almost exclusively marine Triassic focused on the displacement of numerous continental flysch deposits, which may have locally reach a blocks accreted to the Siberian cratonŽ. Fig. 1a dur- maximum thickness of 15 kmwx 4 . The Paleozoic ing the Phanerozoic and now forming the `Asian sequence, 4000±6000 m thick, is represented by puzzle'wx 1 . The Songpan±Ganze Triassic beltŽ Fig. Cambrian to Permian sediments corresponding 1b. of central China is a key area because it is mainly to shales and carbonateswx 4 . The basin is now located at the junction between a number of litho- bounded by the continental blocks of South China spheric plates. Sediments deposited in the Triassic Ž.SCB , North China Ž NCB . , and Tibetan plateau Songpan flysch basin have therefore been supplied Ž.Fig. 1a . The development of the basin has been by the various emergent continental masses subject related to extensional processes resulting in fragmen- to erosion. However, so far, the source of the sedi- tation of Gondwanaland during the early Permianwx 5 ments as well as the mechanism responsible for and responsible for the creation of the Paleo-Tethys. accumulation of such huge amounts of detritus are This period was coeval with the important Emeishan still unclear. Nie et al.wx 2 argued that Middle to basaltic volcanismwx 6 and with the beginning of Upper Triassic flysch sequences of the Songpan± continental rifting in the Panxi-rift area to the east Ganze basin were derived from denudation of the wx7 , in the South China plate. Basaltic volcanics are eastern part of the orogenic belt between North also known in the Songpan±Ganze basin where they China and South China blocks, as a result of ex- intrude Permian sedimentswx 4 . By the Early Triassic, humation of ultra-high pressureŽ. UHP metamorphic the basin enlarged and may have reach a width of at rocks of the Dabie Shan and Shandong regions. This least 500 km. Although the exact nature of the attractive hypothesis has however been questioned basementŽ. oceanic or continental below the Triassic wx3 . The aim of this paper is to examine the age sediments is not well known, significant shortening spectrum preserved by detrital zircons from three suggests that most of the lithosphere was subducted sandstone samples collected in the central part of the during compressional Indo-Sinian tectonics in the basinŽ. Fig. 1c and to determine the source area of late Triassic and early Jurassic. Subduction of the the sediments. Zircon is resistant to physical and Songpan sea under the continental block of North mechanical degradation and ages derived from detri- China occurred during the Middle Triassic, which tal zircon analyses reflect the characteristic age spec- was responsible, in the Late Triassic, for folding and trum of primary source rocks. Combined with the thrusting of the sediments southward, finally leading age information from the various blocks adjacent to to overthrusting of a huge accretionary wedge onto the basin, the new zircon age data can be used for the margin of the South China Blockwx 8 . assessing possible relationships between erosion and tectonics in the source areas feeding the basin. 2.2. Potential sources for the sediments Each block surrounding the basin during the Tri- 2. Geological setting assic period is a priori a potential source for the clastic material accumulatedŽ. Fig. 1a . Finding out 2.1. General presentation which blocks supplied the detritus therefore requires The Songpan±Ganze fold belt represents the most a knowledge of the geochronological history of these important flysch basin in the world and covers a various continental land masses. Although huge triangular area of more than 200,000 km2, geochronological data are relatively scarceŽ except O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 219 notes Triassic unding continental Ž. Ž. .Ž. wx Ž blocks after 8 . c Simplified geology of the eastern part of the Songpan±Ganze area showing the main geological features and sample locations. Blank de sandstones. Fig. 1. Composite geological map of Central China. a Location of the study area in Asia. b Structural sketch map of the Songpan±Ganze fold belt and surro 220 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 maybe for the NCB. the evolution of the various margins by the Hercynian Kunlun fold belt and to blocks is broadly knownŽ for a review, seewx 9. . The the north by the Tianshan Range. NCBŽ. or Sino±Korean craton presents a complex tectonothermal history, starting in the early Archean. In the northeastern part of the craton, Archean and 3. Analytical techniques Early Archean Ž.)3.5 Ga material has been identi- fiedwx 10±12 with ages up to 3.8 Ga in the Liaoning Zircons were separated from 5±10 kg rocks fol- and Hebei provinces. This Archean terrain has a lowing standard techniquesŽ e.g.wx 36. and, in order complex crustal evolution and the time range be- to minimize discordant zircon, the use of air abrasion tween 2.5 and 2.7 Ga, particularly at ca. 2.5 Ga wx37 has been applied to all crystals analysed in this during the Qianxi event, was the most important study. Single grains were subsequently processed period of geological activity and crustal growthw 13± according towx 38 . Isotopic measurements were car- 20x . A large part of the craton, however, comprises ried out on a VG Sector mass spectrometer using a younger material of Proterozoic ageŽ. 1.8±2.0 Ga Daly detector for small sample loads. A mass dis- which was related to a major crust-forming event crimination correction for measurements on the Daly wx9,15,21±23 . This period Ð called the Luliang detector of 0.15"0.05% per amu was determined orogeny Ð is coeval with other crust-forming events from NBS common leadŽ. NBS981 and uranium worldwidewx 24 . On its southern margin, the Sino± Ž.U500 standards. Total Pb blanks over the period of Korean craton is limited by the Caledonian Qilian the analyses range from 10 to 30 pg and uranium Shan belt which can be followed in an E±W direc- blanks were less than 5 pg. The isotopic composition tion along the Tarim block. The NCB is sutured to of radiogenic Pb was determined by subtracting first the SCB by the E±W trending Qinling±Dabie oro- the blank Pb and then the remainder, assuming a genic belt. This collision belt presents a complicated common Pb composition at the time of initial crys- tectonic evolution, probably starting in the Caledo- tallisation determined from the model ofwx 39 . Calcu- nian with the amalgamation of microcontinents and lations were made using the programme ofwx 40 . island arc terraneswx 25±27 . Tectonic activities cul- Analytical uncertainties are listed as 2s and uncer- minated during the late Triassicwx 28±32 , involving tainties in ages as 95% confidence levels. collision and final suturing of the two cratons. The SCB is, with the NCB, the most important unit composing eastern China. It mainly comprises 4. U±Pb isotopic results the Yangtze craton, consisting of Proterozoic material formed during the Yangtze orogenyŽ 750±850 Samples were collected in the central part of the Ma. , although unexposed Archean rocks may also be basinŽ. Fig. 1c and correspond to flyschoid sand- presentwx 23 . The Yangtze orogeny probably repre- stones belonging to the Middle TriassicŽ CDU 21 sents a major event leading to new crust addition and and CDU 27. and to the Upper Triassic sequence continental growth of the cratonwx 33 . Shi and co- Ž.CDU 58 . Zircon grains show a wide range in size workerswx 34 also proposed the subduction of an and morphology. They are dominated by light pink oceanic plate during the Jinningian periodŽ 800±1000 to colourless translucent crystals, the morphology of Ma. , which could have been responsible for calc-al- which range from prismatic euhedral crystals, with kaline plutonism along the northernmost part of the undamaged faces and corners, to rounded zircons Yangtze craton. The occurrence of such material is showing signs of abrasion. Crystals with well pre- supported by a geochemical study of volcanic and served shapes are likely to indicate a short sedimen- plutonic rocks, probably Middle to Late Proterozoic tary transport. Conversely, the occurrence of rounded in age, and covered by platform sedimentary rocks of grains indicates that they have been transported over Sinian agewx 35 . Another prominent unit is the Tarim a long distance andror that they may have survived craton, on the west of the NCB, whose basement more than one cycle of erosion and deposition. U±Pb includes Archean and Proterozoic rocks of Luliang results on 52 air-abraded single zircon grains are agewx 9 and is surrounded on its southern and eastern reported in Table 1 and shown in Figs. 2 and 3. O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 221

4.1. Middle Triassic samples() CDU 27 and CDU 21 230 Ma to ca. 2500 MaŽ. Fig. 2 . This wide range of ages emphasises the great variety of rocks in the The 44 zircon grains from the Middle Triassic source areaŽ. s . Broadly, the grains fall into four sandstones yielded apparent ages ranging from ca. groups according to their apparent Pb±Pb ages and

Fig. 2. Concordia plot for detrital zircons from the Middle Triassic samples. Insets a,b,c and d are enlargements between 1500 and 2000, 380 and 480, 210 and 280, and 600 and 900 Ma, respectively. Polygons are 2s error. 222 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 Pb 02 05 10 07 09 08 03 04 03 03 07 03 01 03 02 02 03 03 02 43 12 21 26 14 08 08 r " " " " " " " " " " " " " " " " " " " " " " " " " " UPb 06 2518 05 2089 09 1874 06 1786 04 1880 07 1762 02 1930 03 1864 08 1829 05 2373 02 1868 05 2535 03 1862 04 2296 07 1909 03 1833 02 2277 07 409 05 1660 02 529 03 1103 03 425 03 266 04 469 02 742 02 447 r " " " " " " " " " " " " " " " " " " " " " " " " " " Ž. UPb 238 207 235 206 12 2480 08 1990 16 1824 08 1690 04 1682 09 1455 04 1796 06 1776 14 1788 09 2193 04 1817 10 2503 06 1836 07 2135 13 1896 06 1632 04 1889 02 390 02 864 01 465 03 729 02 420 01 262 02 438 01 674 02 434 r " " " " " " " " " " " " " " " " " " " " " " " " " " Pb Apparent ages Ma 206 207 24 2434 108 386 35 1896 58 587 45 1781 31 453 52 1615 30 614 22 1528 20 1254 51 419 16 1683 57 261 35 432 22 654 18 1702 43 1754 30 2006 10 1772 23 2464 16 1814 20 432 15 1971 25 1884 17 1481 14 1556 " " " " " " " " " " " " " " " " " " " " " " " " " " r Pb s 2 65 0.16598 55 0.16771 10 0.05491 33 0.12931 10 0.10196 57 0.11465 04 0.05797 32 0.10917 06 0.07630 16 0.11502 26 0.10775 06 0.05531 15 0.11825 04 0.05157 05 0.05642 04 0.06400 21 0.11400 48 0.11181 44 0.15237 12 0.11426 20 0.11387 03 0.05586 29 0.14570 43 0.11689 18 0.11205 15 0.14409 Ž. " " r " " " " " " " " " " " " " " " " " " " " " " " " Pb s 2 09 4.241 28 10.500 03 0.468 16 6.096 05 1.341 33 5.028 02 0.581 15 4.284 04 1.051 17 3.191 03 0.512 08 4.865 01 0.294 04 0.539 03 0.942 12 4.750 29 4.819 19 7.669 07 4.984 23 10.764 12 5.102 02 0.534 14 7.185 26 5.470 11 3.990 07 5.423 " " " " " " " " " " " " " " " " " " " " " " " " " " r Pb s 2 Ž. Ž. Ž. r Pb r 206204 208 206 206 238 207 235 207 206 mg ppm ppm Pb Pb U U Pb Ž. Ž.Ž. Xiaoyin CDU 27 Palang Shan Pass CDU 21 26. Zr. lp, elg 0.007 160 90 1605 0.249 0.4588 25. Zr. co, elg 0.006 233 15 200 0.197 0.0618 24. Zr. lp, rd 0.006 297 107 1119 0.077 0.3419 23. Zr. co, rd 0.006 274 29 469 0.191 0.0954 22. Zr. co, cu 0.004 97 30 713 0.086 0.3181 21. Zr. co, rd 0.018 276 20 683 0.092 0.0727 20. Zr. co, eu 0.017 52 16 571 0.153 0.2846 19. Zr. co, rd 0.016 306 31 752 0.136 0.0999 18. Zr. lp, rd 0.014 230 67 890 0.173 0.2674 17. Zr. co, rd 0.011 391 84 1423 0.066 0.2148 16. Zr. co, elg 0.010 318 23 419 0.230 0.0672 15. Zr. lp, rd 0.009 517 153 1397 0.062 0.2984 14. Zr. co, eu 0.009 591 25 505 0.110 0.0413 13. Zr. co, rd 0.007 490 36 587 0.196 0.0693 12. Zr. co, eu 0.006 808 84 1003 0.072 0.1068 11. Zr. lp, eu 0.006 357 107 963 0.053 0.3022 10. Zr. co, rd 0.006 81 26 475 0.152 0.3126 9. Zr. co, elg 0.006 241 101 671 0.197 0.3651 8. Zr. lp, eu 0.006 679 208 2424 0.022 0.3164 7. Zr. lp, eu 0.006 212 110 1191 0.145 0.4655 6. Zr. lp, rd 0.005 425 143 1975 0.105 0.3249 5. Zr. lp, rd 0.005 1627 121 1352 0.204 0.0693 4. Zr. lp, rd 0.005 896 322 4325 0.033 0.3576 3. Zr. lp, elg 0.005 146 54 716 0.184 0.3394 2. Zr. lp, elg 0.005 443 114 1383 0.060 0.2583 Conventional U±Pb data for detrital zircons from quartzite samples of the Songpan±Garze basin east-central China Table 1 1. Zr. lp, rd 0.004 1154 329 2758 0.069 0.2730 Sample Weight U Pb Pb O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 223 03 06 06 02 03 04 02 03 03 03 02 05 02 02 04 02 10 07 20 11 21 29 12 13 59 12 " " " " " " " " " " " " " " " " " " " " " " " " " " 04 1855 05 2135 05 1818 03 1830 04 1795 04 1721 05 1816 04 1983 05 1978 03 1823 03 1810 05 1918 04 2448 04 1875 05 1906 03 2513 03 758 02 752 06 754 04 965 06 752 03 249 03 914 04 745 05 200 02 452 " " " " " " " " " " " " " " " " " " " " " " " " " " 05 1797 07 1655 05 1756 05 1642 06 1445 09 1724 07 1911 03 1890 05 1780 04 1638 09 1879 09 2358 07 1796 09 1862 06 2425 03 719 02 631 03 683 03 747 03 673 04 1418 01 234 04 887 03 725 01 228 02 409 " " " " " " " " " " " " " " " " " " " " " " " " " " and refer to last digits.

a 33 707 20 598 60 662 37 677 22 1747 60 649 43 990 34 1530 15 1695 64 233 18 1525 25 1265 12 1648 22 1845 25 1811 14 1743 41 877 41 719 17 1507 31 1843 24 2255 127 231 12 1729 31 402 28 1822 19 2322 " " " " " " " " " " " " " " " " " " " " " " " " " " . wx Ž 07 0.06450 04 0.06431 10 0.06436 08 0.07128 19 0.11345 11 0.06430 18 0.13281 26 0.11113 19 0.11184 04 0.05118 15 0.10973 18 0.10539 28 0.11104 26 0.12184 23 0.12148 19 0.11143 10 0.06950 08 0.06409 15 0.11065 33 0.11748 42 0.15929 07 0.05011 21 0.11468 04 0.05600 31 0.11670 34 0.16556 " " " " " " " " " " " " " " " " " " " " " " " " " " 05 1.031 03 0.861 04 0.959 05 1.088 10 4.869 06 0.939 07 3.040 14 4.105 11 4.638 02 0.259 10 4.038 11 3.150 18 4.460 14 5.566 12 5.433 11 4.770 06 1.396 05 1.042 09 4.016 18 5.361 18 9.197 01 0.252 13 4.863 03 0.497 17 5.254 14 9.897 " " " " " " " " " " " " " " " " " " " " " " " " " " elongated with smooth edges. All zircons are taken from the least magnetic fractions and air s euhedral; elg s rounded; eu s colourless; rd s light pink; co s wx zircon; lp s 52. Zr. co, eu 0.009 242 37 707 0.480 0.1159 51. Zr. co, eu 0.007 441 45 1018 0.169 0.0971 50. Zr. co, eu 0.006 242 32 343 0.375 0.1081 49. Zr. lp, eu 0.004 343 43 570 0.277 0.1107 48. Zr. lp, elg 0.004 591 197 1315 0.145 0.3113 47. Zr. co, eu 0.004 193 23 307 0.231 0.1059 46. Zr. lp, rd 0.003 418 77 513 0.145 0.1660 Yajiang CDU 58 45. Zr. lp, eu 0.003 344 97 611 0.141 0.2679 44. Zr. lp, rd 0.019 301 96 2144 0.131 0.3008 43. Zr. co, elg 0.017 272 11 360 0.193 0.0368 42. Zr. co, eu 0.016 203 53 1341 0.043 0.2669 41. Zr. co, rd 0.015 182 43 1412 0.184 0.2168 40. Zr. lp, eu 0.014 528 152 5117 0.046 0.2914 39. Zr. co, rd 0.013 166 56 1314 0.070 0.3313 38. Zr. lp, rd 0.012 214 71 1766 0.084 0.3244 37. Zr. lp, rd 0.010 326 106 2573 0.118 0.3105 36. Zr. co, elg 0.010 170 25 476 0.108 0.1456 35. Zr. lp, rd 0.010 243 30 577 0.168 0.1180 34. Zr. lp, elg 0.010 788 208 1838 0.063 0.2633 33. Zr. lp, eu 0.009 222 83 1748 0.210 0.3309 32. Zr. co, elg 0.009 340 163 2081 0.167 0.4188 31. Zr. co, eu 0.009 425 16 246 0.116 0.0365 30. Zr. lp, elg 0.009 477 144 1846 0.039 0.3076 29. Zr. lp, elg 0.009 604 36 775 0.031 0.0643 28. Zr. lp, rd 0.008 184 71 724 0.279 0.3265 27. Zr. lp, rd 0.008 854 403 3505 0.094 0.4336 zr abraded 37 . Lead isotopic ratios have been corrected for fractionation, blank and initial common Pb after 39 . Errors are 2 224 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 degree of discordance. Almost half of the grains A third group, shown on inset b of Fig. 2, is have minimum 207 Pbr 206 Pb ages ranging from 1700 represented by grains with apparent ages ranging Ma to 2000 Ma indicating that, in the Middle Trias- from 408 to 529 Ma. One colourless zircon grain sic, the main contribution was from source material Ž.analyses 16 gave a concordant data with a Pb±Pb of Luliang ageŽ. inset a in Fig. 2 . The relatively high age of 425"21 Ma and analysis 5, from a light degree of discordance exhibited by some translucent, pink, rounded grain, is sub-concordant at 447"8 low-uranium grainsŽ analyses 10, 20 and 22 with 80, Ma. Placed on a chord through 420 Ma, analyses 13 52 and 97 ppm of U, respectively. is uncommon. and 21, from colourless euhedral and colourless With such low U contents, Pb loss by diffusion is rounded grains, respectively, project to ca. 1800 Ma unlikely. Conversely, this suggests the grains have suggesting that these grains could correspond to suffered lead loss during a metamorphicrtectonic Proterozoic zircons almost completely reset in Cale- event. Detrital zircons constitute a mixture of grains donian times or that they reflect mixing of a Caledo- of different ages and the time for Pb loss is difficult nian and an older, possibly Luliang, zircon compo- to assess with confidence. The Luliang grains, how- nent. ever, are located within a fan-like domain intersect- A fourth group is constituted by three grains ing concordia at 1800, 2000 and ca. 420 Ma, sug- Ž.inset c, Fig. 2 which gave significantly younger gesting that parts of the source region may have ages. The youngest concordant grainsŽ analyses 31 undergone a metamorphic event in Caledonian times. and 43. have, within error margins, overlapping The second group, represented by zircons which U±Pb ages of 231"1 and 233"1 Ma. Their euhe- come evenly from both samples, is constituted by dral shapes are consistent with a plutonicrvolcanic grains with apparent ages older than 2050 Ma and igneous source and the coincidence with age deposi- reaching values of about 2500 Ma for the least tion supports a volcanic origin. A similar grainŽ anal- discordant analyses 7, 26 and 27Ž 2535"2 Ma, ysis 14. gave a concordant data with U±Pb ages of 2517"2 Ma and 2513"2 Ma. . The low level of ca. 261 Ma. discordance indicates that Late Archean rocks with Finally, three grainsŽ. analyses 12, 35 and 36 are ages close to 2500±2600 Ma occur in the source located apart from the previously defined groups region. Younger ages for analyses 1, 4 and 9 suggest Ž.inset d, Fig. 2 . Analyses 12 and 35 have Pb±Pb a Pb loss pattern during at least one subsequent minimum ages of ca. 750 Ma, which suggests an event. origin of these grains from Late ProterozoicŽ. Sinian

Fig. 3. Concordia plot for detrital zircons from the Yajiang Upper Triassic sampleŽ. CDU 58 . `scolourless grains; v slight pink grains; polygonss2s error. O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 225 source material. Analysis 36, from a colourless, less straightforward but grain 48Ž Pb±Pb age of elongated grain with smoothed edges, presents a 1855"3 Ma.Ž and to a lesser degree grain 45 Pb±Pb 207 Pbr 206 Pb age of ca. 915 Ma and a low discor- age of 1818"6 Ma. suggest the occurrence of dance degree Ž.-5% , suggesting that the Pb±Pb age Luliang material, whereas the very discordant grain is a reasonably good approximation to the age of the 46Ž207 Pbr 206 Pb age of 2135"6 Ma. is likely to source material. derive from Archean source material.

4.2. Upper Triassic samples() CDU 58 5. Discussion Analytical results from the Upper Triassic sandstone again demonstrate that more than one age is 5.1. Age spectrum and source area presentŽ. Fig. 3 . Four colourless euhedral grains Ž.analyses 47, 50±52 yield indistinguishable U±Pb analyses obtained on 52 single zircon grains 207 Pbr 206 Pb ages ranging from 752 to 758 Ma, are summarized in Fig. 4. The new data set provides suggesting that they may represent a single popula- insights for the ages of the source rocks and allows tion deriving from a single source region. These reconstruction of a reasonable picture of the source grains are colinear along a line connecting the origin regionŽ. s . Ž.37"75 Ma and an upper intercept of 760"15 Ma The age spectrum for the Middle Triassic sand- Ž.MSWDs6.7 which may represent the age of this stones CDU 21 and CDU 27 is similar, indicating source region. Results from the four other grains are identical source regions for these samples collected

Fig. 4. Frequency diagrams for apparent ages of detrital zircons from the XiaoyinŽ. CDU 27 , Palang Shan Pass Ž. CDU 21 and Yajiang Ž.CDU 58 sandstones. 226 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 in the central part of the basin. The youngest concor- old with minor amounts of Late ArcheanŽ 2.5±2.6 dant grainsŽ. 231"1 Ma and 233"1 Ma provide a Ga. material. The Luliang zircons constitute ca. 52% maximum age for depositionŽ. inset c, Fig. 3 . The Ž.23 grains out of 44 of the detrital grains analysed euhedral shape of these grains suggests a and the ages broadly cover a range of about 300 Ma, magmaticrvolcanic origin and their age is identical from 1660 to 1983 Ma. The least discordant grains to those mentioned for the Kunlun granitic batholiths shown in Fig. 2Ž. inset a present a geometric mean of wx41±43 and to volcanics of the Litang±Batang arc 1848 Ma, similar to the 1840"40 Ma age for the located along the northern and western margin of the amphibolite to granulite facies event recognized in basin, respectively. The calcic to calc-alkaline Kun- the Fuping Group of the Shanxi Province of the lun batholith, on the southern margin of NCB, is Sino±Korean cratonwx 15 . Source rocks for these related to the north-dipping subduction of the Song- grains may have been subjected to metamorphism pan sea under the continental plate of NCB, which about 400 Ma ago. Superimposition of a ca. 400 Ma was active from 260 to 240 Ma in the Late Permian metamorphic event on 1.8±2.0 Ga old rocks is sup- to Middle Triassicwx 41±43 . Although there are today ported by anomalous discordance pattern observed in no obvious eruptive products in this area, swarms of some low-U grains and by ca. 420±450 Ma old basalt±andesite dykes cross-cut the batholith and crystalsŽ. inset c, Fig. 3 . The time for Pb loss is adjacent sedimentary strata. These volcanics have difficult to assess with confidence but, in any case, been related to an active continental margin setting results indicate that Caledonian rocks are present in and assigned a Late Permian to Early Triassic age the source region. wx44 . Evidence for Triassic volcanic activity in the Possible igneousrmetamorphic source areas for Kunlun Terrane is also supported by coarse clastic the 1.8±2.0 Ga zircons can be restricted to three rocks that represent subaerial and possible sub- main blocks: the Qaidam and Tarim blocks and the aqueous volcanic productswx 45 . Sino±Korean craton. The lack of Hercynian detrital The Litang island arc formed in a different tec- grainsŽ. see Fig. 4 has important implications in tonic setting and is related to intra-oceanic subduc- identifying the source regions. Both the Tarim craton tion with emplacement of basalt±andesite flows and the Qaidam block are surrounded by the Hercy- overlying fluvial conglomerates and sandstones and nian Kunlun mountain belt, excluding, therefore, overlain by marine carbonateswx 45 . From strati- these as the source of the Luliang detritus. The most graphic correlations, volcanics have been given an favourable source material for the 1.8±2.0, Ga zir- early Late Triassic ageŽ. ca. 230 Mawx 44 . Based on cons is thus represented by Luliang rocksŽ gneisses biotite and hornblende K±Ar ages, the main period and granites. now forming the bulk of the huge of volcanic activity is thought to occur between 180 continental landmass of the Sino±Korean craton. and 230 Ma, most rocks being emplaced at around The occurrence of Caledonian grains is consistent 200 MaŽ ages quoted inwx 4. . Although grains 31 and with a northeastern to eastern source region because 43 have ages matching those of volcanics from the these grains could have been derived either from the Litang arc, the nature of these rocks, mainly basaltic, Qinling belt between NCB and SCB or from the is not suitable. Moreover, the closeness of the Kun- Qilian Shan belt extending westward along the lun arc makes this latter a more probable source southern margin of the Sino±Korean craton. This region. The Late Saxonian ageŽ. 261"1 Ma for indicates that sediments were transported to the basin grain 14 of similar shape may also be attributed to through the Caledonian belts encircling the Sino± volcanicrplutonic source materialŽ. inset c, Fig. 3 . Korean craton. The lack of Archean Ž.)2.6 Ga to Its age is similar to the first volcanicrmagmatic Early Archean crystals in the age spectrum also period identified in the Kunlun arc, again suggesting suggests that the source region is restricted to the that this region was the most favourable source area. southern margin of the Sino±Korean craton because The age spectrum from the other grains is domi- no very old zirconsŽ. i.e. Archean to Early Archean nated by Luliang crystals; indicating that, by the end have been detected in the samples. In the age spec- of the Middle Triassic, the source region consisted trum, only a few grains depart from the consistent mainly of plutonicrmetamorphic rocks 1.8±2.0 Ga picture showing a Sino±Korean provenance for the O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 227 sediments. Among these, analyses 12 and 35 present process was accompanied by the formation of an ages of about 760 MaŽ. Fig. 2, inset d which cannot Andean-type mountain chain on the southern margin be found in NCB. The most probable sources for of the NCB. The occurrence of possible volcanic these grains can be found in the Yangtze craton to grains with ages similar to that of volcanics from the the southeast, whilst a further potential source area Kunlun arc is consistent with detrital inputs deriving may be found in the Qilian micro-block, north of the from such a mountain belt. However, this pattern basinwx 9 . Analysis 36 records a Pb±Pb age of 915 alone could not account for the huge quantity of Ma, consistent with a Yangtze source region because clastic sediments accumulated in the Songpan±Ganze between 900 and 1000 Ma subduction has been basin. First, if most of the detritus were derived from proposed to occur along the northern margin of the such a Triassic Andean-type mountain chain, the age cratonwx 34,35 . spectrum should indicate a much greater contribution The age spectrum from CDU 58 is quite different, from plutonicrvolcanic material of Permian±Tri- suggesting that the source region changed signifi- assic age. Moreover, from present day equivalents, it cantly during the Late Triassic. Though we acknowl- is suggested that most of the detritus deriving from edge our database for the Upper Triassic is rather erosion of such a belt is transported in the hinterland, limitedŽ. 8 grains analysed the age spectrum from by analogy with South America and the Amazon sandstone CDU 58 is drastically differentŽ. Fig. 4 . river. This scheme could account for deposition of From the eight randomly hand picked grains, fourŽ or clastic material in the Ordos basin and is consistent 50%. are from Sinian source material. This is twice with non-marine fluvial sedimentation over most of as important as for the 44 grains analysed from the NCB during the Permian±Triassicwx 45 . As evi- Middle Triassic sandstones and indicates that the denced by the three Permian and Triassic grains difference is real and not related to the small number from CDU 21 and CDU 27, at least some detrital of grains analysed. The age spectrum therefore shows inputs in the Songpan±Ganze basin were derived that the main contribution was from a source region from such a northern source, but not most of the predominantly constituted of Sinian material. The huge amounts of sediments accumulated. most obvious potential source region for these rocks Flysch deposition in the Songpan±Ganze basin is is the Yangtze craton of SCB. Analysis 49, with a coeval with the collision between the NCB and SCB 207 Pbr 206 Pb age of 965 Ma, is also consistent with a dated at ca. 210±220 Mawx 28±32 . This suggests a source area represented by the northern margin of probable relationship between the two processes and the craton. Variation in the age spectrum, from pre- it has been proposedwx 2 that the huge quantity of dominantly Luliang to predominantly Sinian source detrital material accumulated within the Songpan± rocks during the Middle Triassic and Upper Triassic, Ganze basin derived from exhumation and denuda- respectively, clearly shows that the source region had tion of rocks from the Dabie Shan beltŽ including the changed with time and that detrital inputs from the UHP assemblages. uplifted during the Late Triassic Sino±Korean craton, which were predominant in the by the collision between the NCB and SCB. Contem- Middle Triassic, have been drastically reduced. poraneity between deposition and exhumation of the UHP rocks however would imply very rapid uplift 5.2. Relationship between erosionrsedimentation rates. In a recent series of paperswx 46,47 it has been and tectonism proposed that continental subduction and exhumation of UHP rocks can be synchronous, relief erosion Huge amounts of detrital material accumulated in producing an unloading effect responsible for the the Songpan±Ganze basinŽ ca. 2.0=1063 km. have uplift of subducted continental slices. Thus, there is no equivalent today and this may testify to peculiar no reason to reject a possible origin of the sediments mechanisms in the source region and in relief forma- from this region on the basis of time constraints tion. Mechanisms which could account for the cre- alone. ation of elevated reliefs can be found in the subduc- Although paleocurrent measurements and prove- tion process of the Songpan sea beneath the NCB nance analysis are consistent with inputs of the which occurred during the Middle Triassic. This sediments from a northeastern metamorphic source 228 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 areawx 48 , the lack of high pressure minerals in the Triassic sandstones has cast doubts on this hypothesiswx 3 . Moreover, U±Pb zircon dating on gneisses from the Dabie regionwx 32 gave protolith ages of about 700±800 Ma. Thus, erosion of the UHPM formations should mainly provide 700±800 Ma and 210±220 Ma old detrital zircons. This is not consistent with the age spectra obtained from sandstones CDU 21 and CDU 27, which allows us to propose that detritus was derived from middle Proterozoic source rocks in the Middle Triassic. Similarly to Nie et al.'s hypothesiswx 2 , we propose that source rocks were brought into erosion position as a consequence of collision between the NCB and SCB but, contrary to this model, we suggest that detrital inputs derived from the uplifted margin of the over-riding plateŽ i.e. NCB. and not from the exhumed continental slice of SCB. Physical modelswx 46 are consistent with this proposal and indicate that, in a continental subduction setting, the margin of the over-riding plate is significantly uplifted. Collision was diachronous and started first in the east, migrating westward with the clockwise rotation of the SCBwx 49 . The eastern part of the NCB therefore comprised highlands and was elevated compared to the western part. This region was potentially subjected to erosion and detritus could have migrated along this natural tilt to reach the Songpan sea. Sediments were probably transported westward in the basin by a river network flowing out between the NCB and SCB through the Fig. 5. Model for sediment infilling in the Songpan±Ganze basin Qinling regionŽ. Fig. 5a , which may have acted as a duringŽ. a the Middle Triassic and Ž. b the Late Triassic. White spout. The similarity of the age spectra for the two arrows indicate block movement. Black arrows represent detrital samples of Middle Triassic age suggests an identical inputs. source region but may also plead for homogenisation of the various detrital components during transport. The occurrence of detrital grains as young as the age suggest that folding and relief formation first oc- of depositionŽ. ca. 230 Ma suggests geological activ- curred in the north, along the margin of the NCB ity around the basin and implies that at least part of with uplift an accretionary prismwx 8 . As the deforma- the detritus is first cycle material, transported di- tion migrated southward, it is suggested that this was rectly from the source region to the basin. These accompanied, in the present Qinling region, by a observations suggest that the river network was quite westward migration of the deformation, induced by important and may have been similar to the present the scissor-like movement of the SCBŽ. Fig. 5b . Ganges river system. As the collision continued, Formation of these reliefs has constituted a natural rotation of the SCB progressively closed the Song- barrier for detritus originating from erosion of the pan sea, by subduction under the NCBwx 1,50 and, in NCB, stopping them from reaching the basin. This is the Late Triassic, uplifted the Indo-Sinian part of the thought to be responsible for the drastic change in Qinling belt. The geometry of the basin and its the zircon age spectrum of Middle Triassic and closure by northward subduction under the NCB Upper Triassic sandstones. In the Late Triassic, detri- O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 229 tal inputs were then mainly derived from Sinian these highlands. The rotation of the SCB and thick- source material located on the northern margin of the ening of the accretionary prism along the southern SCBŽ. Fig. 5b , probably at the western end of the margin of the NCB progressively closed the basin in closing spout. These source regions, swamped by the Late Triassic. Sediments were then predomi- Luliang detritus in the Middle Triassic, became pre- nantly derived from eastern to southeastern sources dominant in the age spectra presented by detrital located on the northern margin of the SCB. zircon grains. This infilling model differs from that proposed earlier by Nie et al.wx 2 , in particular with regards to the source material. The lack of 210±220 Acknowledgements Ma detrital zircons indicates that the UHPM terranes of the Dabie region cannot be the source of the This work was supported by the DBT Programme detrital material and therefore that exhumation of and was carried out as part of the Ph.D. thesis of the these rocks is not linked with the huge clastic accu- first author. Samples were collected in 1990 during a mulation in the Songpan sea. Sino±French field trip across the Songpan Belt. The authors are grateful to D. Bosch and M. Mattauer for fruitful discussions and comments on the paper. We 6. Conclusions thank Bor-ming Jahn and An Yin for critical and helpful reviews. []RV Conventional U±Pb analyses on single detrital zircon grains from sandstones of the Songpan±Ganze basin indicate that clastic material accumulated dur- References ing the Middle Triassic was largely derived from Middle Proterozoic source rocksŽ. 1.8±2.0 Ga . Dis- wx1 R.J. Enkin, Z. Yang, Y. Chen, V. Courtillot, Paleomagnetic tinctive components include late ArcheanŽ 2.5±2.6 constraints on the geodynamic history of the major blocks of Ga.Ž. , Caledonian 400±450 Ma and, to a lesser China from the Permian to the Present, J. Geophys. Res. 97 Ž. Ž. Ž 1992 13953±13989. degree, Sinian ca. 760 Ma and younger 230±260 wx2 S. Nie, A. Yin, D.B. Rowley, Y. Jin, Exhumation of the Ma. , possibly volcanic, material from an Andean- Dabie Shan ultra-high-pressure rocks and accumulation of type mountain chain on the northern margin of the the Songpan Ganze flysch sequence, central China, Geology basinŽ. Kunlun arc . The most probable source area 22Ž. 1994 999±1002. wx for these grains is represented by the southern mar- 3 D. Avigad, Exhumation of the Dabie Shan ultra-high-pressure rocks and accumulation of the Songpan Ganze flysch gin of the Sino±Korean craton while sources for the sequence, central China: comment, GeologyŽ. 1995 764±765. Sinian grains may have been from the Yangtze cra- wx4 S. Calassou, Etude tectonique d'une chaõneÃÂ de decollement. ton. Sediments were transported westward through A. Tectonique triasique et Tertiaire de la chaõneÃ de Song- the Qinling belt, wrapping the southern margin of the pan±GanzeŽ.. Est-Tibet . B GeometrieÂÂ et cinematique Â des Sino±Korean craton. Upper Triassic sandstone shows deformationsÂÂÂ dans les prismes d'accretion sedimentaire: modelisationÂ analogique, Ph.D. Thesis, Univ. Montpellier II, a completely different age spectrum with a predomi- 1994. nance of Sinian material, probably derived from wx5 M.W. McElhinny, B.J.J. Embleton, X. Ma, Z. Zhang, Frag- erosion of the northern margin of the Yangtze craton. mentation of Asia in the Permian, Nature 239Ž. 1981 212± This drastic change in the source region is consistent 216. wx with a model in which infilling and subsequent 6 S.L. Chung, B.M. Jahn, Plume±lithosphere interaction in generation of the Emeishan flood basalts at the Permian±Tri- closure of the basin was closely related to the tec- assic boundary, Geology 23Ž. 1995 889±892. tonic evolution of the area and particularly to the wx7 Y. Zhang, Y. Luo, C. Yang, Panxi Rift and Its Geodynamics, Triassic continental collision between the South Geological Publishing House, Beijing, 1990, 415 pp. China BlockŽ. SCB and North China Block Ž NCB . . wx8 M. Mattauer, J. Malavieille, S. Calassou, J. Lancelot, F. In the Middle Triassic this diachronous event up- Roger, H. Ziwen, X. Zhiqin, H. Liwei, La chaõneÃ triasique de Songpan±GanzeŽ. Ouest Sechuan et Est Tibet : une chaõneÃ de lifted a collisional belt in the southeastern part of the plissement±decollementÂ sur marge passive, C. R. Acad. Sci. NCB but left a wide open sea on the west to be filled Paris 314Ž. 1992 619±626. by huge amounts of detrital material originating from wx9 Z.M. Zhang, J.G. Liou, R.G. Coleman, An outline of the 230 O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231

plate tectonics of China, Geol. Soc. Am. Bull. 95Ž. 1984 Qinling Belt: Build-up and evolution of eastern Asia, Nature 295±312. 317Ž. 1985 496±500. wx10 B.M. Jahn, B. Auvray, J. Cornichet, Y.L. Bai, Q.H. Shen, wx26 M.F. Lerch, F. Xue, A. Kroner,È W. Todt, Early Paleozoic D.Y. Liu, 3.5 Ga old amphibolites from eastern Hebei island arc accretion to the North China craton and the Shang Province, China, field occurrence, petrography, Sm±Nd Dan fault zone: a major paleoplate boundary in eastern Asia, isochron age and REE geochemistry, Precambrian Res. 38 J. Geophys. Res. 101Ž. 1996 17813±17826. Ž.1987 381±403. wx27 F. Xue, A. Kroner,È T. Reischmann, F. Lerch, Palaeozoic pre- wx11 D.Y. Liu, A.P. Nutman, W. Compston, J.S. Wu, Q.H. Shen, and post-collision calc-alkaline magmatism in the Qinling Remnants of )3800 Ma crust in the chinese part of the orogenic belt, central China, as documented by zircon ages Sino±Korean craton, Geology 20Ž. 1992 339±342. on granitoid rocks, J. Geol. Soc. London 153Ž. 1996 409±417. wx12 B. Song, A.P. Nutman, D. Liu, J. Wu, 3800 to 2500 Ma wx28 L. Ames, G.R. Tilton, G. Zhou, Timing of collision of the crustal evolution in the Anshan area of Liaoning Province, Sino±Korean and Yangtze cratons: U±Pb zircon dating of northeastern China, Precambrian Res. 78Ž. 1996 79±94. coesite-bearing eclogites, Geology 21Ž. 1994 339±342. wx13 R.T. Pidgeon, 2480 Ma old zircons from granulites facies wx29 E.A. Eide, M.O. Mc Williams, J.G. Liou,40 Arr 39Ar rocks from east Hebei Province, North China, Geol. Rev. 26 geochronology and exhumation of high-pressure to Ž.1980 198±207. ultrahigh-pressure metamorphic rocks in east-central China, wx14 W. Compston, F.D. Zhong, J.J. Foster, K.D. Collerson, J. Geology 22Ž. 1994 601±604. Bai, D.C. Sun, Rubidium±strontium geochronology of Pre- wx30 S. Li, S. Wang, Y. Chen, D. Liu, J. Qiu, H. Zhou, Z. Zhang, cambrian rocks from the Yensha region, North China, Pre- Excess argon in phengite from eclogite: Evidence from dat- cambrian Res. 22Ž. 1983 175±202. ing of eclogite minerals by Sm±Nd, Rb±Sr and 40Arr 39Ar wx15 L. Dunyi, R.W. Page, W. Compston, W. Taishan, U±Pb methods, Chem. Geol. 112Ž. 1994 343±350. zircon geochronology of Late Archean metamorphic rocks in wx31 V. Chavagnac, B.M. Jahn, Coesite-bearing eclogites from the the Taihangshan±Wutaishan area, North China, Precambrian Bixiling Complex, Dabie Mountains, China: Sm±Nd ages, Res. 27Ž. 1985 85±109. geochemical characteristics and tectonic implications, Chem. wx16 A. Kroner, W. Compston, Z. Guo-Wei, G. An-Lin, W. Todt, Geol. 133Ž. 1996 29±51. Age and tectonic setting of Late Archean greenstones±gneiss wx32 L. Ames, G. Zhou, B. Xiong, Geochronology and isotopic terrain in Henan Province, China, as revealed by single-grain character of ultrahigh-pressure metamorphism with implica- zircon dating, Geology 16Ž. 1988 211±215. tions for collision of the Sino±Korean and Yangtze cratons, wx17 D.Y. Liu, Q.H. Shen, Z.Q. Zhang, B.M. Jahn, B. Auvray, central China, Tectonics 15Ž. 1996 472±489. Archean crustal evolution in China: U±Pb geochronology of wx33 A. Kroner, G.W. Zhang, Y. Sun, Granulites in the Tongbai the Qianxi Complex, Precambrian Res. 48Ž. 1990 233±244. area, Qinling belt, China: geochemistry, petrology, single wx18 B.M. Jahn, Z.Q. Zhang, Archean granulite gneisses from zircon geochronology and implications for the tectonic evolu- eastern Hebei Province, China: rare earth geochemistry and tion of eastern Asia, Tectonics 12Ž. 1993 245±255. tectonic implications, Contrib. Mineral. Petrol. 85Ž. 1984 wx34 Y. Shi, C. Jia, D. Jia, L. Guo, Plate tectonics of East Qinling 224±243. Mountains, China, Tectonophysics 183Ž. 1990 25±30. wx19 B.M. Jahn, W.G. Ernst, Late Archean Sm±Nd isochron age wx35 S. Gao, B. Zhang, Z. Li, Geochemical evidence for Protero- for mafic±ultramafic supracrustal amphibolites from the zoic continental arc and continental-margin rift magmatism northeastern Sino±Korean craton, China, Precambrian Res. along the northern margin of the Yangtze craton, South 46Ž. 1990 295±306. China, Precambrian Res. 47Ž. 1991 205±221. wx20 B.M. Jahn, B. Auvray, Q.H. Shen, D.Y. Liu, Z.Q. Zhang, wx36 D. Bosch, O. Bruguier, R.T. Pidgeon, The evolution of an Y.J. Dong, X.J. Ye, Q.Z. Zhang, J. Cornichet, J. Mace, Archean Metamorphic Belt: a conventional and SHRIMP Archean crustal evolution in China: the Taishan complex and U±Pb study on accessory minerals from the Jimperding Belt, evidence for juvenile crustal addition from long-term de- Yilgarn craton, West Australia, J. Geol. 104Ž. 1996 695±711. pleted mantle, Precambrian Res. 38Ž. 1988 381±403. wx37 T.E. Krogh, Improved accuracy of U±Pb zircon ages by the wx21 Z. Wangh, X.F. Qiao, Proterozoic stratigraphy and tectonic creation of more concordant systems using an air abrasion framework of China, Geol. Mag. 121Ž. 1984 599±614. technique, Geochim. Cosmochim. Acta 46Ž. 1982 637±649. wx22 D. Sun, S. Liu, A subdivision of the Precambrian of China, wx38 O. Bruguier, U±Pb ages on single detrital zircon grains from Precambrian Res. 28Ž. 1985 137±162. the Tasmiyele Group: implications for the evolution of the wx23 C.Y. Lan, T. Lee, X.H. Zhou, S.T. Kwon, Nd isotopic study Olekma BlockŽ. Aldan Shield , Precambrian Res. 78 Ž. 1996 of Precambrian basement of South Korea: evidence for Early 197±210. Archean crust?, Geology 23Ž. 1995 249±252. wx39 J.S. Stacey, J.D. Kramers, Approximation of terrestrial lead wx24 P.J. Patchett, N. Arndt, Nd isotopes and tectonics of 1.9±1.7 isotope evolution by a two stage model, Earth Planet. Sci. Ga crustal genesis, Earth. Planet. Sci. Lett. 78Ž. 1986 329± Lett. 6Ž. 1975 15±25. 338. wx40 K.R. Ludwig, Isoplot200, a plotting and regression program wx25 M. Mattauer, P. Matte, J. Malavieille, P. Tapponnier, H. for isotope geochemists, for use with HP series 200 comput- Maluski, X.Z. Qin, L.Y. Lun, T.Y. Qin, Tectonics of the ers, U.S. Geol. Surv. Open-file report 85-513, 1987. O. Bruguier et al.rEarth and Planetary Science Letters 152() 1997 217±231 231 wx41 N.B.W. Harris, X. Ronghua, C.L. Lewis, J. Chengwei, Plu- mechanism for syn-collisional rock exhumation and associ- tonic rocks of the 1985 Tibet Geotraverse, Lhasa to Golmud, ated normal faulting: results from physical modelling, Earth Phil. Trans. R. Soc. London Ser. A 327Ž. 1988 145±168. Planet. Sci. Lett. 132Ž. 1995 225±232. wx42 N.B.W. Harris, X. Ronghua, C.L. Lewis, C.J. Hawkesworth, wx47 A.I. Chemenda, M. Mattauer, A.N. Bokun, Continental sub- Z. Yuquan, Isotope geochemistry of the 1985 Tibet Geotra- duction and a mechanism for exhumation of high-pressure verse, Lhasa to Golmud, Phil. Trans. R. Soc. London Ser. A metamorphic rocks: new modelling and field data from Oman, 327Ž. 1988 263±285. Earth Planet. Sci. Lett. 143Ž. 1996 173±182. wx43 J. Pearce, M. Houjun, Volcanic rocks of the 1985 Tibet wx48 S. Nie, A. Yin, D.B. Rowley, Y. Jin, Exhumation of the Geotraverse: Lhasa to Golmud, Phil. Trans. R. Soc. London Dabie Shan ultra-high-pressure rocks and accumulation of Ser. A 327Ž. 1988 169±201. the Songpan Ganze flysch sequence, central China: reply, wx44 M.R. Leeder, A.B. Smith, Y. Jixiang, Sedimentology, GeologyŽ. 1995 765±766. palaeoecology and palaeoenvironmental evolution of the 1985 wx49 A. Yin, S. Nie, An indentation model for the North and Lhasa to Golmud Geotraverse, Phil. Trans. R. Soc. London South China collision and the development of the Tan-Lu Ser. A 327Ž. 1988 107±143. and Honam fault system, East Asia, Tectonics 12Ž. 1993 wx45 M.P. Watson, A.B. Hayward, D.N. Parkinson, Z.M. Zhang, 801±813. Plate tectonic history, basin development and petroleum wx50 X. Zhao, R.S. Coe, Palaeomagnetic constraints on the colli- source rock deposition onshore China, Mar. Pet. Geol. 4 sion and rotation of North and South China, Nature 327 Ž.1987 205±225. Ž.1987 141±144. wx46 A.I. Chemenda, M. Mattauer, J. Malavieille, A.N. Bokun, A Paper 202 Disc

An early Miocene age for a high-temperature event in gneisses from Zabargad Island (Red Sea, Egypt): mantle diapirism? D. Bosch* and O. Bruguier Laboratoire de Ge´ochimie Isotopique, CNRS-UMR 5567, cc 066,Universite´ de Montpellier II, Place E. Bataillon, 34095 Montpellier Cedex 05, France

ABSTRACT U±Pb zircon data from a felsic gneiss located at the contact zone recrystallization under metamorphic conditions. The 22.4 Myr with the central peridotite body of Zabargad Island (Red Sea, Miocene age is thus interpreted as dating a high-temperature Egypt) provide an age of 23.2 + 5.9 Myr consistent with the metamorphic event. The proximity between the studied sample 238U±206Pb age of the youngest concordant grain (22.4 + 1.3 and the peridotite supports previous conclusions which regard Myr). Concordant grains indicate new zircon growth and/or parts of the peridotites from Zabargad Island as an asthenopheric resetting whereas slightly discordant analyses suggest mantle diapir which intruded the thinned Pan-African continental participation of an older zircon component whose age cannot crust during the early stages of the Red Sea opening. be defined precisely. SEM back-scattered imaging further reveals the occurrence of zoned domains almost completely erased by complex internal structures attributed to extensive Terra Nova, 10, 274±279, 1998

in harzburgite (Trieloff et al., 1997). date (Brueckner et al., 1995) from a Introduction Both ages are identical, within errors, felsic gneiss located a few hundred In the last decade, much attention has of a 18.4 + 1.0 Myr zircon age obtained metres from the contact zone, between been focused on the tectonothermal by Oberli et al. (1987). Although there the central peridotite body and the and geochemical evolution of rocks is a good agreement between the three gneisses. These ages, however, provide outcropping on Zabargad Island (Red methods, Oberli’s zircon value has been no constraints on the timing of juxta- Sea, Egypt). This island, though of presented only in an abstract, thus position of gneisses and peridotites. In limited size (& 4km2), has an almost making it difficult to properly evaluate. order to solve this fundamental pro- unique geodynamic setting, 50 km west K–Ar and Ar–Ar ages have also been blem, and because of the known resis- of the Red Sea axis, and reveals a close questioned for decoupling of parent tance to alteration of the mineral zir- association of mafic/ultramafic rocks and daughter nuclides (Villa, 1990), or con, we focused on the U–Pb dating of Ahed with a metamorphic gneiss complex. for their meaning in terms of a geolo- single zircon grains extracted from a Bhed Understanding the geodynamic evolu- gical event (uplift, hydrothermalism or felsic gneiss located at the contact zone Ched tion of the island was expected to elu- emplacement age) due to their known with the central peridotite body. Dhed cidate the processes operating in a susceptibility to low-grade events. An Ref marker young rift-setting environment and alternative interpretation considers the Geological setting Fig marker crust/mantle interface tectonics. Since peridotites and the gneisses to be Pan- Table marker the first work of Bonatti et al. (1981), African (Brueckner et al., 1988, 1995) Located about 90 km south-east of the Ref end however, genetic models have been lim- and that juxtaposition of both rock Raˆ s Baˆnas peninsula and about 50 km Ref start ited by the the scant data on the timing types occurred shortly after differentia- west of the Red Sea axis, Zabargad of juxtaposition between the perido- tion from a common depleted mantle island (Fig. 1) presents a petrographic tites and the gneisses, as well as the source & 700 Ma. This is apparently association dominated by three bodies nature of the mafic/ultramafic rocks. supported by the so-called SLAP error- of fresh ultramafic rocks. The southern One interpretation considers the peri- chron resulting from alignment of body is constituted by plagioclase peridotites as an asthenospheric mantle whole rock peridotite samples and dotites with a great abundance of dia- diapir intruding a Pan-African conti- CPX separates in the Sm–Nd isochron base dikes at the contact with the over- nental crust during the early stages of diagram (Brueckner et al., 1988). This lying metasedimentary formation. The the Red Sea opening (e.g. Nicolas et al., model, however, has been challenged in central and northern massifs are spinel 1987). The few attempts to date the time two recent papers on the grounds of lherzolites. Three generations of am- of intrusion of the peridotites are con- fluid inclusion (Boullier et al., 1997) phiboles are present in all three perido- sistent with this view. Nicolas et al. and noble gases and argon chronologi- tite bodies and have been classified into (1985), for example, inferred a K–Ar cal studies (Trieloff et al., 1997). In both three generations related to distinct age of 23 + 7 Myr for amphiboles ex- models, gneisses are seen to represent metasomatic events (Agrinier et al., tracted from an amphibolite. This value remnants of the Pan-African continen- 1993). The northern and central massifs concurs with the more precise Ar–Ar tal deep crust left behind during open- are in contact with a complex polygenic age (18.7 + 1.3 Myr) given by hornble- ing of the Red Sea rift. Pan-African assemblage constituted by Pan-African nde extracted from a pegmatoidal ages for the metamorphic gneiss com- granulite-facies gneisses (Bosch, 1990; pocket of a coarse grained hornblendite plex are well established by Sm–Nd and Lancelot and Bosch, 1991; Brueckner Rb–Sr mineral isochrons (Lancelot and et al., 1995) equilibrated at a depth of *Correspondence: Bosch, 1991; Brueckner et al., 1995) as over 30 km (Boudier et al., 1988) and by E-mail: [email protected] well as by a zircon Pb–Pb evaporation igneous pyroxenites and gabbros un-

274 *C 1998 Blackwell Science Ltd Paper 202 Disc

Terra Nova, Vol 10, No. 5, 274±279 Mantle diapirism, Zagarbad Island? . D. Bosch and O. Bruguier ......

SAUDI dard techniques (e.g. Bosch et al., 1996; ZABARGAD ISLAND ARABIA Pidgeon et al., 1996) where only the best

RED SEA, EGYPT quality grains (nonmagnetic, transpar- ZABARGAD ent and free of visible inclusions, frac- RED SEA

tures and cores) were selected for analysis. Single grains were subsequently

Northern EGYPT Massif processed according to Bruguier et al. (1997), but using the dissolution cap-

sules designed by Parrish (1987) which

Central significantly reduced Pb blanks. Isoto- Z1394 Massif

(695±2 Ma) pic measurements were carried out on a (22.4±1.3 Ma) (Pb-Pb zircons)

86Z2d 23¡37 (U-Pb zircons) VG Sector mass spectrometer using a Z2137 ❊

(<20 Ma) Daly detector. Because all common (Ar-Ar amphiboles) (548±10 Ma) (Sm-Nd)

85Za68b lead can be ascribed to blank Pb (15 (Disturbances) (Rb-Sr) pg + 50%), its isotopic composition (669±34 Ma) (Sm-Nd) 86Za47b was measured precisely and used to

(655±8 Ma) (Rb-Sr) Peridot gemstones determine the isotopic composition of mineralized zone radiogenic Pb. Corrected isotopic ra-

**** Southern tios, regression lines and intercepts

Massif were calculated according to Ludwig

(1987). Analytical uncertainties are listed in Table 1 as 2 s and uncertainties in ages as 95% confidence levels. N Results and discussion

The studied sample (86Z2d) was col-

23¡36 lected at the contact zone between the 500 m

gneisses and the central peridotite body. It corresponds to a fine-to med- 36¡1136¡12 ium-grained felsic gneiss containing c. 80% of quartz and plagioclase with

Plagioclase peridotite Spinel peridotite Amphibole peridotite iron oxides. Zircons are typically trans-

lucent, colourless to light pink and with

Gneiss Zabargad Formation Diabase rare inclusions. They present strong rounding of their external shapes sug- Fig. 1 Geological sketch map of Zabargad Island (modified after Bonatti et al., 1981) gesting partial dissolution preferen- indicating location of the studied sample (86Z2d). Samples dated by other authors and tially occurring at the terminations of mentioned in the text are also shown: (&) felsic gneiss Z1394 (Brueckner et al., 1995); the crystals. This rounding is associated (õ¨ õ¨ TMonotype Sorts) dolerite Z2137 (Villa, 1990); (*) mafic granulite 85Za68b and felsic with the development of multifaceted granulite 86Za47b (Lancelot and Bosch, 1991). Cenozoic sediments blank. surfaces, a morphological feature commonly found in high-grade metamorphic zircons (e.g. Kro¨ ner et al., derplated at the base of the crust and Bosch, 1991; Boudier and Nicolas, 1997) but also on zircons affected by recrystallized into mafic granulites be- 1991). Chronological data available high-temperature contact metamorph- fore and during the early stages of the on diabases are rather scarce so that ism (Davis et al., 1968). SEM imaging Red Sea rifting (Bonatti and Seyler, their age is not known precisely. Villa of the grains reveals complex internal 1987; Seyler and Bonatti, 1988). Nico- (1990) suggests an age younger than 20 structures and the presence of two zir- las et al. (1985) and Boudier et al. (1988) Myr, concurring with the tholeiitic con components (see Fig. 2a,b). In most provided evidence that away from the magmatism occurring 18–24 Ma along crystals examined, zoned domains are contact with the peridotites the shear the western coast of Saudi Arabia and erased and survive as relics in restricted strain and associated recrystallizations related to the early history of the Red parts of the grains. The internal struc- decrease rapidly in the metamorphic Sea opening (Coleman et al., 1979; ture of the grain shown in Fig. 2(a) is complex. Finally, diabase dykes and Feraud et al., 1991). Unmetamor- consistent with penetrative migration sills cross-cut all rock units including phosed Middle to Upper Miocene eva- of impurity components within the the gneisses and the peridotites. Dia- porites discordantly overlay the Zabar- crystal lattice and, according to Vavra bases are abundant within the southern gad Metasedimentary Formation. et al. (1996), is attributed to resorption peridotitic body and at the contact at the surface of the grains. The grain between this massif and the Zabargad presented in Fig. 2(b) is slightly differ- Analytical procedure Metasedimentary Formation (see Fig. ent as it shows a massive, structureless 1) where they promoted the growth of The minerals were extracted and pre- outer part replacing the original zircon. gem peridots (Clochiatti et al., 1981; pared for analysis according to stan- Although affected by a strong round-

*C 1998 Blackwell Science Ltd 275 Paper 202 Disc

Mantle diapirism, Zagarbad Island? . D. Bosch and O. Bruguier Terra Nova, Vol 10, No. 5, 274±279 ......

Table 1 Conventional single grain U–Pb isotopic data for zircons from the felsic gneiss 86Z2d, Zabargad Island, Red Sea, Egypt

Weight U Pb 206Pb/238U 207Pb/235U 207Pb/206Pb Apparent Ages (Myr) Sample1 (mg) (ppm)2 (ppm)3 206Pb/204Pb4 208Pb/206Pb4 (+ 2 s)4 (+ 2 s)4 (+ 2 s) 206Pb/238U 207Pb/235U 207Pb/206Pb r

Felsic Gneiss 86Z2d 1.Zr, mf, elg, c 0.0147 436 1.5 26.75 0.1420 0.00348 + 20 0.0219 + 30 0.0457 + 53 22.4 + 1.3 22.0 + 3.0 ±18.5 0.50 2.Zr, mf, rd, c 0.0128 568 2.2 43.89 0.1927 0.00354 + 69 0.0228 + 46 0.0467 + 35 22.8 + 4.5 22.9 + 4.6 32.0 0.93 3.Zr, mf, rd, lp 0.0243 399 2.4 102.34 0.3319 0.00403 + 03 0.0266 + 13 0.0480 + 22 25.9 + 0.2 26.7 + 1.3 96.8 0.48 4.Zr, mf, rd, lp 0.0132 609 2.7 69.20 0.2937 0.00380 + 53 0.0258 + 37 0.0487 + 22 24.5 + 3.4 25.9 + 3.7 134.5 0.98 5.Zr, mf, elg, lp 0.0104 689 3.7 69.75 0.4552 0.00417 + 29 0.0288 + 21 0.0501 + 10 26.8 + 1.9 28.8 + 2.1 198.4 0.96

Notes: 1grains were selected from nonmagnetic separates at full magnetic field in Frantz Magnetic Separator and air abraded following Krogh (1982); mf = multifaceted; rd: rounded; elg = elongated; c = colourless; pk: light pink. 2Radiogenic lead 3Measured, uncorrected ratio. 4Ratio corrected for fractionation and blank. Blank lead isotopic composition, 204: 206: 207: 208 = 1, 18.31, 15.59, 37.88), + 0.10, 0.11 and 0.13% for the 206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb ratios, respectively. Pb and U fractionation correction = 0.15%/amu (+ 0.05, 2 \'73). Pb blanks = 15 pg (+ 50%); U blanks = 5 pg (+ 50%). Isotopic ratios and absolute uncertainties in the Pb/U and 207Pb/206Pb ratios calculated following Ludwig (1987). Errors refer to last digits. U half-lives and isotopic abundance ratios from Jaffey et al. (1971). ing, the crystal roughly preserved a that the analysed zircons reflect new lization were brought about by contact prismatic shape and the observed per- growth or recrystallization during con- effect. Independent arguments for con- ipheral domain more likely resulted tact metamorphism developed at the tact metamorphism are provided by the from a resorption-induced recrystalli- vicinity of an ascending mantle diapir. 695 + 2 Myr age obtained by Brueck- zation. These morphological features In this view, emplacement of the peri- ner et al. (1995) on zircons extracted are in agreement with the grains being dotites in the Pan-African granulite from a felsic gneiss (Z1394) located c. originally magmatic zircons strongly gneisses took place early in the Red 300 m westward of the central perido- disturbed during a subsequent meta- Sea opening. Previous studies have tite body. The lack of isotopic distur- morphic event. In the concordia dia- highlighted the possible growth of bances in the evaporation age spectrum gram (Fig. 3), all points are close to new zircons during contact meta- obtained by these authors is in agree- concordant and fall on a chord with a morphism (Gulson and Krogh, 1975) ment with high temperatures restricted lower intersection at 23.2 + 5.9 Myr and the progressive resetting of U–Pb to the vicinity of the peridotite. While and an upper intersection, which is the ages towards hot intrusive bodies (Da- Pan-African Rb–Sr and Sm–Nd chron- result of a long extrapolation, at vis et al., 1968). In the latter case, ological information could be deter- 949 + 691 Myr (MSWD = 0.7). The evidence has been presented that Pb mined from a felsic granulite located latter is considered without any age loss can reach up to 80% of the un- 250 m away from the contact (sample signification. Analysis 5, however, is affected material for zircons located 86Za47b from Lancelot and Bosch, discordant and points to an old zircon 5 1 m from the contact. Experimental 1991), a mafic granulite 85Za68b (op component with a minimum investigations of the disturbance of the cit.) sampled 115 m from the central 207Pb/206Pb age of & 200 Myr. The U–Pb isotopic systems under hydro- peridotite body gave no Rb–Sr age 22.4 + 1.3 Myr U–Pb age of the con- thermal conditions (Pidgeon et al., information but still preserved a Pan- cordant analysis 1 is therefore tenta- 1966; Sinha et al., 1992) also revealed African Sm–Nd age (see Fig. 1); both tively taken as our best estimate for new the ability of hydrothermal and meta- are consistent with a thermal gradient growth and recrystallization of the zir- morphic fluids to dissolve and trans- towards the peridotite. Interestingly, con grains. The reason for this perva- port elements, and that up to 61% of the two apparently conflicting hypoth- sive recrystallization is not straightfor- the radiogenic lead can be leached out eses mentioned above (contact meta- ward, however. Two alternative of the zircons. The morphologies of the morphism vs. hydrothermalism) could interpretations can be proposed which zircons analysed herein, however, are be combined in the sense that contact have different implications for the geo- not consistent with those observed for metamorphism is classically associated dynamic evolution of the Red Sea area. natural or synthetic hydrothermally with fluid production emanating from Brueckner et al. (1995) proposed that grown zircons (Caruba et al., 1975; both the intrusive bodies and wall- the huge diabase sill at the contact with Kurat et al., 1982; Rubin et al., 1989). rocks (Hanson, 1995). In peridotites the southern peridotite massif was re- Such zircons occur in veins and present the occurrence of two amphibole gen- sponsible for a pervasive hydrothermal very irregular or even amoeboid shapes erations, whose growth was linked to event and suggested that the 18.4 + 1.0 (see fig. 4 of Rubin et al., 1989) or, hot hydrous fluids released during the Myr U–Pb zircon age obtained by alternatively, they may be well crystal- diapiric uplift (Agrinier et al., 1993), is Oberli et al. (1987) dates zircon growth lized, euhedral, probably zoned and consistent with such a model. In any under hydrothermal conditions, mm–cm in size (see figs 6 and 7 in case, the possibility that the large dia- although the latter authors interpreted Caruba et al., 1975; fig. 9 in Kurat et base sill could be responsible for the it as corresponding to the emplacement al., 1982). Thus, the observed morpho- & 22 Myr zircon age can be dismissed age of the peridotites. This view is con- logical features suggest a metamorphic as the diabase crosscuts and is therefore sistent with a Pan-African origin of the origin and the proximity of the studied younger than the southern peridotite Zabargad peridotites (Brueckner et al., sample to the peridotite body implies body which has been dated at 1988). Conversely, it could be proposed that zircon growth and recrystal- 18.7 + 1.3 Myr on magmatic horble-

276 *C 1998 Blackwell Science Ltd Paper 202 Disc

Terra Nova, Vol 10, No. 5, 274±279 Mantle diapirism, Zagarbad Island? . D. Bosch and O. Bruguier ......

line intersecting concordia at 143 + 11 Ma and 18.4 + 1.0 Ma. The significance of the Early Cretaceous upper intercept is not yet clear but suggests that there is more than just one Pan- African and one early Miocene time component present in the gneisses. Oberli’s analyses, however, have individual 238U–206Pb ages narrowly restricted to values ranging from 19 to 21 Myr which are in good agreement with our proposed age. The 18.7 + 1.3 Myr Ar–Ar age from a hornblende magmatic cumulate collected in the southern body has been related to magmatic crystallization during the final stages of diapiric uplift (Trieloff et al., 1997) and may therefore be regarded as a minimum age for emplacement of the peridotite body into the metamorphic gneiss complex. In a recent paper, Boullier et al. (1997) summarized the tectonic evolution of the mafic/ultramafic rocks of Zabargad Island and associated gneissic complex as resulting from at least Fig. 2 SEM back-scattered electron images of zircons from the felsic gneiss 86Z2d. Prior five stages. There is a consensus con- to SEM analyses, grains were etched by HF vapours for one minute. (a) Translucent cerning the first stage which is attribu- colourless rounded grain showing flow structures accompanying resorption. Faint relics ted to the Pan-African cycle (550–700 of magmatic zoning survive in the left part of the grain (see enlargement). (b) Resorbed Ma) and which is regarded as a major broken grain showing an outer structureless domain resulting from resorption-induced event leading to crust production and recrystallization. high-grade metamorphism with temperatures of 800–8508C and pressure nde (Trieloff et al., 1997). Thus, new related to a metamorphic/hydrother- c.10 kb (Boudier et al., 1988). It is still zircon growth and recrystallization at mal episode associated with intrusion unclear, however, whether the subse- 22.4 Ma cannot be linked to an hydro- of the Zabargad peridotites. Although quent evolution of the Zabargad peri- thermal event induced by diabase em- slightly older, this early Miocene age is dotites is related to the Pan-African placement. The Ar–Ar age mentioned in good agreement with that provided cycle (Brueckner et al., 1988, 1995) or above further indicates that the perido- by Oberli et al. (1987). These authors to the opening of the Red Sea rift tite was still hot during the Miocene. reported U–Pb analyses from zircons (Nicolas et al., 1987; Boudier et al., This strongly suggests that the U–Pb collected near the central peridotite 1988). Boullier et al. (1997), following zircon age reported herein could be body that spread along a discordant Dupuy et al. (1991), proposed an alter-

*C 1998 Blackwell Science Ltd 277 Paper 202 Disc

Mantle diapirism, Zagarbad Island? . D. Bosch and O. Bruguier Terra Nova, Vol 10, No. 5, 274±279 ......

peridotite, Zabargad Island (Red Sea). In: Orogenic Lherzolites and Mantle Processes (M.A. Menzies et al., Eds). J. Petrol., Spec. Issue. 243–253. Boudier, F., Nicolas, A., Ji, S., Kienast, J.R. and Mevel, C., 1988. The gneiss of Zabargad Island: deep crust of a rift. Tectonophysics, 150, 209–227. Boullier, A.M., Firdaous, K. and Boudier, F., 1997. Fluid circulation relatedtodeformationintheZabargad gneisses (Red Sea rift). Tectonophysics, 279, 281–302. Brueckner, H.K., Elhaddad, M.A., Hamelin, B., Hemming, S., Kro¨ ner, A., Reisberg, L. and Seyler, M., 1995. A Pan African origin and uplift for the gneisses and peridotites of Zabargad Island, Red Sea: a Nd Sr, Pb and Os isotope study. J. Geophys. Res., 100, 22,283–22,297. Fig. 3 U–Pb concordia diagram for single zircon grains from the felsic gneiss 86Z2d. All Brueckner, H.K., Zindler, A., Seyler, M. and samples represented by symbols sized according to their 2 s error. Bonatti, E., 1988. Zabargad and the isotopic evolution of the sub-Red Sea mantle and crust. Tectonophysics, 150, native model where the Northern and and for constant stimulating scientific 163–176. Central peridotite bodies represent exchanges. The paper greatly benefited for Bruguier, O., Lancelot, J.R. and Malavieille, parts of a metasomatized Pan-African constructive reviews by H.K. Brueckner, F. J., 1997. U-Pb dating on single detrital lithospheric mantle intruded by an Oberli, and one anonymous reviewer. zircon grains from the Triassic Songpan- asthenospheric mantle diapir repre- Ganze flysch (Central China): provenance and tectonic correlations. Earth Planet. sented by the Southern massif. Our References Sci. Lett., 152, 217–231. new U–Pb age substantiates the propo- Agrinier, P., Mevel, C., Bosch, D. and Caruba, R., Baumer, A. and Turco, G., sition that the Central peridotite body Javoy, M., 1993. Metasomatic hydrous 1975. Nouvelles syntheses is also part of the mantle diapir which fluids in amphibole peridotites from hydrothermales du zircon: substitutions intruded the thinned Pan-African litho- Zabargad Island (Red Sea). Earth Planet. isomorphiques; relation morphologie- sphere during early Miocene times. Fi- Sci. Lett., 120, 187–205. milieu de croissance. Geochim. nally, Zabargad ultramafics from the Bonatti, E., Hamlyn, P.R. and Ottonello, Cosmochim. Acta, 39, 11–26. Southern and Central massifs also re- G., 1981. The upper mantle beneath a Clochiatti, R., Massare, O. and Jehanno, C., cord a large range of P–T conditions, young oceanic rift. Peridotites from the 1981. Origine hydrothermale des olivines with the highest values being 12808C island of Zabargad (Red Sea). Geology, 9, gemmes de l’Ile de Zabargad (St John), 474–479. Mer Rouge, par l’e´ tude de leurs and 27 kb (Kurat et al., 1993). These are Bonatti, E. and Seyler, M., 1987. Crustal inclusions. Bull. Mineral., 104, 354–360. very different to the equilibration con- underplating and evolution in the Red Sea Coleman, R.G., Hadley, D.G., Fleck, R.G., ditions of the Pan-African granulites rift: uplifted gabbro/gneiss crustal Hedge, C.T. and Donato, M.M., 1979. (see above; Boudier et al., 1988) and complexes on Zabargad and Brothers The Miocene Tihama Asir ophiolite and further supports the fact that the Cen- Islands. J. Geophys. Res., 92, 12803–12821. its bearing on the opening of the Red Sea. tral and Southern peridotite bodies Bosch, D., 1990. Evolution ge´ ochimique In: Proc. Symp. Evol. Mineralization more likely constitute an astheno- initiale et pre´ coce d’un rift. Syste´ matique Arabian Nubian Shield. Inst. Appl. Geol. spheric mantle diapir which rose to isotopique Pb Sr et Nd du diapir Bull. (Jeddah), 1, 173–179. shallow levels in the crust during the mantellique de Zabargad, de son Davis, G.L., Hart, S.R. and Tilton, G.R., 1968. Some effects of contact early stages of the Red Sea opening as encaissant gneissique et de son hydrothermalisme. Conse´ quences metamorphism on zircon ages. Earth already proposed on the grounds of ge´ odynamiques et me´ talloge´ niques. Planet. Sci. Lett., 5, 27–34. geophysical data (Styles and Gerdes, Unpubl. doctoral dissertation, University Dupuy, C., Mevel, C., Bodinier, J.L. and 1983). However, the question remains of Montpellier II, 231pp. Savoyant, L., 1991. Zabargad peridotite: open whether the Northern peridotite Bosch, D., 1991. Introduction d’eau de mer evidence for multistage metasomatism body is also part of the mantle diapir or dans le diapir mantellique de Zabargad during Red Sea rifting. Geology, 19, constitutes a piece of a Pan-African (Mer Rouge) d’apre` s les isotopes du Sr et 722–725. lithospheric mantle wedge dredged up- du Nd. C. R. Acad. Sci. Paris, 313, 49–56. Feraud, G., Zumbo, V., Sebai, A. and 40 39 ward with the gneissic complex as sug- Bosch, D., Bruguier, O. and Pidgeon, R.T., Bertrand, H., 1991. Ar/ Ar age and gested earlier (Brueckner et al., 1988, 1996. The evolution of an Archean duration of tholeiitic magmatism related Metamorphic Belt: a conventional and to the early opening of the Red Sea rift. 1995; Dupuy et al., 1991). SHRIMP U-Pb study on accessory Geophys. Res. Lett., 18, 195–198. minerals from the Jimperding Belt, Gulson, B.L. and Krogh, T.E., 1975. Acknowledgements Yilgarn craton, West Australia. J. Geol., Evidence for multiple intrusion, possible 104, 695–711. resetting of U-Pb ages, and new A. Nicolas and F. Boudier are greatly Boudier, F. and Nicolas, A., 1991. High- crystallization of zircons in the post- thanked for providing sample for this study temperature hydrothermal alteration of tectonic intrusions (`Rapakivi granite’)

278 *C 1998 Blackwell Science Ltd Paper 202 Disc

Terra Nova, Vol 10, No. 5, 274±279 Mantle diapirism, Zagarbad Island? . D. Bosch and O. Bruguier ......

and gneisses from South Greenland. and regression program for isotope Seyler, M. and Bonatti, E., 1988. Petrology Geochim. Cosmochim. Acta, 39, 65–82. geochemists, for use with HP series 200 of gneiss-amphibolite lower crustal unit Hanson, R.B., 1995. The hydrodynamics of computers. U.S.G.S. Open-File Report, from Zabargad Island, Red Sea. contact metamorphism. Bull. Geol. Soc. 85–513. Tectonophysics, 150, 177–207. Am., 107, 595–611. Nicolas, A., Boudier, F., Lyberis, N., Sinha, A.K., Wayne, D.M. and Hewit, D.A., Jaffey, A.H., Flynn, D.F., Glendenin, L.E., Montigny, R. and Guennoc, P., 1985. 1992. The hydrothermal stability of Bentlye, W.C. and Essling, A.M., 1971. L’õˆle de Zabargad (Saint-Jean): te´ moin cle´ zircon: Preliminary experimental and Precision measurement of the half-lives de l’expansion pre´ coce en Mer Rouge. C. isotopic studies. Geochim. Cosmochim. and specific activities of 235U and 238U. R. Acad. Sci. Paris, 301, 1063–1068. Acta, 56, 3551–3560. Phys. Rev., 4, 1889–1906. Nicolas, A., Boudier, F. and Montigny, R., Styles, P. and Gerdes, K.D., 1983. St. John’s Krogh, T.E., 1982. Improved accuracy of U- 1987. Structure of Zabargad Island and Island (Red Sea): a new geophysical Pb zircon ages by the creation of more early rifting of the Red Sea. J. Geophys. model and its implications for the concordant systems using an air abrasion Res., 92, 461–474. emplacement of ultramafic rocks in technique. Geochim. Cosmochim. Acta, Oberli, F., Ntaflos, T., Meier, M. and Kurat, fracture zones and at continental margins. 46, 637–649. G., 1987. Emplacement age of the Earth Planet. Sci. Lett., 65, 353–368. Kro¨ ner, A., Jaeckel, P., Reischmann, T. and peridotites from Zabargad Island (Red Trieloff, M., Weber, H.W., Kurat, G., Kroner, U., 1997. Further evidence for an Sea): a zircon U-Pb isotope study. Terra Jessberger, E.K. and Janicke, J., 1997. early Carboniferous (c. 340 Ma) age of Cognita (Abstr.), 7, 334. Noble gases, their carrier phases, and high-grade metamorphism in the Parrish, R.R., 1987. An improved micro- argon chronology of upper mantle rocks Saxonian granulite complex. Geol. capsule for zircon dissolution in U-Pb from Zabargad Island, Red Sea. Geochim. Rndsch., 86, 751–766. geochronology. Chem. Geol., 66, 99–102. Cosmochim. Acta, 61, 5065–5088. Kurat, G., Niedermayr, G. and Prinz, M., Pidgeon, R.T., Bosch, D. and Bruguier, O., Vavra, G., Gebauer, D., Schmid, R. and 1982. Peridot von Zabargad, Rotes Meer. 1996. Inherited zircon and titanite U-Pb Compston, W., 1996. Multiple zircon Der Aufschluss, 33, 169–182. systems in an Archean Katrine syenite growth and recrystallisation during Kurat, G., Palme, H., Embey-Isztin, A. et from southwestern Australia: polyphase Late Carboniferous to triassic al., 1993. Petrology and geochemistry of implications for U-Pb stability of titanite. metamorphism in granulites of the Ivrea peridotites and associated vein rocks of Earth Planet. Sci. Lett., 141, 187–198. Zone (Southern Alps): an ion microprobe Zabargad Island, Red Sea. Egypt. Min. Pidgeon, R.T., O’Neil, J.R. and Silver, L.T., (SHRIMP) study. Contr. Miner. Petrol., Petrol., 48, 309–341. 1966. Uranium and lead isotopic stability 122, 337–358. Lancelot, J.R. and Bosch, D., 1991. A Pan in a metamict zircon under experimental Villa, I.M., 1990. 40Ar/39Ar dating of African age for the HP-HT granulite hydrothermal conditions. Science, 154, amphiboles from Zabargad Island (Red gneisses of Zabargad Island: 1538–1540. Sea) is precluded by interaction with implications for the early stages of the Rubin, J.N., Henry, C.D. and Price, J.G., fluids. Tectonophysics, 180, 369–373. Red Sea rifting. Earth Planet. Sci. Lett., 1989. Hydrothermal zircons and zircon 107, 539–549. overgrowths, Sierra Blanca Peaks, Texas. Received 16 April 1998; revised version ac- Ludwig, K.R., 1987. Isoplot200, a plotting Am. Miner., 74, 865–869. cepted 17 February 1999

*C 1998 Blackwell Science Ltd 279

Contrib Mineral Petrol (1999) 136: 258 – 272 Ó Springer-Verlag 1999

O. Bruguier á D. Bosch á R.T. Pidgeon D.I. Byrne á L.B. Harris U-Pb chronology of the Northampton Complex, Western Australia ± evidence for Grenvillian sedimentation, metamorphism and deformation and geodynamic implications

Received: 7 January 1998 / Accepted: 10 March 1999

Abstract Conventional and SHRIMP U-Pb analyses of retrograde metamorphism. Granitic activity at zircon, monazite, titanite and apatite from the high 1068 13 Ma was followed by intrusion of post-D2 grade rocks of the Northampton Complex in Western pegmatite (989 2 Ma), which constrains the end of Australia provide constraints on the timing of meta- metamorphism and associated deformation. Cooling of morphic processes and deformation events in the the complex to about 500 °C is timed by the apatite age northern Darling Mobile Belt (western margin of the of 921 23 Ma. SHRIMP U-Pb ages of detrital zir- Archean Yilgarn Craton). Paragneisses and mafic vol- cons from a paragneiss sample yield a maximum age of canics and/or intrusions have undergone granulite facies 2043 Ma, with no evidence of an Archean Yilgarn sig- metamorphism in a probable extensional tectonic setting nature. A majority of ages between 1.6 and 1.9 Ga are prior to formation of W- to NW-verging folds and consistent with derivation from the Capricorn Orogen thrusts cut by normal shears (interpreted as late collapse on the northern margin of the Yilgarn Craton. Younger structures) during the main deformation event (D1). detrital zircons with 1150–1450 Ma ages, however, in- These structures are folded by open to tight folds with dicate an additional source that had undergone early NW-striking axial surfaces developed in a second, Grenvillian igneous or metamorphic event(s) and also NE-SW contractional event (D2). Zircons from a mafic places a maximum age constraint upon deposition. The granulite provide an age of 1079 3 Ma attributed to source of this clastic material may have been from within new zircon growth prior to, or at the peak of regional the southern Darling Mobile Belt or from Greater India granulite facies metamorphism. Metamorphic monazites (adjacent to the Northampton Complex in Rodinia re- extracted from a paragneiss yield an identical age of constructions). This study documents an extended 1083 3 Ma. The similarity of ages between zircons Grenvillian history, with basin formation, sedimenta- from the mafic granulite (1079 3 Ma) and monazites tion, granulite facies metamorphism, contractional tec- from the paragneiss (1083 3 Ma) is interpreted to tonics (two periods with orthogonal directions of reflect fast cooling and/or rapid uplift, which is consis- shortening) and late pegmatite emplacement taking tent with thrusting of the gneissic units during the first place between 1150–989 Ma on the western margin of deformation event (D1) associated with the onset of the Yilgarn Craton. Ages recorded in this study indicate that the proposed global distribution of Grenvillian belts during assembly of the Rodinia supercontinent should O. Bruguier (&) á D. Bosch be reassessed to include the Darling Mobile Belt. ISTEEM, Universite´ Montpellier 2-CNRS, cc 066, Place Euge` ne Bataillon, 34095 Montpellier Cedex 5, France e-mail: [email protected] Introduction R.T. Pidgeon School of Applied Geology, The Northampton Complex has been included in the Curtin University of Technology, GPO Box U 1987, Perth, WA 6001, Australia Darling Mobile Belt (or Pinjarra Orogen after Myers et al. 1996) and represents an isolated Proterozoic D.I. Byrne á L.B. Harris Tectonics Special Research Centre, basement inlier within the Phanerozoic Perth Basin Department of Geology and Geophysics, (Fig. 1a). It is dominated by granulite facies para- The University of Western Australia, gneisses that show polyphase deformation, and is in- Nedlands, WA 6907, Australia truded by granitoids. A major question concerns the Editorial Responsibility: T.L. Grove relationships between the Complex and the adjacent 259

Fig. 1a, b Location of study area. a Composite geological sketch map of Western Australia (after Myers 1990). Proterozoic complexes Geological setting within the Darling Mobile Belt/Pinjarra Orogen (Northampton, Mullingarra and Leeuwin) are shown in black. b Simplified geology of the Northampton Complex showing the main geological features and General presentation sample locations. Phanerozoic rocks of the Perth Basin blank The Northampton Complex (Fig. 1a) covers an area of about 100 km · 30 km (Mathur and Shaw 1982) and is thought to represent the exposed top of a horst, Yilgarn Craton. Is the Complex formed from reworking bounded by numerous faults (Fig. 1b) such as the of Yilgarn crust, or, alternatively, is it part of an inde- Hardabut and Geraldton faults on the west and the pendent crustal remnant related to Proterozoic material Yandi and other unnamed faults on the east. It consists presently under Greater India? In this contribution, we of high-grade metamorphic rocks, mainly paragneisses, present the first detailed conventional U-Pb geochro- but also mafic granulites, granites and pegmatites. nological results on zircon, monazite, titanite and apa- Charnockites, graphitic schists, dolerite dykes and tite from the Northampton Complex. The objective of quartzites have also been described within the gneisses. the study is to provide a geochronological framework The complex experienced granulite facies metamorphism for the timing of metamorphism and associated defor- with peak temperatures and pressures of 800–900 °C mation and, in combination with SHRIMP U-Pb ages and 5 to 6 kbar. Limited outcrop has prevented attempts on detrital zircon grains, to examine the question of the to establish a pattern of isograds in the complex. Most origin of the complex and possible correlations with rocks contain textures reflecting retrograde metamor- other Proterozoic mobile belts with implications for phism from anhydrous granulite facies assemblages, to supercontinent reconstructions. more hydrous, hornblende dominated, assemblages of 260 the amphibolite facies. The onset of retrograde meta- tectonic correlations. To reach this degree of under- morphism is associated with the development of a foli- standing of the evolution of the Complex requires de- ation (S1) parallel to original bedding during tailed geochronology of mineral phases combined with deformation D1. The P-T path during this event (D1), petrographic interpretation of their metamorphic sig- and the formation of a metamorphic dierentiated lay- nificance. In the following we present U-Pb results on ering in the gneisses, involved retrogression of garnet to accessory minerals from a mafic granulite (W405), a cordierite, which represents a decrease in pressure, and psammitic gneiss (W404), a porphyritic granite (W412), eventually to biotite, which represents a decrease in a calc-silicate (W410), and a pegmatite (W411) selected temperature. P-T conditions of about 600 °C and 4 kbar to provide a range of geochronological information on were maintained during the late stage of D1. A second the timing of metamorphism of the Complex and asso- deformation event (D2) is characterized by open to ciated igneous activity (sketch map, Fig. 1b). gentle folds with NW-SE trending, double-plunging axes and was responsible for the development of a second, sub-vertical foliation (S2) defined by biotite. Abundant pegmatites, quartz-feldspar veins and granites with Petrographic descriptions variable timing relationships to deformation indicate that melting occurred at dierent stages during the Mafic granulites are rare in the Northampton Complex, metamorphism. The first generation of melt occurred and the field relationship between these mafic bodies and during prograde metamorphism and produced concor- the gneisses is uncertain due to the lack of suitable dant garnetiferous pegmatites, sometimes bounded by outcrop. On a TAS diagram (Lemaitre 1984) the sample cordierite-bearing melanosomes, and quartz-feldspar collected (Woodbine granulite W405) plots in the ba- veins. A second stage of melting is associated with ret- saltic andesite field suggesting that the mafic granulites rograde metamorphism and produced pegmatites local- represent basic volcanic rocks which were intruded or ised along extensional shear bands and coarse grained extruded onto sediments prior to or during metamor- granite. Post D2 pegmatites also intrude the Complex phism. Sample W405 is granoblastic with a possible S1 along late shear zones. foliation defined by hornblende, phlogopite, and ilmenite. Hornblende and phlogopite also occur as partial coronas around pyroxene and ilmenite. Quartzo-fe- Previous geochronology ldspathic gneisses are the dominant rock type outcropping in the complex. Based on the whole rock chemical The western margin of the Archean Yilgarn craton has analyses, mineralogy, stratigraphic relationships and the experienced repeated tectonic activity and crustal ag- presence of probable sedimentary structures such as gregation since the Archean (Myers 1990) but little is graded bedding, cross-bedding, load casts and channels, known about the crystalline basement west of the Dar- most of the quartzo-feldspathic gneisses probably rep- ling fault. Early Rb-Sr results from the Northampton resent a metamorphosed turbidite sequence. Calc-silicate Complex of Wilson et al. (1960) indicate an age of about rocks occur as small ellipsoidal bodies (0.4 to 1.6 m in 1000 Ma for a pegmatite dyke, and a Rb-Sr whole rock length) within the gneisses and may correspond to in- age of 1020 50 Ma was reported for granulites by traformational fragments of impure carbonate material Compston and Arriens (1968). Richards et al. (1985) (such as marl). Alternatively, they could correspond to interpreted a Rb-Sr whole rock age of 1037 146 Ma carbonate concretions, where greywacke particles have for partially remelted granulites as the age of granulite been cemented together by calcium-magnesium-iron ce- facies metamorphism. Sm-Nd whole rock analyses by ment. The gneisses commonly contain two foliations, Fletcher et al. (1985) on rocks of the Northampton both defined by sparsely distributed biotite and ilmenite. Complex (including one paragneiss sample) and the The first foliation, S1, is, in low strain area, parallel to Mullingarra Inlier indicates that the basement in the the lithological layering and wraps around garnet pop- Pinjarra Orogen essentially comprises material extracted hyroblasts. The second, S2, is oblique or perpendicular from the mantle between 1.8 and 2.0 Ga. The complex to S1 and presents a lower degree of intensity. has been intruded by a 650–800 Ma-old doleritic dyke The porphyritic granite (W412) is dominated by a swarm (Embleton and Schmidt 1985) and is overlain by foliation (parallel to the regional S2) defined by align- Phanerozoic sediments of the Perth and Carnarvon ba- ment of K-feldspar phenocrysts and orientation of sins. Whereas these geochronological results demon- biotite and coarse sillimanite which wraps around K- strate a Grenvillian age for the Complex, they are not feldspar. The granite also contains rounded elongate capable of elucidating the tectonic temperature-time xenoliths of garnet granulite and pegmatite which are history of the metamorphism or the relationship be- aligned parallel to the foliation and which indicate that tween metamorphism and igneous activity. Most im- intrusion post-dates formation of the granulites. portantly, previous studies have not determined the age The sample of biotite-bearing pegmatite (W411) is (and hence provenance) of source material for the from one of the numerous post-D2 pegmatitic dykes Complex. These factors are crucial in determining the which intrude the gneisses and garnetiferous peg- relationship of the Complex to adjacent terrains and for matites. 261 uncertainties. Analysis 6 is slightly discordant, falling Analytical techniques below the other six points. This grain does not have an unusually high U content so the reason for the discor- Separation of minerals was carried out using standard techniques dance is not readily explained. A weighted mean (e.g. Bosch et al. 1996). Single zircon grains were air-abraded 207Pb/206Pb age for the six concordant analyses (open (Krogh 1982) prior to washing in dilute 6 N HNO3. Other accessory minerals were not abraded and were washed using dilute 0.5 N symbol in Fig. 3) is 1079 3 Ma (MSWD = 0.7). On HNO3 (monazite and titanite) or tridistilled water (apatite). Single the basis of the morphological evidence and lack of an grains and small fractions were then weighed with an electronic inherited component, this is interpreted as the best es- micro-balance and placed in a Teflon dissolution capsule (Parrish 1987) with a few microlitres of 205Pb-235U spike and a drop of acid. timate of the age of zircon growth during granulite facies Dissolution reagents included HF 48% (13 N HNO3) for zircon metamorphism. and titanite and 6 N HCl for monazite and apatite. After disolution at 195 °C for three days, lead and uranium were separated on anion exchange resin following the method of Krogh (1973). For titanite, monazite, and apatite, an additional purification step for the ura- Paragneiss (W404) nium was performed following the procedure described by Manhes et al. (1978). All isotopic measurements were carried out on the Monazite results Curtin University VG 354 mass spectrometer using a Daly detector. Mass discrimination for the Daly detector has been determined Five single monazite grains were analysed from the pa- at 0.20 0.05%/a.m.u., using NBS common lead (NBS981) and radiogenic lead (NBS983) standards. Total Pb blanks over the ragneiss. The grains are interpreted to have grown dur- period of the analyses range from 5 to 20 pg and uranium blanks ing metamorphism. They have rounded shapes typical of were less than 5 pg. The calculation of common Pb was made by growth under metamorphic conditions (Parrish 1990). subtracting blank and then assuming the remaining common Pb to Back-scattered electron images indicate that some mo- be Pb incorporated during crystallisation with a composition determined from the model of Stacey and Kramers (1975). Corrected nazite grains have complex internal structures consisting isotopic ratios, regression lines and intercepts were calculated ac- of dark, irregular-shaped domains invading a brighter cording to Ludwig (1987). Analytical uncertainties are listed as 2r matrix (Fig. 2c, d) whereas other grains consist almost and uncertainties in ages as 95% confidence levels. Decay constants entirely of a bright, U-rich, domain. On a concordia plot are those recommended by the IUGS Subcommission on Geo- chronology (Steiger and Jager 1977). (Fig. 4), data points are aligned on a discordia For the SHRIMP analyses, zircon grains were mounted in ep- (MSWD = 0.6) which has intersections with concordia oxy resin and polished to approximately half their thickness. U-Th- at 1080 5 Ma and 664 278 Ma. The alignment is Pb analyses were performed on the SHRIMP II ion microprobe at convincing but is dependent on the rather large uncer- Curtin University of Technology (Perth), following the techniques tainties of analyses 8, 9 and 10. From data on Table 1, it of Compston et al. (1984), Williams et al. (1984) and Nelson (1997). Isotopic ratios were measured with a mass resolution of can be seen that the lower U grains are more discordant ca. 5000, and Pb/U ratios were normalised via quadratic working and have 207Pb/206Pb ages ranging from 1070 to curves to those measured on the standard CZ3 1080 Ma compared with higher uranium grains which 206 238 ( Pb/ U = 0.0914) prepared from a gem quality Sri Lankan are concordant at ca. 1082 Ma. The Th/U ratios of the zircon (Pidgeon et al. 1994). Common Pb corrections were based on the measured 204Pb and for all data, the assumed common Pb high U grains are ca. 5 and 10 (grains 12 and 11) and composition was modelled as contemporaneous Pb (Cumming and that of the low U grains range from 20 to 30. This Richards 1975). suggests a possible loss of uranium from the low U grains. We relate this to the presence of dark (low U) finger-like domains which are interpreted to represent U-Pb isotopic results recrystallisation of the original monazite accompanied by loss of U, Th, and Pb. On this basis, our best estimate Woodbine mafic granulite (W405) of the age of the monazite is given by the concordant age of high uranium grain 12 of 1083 3 Ma. Mafic granulite sample (W405) contains a uniform population of translucent, colourless anhedral to ellipsoidal zircon grains which are typical of zircons from Zircon results high grade metamorphic rocks (Pidgeon and Aftalion 1972). Back-scattered Scanning Electron Microscope Zircons from the paragneiss have rounded shapes simi- (SEM) images of polished surfaces of two grains lar to those commonly observed for detrital zircons, but (Fig. 2a, b) show that these zircons contain irregular such shapes could also result from corrosion during high shadowy domains of diuse SEM intensity, marking grade metamorphism. Results of SHRIMP II analyses irregularities in trace element concentrations. The lack of thirty one grains are presented in Table 2 and on a of continuity and symmetry of dark patches argues concordia plot (Fig. 5). The 207Pb/206Pb age spectrum, against a possible later overgrowth forming on these ranges from 1059 32 Ma (2r) to 2043 136 Ma zircons and we interpret the internal structures as rep- (2r), and confirms that the zircons are detrital grains resenting trace element diusion during formation of the from a heterogeneous provenance. The most striking zircons during the granulite facies metamorphism. Six of feature of the age spectrum is that grains were derived the seven abraded single zircon grains analysed (Table 1) from Proterozoic rocks (ca. <2.2 Ga). No Archean cluster on the concordia curve (Fig. 3) with overlapping grains were detected. Broadly, within the 207Pb/206Pb 262

50 µm 50 µm

e f

Fig. 2a±f SEM backscattered images of zircons and monazites from age spectrum, two age groups can be distinguished the mafic granulite W405, from the paragneiss W404 and from the (Fig. 5, Table 2). The first group includes grains with porphyritic granite W412. a and b typical anhedral, unzoned, inclusion-free grains from the mafic granulite (W405) reflecting ages ranging from 1150 to 1450 Ma and the second growth under high grade metamorphism; c and d rounded anhedral group is made up of grains with a spread in age of 1600– metamorphic monazites showing saw-tooth edges and irregular, finger 1900 Ma. The two groups are interpreted as represent- like, domains (dark) interpreted as recrystallisation of the original ing dominant ages of the source rocks. With the excep- (bright) monazite; e and f composite grains from the porphyritic tion of analysis 16-1, the youngest grain (19-1) is granite W412. Grains consist of a central euhedral domain with faint zoning and inclusions (the white inclusion in the light zoned part of concordant at 1151 18 Ma (2r) which constitutes a grain 2f is monazite) reflecting growth in a magma. Each grain is maximum value for the age of deposition of the clastic surrounded by a massive, unzoned, inclusion free metamorphic sediments. Isotopic disturbance during the ca. 1080 Ma overgrowth. Grain 2f in addition shows a central inherited core where high grade event is well supported both by the minimum the rounded shape suggests a detrital origin ages of discordant experimental points and by one concordant zircon grain (16-1) at 1059 32 Ma. The rounded shape of this grain, and the lack of visible in- Table 1 Conventional U-Pb data for accessory minerals from rocks of the Northampton Complex (Western Australia) (zr zircon, ti titanite, mo monazite, ap apatite, mltgr multigrain analysis; numbers in bracket refer to the number of grains analysed)a

Sample Weight U Pb* Pbc 208 Th/U 206 Pb/ 208 Pb/ 206 Pb/ 207 Pb/ 207 Pb/ Apparent ages (Ma) Disc (mg) (ppm) (ppm) (pg) mol% 204 Pb 206 Pb 238 U 235 U 206 Pb 206 Pb/ 207 Pb/ 207 Pb/ (%) (2r) (2r) (2r) 238 U 235 U 206 Pb

Ma®c granulite W405 1. Zr (1) 0.006 184 37 3 15.1 0.63 1177 0.19 0.1842 12 1.919 18 0.07556 44 1090 07 1088 06 1084 12 )0.6 2. Zr (1) 0.018 258 53 2 17.7 0.76 3482 0.23 0.1820 08 1.893 10 0.07543 19 1078 04 1079 04 1080 05 0.2 3. Zr (1) 0.004 330 70 2 20.0 0.88 1416 0.27 0.1812 11 1.879 15 0.07523 38 1074 06 1074 05 1075 10 0.0 4. Zr (1) 0.006 502 98 1 13.9 0.57 1611 0.17 0.1822 11 1.900 15 0.07562 39 1079 06 1081 05 1085 10 0.6 5. Zr (1) 0.008 695 150 1 21.6 0.97 4548 0.30 0.1831 05 1.902 07 0.07534 17 1084 03 1082 02 1078 05 )0.6 6. Zr (1) 0.005 852 179 2 20.9 0.94 4218 0.28 0.1789 05 1.849 06 0.07496 13 1061 03 1063 02 1068 04 0.6 7. Zr (1) 0.004 1209 235 4 13.4 0.54 3654 0.17 0.1810 05 1.881 06 0.07536 14 1073 03 1074 02 1078 04 0.5 Paragneiss W404 8. Mo (1) 0.004 1893 3050 105 89.6 30.35 694 9.21 0.1791 05 1.863 17 0.07544 62 1062 03 1068 06 1080 17 1.7 9. Mo (1) 0.009 2363 3511 378 88.5 27.25 590 8.27 0.1811 04 1.877 23 0.07515 84 1073 02 1073 08 1073 22 0.0 10. Mo (1) 0.006 2994 3469 44 85.5 20.88 2721 6.34 0.1800 04 1.865 06 0.07514 16 1067 02 1069 02 1072 04 0.5 11. Mo (1) 0.007 4998 3307 41 74.3 10.23 4871 3.11 0.1825 03 1.898 04 0.07543 11 1081 02 1080 02 1080 03 0.0 12. Mo (1) 0.009 11537 4707 68 58.2 4.94 11669 1.50 0.1831 03 1.907 04 0.07553 10 1084 02 1084 02 1083 03 )0.1 Granite W412 13. Zr (1) 0.005 480 92 3 4.7 0.17 2386 0.05 0.1951 13 2.125 16 0.07901 23 1149 07 1157 05 1172 06 2.0 14. Zr (1) 0.003 525 159 3 35.1 1.93 2086 0.58 0.2093 08 2.391 13 0.08286 28 1225 05 1240 04 1266 07 3.2 15. Zr (1) 0.004 868 152 10 5.1 0.19 2159 0.06 0.1765 05 1.852 07 0.07612 20 1048 03 1064 03 1098 05 4.6 16. Zr (1) 0.003 946 178 7 9.3 0.36 1872 0.11 0.1819 07 1.945 10 0.07784 25 1078 04 1097 03 1135 06 5.1 17. Zr (1) 0.003 1285 233 4 7.4 0.28 3134 0.09 0.1806 05 1.926 08 0.07735 18 1070 03 1090 03 1130 05 5.3 18. Zr (1) 0.003 1372 252 1 6.6 0.25 4158 0.08 0.1847 05 1.994 07 0.07829 15 1093 03 1113 02 1154 04 5.3 19. Zr (1) 0.002 1387 242 1 3.7 0.14 1468 0.04 0.1812 11 1.892 17 0.07571 43 1074 06 1078 06 1088 11 1.3 20. Zr (1) 0.002 1450 254 11 5.0 0.19 1709 0.06 0.1756 05 1.848 08 0.07634 24 1043 03 1063 03 1104 06 5.6 21. Zr (1) 0.005 1516 262 29 6.9 0.26 1975 0.08 0.1702 06 1.787 08 0.07616 21 1013 03 1041 03 1099 05 7.8 22. Mo (1) 0.005 3409 5461 178 89.4 29.85 808 9.06 0.1808 04 1.883 16 0.07553 58 1072 02 1075 06 1083 16 1.0 23. Mo (1) 0.005 4353 6608 412 88.8 27.97 559 8.49 0.1808 05 1.890 25 0.07583 91 1071 03 1078 09 1091 24 1.7 Calc-Silicate W410 24. Ti (2) 0.048 77 16 22 23.0 1.06 961 0.32 0.1737 04 1.788 10 0.07463 34 1033 03 1041 04 1059 09 2.4 25. Ti (2) 0.017 83 17 15 17.1 0.73 453 0.22 0.1724 09 1.776 20 0.07468 70 1026 05 1037 07 1060 19 3.2 26. Ti (2) 0.025 101 23 15 30.8 1.57 738 0.48 0.1635 06 1.681 12 0.07457 44 976 03 1001 05 1057 12 7.7 27. Ti (2) 0.021 102 27 28 33.5 1.79 503 0.54 0.1780 07 1.840 20 0.07497 71 1056 04 1060 07 1068 19 1.1 28. Ti (4) 0.075 105 31 264 34.6 1.87 308 0.57 0.1812 05 1.879 41 0.07519 151 1074 03 1074 15 1074 41 0.0 29. Ap, mltgr 0.200 49 17 285 55.1 4.32 311 1.31 0.1514 05 1.450 34 0.06947 152 909 03 910 14 913 45 0.4 30. Ap, mltgr 0.200 60 24 992 55.3 4.35 124 1.32 0.1502 08 1.448 96 0.06996 432 902 04 909 41 927 127 2.8 31. Ap, mltgr 0.420 95 32 756 55.7 4.43 494 1.35 0.1528 03 1.471 22 0.06982 97 917 02 919 09 923 29 0.7 Pegmatite W411 32. Zr (1) 0.006 328 50 1 0.1 0.005 32468 0.001 0.1644 03 1.636 04 0.07217 07 981 02 984 01 991 02 1.0 33. Zr (1) 0.004 2328 364 29 0.4 0.01 2383 0.004 0.1653 12 1.656 13 0.07266 17 986 07 992 05 1005 05 1.8 34. Zr (1) 0.021 4511 676 11 0.1 0.004 44052 0.001 0.1624 03 1.611 04 0.07195 06 970 02 975 01 985 02 1.5 35. Zr (1) 0.015 4593 746 6 8.1 0.31 42373 0.09 0.1619 04 1.606 04 0.07194 06 967 02 973 02 984 02 1.7 36. Zr (1) 0.015 5105 765 4 0.1 0.004 53476 0.001 0.1625 04 1.614 04 0.07203 06 971 02 976 02 987 02 1.6 37. Zr (1) 0.006 5288 808 21 0.1 0.004 10131 0.001 0.1650 04 1.640 04 0.07211 07 984 02 986 02 989 02 0.5 38. Zr (1) 0.007 6504 525 2 0.1 0.003 26810 0.001 0.1623 04 1.608 05 0.07184 06 970 03 973 02 981 02 1.2 39. Zr (1) 0.004 12062 1841 39 0.1 0.005 9569 0.001 0.1645 03 1.635 04 0.07208 08 982 02 984 01 988 02 0.6 a All zircon grains have been taken from the least magnetic bulk fractions and air abraded (Krogh 1982). Lead isotopic ratios have been corrected for fractionation, blank and initial 208 206 common Pb (after Stacey and Kramers 1975). Th/U ratios calculated from the radiogenic Pb/ Pb assuming concordance between the U-Pb and Th-Pb systems. The right hand 263 column is percentage discordance assuming recent lead losses. Errors are 2r and refer to last digits 264 surrounded by an unzoned overgrowth (Fig. 2e) al- 0.1854 NORTHAMPTON COMPLEX though most grains in addition contain a central 1090 W405 MAFIC GRANULITE rounded core unconformably surrounded by a rim of 0.1836 Zircons zoned and unzoned zircon (Fig. 2f). The euhedral 1080 zoned domains are interpreted to reﬂect igneous crys-

8 0.1818

2 ± tallisation in the porphyritic granite magma whereas the / 1070 1079 7 Ma b MSWD= 1.4 P unzoned outer part is interpreted as a metamorphic 6 0.1800

2 Weighted average overgrowth. On a concordia diagram (Fig. 6), analyses 1060 1079±3 Ma are scattered as a result of a combination of inheri- 6 MSWD= 0.7 0.1782 tance, new growth and subsequent Pb loss. From Ta- To 654±225 Ma ble 1, it can be seen that only one grain (analysis 19) is 0.1764 concordant within the uncertainties and has a 1.82 1.84 1.86 1.88 1.90 1.92 1.94 207Pb/206Pb age of 1088 11 Ma. Assuming that this 207 235 Pb/ U zircon grain does not have significant inheritance, and Fig. 3 Concordia diagram for conventional single zircon grain since it has been abraded to remove the outer rim, its analyses from the mafic granulite W405. Boxes are 2r error age is tentatively regarded as dating crystallisation of the porphyritic granite. However, as most zircons were observed to contain rounded cores (Fig. 2f), it cannot be ruled out that the single grain conventional mea- 0.1854 NORTHAMPTON COMPLEX W404 PARAGNEISS 1090 surements were influenced by a component of inherited Monazites Pb which has biased the ages to be too old. To in- 0.1836 1080 vestigate this, U-Pb analyses were made on cores and 12

U rims of ﬁve zircon grains using the SHRIMP II ion

8 0.1818 11

2 microprobe. The cores and rims were found to have / 1070

b 10

P Upper intercept very dierent U-Th contents and ages (Table 2 and 6 0.1800 9

0 2 1080±5 Ma Fig. 7). Cores had variable U and Th contents of 70– 1060 MSWD = 0.6 2000 ppm and Th/U ratios of 0.24–0.91, whereas rims 0.1782 8 had relatively uniform U contents of 500–750 ppm and To 664±278 Ma 0.1764 Th/U ratios of 0.03–0.14. The ages of cores ranged from 1162–2213 Ma (excluding the analysis of the core 1.82 1.84 1.86 1.88 1.90 1.92 1.94 from grain 2 which appears to have struck a crack in 207Pb/235U the grain or an inclusion). The age of rims was quite Fig. 4 Concordia diagram for conventional single monazite grain uniform with a weighted mean 207Pb/206Pb of analyses from the paragneiss W404. Boxes are 2r error 1068 13 Ma and a weighted mean 206Pb/238U age of 1068 22 Ma (Fig. 7). The lack of any resolvable age ternal structure, is consistent with growth or recrystal- dierence between the dierent zircon components in lisation of this zircon in the solid state during high grade the rim (magmatic and metamorphic, see Fig. 2f) indi- metamorphism. cates a short time span between emplacement of the porphyritic granite and its subsequent deformation. This zircon rim age is younger but within error of the Porphyritic granite conventional single grain analysis 19 (1088 11 Ma) which may result from a small inherited component in Porphyritic granite sample W412 has a strong NW fo- the single grain. liation defined by the alignment of K-feldspar pheno- Two measurements made on single grains of mona- crysts, biotite and sillimanite, which parallels regional zite (analyses 22 and 23) yield close to concordant ana- 207 206 S2. The shape-prefered alignment of phenocrysts is lyses. They have Pb/ Pb ages of 1083 16 Ma and similar to textures attributed by Blumenfeld and Bou- 1091 24 Ma but consistent 206Pb/238U ages of about chez (1988) to deformation whilst the granite was in a 1071 Ma (Table 1). The high uncertainties reflect the partly molten state. The granite also contains rounded high common lead contents. The zircon rim age is not elongate xenoliths of garnet granulite and pegmatite significantly dierent from the 206Pb/238U ages of mo- which are aligned parallel to the S2 foliation. The nazites at ca. 1071 Ma. Since the granite has been de- porphyritic granite intrusion appears to post-date peak formed during the second deformation event D2, it is metamorphism and formation of the S1 foliation that likely that monazites date this event. These relationships predates D2 deformation. Granite intrusion may have confirm geological evidence, such as the presence of taken place either during a stage of extensional collapse xenoliths with high grade mineral assemblages, that the late in D1 during which time pegmatite-filled normal porphyritic granite was emplaced late in the metamor- shear zones developed in the gneisses, or early in D2. phic event and we conclude that emplacement, crystal- Zircons from the porphyritic granite have rounded lisation and metamorphism took place within the external shapes and consist of a euhedrally zoned centre uncertainties of the zircon rim-monazite age. 265

Table 2 SHRIMP II U-Th-Pb results for zircons extracted from the paragneiss W404 and from the porphyritic granite W412a

Grain U Th Th/U 206Pb/ 208Pb/ 206Pb/238U 207Pb/235U 207Pb/206Pb 7/6 age (Ma) Disc area (ppm) (ppm) 204Pb 206Pb (1r error) (1r error) (1r error) (1r error) (%)

Paragneiss W404 1-1 150 194 1.29 425 .3786 .193 4 2.20 10 .0828 31 1265 74 10 2-1 3615 96 0.03 31957 .0077 .213 4 2.31 05 .0788 02 1167 06 )07 3-1 44 63 1.41 1256 .4096 .340 8 5.90 28 .1260 48 2043 68 08 4-1 382 323 0.85 4667 .2385 .320 7 4.75 12 .1076 12 1759 21 )02 5-1 910 368 0.40 2741 .1193 .259 5 3.58 08 .1003 08 1630 15 09 5-2 862 313 0.36 16422 .1023 .287 6 4.07 09 .1031 05 1681 10 03 6-1 223 442 1.98 4896 .5874 .221 5 2.48 08 .0814 17 1231 40 )05 7-1 734 440 0.60 8987 .1733 .238 5 2.79 06 .0849 07 1314 16 )05 8-1 356 277 0.77 3708 .2231 .251 5 3.13 09 .0903 15 1432 31 )01 9-1 62 44 0.70 696 .1944 .288 6 4.26 23 .1070 50 1750 85 07 10-1 32 24 0.76 490 .2022 .332 8 5.31 24 .1159 41 1895 64 02 11-1 331 193 0.58 882 .1580 .310 6 4.49 13 .1051 19 1716 33 )01 12-1 377 86 0.23 3452 .0713 .212 4 2.57 07 .0876 12 1375 25 10 13-1 117 131 1.12 2675 .3369 .198 4 2.32 10 .0851 29 1318 67 12 14-1 293 119 0.40 1838 .1044 .272 6 3.70 10 .0989 13 1603 24 03 15-1 230 116 0.50 1398 .1575 .291 6 4.43 13 .1103 20 1804 32 09 16-1 770 144 0.19 4584 .0529 .180 4 1.86 04 .0747 06 1059 16 )01 17-1 188 217 1.16 1438 .3263 .205 4 2.28 10 .0804 30 1208 73 00 18-1 124 96 0.78 763 .2143 .195 4 2.15 09 .0802 25 1202 61 05 19-1 1490 116 0.08 4309 .0236 .196 4 2.11 04 .0782 03 1151 09 00 20-1 674 278 0.41 2440 .1184 .258 5 3.64 08 .1022 09 1665 17 11 20-2 1207 1079 0.89 9990 .2472 .292 6 4.11 09 .1021 05 1662 09 01 21-1 351 184 0.52 2494 .1386 .295 6 4.08 11 .1004 14 1631 27 )02 22-1 370 486 1.31 3033 .3711 .235 5 2.77 07 .0855 12 1326 28 )03 23-1 158 164 1.03 1559 .3003 .182 4 1.98 09 .0795 31 1184 78 09 24-1 426 401 0.94 2632 .2695 .326 7 5.14 12 .1144 10 1870 17 03 25-1 422 320 0.76 3446 .2233 .202 4 2.32 06 .0832 12 1275 28 07 26-1 393 332 0.85 1794 .2486 .234 5 2.74 08 .0848 14 1311 32 )04 27-1 1114 2964 2.66 2787 .8318 .306 6 4.58 10 .1084 07 1773 11 03 28-1 196 99 0.50 2056 .1490 .236 5 3.09 10 .0950 20 1528 39 11 29-1 321 39 0.12 3010 .0368 .231 5 3.18 08 .0998 14 1621 25 17 30-1 245 166 0.68 1357 .1944 .215 4 2.43 07 .0819 15 1243 36 )01 31-1 32 42 1.33 503 .3861 .329 8 5.02 31 .1106 58 1809 96 )01 Granite W412 1-1R 520 35 0.07 3571 .0185 .181 3 1.86 03 .0745 06 1055 15 )01 1-2C 68 62 0.91 1389 .2599 .237 4 2.92 10 .0895 25 1415 54 03 2-1R 591 59 0.10 50000 .0298 .177 3 1.82 03 .0749 04 1064 12 02 2-2C 455 110 0.24 1299 .0700 .166 3 1.68 04 .0734 12 1026 33 04 2-3R 527 50 0.10 4545 .0276 .176 3 1.80 03 .0745 06 1054 17 01 3-1R 741 98 0.13 100000 .0376 .184 3 1.92 03 .0756 04 1084 09 )01 3-2C 1983 909 0.46 4167 .1363 .203 3 2.23 04 .0798 03 1191 08 00 4-1R 506 72 0.14 20000 .0422 .185 3 1.90 04 .0746 06 1058 16 )03 4-2C 144 75 0.52 14286 .1592 .333 6 6.37 12 .1389 10 2213 13 16 5-1R 726 22 0.03 16667 .0087 .181 3 1.87 03 .0749 04 1067 11 00 5-2C 270 235 0.87 14286 .2577 .197 3 2.13 04 .0786 08 1162 19 00 a Errors are 1r and refer to last digits. The right hand column is percentage discordance assuming recent lead losses

Calc-silicate fractions appear to be slightly discordant on the concordia plot (Fig. 8) although this is masked to some The calc-silicate W410 contains abundant titanites and extent by the large uncertainties in the 207Pb/206Pb ra- apatites which are part of the high grade equilibrium tios. The spread of the data does not allow calculation of mineral assemblage. Titanites and apatites have low U a discordia line but the 207Pb/206Pb weighted mean (from 49 to 105 ppm) and Pb (from 16 to 32 ppm) contents and relatively high common Pb with corre- provides an age of 921 23 Ma (MSWD = 0.1). spondingly increased uncertainty in the determination of 207Pb/206Pb ages. On a Concordia plot (Fig. 8), titanite Pegmatite analyses fall on a chord (MSWD = 0.3) with an upper intercept of 1063 13 Ma and a lower intersection not Biotite-bearing pegmatite sample W411 is representa- signiﬁcantly dierent from zero million years tive of the numerous post-D2 pegmatite dykes, which (95 275 Ma), suggesting that grains have not under- intrude the gneisses and cut across earlier garnet- gone ancient isotopic disturbance. The three apatite iferous pegmatites. Zircons from the pegmatite sample 266

Fig. 5 Concordia diagram for SHRIMP analyses of detrital zircons extracted from the paragneiss W404. Boxes are 1r error

Fig. 6 Concordia diagram for conventional single zircon and 0.212 monazite grain analyses from NORTHAMPTON COMPLEX the porphyritic granite W412. W412 Porphyritic Granite 1200 > 1266 Ma Shaded boxes are monazites, 0.204 white boxes are zircon. Boxes are 2r error 0.196 1150 U

238 0.188 1088±11 Ma 1100 (analysis 19) Pb / 1080 206 0.180 23 1050 Monazites 1060 22 and 23 22 0.172

0.164 1.68 1.82 1.96 2.10 2.24 2.38 207Pb / 235U

W411 have euhedral shapes and cores have been observed in some grains. From Table 1, six of the eight Discussion analyses have U contents between 4500 ppm to 12000 ppm which is typical of pegmatites. Analyses 34 Timing of metamorphic and igneous activity to 36 and 38 are slightly discordant (Fig. 9) and may be explained by a combination of ancient and recent In high grade metamorphic terranes, zircon growth or Pb loss. Our best estimate for the age of emplacement recrystallisation can be achieved during the prograde of the pegmatite is given by the 207Pb/206Pb weighted stage of metamorphism as protolith minerals are con- mean of the three least discordant analyses (32, 37 and verted to granulite mineral assemblages. At this stage, 39) which yields an age of 989 2 Ma. This is similar igneous pyroxene and amphibole may expel incompatible to the ca. 980 Ma Rb-Sr muscovite age from a mus- elements such as zirconium to form zircon (Nemchin covite bearing pegmatite reported by Wilson et al. et al. 1994). Such zircon generally forms with a low (1960). uranium content, typically in the range 5–50 ppm (e.g. 267

Fig. 7 Concordia diagram for SHRIMP analyses of zircons from the porphyritic granite W412. White boxes are rim analyses, black boxes are core analyses. Boxes are 1r error

Fig. 8 Concordia diagram for 1080 conventional multigrain ana- NORTHAMPTON COMPLEX 0.180 lyses on titanites and apatites W410 Calc-Silicate extracted from the calc-silicate 17 W410. Boxes are 2r error 1020 0.170

U 960

2 0.160 Titanite line: / Intercepts at

P 1063 ± 13 Ma

0 900 ± 2 95 275 Ma 0.150 (MSWD = 0.3) Apatite weighted 840 average: 0.140 921 ± 23 Ma (MSWD = 0.1)

0.130

1.26 1.40 1.54 1.68 1.82 207Pb / 235U

Krogh 1993) but up to 200 ppm (Nemchin et al. 1994). (Fig. 2c, d) of these monazites shows that the grains Zircons from the Woodbine mafic granulite have mor- exhibit complex shapes and internal structures charac- phologies characteristic of grains grown under high- terized by the occurrence of two distinct domains (U grade metamorphic conditions (see Fig. 2a, b) but their depleted and U undepleted) with complex relationships. U contents (180–1200 ppm) are higher than those ex- Such features have been reported from metamorphic pected in typical granulitic zircons. The parent material monazites and have been related to mineral-fluid inter- was possibly a basaltic andesite, which is unlikely to action (De Wolf et al. 1993; Hawkins and Bowring contain high U zircons such as those presently observed. 1997). Our present interpretation is that monazite grew This suggests that zircon crystallised in the presence of a or became a closed system at 1083 3 Ma, the high U fluids such as melts or metamorphic fluids re- 207Pb/206Pb age of the concordant grain 12 (Table 1). It leased from lower units during high grade metamor- has been reported that monazites in metamorphosed phism. The zircon age (1079 3 Ma) is therefore pelites form by destabilization of precursors such as interpreted as dating a time point within the high grade apatite and allanite or Th- and REE-rich oxides during metamorphism of the Northampton Complex. the prograde stages of metamorphism at temperatures of A similar U-Pb age of 1080 5 Ma was determined about 500 °C (Smith and Barreiro 1990). The peak on monazites from the paragneiss W404. SEM imaging metamorphic temperature, estimated at 800–900 °C, is 268

Fig. 9 Concordia diagram for conventional single zircon grain analyses from the pegmatite W411. Boxes are 2r error

in excess of the nominal closure temperature of monazite granite having a predominantly igneous texture but also of 725 25 °C (Copeland et al. 1988). More recently, containing textures indicative of mineral adjustments Smith and Giletti (1997) modelled the Pb diusion be- under metamorphic conditions where garnet is stable. haviour in monazite as a function of temperature-time The nature of the S2 foliation in the granite also indi- conditions and have shown that crystals of about cates that it was still in a plastic state during D2 and 100 lm in diameter can loose all their radiogenic lead therefore that any time gap between D2 and granite when exposed to temperatures of about 750 °C during emplacement should have been very short. The lack of ca. 10 Ma. Peak metamorphic temperatures of 800– resolvable age dierence between magmatic and meta- 900 °C imply a correspondingly shorter time for reset- morphic domains in the zircon grains further indicates ting by diusion and the ca. 1083 Ma age could be in- that deformation took place within the uncertainties of terpreted as reflecting initiation of the monazite the zircon age (1068 13 Ma). Interestingly, this is not radiometric clock on the downside of the metamorphic dierent from apparent ages of monazites from the path. The similarity of this age with the zircon age from granite (ca. 1071 Ma) and from the ages of the U-de- the Woodbine granulite (1079 3 Ma) suggests there- pleted monazites from the paragneiss (ca. 1072 Ma). fore an episode of fast cooling and/or rapid uplift as we The titanite age of 1063 13 Ma from the calc-sili- interpret the granulite facies zircons to form either prior cate is similar to the age of the porphyritic granite to, or at the peak of regional metamorphism. This is (1068 13 Ma), just within the uncertainty of the zir- consistent with W to NW thrusting of the gneissic units con age (1079 3 Ma) from the mafic granulite W405, at upper levels in the crust during the first deformation but significantly younger than the age attributed to D1 event D1. Determining the age of recrystallisation and U on the basis of monazites (1083 3 Ma) from the pa- depletion revealed by SEM imaging on some grains is ragneiss W404. Because titanite is present in the high not possible with the present data set. The youngest grade equilibrium assemblages, this younger age may 207Pb/206Pb age of 1072 4 Ma (analysis 10, Table 1) record a cooling point along the P-T-t path or corres- could be interpreted as a maximum age for monazite pond to recrystallisation and updating of this mineral recrystallisation but such a conclusion needs further during the second deformation event D2. Cases have substantiation. been reported (e.g. Corfu 1996) where titanite appears to The estimated age of the porphyritic granite zircons recrystallize or to form from breakdown of earlier (1068 13 Ma) is younger, but within the uncertainty phases during retrograde reactions. However, the equi- of the age of the granulite zircons (1079 3 Ma). The librium textural relationship of titanite to the M1 as- granite contains xenoliths of garnet granulite, and field semblage argues against a later recrystallisation of relationships show that granite intrusion postdated the titanite as an explanation in the present case. Studies of peak of regional metamorphism and the main D1 de- magmatic titanites (Pidgeon et al. 1996; Zhang and formation. On the other hand, zircons from the granite Scha¨ rer 1996) have emphasised that the U-Pb system of clearly show a massive, unzoned rim (Fig. 2e, f) attrib- this mineral can remain closed up to temperatures of uted to a metamorphic growth indicating that the about 650–700 °C. The textural compatibility of titanite porphyritic granite has also undergone high grade in the high grade equilibrium assemblage indicates this metamorphism. This can be explained if the porphyritic mineral formed at a temperature close to the peak of granite was emplaced late in the main metamorphic metamorphism (ca. 800–900 °C). This is well above the event. This explanation is in line with the porphyritic titanite closure temperature and the titanite age of 269 1063 13 Ma most likely records the closure temper- (1.6–2.0 Ga) on the northern margin of the Yilgarn ature age of ca. 650–700 °C. Craton (Fig. 1), a further source for the younger (1.15– In summary, it appears that peak metamorphic con- 1.45 Ga) detrital material is required. Sm-Nd isotope ditions were reached at ca. 1080 Ma. This was followed studies (McCulloch 1987) suggest that components of by fast cooling and/or rapid uplift, resulting in a fall in the Leeuwin Complex were extracted from the mantle at temperature to the titanite closure temperature of 650– 1130–1160 Ma (TDM ages) as well as at ca. 600 Ma. 700 °C by 1063 13 Ma. A further long time interval Mantle extraction ages of ca. 1800 Ma were also re- of about 140 Ma was needed for the temperature to corded for basement to the Perth Basin intersected in reach about 500 °C (at 921 23 Ma), the closure drill-core (Fletcher et al. 1985). Inherited components temperature of apatite (Cherniak et al. 1991). An ap- are widely distributed in the porphyritic granite zircons proximate mean cooling rate of about 1 °C Ma)1 be- and indicate large proportions of crustal material in the tween 1060 and 920 Ma can be calculated over a magma source regions. Excluding analysis 4–2 (Table 2), temperature drop of ca. 650–700 °C (Pidgeon et al 1996; which is significantly older (>2.2 Ga), the range in age Zhang and Scha¨ rer 1996) to 500 °C (Cherniak et al. of these inherited components (1100–1400 Ma) is 1991). This assumes no additional heating of the system broadly similar to the ages of the youngest group of occurred between 1060 and ca. 920 Ma. This may be an detrital zircons from the paragneiss W404. This suggests oversimplification as, whereas the intrusion of peg- that the paragneisses could be the source rocks of the matites at ca. 990 Ma may represent a final consolida- porphyritic granite melt. However, the age spectrum tion of fluids, this could also indicate a late reheating. does not match completely that of the detrital grains. In Shear zones cutting the Northampton Complex have particular, the lack of 1.6–1.9 Ga old grains (which been correlated with 655–550 Ma tectonothermal events dominate in the detrital zircon age distribution) casts (Compston and Arriens 1968), which in turn could be doubts on this assumption. These inherited components related to events in the Leeuwin Complex and reflect a may therefore reflect the ages of basement rocks to the Pan-African overprint (Harris 1994) in the southern part Northampton Complex. These results suggest that a of the Darling Mobile Belt. The non-zero lower inter- possible source for detrital material may have been from cepts (ca. 600 Ma) for zircons from the mafic granulite within the Darling Mobile Belt. It is also possible how- and monazites from the paragneiss may reflect the im- ever that sediments were derived from Greater India print of this event. However, preservation of the 920 Ma (placed next to the Northampton Complex in recon- age for apatite indicates that temperatures after this age structions of Rodinia and Gondwanaland; Fig. 10) or did not exceed 500 °C. This substantiates our conclusion that the last regional high grade event to occur in the Northampton Complex took place in Grenvillian times and that a significant Pan-African thermal event did not GREATER TIBET Pinjarra Orogen INDIA aect this area. EG AUSTRALIA AF Ms Correlation with other Grenvillian belts, source for Mirnyy TL metasediments and geodynamic implications EAST ANTARCTICA SIBERIA The age of the granulite facies metamorphism in the Northampton Complex is in the range of the worldwide KALAHARI Grenvillian orogeny and it is useful to compare the LAURENTIA O° chronology established for the Northampton Complex with other similar-aged belts of Gondwanaland. High grade metamorphic conditions were reached earlier in other Australian Proterozoic complexes such as the Musgrave Ranges of central Australia (>1200 Ma), the CONGO Albany Mobile Belt (ca. 1190 Ma) and the Fraser Mo- AMAZONIA bile Belt (ca. 1300 Ma) in southwestern Western Aus- tralia (Maboko et al. 1991, Pidgeon 1990, and Black WEST BALTICA et al. 1992 respectively). The ages presented herein for AFRICA the Northampton Complex indicate that at least part of the Darling Mobile Belt along the western margin of the Yilgarn Craton has also undergone Grenvillian events. In the paragneiss, the age spectrum presented by the pre-Grenville cratons Grenville-age belts detrital zircon populations indicates that detritus was derived from Proterozoic source material without con- Fig. 10 Global reconstruction at ca. 0.7 Ga (modified after Homan tribution from the Yilgarn Craton. Whilst a component 1991). AF Albany Fraser Orogen, EG Eastern Ghats, Ms Musgrave of sediments may be derived from the Capricorn Orogen Block, TL Tasman line 270 from mixed sources but not from the Yilgarn Craton, as demonstrated by the preservation of the ca. 920 Ma which may not have been a topographic high at this apatite age. time. Identification of Grenvillian rocks along the Western margin of the Yilgarn Craton necessitates a modification of commonly used Gondwanaland recon- Conclusions structions (e.g. Moores 1990; Dalziel 1991; Borg and DePaolo 1994), which portray a continuous Grenvillian Conventional and SHRIMP isotopic studies on the U- belt wrapping around southern Australia into the East- Pb systems of zircon, monazite, titanite and apatite in- ern Ghats of India, following the East Antarctic Shield. dicate that rocks of the Northampton Complex record A more complex Grenvillian reconstruction is envisaged an extended Grenvillian history during a time span ex- (Fig. 10), with a Grenvillian belt being developed along tending from 1150 Ma to 920 Ma but with a rapid se- the western margin of the Yilgarn Craton in addition to quence of events (including granulite facies the southeastern (Fraser Mobile Belt) and southern metamorphism, major deformation events and granite (Albany Mobile Belt) margins. The Albany Mobile Belt, intrusion) between 1080 and 1060 Ma. SHRIMP ana- in which the high grade event occurred at ca.1200 Ma, is lyses on detrital zircons from a psammitic gneiss indicate thought to continue through Meghalaya in NE India source rocks with ages of 1.15–1.45 Ga and 1.60– across the central India on the northern margin of the 1.90 Ga. No Archean zircons from the nearby Yilgarn Singhbhum Craton (Harris 1993, 1995), with another Craton were found in the detrital zircon population. branch continuing into the Eastern Ghats of India, Sri Sources of the detrital zircons may have been from the Lanka and contiguous East Antarctic terrains. The Capricorn Orogen on the northern margin of the Yil- present study raises the possibility that this belt bifur- garn Craton although other sources for detrital material cates into two branches, wrapping around the Yilgarn may have been from within the Darling Mobile Belt or Craton. Moreover, as the tectonic transport direction from Greater India. Deposition of these clastic sedi- during D1 is away from the Yilgarn Craton and towards ments, possibly corresponding to turbiditic sequences, Greater India (Byrne and Harris 1993), a Grenville-aged occurred between 1150 and 1080 Ma. Geochronological counterpart to the Northampton Complex is likely to be results from metamorphic rocks of the Complex dem- present within Greater India. This terrain may have onstrate that high grade metamorphism occurred at ca. undergone an Early Grenvillian event to provide the 1080 Ma and was quickly followed by the onset of ret- younger detrital 1150–1450 Ma zircons found in the rograde metamorphism associated with the first defor- Northampton Complex. It is not possible to positively mation event D1. Granitic activity, which took place identify this as the source region but U-Pb studies on between D1 and a second contractional event D2, oc- zircons from plutonic rocks in the Himalayas and in the curred at 1068 13 Ma and was followed by further Tibetan Plateau (Scha¨ rer and Alle` gre 1983; Scha¨ rer cooling to reach the apatite closure temperature at ca. et al. 1984, 1986) demonstrate that melting of Paleozoic 920 Ma. Post-D2 pegmatitic activity at 989 2 Ma to Precambrian rocks was involved in magma genesis. represents the waning stages of regional metamorphism. This is consistent with the view that a substantial part of This sequence of event is slightly younger than the main the gap between India and Australia can be filled with Grenvillian tectonothermal events on the southern and material now lost by subduction processes accompany- south-eastern margins of the Yilgarn Craton indicating ing the collision of India with Eurasia. amalgamation and orogeny around the margins of the On the other hand, the age of the granulite facies craton at dierent times. In contrast to the southern event in the Northampton Complex is significantly older Darling Mobile Belt and to East Antarctica, which un- than the age of high-grade rocks from the Leeuwin derwent granulite facies metamorphism and deforma- Complex (ca. 600 Ma after Wilde and Murphy 1990). tion during the ca. 500–600 Ma assembly of the Interpretation of pseudo-gravity imagery based on sat- Gondwanaland supercontinent, only minor resetting ellite altimeter data (Nash et al. 1996), combined with was recorded in the Northampton Complex, which was the results of dating basement encountered in bore-holes cut by discrete shear zones during this event. The in the Perth Basin (Fletcher et al. 1985), indicates a clear Northampton Complex therefore constitutes a remnant separation between a northern area representing an of a Grenvillian belt developed along the western margin oshore extension of the Northampton complex and a of the Yilgarn craton, and may have a counterpart in southern area made of a basement dominantly akin to Greater India. Global reconstruction of the Rodinia the Leeuwin Complex with a small sliver of older base- supercontinent should be modified to include the Dar- ment adjacent to the Yilgarn Craton. In a recent model ling Mobile Belt. (Harris 1994) it is suggested that formation of the Leeuwin Complex is related to localised crustal exten- Acknowledgements This study has been carried out in the labo- sion occurring within a sinistral transcurrent to trans- ratory of R.T. Pidgeon at the Curtin University of Technology pressional belt that may be linked into the Mirnyy area (Perth, Western Australia). O.B. was supported by an A.R.C. In- of the East Antarctic Shield. The Northampton Com- ternational Fellowship. D.B. acknowledges support from the CNRS and INSU. Fieldwork in the Northampton Complex was plex, although aected by discrete shear zones, escaped funded by Western Australian Metals and an ARC grant to L.H. regional metamorphism during this Pan-African event and carried out by D.I.B whilst the holder of an UWA postgrad- 271 uate scholarship. Helpful advice from Allen Kennedy, Peter Kinny Proterozoic Gondwanaland reconstructions. Geol Soc India and Sasha Nemchin while running the WA Consortium SHRIMP Mem 34: 47–71 II was greatly appreciated. Special thanks to Dom Furfaro and Hawkins DP, Bowring SA (1997) U-Pb systematics of monazite Warrick Chislholm for their help in the lab and when running the and xenotime: case studies from the Paleoproterozoic of the mass spectrometer. Constructive and helpful reviews by D.P. Grand Canyon, Arizona. Contrib Mineral Petrol 127: 87–103 Hawkins and an anonymous reviewer helped improve this manu- Homan PF (1991) Did the breakout of Laurentia turn Gondw- script. analand inside out? Science 252: 1409–1412 Krogh TE (1973) A low-contamination method for hydrothermal decomposition of zircon and extraction of U and Pb for isotopic age determinations. Geochim Cosmochim Acta 37: 485– References 494 Krogh TE (1982) Improved accuracy of U-Pb zircon ages by the Black LP, Harris LB, Delor CP (1992) Reworking of Archean and creation of more concordant systems using an air abrasion Early Proterozoic components during a progressive, Middle technique. Geochim Cosmochim Acta 46: 637–649 Proterozoic tectonothermal event in the Albany Mobile Belt, Krogh TE (1993) High precision U-Pb ages for granulite meta- Western Australia. Precambrian Res 59: 95–123 morphism and deformation in the Archean Kapuskasing Blumenfeld P, Bouchez J (1988) Shear criteria in granite and structural zone, Ontario: implications for structure and devel- migmatite deformed in the magmatic and solid states. J Struct opment of the lower crust. Earth Planet Sci Lett 119: 1–18 Geol 10: 361–372 LeMaitre RW (1984) A proposal by the IUGS Subcommission on Borg SG, DePaolo DJ (1994) Laurentia, Australia and Antarctica the Systematics of Igneous Rocks for a chemical classification as a late Proterozoic supercontinent: constraints from isotopic of volcanic rocks based on the total alkali silica (TAS) diagram. mapping. Geology 22: 307–310 Aust J Earth Sci 31: 243–255 Bosch D, Bruguier O, Pidgeon RT (1996) The evolution of an Ludwig KR (1987) Isoplot200, a plotting and regression program Archean metamorphic belt: a conventional and SHRIMP U-Pb for isotope geochemists, for use with HP series 200 computers. study on accessory minerals from the Jimperding Belt, Yilgarn US Geol Surv Open-file report 85–513 craton, West Australia. J Geol 104: 695–711 Maboko MAH, Williams IS, Compston W (1991) Zircon U-Pb Byrne DR, Harris LB (1993) Structural controls on the base-metal chronometry of the pressure and temperature history of gran- vein deposits of the Northampton Complex, Western Australia. ulites in the Musgrave Ranges, Central Australia. J Geol 99: Ore Geol Rev 8: 89–115 675–697 Cherniak DJ, Lanford WA, Ryerson FJ (1991) Lead diusion in Manhes G, Minster JF, Alle` gre CJ (1978) Comparative uranium- apatite and zircon using ion implantation and Rutherford thorium-lead and rubidium-strontium study in Saint-Severin backscattering techniques. Geochim Cosmochim Acta 55: amphoterite – consequences on early solar system chronology. 1663–1673 Earth Planet Sci Lett 39: 14–24 Compston W, Arriens PA (1968) The Precambrian geochronology Mathur SP, Shaw RD (1982) Australian orogenic belts: evidence of Australia. Can J Earth Sci 5: 561–583 for evolving plate tectonics. Earth Planet Sci Lett 4: 281–308 Compston W, Williams IS, Meyer C (1984) U-Pb geochronology of McCulloch MT (1987) Sm-Nd isotopic constraints on the evolution zircons from lunar breccia 73217 using a sensitive high-resolu- of Precambrian crust in the Australian continent. In: Protero- tion ion-microprobe. Proc 14th Lunar Planet Sci Conf, J zoic Lithospheric Evolution (Geodynamics Series 17), Int Geophys Res 89: B525–534 Lithospheric Prog Publ 0130, American Geophysical Union, Copeland P, Parrish RR, Harrison TM (1988) Identification of pp 115–130 inherited radiogenic Pb in monazite and its implications for U- Moores EM (1990) Southwest U.S.-East Antarctic (SWEAT) Pb systematics. Nature 333: 760–763 connection: a hypothesis, Geology 19: 425–428 Corfu F (1996) Multistage zircon and titanite growth and inheri- Myers JS (1990) Precambrian tectonic evolution of part of Go- tance in an Archean gneiss complex, Winnipeg River Sub- ndwana, southwestern Australia. Geology 18: 537–540 province, Ontario. Earth Planet Sci Lett 141: 175–186 Myers JS, Shaw RD, Tyler IM (1996) Tectonic evolution of Pro- Cumming GL, Richards JR (1975) Ore lead isotopes in a contin- terozoic Australia. Tectonics 15: 1431–1446 uously changing Earth. Earth Planet Sci Lett 28: 155–171 Nash C, Rankin L, Leeming P, Harris LB (1996) Delineation of Dalziel IWD (1991) Pacific margins of Laurentia and East Ant- lithostructural domains in northern Orissa, India from Landsat arctica-Australia as a conjugate rift pair: evidence and impli- Thematic Mapper imagery. Tectonophysics 260: 245–257 cations for an Eocambrian supercontinent. Geology 19: 598– Nelson DR (1997) Compilation of SHRIMP U-Pb zircon geo- 601 chronological data, 1996. Geol Surv West Aust Record 1997 De Wolf CP, Belshaw N, O’Nions RK (1993) A metamorphic Nemchin AA, Pidgeon RT, Wilde SA (1994) Timing of Late history from micron-scale 207Pb/206Pb chronometry of Archean Archean granulite facies metamorphism in the southwestern monazite. Earth Planet Sci Lett 120: 207–220 Yilgarn Craton of Western Australia: evidence from U-Pb Embleton BJJ, Schmidt PW (1985) Age and significance of mag- ages of zircons from mafic granulites. Precambrian Res 68: netizations in dolerite dykes from the Northampton Block, 307–321 Western Australia. Aust J Earth Sci 32: 279–286 Parrish RR (1987) An improved micro-capsule for zircon dissolu- Fletcher IR, Wilde SA, Rosman KJR (1985) Sm-Nd model ages tion in U-Pb geochronology. Chem Geol 66: 99–102 across the margins of the Archean Yilgarn Block, Western Parrish RR (1990) U-Pb dating of monazite and its application to Australia – III. The western margin. Aust J Earth Sci 32: 73–82 geological problems. Can J Earth Sci 27: 1431–1450 Harris LB (1993) Correlations between the Central Indian Tectonic Pidgeon RT (1990) Timing of the plutonism in the Proterozoic Zone and the Albany Mobile Belt of Western Australia: evi- Albany Mobile Belt, southwestern Australia. Precambrian Res dence for a continuous Proterozoic orogenic belt. In: Findlay 47: 157–167 RH, Unrug R, Banks MR, Veevers JJ (ed) Gondwana 8: as- Pidgeon RT, Aftalion M (1972) The geochronological significance sembly, evolution and dispersal. Balkema, Rotterdam, pp 165– of discordant U-Pb ages of oval-shaped zircons from a Lewisian 180 gneiss from Harris, Outer Hebrides. Earth Planet Sci Lett 17: Harris LB (1994) Neoproterozoic sinistral displacement along the 269–274 Darling Mobile Belt, Western Australia. J Geol Soc London Pidgeon RT, Furfaro D, Kennedy AK, Nemchin AA, Van 151: 901–904 Bronswjk W (1994) Calibration of zircon standards for the Harris LB (1995) Correlations between the Albany, Fraser and Curtin SHRIMP II (8th Int Conf on Geochronology, Cosmo- Darling mobile belts of Western Australia and Mirnyy to chronology and Isotope Geology, Berkeley). US Geol Surv Windmill Islands in the East Antarctic Shield: implications for Circular 1107: 251 272

Pidgeon RT, Bosch D, Bruguier O (1996) Inherited zircon and Stacey JS, Kramers JD (1975) Approximation of terrestrial lead titanite U-Pb systems in an Archean Katrine syenite from isotope evolution by a two stage model. Earth Planet Sci Lett 6: southwestern Australia: implications for U-Pb stability of tita- 15–25 nite. Earth Planet Sci Lett 141: 187–198 Steiger RH, Jager E (1977) Subcommission on geochronology: Richards JR, Blockley JG, De Laeter JR (1985) Rb-Sr and Pb convention on the use of decay constants in geo- and cosmo- isotope data from the Northampton Block, Western Australia. chronology. Earth Planet Sci Lett 36: 359–362 Bull Proc Australia Inst Min Metall 290(5): 43–55 Wilde SA, Murphy DMK (1990) The nature and origin of Late Scha¨ rer U, Alle` gre CJ (1983) The Palung granite (Himalaya): high- Proterozoic high-grade gneisses of the Leeuwin Block, Western resolution U-Pb systematics in zircon and monazite. Earth Australia. Precambrian Res 47: 251–270 Planet Sci Lett 63: 423–432 Williams IS, Compston W, Black LP, Ireland TR, Foster JJ (1984) Scha¨ rer U, Xu RH, Alle` gre CJ (1984) U-Pb geochronology of Unsupported radiogenic Pb in zircon: a cause of anomalously Gangdese (Transhimalaya) plutonism in the Lhasa-Xigaze re- high Pb-Pb, U-Pb and Th-Pb ages. Contrib Mineral Petrol 88: gion, Tibet. Earth Planet Sci Lett 69: 311–320 322–327 Scha¨ rer U, Xu RH, Alle` gre CJ (1986) U-(Th)-Pb systematics and Wilson AF, Compston W, Jeery PM, Riley GH (1960) Radio- ages of Himalayan leucogranites, South Tibet. Earth Planet Sci active ages from the Precambrian rocks of Australia. J Geol Soc Lett 77: 35–48 Aust 6: 179–195 Smith HA, Barreiro B (1990) Monazite U-Pb dating of staurolite Zhang LS, Scha¨ rer U (1996) Inherited Pb components in magmatic grade metamorphism in pelitic schists. Contrib Mineral Petrol titanite and their consequence for the interpretation of U-Pb 105: 602–615 ages. Earth Planet Sci Lett 138: 57–65 Smith HA, Giletti BJ (1997) Lead diusion in monazite. Geochim Cosmochim Acta 61: 1047–1055 C. R. Geoscience 335 (2003) 1081–1089

Geochemistry (Isotope Geochemistry) An Early-Cambrian U–Pb apatite cooling age for the high-temperature regional metamorphism in the Piancó area, Borborema Province (NE Brazil): initial conclusions

Bruno Dhuime a, Delphine Bosch a, Olivier Bruguier b, Renaud Caby a,∗, Carlos Archanjo c

a Laboratoire de tectonophysique, université Montpellier-2, UMR 5568–CNRS/UMII, place Eugène-Bataillon, 34095 Montpellier cedex 05, France b Service ICP–MS, ISTEEM, cc 056, université de Montpellier-2, place Eugène-Bataillon, 34095 Montpellier cedex 05, France c Instituto de Geociências, University of São Paulo, Brazil

Received 28 February 2003; accepted 24 September 2003

Presented by Georges Pédro

Abstract The Borborema Province (BP) of northeastern Brazil is a complex crustal assemblage, which has undergone a polycyclic evolution during the Proterozoic. In the Piancó-Alto Brígida belt, a metamorphosed leucosome vein inserted in amphibolites has a trace element pattern suggesting a T-MORB protolith. Apatites yield a REE pattern indicating growth in equilibrium with garnet, thus pointing to its metamorphic origin. U–Pb analyses yield an age of 540±5 Ma interpreted as a cooling age following amphibolite facies regional metamorphism associated with granitic emplacement at ca. 580 Ma. The resulting slow cooling rates ◦ − (ranging from ca. 2.5 to 5 CMa 1) are consistent with underplating of maﬁc magmas, or crustal thickening caused by nappe stacking, as possible processes governing the metamorphic evolution of the BP. To cite this article: B. Dhuime et al., C. R. Geoscience 335 (2003).  2003 Académie des sciences. Published by Elsevier SAS. All rights reserved. Résumé Âge de refroidissement U–Pb sur apatite du Cambrien inférieur pour le métamorphisme de haute température de la région de Piancó, province de Borborema (Nord-Est du Brésil) : premières conclusions. La Province Borborema (BP), dans le Nord-Est du Brésil, est un assemblage complexe ayant subi une évolution polycyclique durant le Protérozoïque. Dans la ceinture de Piancó-Alto Brígida, un leucosome métamorphisé, associé à des amphibolites, présente des caractéristiques d’éléments en trace, suggérant un protolithe de type T-MORB. Les apatites présentent un spectre de terres rares indiquant une croissance à l’équilibre avec le grenat, ce qui traduit une origine métamorphique. Les analyses U–Pb fournissent un âge de 540 ± 5 Ma, interprété comme un âge de refroidissement suivant le pic du métamorphisme régional daté à 580 Ma. Les taux ◦ − de refroidissement faibles, compris entre 2.5 et 5 CMa 1, sont compatibles avec des processus de sous-placage magmatique

* Corresponding author. E-mail addresses: [email protected] (B. Dhuime), [email protected] (R. Caby).

1631-0713/$ – see front matter  2003 Académie des sciences. Published by Elsevier SAS. All rights reserved. doi:10.1016/j.crte.2003.09.012 1082 B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089 ou d’épaississement crustal pour expliquer l’évolution métamorphique de la BP. Pour citer cet article : B. Dhuime et al., C. R. Geoscience 335 (2003).  2003 Académie des sciences. Published by Elsevier SAS. All rights reserved.

Keywords: Borborema Province; apatite; U–Pb dating; cooling age

Mots-clés : province Borborema ; apatite ; datation U–Pb ; âge de refroidissement

Version française abrégée d’une séquence métabasique mésocrate à amphibole– clinopyroxène, associée à des niveaux calco-silicatés. 1. Introduction Elle est localisée à une dizaine de kilomètres au nord du massif granitique d’Itaporanga, daté à 584 ± 2Ma La province Borborema (BP), dans le Nord-Est [6]. L’échantillon renferme 50% de grenat almandi- du Brésil, est constituée par un socle paléoprotéro- neux, 30% de quartz, 11% de plagioclase calcique, 6% zoïque incluant quelques noyaux archéens et des for- d’amphibole verte et 1% de biotite. Les minéraux ac- mations métasédimentaires protérozoïques. Le « do- cessoires observés sont les sulfures, l’apatite, le zircon maine transversal », délimité par les deux grands ci- et le sphène. Le thermomètre amphibole–plagioclase saillements dextres de Patos et de Pernambuco, se ca- [3] fournit une température d’équilibre de l’ordre de ractérise par la présence d’assemblages volcaniques 640 ◦C à 4 kbar. et plutoniques d’âge Mésoprotérozoïque et de bassins turbiditiques du Néoprotérozoïque terminal [2, 3. Techniques analytiques 26]. L’existence de plutons synorogéniques (Itapo- ranga, Campina Grande, Acari, Caraubas, S. Rafael, Les apatites ont été dissoutes sur plaque chauffante Faz Nova...), datés par la méthode U–Pb sur zir- en acide chlorhydrique 6 N. La séparation du plomb cons autour de 580 Ma [6,14] indique que l’ensemble et de l’uranium a été effectuée selon la méthode des structures et le métamorphisme régional datent de décrite par Krogh [17]. Les blancs de procédure au cette époque. Ce travail se propose de dater, au sein du cours de la période d’analyse ont varié entre 15 domaine de Piancó-Alto Brígida (PAB), le refroidis- et 30 pg pour le Pb et moins de 5 pg pour l’U. sement associé au métamorphisme régional de haute La composition isotopique du Pb radiogénique a été température en utilisant le géochronomètre U–Pb sur déterminée en retirant le blanc, puis en supposant pour apatite, combiné avec des analyses d’éléments en trace le Pb commun restant une composition déterminée et de microscopie électronique à balayage. Les résul- d’après le modèle de Stacey et Kramers [24]. Le calcul tats obtenus sont comparés avec ceux connus dans des points expérimentaux et des âges a été effectué d’autres parties de la BP. en utilisant le programme de Ludwig [18]. La mesure des éléments en traces a été réalisée par ICP–MS, 2. Cadre géologique suivant une méthode de calibration externe en vigueur à Montpellier et décrite en détails par Ionov et al. [15]. Au sein du domaine transversal, la ceinture PAB (Fig. 1) comprend des métasédiments, des méta- 4. Résultats et discussion volcanites acides, des orthogneiss à foliation sub- horizontale replissée et des gneiss anatectiques d’âge L’échantillon de roche totale présente un spectre Paléoprotérozoïque présumé. Le métamorphisme ré- plat (Fig. 2), légèrement appauvri en terres rares lé- gional a culminé dans le faciès amphibolite, avec ana- gères, comparable au N-MORB. Les teneurs en terres texie à proximité des plutons granitiques. La plupart de rares sont cependant élevées (environ 50 fois les ces plutons se sont mis en place en conditions synciné- chondrites), ce qui suggère une affinité de type T- matiques à 580 Ma [6,22], qui est l’âge du pic du méta- MORB [11]. L’anomalie négative en Eu suggère que morphisme régional. La roche étudiée provient d’une le liquide initial a subi un processus de cristallisa- bande leucocrate de 30 cm d’épaisseur, située au sein tion fractionnée. L’échantillon est interprété comme B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089 1083 un leucosome issu de la fusion partielle de basaltes al- vés le long du cisaillement de Patos. Ces bassins sont térés, similaire aux veines leucosomiques observées en contemporains de la mise en place de plutons grani- domaine océanique (par exemple, [13]). Le spectre de tiques de type A, d’âge Cambrien supérieur. La préser- terres rares des apatites est marqué par un appauvrisse- vation de l’âge U–Pb sur apatite indique que cet évé- ment prononcé en terres rares légères et lourdes. Ceci nement n’a pas été associé à une anomalie thermique indique une croissance à l’équilibre avec des phases d’ampleur régionale. métamorphiques majeures (telles que le plagioclase et le grenat) qui incorporent ces éléments de façon pri- 5. Conclusions vilégiée. La structure interne des apatites révèle l’ab- sence de figures magmatiques telles que zonage ou L’âge de 540 ± 5 Ma, obtenu par la méthode U–Pb surcroissance (Fig. 3). Ces critères permettent de défi- sur les apatites métamorphiques d’un leucosome pré- nir une origine métamorphique pour les apatites analy- levé au sein d’un complexe métabasique, est interprété sées. Dans le diagramme Concordia (Fig. 4), les quatre comme celui du refroidissement lent de la ceinture fractions d’apatites définissent un âge 206Pb/238Ude PAB au début du Cambrien. Cet âge fait le lien avec 540 ± 5Ma(MSWD= 2,9). La température de ferme- ceux connus au Nord et au Sud et indique des taux de ture de l’apatite vis-à-vis de la diffusion du Pb est de refroidissement lents (globalement compris entre 2,5 ◦ l’ordre de 450–550 C pour des vitesses de refroidis- et 5 ◦CMa−1) identiques dans l’ensemble de la BP, ◦ − sement comprises entre 1–10 CMa 1 [8]. Dans cette bien que le refroidissement soit plus précoce au Sud. partie du domaine PAB, le pic du métamorphisme ré- gional culmine à environ 580 Ma et 640 ◦C. Une baisse de température depuis 640 ◦C à 580 Ma jusqu’à 450– 1. Introduction 550 ◦C à 540 Ma, permet de calculer des taux de refroidissement compris entre 2,5 à 5 ◦CMa−1.Cesva- The Borborema Province (BP) of northeastern Bra- leurs sont similaires à celles obtenues le long du ci- zil is a remote area of about 450 000 km2, located saillement de Patos par Corsini et al. [10] et suggèrent between the São Francisco Craton to the south and une évolution rétrograde et un refroidissement gouver- the Amazon Craton to the northwest. It constitutes the nés par des processus de remontée isostatique et de continuation in South-America of the Pan-African mo- faibles taux d’érosion. Les parties les plus métamor- bile belts of Western Africa that occupied the north- phiques et granitisées de la ceinture PAB ont donc subi ern part of Gondwana [7] (see Fig. 1). This com- un refroidissement lent comparable et synchrone de plex domain is made up of several displaced ter- celui observé dans d’autres parties de la BP situées au ranes, which include large areas floored by Archean nord du cisaillement de Patos [20]. Cette évolution est and Proterozoic crustal units and Late Neoprotero- cependant différente de celle connue pour le Sud de la zoic turbiditic basins [25]. Growing geochronologi- BP, qui a subi des taux de refroidissement comparables cal evidence suggests a polycyclic evolution during mais de façon plus précoce, comme l’indiquent les the Paleoproterozoic (2.0–2.2 Ga), Mesoproterozoic âges Ar de 580 Ma obtenus sur amphiboles par Neves (0.95–1.10 Ga) and Neoproterozoic (0.58–0.64 Ga) et al. [21]. Les taux de refroidissement étant identiques periods [2,26]. The Paleoproterozoic event is proba- sur l’ensemble de la BP (environ 3 à 5 ◦CMa−1), ceci bly the most important and corresponds, in the stud- suggère un diachronisme avec migration de la défor- ied area, to a major period of crust formation, asso- mation, du magmatisme et du métamorphisme associé ciated metamorphism and granite emplacement. U–Pb du Sud vers le Nord. Les faibles taux de refroidisse- zircon dating of gneisses from the Transversal Domain ment observés ne sont pas compatibles avec des pro- gave ages ranging from 1.97 to 2.20 Ga [2,22]. Two cessus de délamination et suggèrent une évolution mé- major periods of felsic magmatism have been recog- tamorphique liée, soit à des processus de sous-placage nized, around 1.5 Ga [22] and during the 1060–930 Ma de magmas basiques, qui réduisent l’uplift et réchauf- time span (the so-called Cariris Velhos event) [25]. fent le système, soit à un épaississement crustal causé Syn- to late-kinematic granites (Itaporanga, Campina par l’empilement de nappe. Enfin, de petits bassins de Grande, Acari, Caraubas, S. Rafael, Faz Nova...)were molasse continentale non métamorphique sont préser- intruded around 580 Ma [6,14], suggesting that all tec- 1084 B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089

Fig. 1. A. Location of the Borborema Province (NE Brazil) in a pre-drift reconstruction. B. Geological sketch map of the Transversal zone of the Borborema Province. indicates location of the studied sample. Neoproterozoic granites not represented. Abbreviations as follows: AM, Alto Moxoto; AP, Alto Pajeu; GAC, Guyana–Amazon Craton; PAB, Piancó-Alto Brígida; RC: Rio Capibaribe; SFC: São Francisco Craton; SLC: São Luis Craton, PaL: Patos Lineament, PeL: Pernambuco Lineament. Fig. 1. A. Localisation de la province Borborema (Nord-Est du Brésil) dans une reconstruction ante-ouverture de l’Atlantique. B. Carte géologique simplifiée de la Zone transversale de la province Borborema. indique la localisation de l’échantillon étudié. Les granites néoprotérozoiques n’ont pas été représentés. Les abréviations correspondent à : AM, Alto Moxoto ; AP, Alto Pajeu, GAC,craton Guyano-Amazonien ; PAB, Piancó-Alto Brígida ; RC, Rio Capibaribe ; SFC, craton de São Francisco ; SLC, craton São Luis ; PaL, linéament de Patos ; PeL, linéament de Pernambuco. tonic features as well as the regional syn-kinematic lite facies, low-pressure anatexis being restricted to metamorphism are of Neoproterozoic age [7]. The domains of coalescent Neoproterozoic plutons such present study combines SEM images, ICP–MS trace as the Itaporanga granite, the root of which includes element analyses and U–Pb dating on apatite in or- sillimanite-bearing diatexites free of low tempera- der to date cooling down to ca. 500 ◦C associated with ture retrogression [1]. The selected sample is a 30- the last high-temperature regional metamorphic event cm-thick garnetiferous leucocratic band interlayered that affected this area. Analyses were performed on within a melanocratic Fe- and Ca-rich amphibole– apatites from a metamorphosed mafic rock association clinopyroxene rock from the Tito Braz gold prospect outcropping in the Piancó-Alto Brígida (PAB) domain located few kilometres north of the 584 ± 2MaIta- (Fig. 1) and the results are compared with data from poranga pluton [6]. The sample is composed of garnet other parts of the BP. (50%), quartz (30%), labradorite (11%), green amphibole (6%), biotite (1%). Fe sulphides, zircon, titanite and apatite occur as accessory minerals. Electron mi- 2. Geological setting croprobe analyses and the amphibole-plagioclase ther- mometer of Blundy and Holland [3] gave an equi- The PAB domain is located in the southern part of libration temperature of about 640 ◦C for an esti- the BP, between two major shear zones: the Patos and mated pressure of 4 kbar, which we interpret as the Pernambuco lineaments [26]. Rocks of the PAB domain stage of regional amphibolite facies metamor- main mainly consist of undifferentiated Mesoprotero- phic overprint (data are available upon request from zoic and Late Neoproterozoic metasediments, metavol- the first author: [email protected]). canics and polymetamorphic gneisses, all affected by a recumbent foliation. Most granitoids (about 30% in outcrop) were intruded during the regional deforma- 3. Analytical techniques tion as their thermal aureoles are syn-kinematic and coeval with the regional metamorphism [6]. Regional Apatites were separated by conventional mineral metamorphism grades from green-schist to amphibo- separation methods including high-density liquids and B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089 1085

Table 1 ICP–MS analytical data for the whole rock sample and the apatite fraction from the PAB 50 sample Tableau 1 Résultats des analyses par ICP–MS de la roche totale et de la fraction d’apatite de l’échantillon PAB 50 Sample name PAB 50 Analysis Whole rock Apatite Weight (mg) 50 0.032 Rb 2.00.31 Sr 69 196 Y8984 Zr 22 4.4 Nb 9.6– Fig. 2. Chondrite-normalized rare-earth element patterns for whole Cs 0.22 0.05 rock and apatites from the PAB 50 sample. Chondrite values are Ba 31 4.0 from McDonough and Sun [19]. La 8.323 Fig. 2. Spectre de terres rares normalisées aux chondrites pour la Ce 28 134 roche totale et la fraction d’apatite de l’échantillon PAB 50. Les Pr 4.035 valeurs des chondrites sont de McDonough et Sun [19]. Nd 22 309 Sm 7.9 140 Eu 2.624 a Frantz magnetic separator (e.g., [4]). Apatites were Gd 12.5 118 dissolved overnight in a Savillex teflon beaker on a Tb 2.29.2 hot plate with 6 N HCl. For isotopic analyses, lead Dy 15.432 and uranium were separated following the chemistry Ho 3.13.3 . . described by Krogh [17] and measurements were car- Er 8 950 Tm 1.30.40 ried out on a VG Sector mass spectrometer. Total Pb Yb 8.31.4 blanks over the period of the analyses ranged from 15 Lu 1.40.12 to 30 pg and uranium blanks were less than 5 pg. The Hf 0.67 – isotopic composition of radiogenic Pb was determined Ta 0.71 – . . by subtracting the blank Pb and then the remainder, Pb 3 214 Th 0.62 0.13 assuming a common Pb composition at the time of U0.66 5.0 initial crystallisation [24]. Calculations were made us- Nb/Ta 13.4 ing the program of Ludwig [18]. For trace element de- Th/U 0.95 0.03 termination, after dissolution, samples were diluted in (La/Sm)CN 0.68 0.11 . . HNO3 shortly before analyses. Concentrations were (La/Yb)CN 0 72 11 4 determined on a VG Plasmaquad II ICP–MS (preci- Eu/Eu* 0.81 0.55 sion of 3 to 5%) by external calibration using multi- element solutions prepared from 10 mg ml−1 single elemental solutions [15]. displays a flat to slightly LREE depleted pattern with (La/Sm)CN and (La/Yb)CN ratio close to 0.7, similar to normal MORB derived from a depleted asthenospheric 4. Results and discussion mantle. The rock however is characterized by high REE content of about 50 times the chondrites, which 4.1. Trace element analyses suggests affinity with a T-MORB protolith [11]. In addition, the pattern displays a slight negative Eu Trace-element analyses of the whole rock sample anomaly despite the high modal plagioclase content, and of a small apatite fraction (ca. 30 mg) are suggesting the sample has undergone some degree reported in Table 1 and shown in the chondrite- of differentiation and mineral segregation. The Nb/Ta normalized diagram of Fig. 2. The whole rock sample ratio (13.4) is within error of the MORB value (14.4 1086 B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089

Fig. 3. SEM backscattered image of apatite from the PAB 50 sample. Fig. 3. Image de microscopie électronique à balayage (électrons Fig. 4. U–Pb Concordia diagram for apatite fractions of the PAB 50 rétrodiffusés) d’une apatite de l’échantillon PAB 50. sample. Fig. 4. Diagramme Concordia pour les fractions d’apatite de ±1.7 after [16]) again suggesting a depleted mantle l’échantillon PAB 50. source. The sample is then tentatively regarded as a leucosome vein resulting from hydrous partial melting ses are close to concordant (Fig. 4) and show a spread 207 235 206 238 of altered basalts by similarity to leucocratic melts in the Pb/ U ratios but identical Pb/ Ura- 206 238 found within the roof zones of axial magma chamber tios. The four fractions provide a Pb/ U weighted [13, e.g.]. The apatite REE-normalised pattern shows mean age of 540 ± 5Ma(MSWD= 2.9) interpreted a pronounced depletion in LREE and HREE with a as our best estimate for the age of the metamorphic negative Eu anomaly. The latter is typical of apatites apatite. from various geodynamical settings and host rocks [23]. The marked depletion in both LREE and HREE 4.3. Implication on the post-metamorphic evolution indicates that growth of the apatite occurred after or during crystallisation of LREE- and HREE-rich Although its application is limited by relatively low phases, such as plagioclase and garnet respectively. U and high common Pb contents, apatite is an im- These phases are major metamorphic components of portant accessory phase in U–Pb geochronology as it the rock as they occur with a modal percentage of is common in many igneous and metamorphic rocks. 11 and 50%, respectively, and can therefore seriously Due to its low ionic porosity, apatite is thought to have compete with apatite for REE during its growth. a simple U–Pb system. This contrasts with other phos- phatic phases such as monazites for example which 4.2. U–Pb analyses may contain multiple growth zones within individual crystals [5] and preserve an older age in an inher- Apatites occur as a uniform population of small ited core [9]. Thus, important information on the post- (100–150 µm) colourless translucent grains. The grains peak metamorphic evolution can be obtained from the have rounded shapes and are generally free of inclu- apatite results. Cherniak et al. [8] proposed a clo- sions. SEM images of polished surfaces (Fig. 3) show sure temperature for Pb diffusion in apatite in the that they exhibit structureless internal domains. No range 450–550 ◦C for cooling rates of 1–10 ◦CMa−1 cores, oscillatory zoning patterns or magmatic over- and crystal sizes of 100–500 µm. Electron microprobe growths have been observed which is consistent with study indicates that the sample underwent a metamor- a metamorphic origin. The four fractions analysed phic overprint of about 640 ◦C, therefore exceeding present low radiogenic Pb and U contents and conse- the nominal closure temperature of apatite for the U– quently have low 206Pb/204Pb ratios (Table 2). Analy- Pb system. The 540 ± 5 Ma apatite age is thus in- B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089 1087 ± U 238 Pb/ Apparent age (Ma) 206 error) ± 0078 539 7 0047 542 3 0054 541 4 0061 535 4 . . . . σ Pb 206 0686 0 0547 0 0763 0 0567 0 . . . . Pb/ 207 error) (2 098 0 060 0 070 0 077 0 ± . . . . σ U 235 825 0 661 0 920 0 667 0 . . . . Pb/ 207 error) (2 ± 0011 0 0005 0 0006 0 0006 0 . . . . σ U 238 0873 0 0877 0 0875 0 0866 0 . . . . Pb/ 206 Pb 206 01 0 94 0 05 0 08 0 . . . . Pb/ 208 Pb 21 00 21 61 204 . . . . Pb/ * 206 836 938 735 833 . . . . 82 44 51 54 . . . . 88 7 43 14 95 4 04 11 . . . . Sample WeigthnamePAB U 50 (mg)Ap1 (ppm) Pb (ppm) 0 (2 Ap2 2 Ap3 5 Ap4 7 Table 2 U–Pb isotopic analyses for apatite fractionsTableau from 2 the PAB 50 sample Résultats isotopiques U–Pb des fractions d’apatite de l’échantillon PAB 50 1088 B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089 terpreted as recording a cooling point on the down- sition was coeval with felsic volcanism and with em- side of regional metamorphism and it is suggested placement of A-type hypo-volcanic granites of Late- that the temperature has dropped from 640 ◦C at ca. Cambrian age in the northern part of the BP [12]. 580 Ma (the age of peak metamorphism) to 450– Preservation of the 540-Ma apatite age indicates that 550 ◦C at 540 Ma. This indicates that after peak meta- this magmatic activity was not associated with a sig- morphism, the temperature decreased slowly at a rate nificant thermal anomaly of regional extent. of 2.5 to 5 ◦CMa−1. Similar values (ranging from 3 to 4 ◦CMa−1) were calculated from Ar datings along the Patos shear zone and in the Seridó area [10]. Such 5. Conclusions slow cooling rates suggest that denudation occurred by processes consistent with isostatic recovery and low Conventional U–Pb analyses of apatite extracted erosion rates. This contrasts with tectonically assisted from a leucosome vein of T-MORB affinity outcrop- retrograde evolution associated, for example, with late ping in the Piancó-Alto Brigida (PAB) domain of the metamorphic thrusting and nappe stacking and with Borborema Province (BP) provide an age of 540 ± the differential uplift and exhumation of deep crustal 5 Ma. This value is interpreted as a cooling age follow- units along major shear zones. It is also noteworthy ing regional amphibolite facies metamorphism with that the 540 Ma apatite age is identical to the 540– cooling rates of about 2.5 to 5 ◦CMa−1.Theseare 550 Ma Ar amphibole ages provided by rocks from similar to cooling rates obtained from other parts of the the western termination of the Patos lineament and BP although cooling occurred earlier in the south sug- from the Cedro belt, north of this lineament [20]. Since gesting a possible north–south diachronic evolution. the amphibole closure temperature for Ar is similar to the U–Pb closure temperature of apatite, it is con- cluded that crustal units on each side of the linea- References ment undergone the same post-metamorphic cooling history. Conversely, rocks from the southern part of [1] C.J. Archanjo, E.R. Da Siva, R. Caby, Magnetic fabric and the BP (Caruaru area) yielded identical cooling rates pluton emplacement in a transpressive shear zone system: ◦ − ◦ (ca. 5 CMa 1), but cooling down to 500 C occurred the Itaporanga porphyritic granitic pluton (northeast Brazil), earlier, at ca. 580 Ma [21]. This suggests a possi- Tectonophysics 312 (1999) 331–345. ble diachronous evolution for the tectonometamorphic [2] P.R. Bastos Leite, J.-M. Bertrand, E.S. Lima, J. Leterrier, Timing of granitic magmatism in the northern Borborema event in the BP during the Braziliano/PanAfrican time. Province, Brazil: a U–Pb study of granitoids from the Alto This period of tectonism, metamorphism and igneous Pajeú Terrain, J. South Am. Earth Sci. 13 (2000) 549–559. emplacement represents a major period of plate tec- [3] J.D. Blundy, T.J.B. Holland, Calcic amphibole equilibria tonic activity resulting in the final assembly of the and a new amphibole–plagioclase geothermometer, Contrib. Gondwana supercontinent. The low cooling rates ob- Mineral. Petrol. 104 (1990) 208–224. served in the whole BP, except in the vicinity of ma- [4] D. Bosch, O. Bruguier, R.T. Pidgeon, The evolution of an Archean Metamorphic Belt: a conventional and SHRIMP U– jor shear zones where cooling had reached values Pb study on accessory minerals from the Jimperding Belt, ◦ −1 of 20 CMa (e.g., [20]), and the possible diachro- Yilgarn Craton, West Australia, J. Geol. 104 (1996) 695–711. nism are not compatible with delamination phenom- [5] D. Bosch, D. Hammor, O. Bruguier, R. Caby, J.M. Luck, ena which are typically associated with uplift and high Monazite ‘in-situ’ 207Pb/206Pb geochronology using a small erosion rates. Underplating of mafic magmas, which geometry high-resolution ion probe. Application to Archean and Proterozoic rocks, Chem. Geol. 184 (2002) 151–165. reduces uplift while adding heat to the system, or [6] B.B. Brito Neves, C.R. Passarelli, M.A.S. Basei, E.J. San- crustal thickening caused by nappe stacking and re- tos, Idades U–Pb em zircao de alguns granitos classicos da sponsible for partial melting and granite emplacement, provincia Borborema, Boletim Geol. USP, serie cientifica, 3 are more likely processes to explain the Neoprotero- (in press). zoic metamorphic evolution of the BP and the asso- [7] R. Caby, A. Sial, M. Arthaud, A. Vauchez, Crustal evolution and the Brasiliano orogeny in Northeast Brazil, in: ciated long-lived thermal anomaly. Lastly, relics of R.D. Dallmeyer, J.-P. Lécorché (Eds.), The West African Oro- non-metamorphic continental molassic basins are pre- gens and Circum Atlantic Correlatives, Springer-Verlag, 1991, served along the Patos lineament [7]. Molasse depo- pp. 373–397. B. Dhuime et al. / C. R. Geoscience 335 (2003) 1081–1089 1089

[8] D.J. Cherniak, W.A. Lanford, F.J. Ryerson, Lead diffusion [17] T.E. Krogh, A low-contamination method for hydrothermal in apatite and zircon using ion implantation and Rutherford decomposition of zircon and extraction of U and Pb for isotopic backscattering techniques, Geochim. Cosmochim. Acta 55 age determinations, Geochim. Cosmochim. Acta 37 (1973) (1991) 1663–1673. 485–494. [9] P. Copeland, R.R. Parrish, T.M. Harrison, Identiﬁcation of [18] K.R. Ludwig, Users manual for ISOPLOT/EX version 2. inherited radiogenic Pb in monazite and its implications for A geochronological toolkit for Microsoft Excel, Berkeley U–Pb systematics, Nature 333 (1988) 760–763. Geochronology Center, Spec. Publ. 1a (1999) 1–30. [10] M. Corsini, L. Lambert de Figueiredo, R. Caby, G. Féraud, [19] W.F. McDonough, S.S. Sun, The composition of the Earth, G. Ruffet, A. Vauchez, Thermal history of the Bor- Chem. Geol. 120 (1995) 223–253. borema Province deduced from 40Ar/39Ar analysis, Tectono- [20] P. Monié, R. Caby, M.H. Arthaud, The Neoproterozoic Brasil- 40 39 physics 285 (1998) 103–117. iano orogeny in northeast Brazil: Ar/ Ar and petrostruc- [11] J.P. Eissen, T. Juteau, J.L. Joron, B. Dupré, E. Humler, tural data drom Cearà, Precambrian Res. 81 (1997) 241–264. A. Al’Mukhamedov, Petrology and geochemistry of basalts [21] S.P. Neves, A. Vauchez, G. Féraud, Tectono-thermal evolution, ◦ from the Red Sea axial rift at 18 north, J. Petrol. 30 (1989) magma emplacement, and shear zone development in the 791–839. Caruaru area (Borborema Province, NE Brazil), Precambrian Res. 99 (2000) 1–32. [12] A.H. Fetter, U–Pb and Sm–Nd geochronological constraints on [22] J.-M. Sá, J.-M. Bertrand, J. Leterrier, M.H.F. Macedo, Geo- the crustal framework and geologic history of the Cear’state, chemistry and geochronology of the pre-Brasiliano rocks from NW Borborema Province, NE Brazil: implications for the the Transversal Zone, Borborema Province, Northeast Brazil, assembly of Gondwana, PhD Thesis, University of Kansas, J. South Amer. Earth Sci. 14 (2002) 851–866. 1999, 145 p. [23] L.K. Sha, B.W. Chappell, Apatite chemical composition deter- [13] K.M. Gillis, L.A. Coogan, Anatectic migmatites from the roof mined by electron microprobe and laser ablation inductively of an ocean ridge magma chamber, J. Petrol. 43 (1991) 2075– coupled plasma mass spectrometry as a probe into granite pet- 2095. rogenesis, Geochim. Cosmochim. Acta 63 (1999) 3861–3881. [14] I.P. Guimarães, C.N. Almeida, A.F. Da Silva Filho, [24] J.S. Stacey, J.D. Kramers, Approximation of terrestrial lead J.M. Araújo, Granitoids marking the end of the Brasiliano isotope evolution by a two-stage model, Earth Planet. Sci. (Pan-African) orogeny within the Central Tectonic Domain of Lett. 26 (1975) 206–221. the Borborema Province, Rev. Bras. Geociências 30 (2000) [25] W.R. Van Schmus, B.B. Brito Neves, P. Hackspacher, 117–181. M. Babinski, U/Pb and Sm/Nd geochronological studies of the [15] D.A. Ionov, L. Savoyant, C. Dupuy, Application of the ICP– Eastern Borborema Province, Northeast Brazil: initial conclu- MS technique to trace element analysis of peridotites and their sions, J. South Am. Earth Sci. 8 (1995) 267–288. minerals, Geostandards Newslett. 16 (1992) 311–315. [26] A. Vauchez, S. Pacheco Neves, R. Caby, M. Corsini, [16] F. Kalfoun, Géochimie de Nb–Ta: contraintes sur la différen- M. Egydio-Silva, M. Arthaud, V. Amaro, The Borborema shear ciation terrestre, PhD thesis, University Montpellier-2, France, zone system, J. South Am. Earth Sci. 8 (1995) 247–266. 2001, 300 p. Int J Earth Sci (Geol Rundsch) (2003) 92:338–347 DOI 10.1007/s00531-003-0321-3

ORIGINAL PAPER

O. Bruguier · J. F. Becq-Giraudon · N. Clauer · H. Maluski From late Visean to Stephanian: pinpointing a two-stage basinal evolution in the Variscan belt. A case study from the Bosmoreau basin (French Massif Central) and its geodynamic implications

Received: 3 April 2002 / Accepted: 9 February 2003 / Published online: 17 April 2003 Springer-Verlag 2003

Abstract Post-convergence evolution of the Variscan Keywords Delamination · French Massif Central · belt is characterized by the development of intramontane Intramontane basins · Stephanian · Visean coal-bearing basins containing volcano-sedimentary successions. In the French Massif Central, K–Ar ages on clay particles from fine-grained sediments of the Bosmoreau Introduction basin (Limousin area), help pinpoint the evolution of the basin. In the lower part of the sedimentary pile, illite in a Extensional tectonics is preferentially located along siltstone underlying a volcanic layer previously dated at orogenic belts with a thickened crust and collapse of 332€4 Ma by the U–Pb method on zircon, yields a mountain belts represents an important feature of post- consistent K–Ar age of ca. 340 Ma. Upward in the collisional orogenic stages (e.g. Ratschbacher et al. 1989). sedimentary succession, illite yields Stephanian K–Ar Implicit to this is the creation of a pervasive series of ages, which can be combined to provide a mean continental basins accompanying extension in the upper deposition age of 296.5€3.5 Ma. The Bosmoreau basin, crust. Whereas most studies addressing evolution of albeit mainly filled with Stephanian deposits, was initi- orogenic belts usually deal with basement and deep ated during the late Visean, i.e. ca. 30 Ma earlier than crustal processes (Faure and Pons 1991), an alternative inferred from biostratigraphical constraints. During the approach, whenever possible, is to focus on sedimentary Stephanian, the same structure was reactivated and late records and to use sedimentary basins as tectonic markers, Visean deposits were eroded and subsequently blanketed considering that their formation mirrors deeper processes by thick clastic sediments. These results emphasise a two- (Zoback et al. 1993). Basin opening and sedimentary stage evolution for the Bosmoreau basin, which is closely infilling are often tectonically controlled (e.g. Bruguier et related to extensional tectonics identified on basement al. 1997). A key issue, therefore, is to determine precisely country rocks, and they are used to propose a geodynamic both the age of basin formation and the main stages of evolution of the studied area. infilling, and the relationships between basin initiation and tectonic structures such as major fault systems. These parameters potentially carry important information that have implications for understanding the tectonic control O. Bruguier ()) on the sedimentary record, and to punctuate the different Service ICP-MS, cc 056, ISTEEM, stages of extensional tectonics. The implied requirements Universit de Montpellier II, Place E. Bataillon, for such an approach to be fully operative are tightly 34095 Montpellier, France related to the possibility of selecting suitable sedimentary e-mail: [email protected] records and appropriate chronological methods. J. F. Becq-Giraudon The French Massif Central is one of the most BRGM, 3 Avenue C. Guillemin, BP 6009, important exposures of the Internal Zone of the Variscan 45060 Orlans, France Belt, which extends along ca. 3,000 km from the Iberian Massif in the West to the Bohemian Massif in the East N. Clauer (Fig. 1). In the whole belt, the Late Carboniferous–Early CGS-EOST, Universit Louis Pasteur-CNRS, 1 rue Blessig, 67084 Strasbourg, France Permian time interval is characterized by numerous coal- bearing intramontane basins corresponding to isolated H. Maluski troughs closely associated with fault-zones and filled with Laboratoire de Gophysique, Tectonique et Sdimentologie, coarse, clastic, non-marine sediments deposited uncon- CNRS-UMR 5573, ISTEEM, Universit de Montpellier II, formably on the metamorphic and igneous basement. The Place E. Bataillon, 34095 Montpellier, France 339 Fig. 1 Position of the French Massif Central within the Eu- ropean Variscides (after Matte 1986)

structural control of the opening and further infilling of Geological setting these basins is obvious from their association with faults, at least along one of their borders, and it has been widely The Bosmoreau coalfield is a small graben filled in by documented (e.g. Faure 1995, and references therein). continental Carboniferous deposits and located in the In this study, we have focused on the Bosmoreau basin north-western part of the French Massif Central (Limou- located in the northern part of the French Massif Central sin area). It is entirely bounded by normal brittle faults (Fig. 2). The sedimentary successions are well preserved and developed in the hanging-wall block of the northern in the basin, and they are known in detail. In addition, a end of the Argentat fault (Fig. 2). Sedimentary rocks biostratigraphic support provided by floral records allows unconformably overlie the Late Devonian Guret granitic age constraints to be placed (Becq-Giraudon 1985). This massif dated at 356€10 Ma by the Rb–Sr method on basin is located in the hanging-wall blocks of the Argentat whole rocks (Berthier et al. 1979). Basin fill is estimated fault, a major fault system of this part of the Variscan to be about 600 m thick and consists mainly of orogen (Fig. 2). Following our previous work on this siliciclastic fluvio-lacustrine to palustrine sediments, basin (Bruguier et al. 1998), our goal is herein to date the including coal seams. The lithostratigraphy of the basin sedimentary succession, by selecting illite particles for K– can be divided into four third-order sequences (Fig. 3). Ar analyses from fine-grained sedimentary rocks sampled Sequence 0 is only exposed in the south-western part of at different levels in the sedimentary pile. This approach the basin and consists of about 30 m of fine-grained was thought to help to pinpoint basin infilling and sediments, which include a 60-cm-thick volcanic ash relationships with nearby structuring faults, which may be layer. U–Pb zircon results from this volcanic ash layer viewed as a basin response to more global, crustal scale, yielded a Late Visean upper intercept age of 332€4 Ma, processes. interpreted as the age of eruption of the magma and, 340

Fig. 2.a Location of the main Stephanian–Autunian basins of the French Massif Central. b Simplified geologic sketch map of the Bosmoreau basin showing the main features (after Becq-Giraudon 1985) therefore, of deposition of the airborne ash in the radiogenic 40Ar for five determinations. The blank of the Bosmoreau basin (Bruguier et al. 1998). Sequence 0 is extraction lines and the mass spectrometers was also separated from sequence 1 by an erosional unconformity determined repetitively. During the course of the study, and the ages of sequence 1 to 3 are upper Stephanian the amount of residual 40Ar was always below 110-7 cm3 according to their palaeobotanical and palynological and the 40Ar/36Ar ratio of the atmospheric Ar averaged records (Becq-Giraudon 1985). 293€6 (2s). The usual decay constants were used for the age calculations (Steiger and Jger 1977), and the overall error of the K–Ar determinations was evaluated to be Analytical techniques systematically better than €2% (2s).

K–Ar isotopic determinations were made following a procedure close to that reported by Bonhomme et al. Results (1975). K was measured by flame spectrophotometry with a global accuracy of €1.5%, based on systematic controls Three fine-grained sediments were sampled at different of international standards. For Ar analysis, the samples levels in the sedimentary pile for K–Ar analyses (see were pre-heated under high vacuum at 100 C for at least Fig. 3). Sample ST1 belongs to sequence 0 and was taken 12 h to reduce the amount of atmospheric Ar adsorbed on from a micaceous siltstone underlying the volcanic ash the mineral surface during sample preparation and CI1. Sample ST4 is from the base of sequence 1 and was handling. The Ar isotopic results were periodically taken from a 0.5-m-thick micaceous siltstone located 3 m controlled by analysis of the international GL-O standard, above sequence 0. Sample ST7 is a brownish micaceous which averaged 24.40€0.15x10-6 cm3/g STP (2s)of siltstone located at the base of sequence 3. The clay 341 Fig. 3 Lithostratigraphic column of the sedimentary sequence accumulated in the Bosmoreau basin. A symbol denotes sample locations 342 Table 1 K–Ar isotopic results and mineralogical composition of Ar; the amounts in 40Ar* are given in the STP system. I/S Illite/ 40 -10 the investigated clay fractions. Constants used: l Kb=4.96210 smectite mixed layers; chl chlorite; sm smectite; fsp feldspar -1 40 -10 -1 40 -4 an , l Ke=0.58110 an , K/Ktot=1.16710 . Ar* radiogenic 40 Sample Grain size Mineralogy I/S Illite K2O Ar* Ar* Age (m) by XRDa (%) crystallinity (%) (%) (106 cm3/g) (Ma€2s) ST1 <0.2 I/S 100 9.5 4.06 87.24 48.84 342.4€8.7 duplicate – – – 35.70 49.40 345.9€10.9 ST4 <0.2 I/S, chl, fsp 85 11.4 6.14 95.09 63.41 295.0€6.7 0.2–0.4 I/S, chl, fsp 80 10.5 6.36 96.30 65.52 299.9€6.5 ST7 <0.2 I/S, chl, sm 7.20 96.50 75.34 298.6€6.5 0.2–0.4 I/S, chl, sm, fsp 7.88 98.28 80.63 292.5€6.2 a Ordered from highest to lowest concentrations separates provide K–Ar ages ranging from 293€6 to commonly observed when participation of detrital com- 346€11 Ma (2s), and can be divided in two groups. ponents occurs in clay concentrates (Clauer and Chaud- Sample ST1 yields the oldest age of 342€9 Ma (Table 1). huri 1995; Clauer et al. 1995; Schaltegger et al. 1995). This value was verified by a replicate analysis providing a value in close agreement (346€11 Ma) and giving an average value of 344€10 Ma. Although slightly older, Discussion and implications these Visean ages are similar within analytical uncertainty to the U–Pb zircon age of 332€4 Ma for the overlying Fine-grained sedimentary rocks preserved in the lower volcanic ash layer Ci1. A second group of ages was found part of the sedimentary pile that accumulated in the for illite of siltstones ST4 and ST7 sampled upward in the Bosmoreau basin have K–Ar ages of ca. 340 Ma, sedimentary pile. These two samples provide a narrow consistent with the 332€4 Ma U–Pb zircon age of a cluster of younger ages ranging from 293 to 300 Ma. volcanic ash layer (Bruguier et al. 1998). These results Individual error margins do not allow a distinction to be indicate that parts of the sedimentary succession are older made between the two samples and preclude any than Stephanian and date back to late Visean, a conclu- sedimentation rate to be calculated. However, the consis- sion that could not be drawn from palaeobotanical studies. tency between samples ST4 and ST7 makes it possible to Typical, Stephanian floral records were identified in the combine the data for an average upper Stephanian age of upper part of the basin and were used to attribute to 296.5€3.5 Ma (MSWD=1.1). Clauer and Chaudhuri successions accumulated in this basin an upper Stepha- (1995, 1998) and Dong et al. (1997) showed that thermal nian age (Becq-Giraudon 1985). These observations agree conditions necessary to reset the K–Ar systems of detrital with the mean K–Ar age of 296.5€3.5 Ma exhibited by and diagenetic micas correspond to the anchizone– clay particles recovered from siltstones sampled in epizone boundary and that diagenetic ages can be sequences 1 and 3 of the sedimentary pile. This age falls preserved for lower grade samples. A recent study on in the Gzelian stage of the Stephanian series (Odin 1994). coal maturation of the Bosmoreau coal indicates that the Radiometric dating results are scarce for Stephanian Carboniferous sediments were submitted to a palaeo- occurrences in the French Massif Central, including the geothermal gradient of 90 C/km during Permian times continental reference series of the St-Etienne basin, and (Copard 1998). As the sediments were never buried comparisons, thus, are limited. However, the ca. 297 Ma deeper than 1,500 m (Copard et al. 2000), the samples K–Ar age for sedimentation in the Bosmoreau basin is in were not heated above 150 C subsequent to deposition, close agreement with muscovite and biotite 40Ar/39Ar and the close agreement between the K–Ar results from ages (297€3 Ma) from the southern and northern part of sample ST1 and the U–Pb zircon age from the nearby ash the Montagne Noire area. These 40Ar/39Ar ages have been layer CI1, substantiates that the K–Ar ages can be interpreted as marking movement along an active detach- interpreted as corresponding to the depositional age of the ment (Maluski et al. 1991) and the contemporaneous sediments. The illite crystallinity value, which measures development of the Graissessac basin (see Fig. 2). Lastly, the diagenetic evolution of authigenic clay, is also a this age compares very well with those determined for valuable tool to distinguish between metamorphic and basins throughout the Variscan Belt (e.g. Breitkreuz and diagenetic ages. Diagenetic illites show crystallinity Kennedy 1999; Kninger et al. 2002). This is taken as values above 9, whereas metamorphic ones have lower evidence for a synchronous basin-forming event that values (Kbler 1984); this parameter is also grain-size occurred at orogenic belt scale, at the end of the dependent. From Table 1, it can be seen that all clay Carboniferous. separates analysed have illite crystallinity values above 9, The data presented here have general implications on again substantiating that the K–Ar values date a low- the evolution of the Internal Zone of the Variscan Belt, temperature crystallization. Finally, the K–Ar age deter- especially when compared with the available structural minations do not show any relationships of increasing and geochronologic data. The Variscan belt of Europe has ages with increasing grain-sizes, which is a feature been the site, from Visean to Permian, of repeated post- 343 collisional magmatic pulses that are characterized by The occurrence of the late Visean volcanic ash layer distinct geochemical signatures and geodynamic settings interbedded within sedimentation is also evidence that (Schaltegger 1997). In the French Massif Central, large magmatism in the deep crust and the associated surface volumes of granitoids were emplaced in the middle crust, volcanism were synchronous with basin formation. from 330 to 290 Ma (see review in Pin and Duthou 1990; Although volcanic sources have not been identified for Ledru et al. 1994), and basement studies have demon- this material, the similarity in age suggests a likely origin strated that extensional tectonics occurred during two from the neighbouring Late Visean ‘Tufs Anthracifres’ distinct periods (Faure 1995). The first, apparently located about 60 km east of the Bosmoreau basin (Scott et protracted, extensional event took place from late Visean al. 1984). Late Visean granitoids (340–325 Ma) are to Westphalian, while convergence prevailed at the scale numerous in the French Massif Central (Ledru et al. 1994; of the whole orogen (Van der Voo 1982; Matte 1986). It Faure et al. 2002), indicating widespread magmatism that propagated diachronously southward across the French is not restricted to this part of the Variscan belt, but is also Massif Central, starting in Visean times in the north to well documented in other segments of the Internal reach the southern margin during the Westphalian when Variscides, such as the northern Bohemian massif of the nappe stacking and ductile deformation were still pro- Saxothuringian domain (Wenzel et al. 1997; Krner et al. ceeding in the southern foreland of the Montagne Noire 1998), the Southern Vosges (Schaltegger et al. 1996) and area (Maluski et al. 1991). The second extensional event, the Central Alps (Schaltegger and Corfu 1992). The from late Stephanian to Autunian, has been related to widespread occurrence of this event suggests orogen-wide collapse during late stage evolution of the Variscan belt processes and was attributed to thermal relaxation during (Malavieille 1993; Becq-Giraudon and Van den Driessche episodic thinning of the lithosphere (Schaltegger 1997). 1994; Faure 1995); it is characterized by numerous half- In the French Massif Central, where the crystalline graben geometry basins. In the case of the Bosmoreau basement broadly consists of ca. 50% granitoids, the basin, U–Pb and K–Ar results substantiate that initiation ages available (Ledru et al. 1994; Faure 1995) suggest, in of opening occurred as early as the late Visean, contem- contrast, a continuum of magmatic activity from Visean poraneously with the first period of extensional tectonics. to Permian times, although large error margins often Earlier work (Becq-Giraudon 1985) and additional field result in dates straddling the magmatic pulses identified in observations collected during this study are summarized other parts of the Variscan belt (Schaltegger 1997). in the synthetic lithostratigraphic column of Fig. 2. The Therefore, this may hamper identification of distinct upper Visean ages are restricted to sequence 0 cut by the events occurring in a rather limited range of time. Results overlying conglomeratic sandstones of sequence 1. from the Bosmoreau basin point to two episodic periods Because of this erosional unconformity, it is unknown of basin evolution. The older, upper Visean, volcano- whether sequence 0 was initially restricted to the upper sedimentary sequence reflects basin opening and was Visean or extended through the Namurian or even coeval with explosive volcanism and extensional faulting Westphalian. Siltstone ST4 belonging to sequence 1, is along major crustal-scale faults. The second, Stephanian, located about 3 m above sequence 0 and gave a although not dated on volcanic material in the case of the Stephanian age (see Table 1) concordant with biostrati- Bosmoreau basin, is otherwise characterized by wide- graphical data. This indicates that the Stephanian sedi- spread pyroclastic tuffs (Bouroz 1966). This observation mentation started in the Bosmoreau basin with deposition indicates episodicity of uplift and basement erosion, of the thick conglomeratic sandstones separated from synchronous to magmatic activity and suggests a rela- sequence 0 by an erosional unconformity. Deciphering tionship between uplift and thermal pulses. whether sedimentation was continuous, from upper What caused these episodic events and this ‘accor- Visean to Stephanian is not possible in the case of the dion’-like evolution is probably one of the key issues Bosmoreau basin. However, Namuro-Westphalian sedi- challenging our understanding of the post-collision ments are not known in the French Massif Central and, Variscan evolution. Any proposed model should reconcile thus, we suggest that the two sedimentary cycles were the widespread and apparently pervasive post-collisional episodic and related to two distinct phases of uplift and magmatism in the inner part of the belt, together with the basement erosion. Identification of these two successions, occurrence of episodic magmatic/volcanic flare-up at moreover, indicates a long, although episodic, lifetime for distinct periods, associated with uplift and basin forma- the basin of over 35 Ma. The Argentat fault cuts across tion. These magmatic pulses clearly imply that significant the whole French Massif Central and borders the perturbations in the thermal regime have occurred. It is Bosmoreau basin on its eastern side. 40Ar/39Ar muscovite also probably noteworthy that the Carboniferous basins of ages (Roig et al. 1997) obtained on rocks located in the the French Massif Central are intramontane basins with shear zone developed along the Argentat fault at the high altitude sedimentological and geomorphological expense of metamorphic rocks, yield apparent age evidence (e.g. Becq-Giraudon et al. 1996). Because the patterns in the range of 335–337 Ma. The close agreement Stephanian sediments were unconformably deposited on between these 40Ar/39Ar ages and sedimentation in the metamorphic rocks of various grade, this indicates an basin, as inferred from K–Ar and U–Pb ages, is taken as important erosion of the reliefs sometime before the evidence for a genetic link between fault movement and Stephanian. Conceivably, long-lived magmatism in the basin opening. deep crust and episodic pulses can be combined. The 344 long-lived extensional tectonics observed in the present asthenospheric melting can occur) as can be expected in a outcropping late Visean to Westphalian mid-crustal collisional setting with a thickened crust (Davies and von granitoids (Faure 1995) suggests that the crust had been Blanckenburg 1995). Extensional tectonics experienced softened and had flowed over time, which, in turn, implies by mid crustal granitoids (Faure 1995) indicates thermal continuous heat advection. This is a phenomenon that can softening and a possible decoupling of the upper crust. be explained by lithospheric delamination (Nelson 1992) The above-envisioned mechanisms could explain the first and which has been already suggested to explain the post- Visean pulse and the long-lived magmatism observed in collision evolution of the Variscan belt (Pin and Duthou this part of the Variscan belt during the Namuro- 1990; Schaltegger 1997). In agreement with this model, Westphalian period because lithospheric delamination we envision that, after the continental collision, the has a characteristic time of ca. 60 Ma (Nelson 1992). A subducting lithosphere underwent extension related to second phenomenon, however, is needed to explain the buoyancy contrast between the continental and oceanic upper Stephanian (300–295 Ma) pulse, which is charac- part of the slab (see Fig. 4a). Due to opposite internal terized by instantaneous and widespread development of forces, detachment may have occurred at the weak point basins and intense volcanism in the whole Variscan belt of the system, i.e. at the junction between oceanic and (e.g. Breitkreuz and Kennedy 1999). As pointed out continental crust and initiated delamination of the litho- above, such a flare-up implies a strong perturbation of the spheric mantle (Fig. 4b). The detached continental crust thermal regime. Although this needs further substantia- then rotated upwards and started underplating to the tion, we suggest it may result from slab break-off and Armorica/Laurussia block. This was responsible for uplift detachment of the sinking mantle part of the lithosphere and the formation of extensional volcano-sedimentary (Fig. 4d). A large influx of asthenospheric material would basins such as those now found in the Limousin area (this then be allowed to replace the lithospheric roots, while the study), the north-eastern part of the French Massif Central orogen uplifted and extended due to gravitational insta- (Bertaux et al. 1993; Faure et al. 2002), the southern bilities. Because the lower crust was hot and already Vosges (Schaltegger et al. 1996), the southern Schwarz- softened by about 30 Ma of heat advection and/or wald (Schneider et al. 1989) and in some fragments of the production, it was able to flow rapidly. Mechanical Variscan basement preserved in the Central Alpine area extension was thus predominant over erosion as demon- (Schaltegger and Corfu 1995; Von Raumer 1998). The strated by the preservation of the Stephanian intramontane coincident thinning of the mantle part of the lithosphere basins. Slab break-off can be explained by gradual and rise of the asthenosphere triggered the first, late- thinning and final detachment of the sinking lithospheric Visean magmatic event, now represented by granitoids mantle, but also by the Late Carboniferous–Early Permian and associated volcanic expressions such as those found clockwise rotation of Gondwana (Matte 2001) and in the Bosmoreau basin or the so-called ‘Tufs An- resulting large strike-slip faults cutting across Central thracifres’ further east (Scott et al. 1984). This event was Europe and Northern Africa (Bard 1997). These structures synchronous to granulitic metamorphism identified in may have intersected parts of the sinking slab and, thus, some segments of the southern side of the Variscan belt facilitated slab break-off before the general N–S Permian (e.g. Krner et al. 1998, 2000) and suggests that collapse of the whole orogen. From this standpoint, this granulitization of the lower crust occurred at that time. two-stage model (delamination and slab break-off) may From late Visean to Stephanian (325–305 Ma), delami- explain the episodicity of volcanic and basin-forming nation proceeded by peeling away and underplating of the events along with the apparently long-lived magmatism. continental crust (Fig. 4c) resulting in uplift and the formation of high Namuro-Westphalian topographic reliefs whose erosional products accumulated in foreland Conclusions basins of the External Zones such as the Zone Houillre Brianonnaise (Brousmiche-Delcambre et al. 1995), the The Bosmoreau basin in the Limousin area of the French Montagne Noire area and the Ardennes (Matte 1986). As Massif Central records two main phases of basin infilling, the sinking slab looses its lightest part, it steepens, thus both coeval with extensional tectonics affecting the allowing a continuous thinning of the lithosphere and basement country rocks. uplift of asthenospheric material. The resulting mantle heat advection, along with the radiogenic heat accumu- – The first sedimentary cycle started in the upper Visean lated in the thickened crust, was responsible for melting with a continental succession deposited in troughs of the lower crust. Rising magmas were trapped in the developed in the hanging wall blocks of the Argentat middle crust to produce the numerous Namuro-West- fault. K–Ar analyses of fine-grained sedimentary rocks phalian granitoids or erupted on the surface. High erosion gave upper Visean ages (ca. 340 Ma), concurring with rates during this time interval suggest that volcanic the U–Pb zircon age of an overlying volcanic ash layer products were probably removed. As demonstrated by Sr previously dated at 332€4 Ma (Bruguier et al. 1998). and Nd isotope studies (Pin and Duthou 1990), the mantle These ages are similar to 40Ar/39Ar muscovites ages component involved in the source of these granitoids was (335–337 Ma) from the shear zone associated with rather limited. This is consistent with rise of the ductile deformation along the Argentat fault (Roig et asthenosphere at depths not shallower than 50 km (where al. 1997). Motion and related deformation along major 345 Fig. 4 Simplified ‘accordion’- like model illustrating the proposed post-collision (340– 290 Ma) evolution of the Var- iscan belt. Vertical scale exag- gerated

crustal-scale faults were coeval with basin opening, Erosional unconformity between the two successions suggesting it formed by accommodation of the move- suggests two distinct periods of basement uplift and ment and, thus, substantiating a genetic link. erosion. This two-stage behaviour closely mimics episod- – The second sedimentary cycle took place during the icity of magmatic/volcanic events in the basement and Stephanian when, at least in the case of the pre- concurs with emerging models that depict the late-stage existing Bosmoreau basin, inherited, structures were evolution of the Variscan belt as being related to reactivated. Previous deposits were eroded and subse- lithospheric thinning and delamination. In addition, we quently blanketed by thick detritus dated at propose that the Stephanian magmatic flare-up reflects 296.5€3.5 Ma by the K–Ar method on clay particles. slab break-off, possibly related to clockwise rotation of Gondwana and associated strike-slip faulting. 346 Acknowledgements We acknowledge funding of this study by the Copard Y, Disnar JR, Becq-Giraudon JF, Boussafir M (2000) BRGM-Geofrance 3D program. N.C. also thanks the technical Evidence and effects of fluid circulation on organic matter in assistance of D. Tisserant and R. Wendling (CGS). Helpful intramontane coalfields (Massif Central, France). Int J Coal comments by U. Schaltegger and J. Von Raumer on a previous Geol 44:49–68 version of the manuscript are greatly appreciated. Constructive Davies JH, von Blanckenburg F (1995) Slab break-off: a model of reviews by P. Bankwitz and an anonymous reviewer helped lithosphere detachment and its test in the magmatism and improve this manuscript. deformation of collisional orogens. Earth Planet Sci Lett 129:85–102 Dong H, Hall CM, Halliday AN, Peacor DR, Merriman RJ, Roberts B (1997) 40Ar/39Ar illite dating of Late Caledonian (Acadian) References metamorphism and cooling of K-bentonites and slates from the Welsh Basin, UK. Earth Planet Sci Lett 150:337–351 Bard JP (1997) Dmembrement ant-msozoique de la chane Faure M (1995) Late orogenic carboniferous extensions in the varisque d’Europe Occidentale et d’Afrique du Nord: rle Variscan French Massif Central. Tectonics 14:132–153 essentiel des grands dcrochements transpressifs dextres ac- Faure M, Pons J (1991) Crustal thinning recorded by the shape of compagnant la rotation-translation horaire de l’Afrique durant the Namurian–Westphalian leucogranite in the Variscan belt of le stphanien. C R Acad Sci Paris 324:693–704 the northwest Massif Central, France. Geology 19:730–733 Becq-Giraudon JF (1985) Synthse des nouvelles donnes sur le Faure M, Moni P, Pin C, Maluski H, Leloix C (2002) Late Visean Stphanien de Bosmoreau-les-Mines (Creuse) et ses relations thermal event in the northern part of the French Massif Central: avec les tmoins de mÞme ge au sud de Bourganeuf. Gol new 40Ar/39Ar and Rb–Sr isotopic constraints on the Hercynian France 4:365–380 syn-orogenic extension. Int J Earth Sci 91:53–75 Becq-Giraudon JF, Van den Driessche J (1994) Dpts prigla- Kninger S, Lorenz V, Stollhofen H, Armstrong RA (2002) Origin, ciaires dans le Stphano-Autunien du Massif Central: Tmoin age and stratigraphic significance of distal fallout ash tuffs from de l’effondrement gravitaire d’un haut plateau hercynien. C R the Carboniferous–Permian continental Saar-Nahe Basin (SW Acad Sci Paris 318:675–682 Germany). Int J Earth Sci 91:341–356 Becq-Giraudon JF, Montenat C, Van den Driessche J (1996) Krner A, Jaeckel P, Reischmann T, Kroner U (1998) Further Hercynian high-altitude phenomena in the French Massif evidence for an early Carboniferous (~340 Ma) age of high- Central: tectonic implications. Palaeogeogr Palaeoclim Pa- grade metamorphism in the Saxonian granulite complex. Geol laeoecol 122:227–241 Rundsch 86:751–766 Bertaux J, Becq-Giraudon JF, Jacquemin H (1993) Les bassins Krner A, O’Brien PJ, Nemchin AA, Pidgeon RT (2000) Zircon anthracifres de la Rgion de Roanne (Loire, Massif central), ages for high pressure granulites from South Bohemia, Czech marqueurs d’une tectonique active durant le Visen suprieur. Republic, and their connection to Carboniferous high temper- Gol France 4:3–10 ature processes. Contrib Mineral Petrol 138:127–142 Berthier F, Duthou JL, Roques M (1979) Datation gochronologi- Kbler B (1984) Les indicateurs des transformations physiques et que Rb/Sr sur roches totales du granite de Guret (Massif chimiques dans la diagense, temprature et calorimtrie. In: Central). Age fini-dvonien de mise en place de l’un de ses Lagache M (ed) Thermomtrie et baromtrie gologiques. facis types. Bull Bur Rech Gol Mineral 1:59–72 Socit Franaise de Minralogie et Cristallographie, Paris, Bonhomme MG, Thuizat R, Pinault Y, Clauer N, Wendling R, pp 489–596 Winkler R (1975) Mthode de datation potassium–argon. Ledru P, Costa S, Echtler H (1994) Structures in the Massif Central. Appareillage et technique. Report Univ Strasbourg In: Keppie JD (ed) Pre-Mesozoic geology in France and related Bouroz A (1966) Frquence des manifestations volcaniques au areas. Springer, Berlin Heidelberg New York, pp 305–323 Carbonifre suprieur en France. C R Acad Sci Paris 263:1025– Malavieille J (1993) Late orogenic extension in mountain belts: 1028 insights from the Basin and Range and the late Paleozoic Breitkreuz C, Kennedy A (1999) Magmatic flare-up at the Variscan belt. Tectonics 12:1115–1130 Carboniferous/Permian boundary in the NE German Basin Maluski H, Costa S, Echtler H (1991) Late Variscan tectonic revealed by SHRIMP zircon ages. Tectonophysics 302:307– evolution by thinning of earlier thickened crust. An 40Ar–39Ar 326 study of the Montagne Noire, southern Massif Central, France. Brousmiche-Delcambre C, Mercier D, Coquel R (1995) Implica- Lithos 26:287–304 tions stratigraphiques de la rvision de la flore Carbonifre au Matte P (1986) Tectonics and plate tectonics model for the Sud de Brianon. C R Acad Sci Paris 320:335–340 Variscan belt of Europe. Tectonophysics 126:329–374 Bruguier O, Lancelot JR, Malavieille J (1997) U–Pb dating on Matte P (2001) The Variscan collage and orogeny (480–290 Ma) single detrital zircon grains from the Triassic Songpan-Ganze and the tectonic definition of the Armorica microplate: a flysch (Central China) - provenance and tectonic correlations. review. Terra Nova 13:122–128 Earth Planet Sci Lett 152:217–231 Nelson KD (1992) Are crustal thickness variations in old mountain Bruguier O, Becq-Giraudon JF, Bosch D, Lancelot JR (1998) Late belts like the Appalachians a consequence of lithospheric Visean hidden basins in the internal zones of the Variscan Belt: delamination? Geology 20:498–502 U–Pb zircon evidence from the French Massif Central. Geology Odin GS (1994) Geological time scale (1994). CR Acad Sci Paris 26:627–630 312:59–71 Clauer N, Chaudhuri S (1995) Clays in crustal environments— Pin C, Duthou JL (1990) Sources of Hercynian granitoids from the isotope dating and tracing. Springer, Berlin Heidelberg New French Massif Central: inferences from Nd isotopes and York consequences for crustal evolution. Chem Geol 83:281–296 Clauer N, Chaudhuri S (1998) Isotopic dating of very low-grade Ratschbacher L, Frisch W, Neubauer F, Schmid SM, Neugebauer J metasedimentary and metavolcanic rocks: techniques and (1989) Extension in compressional orogenic belts: the eastern methods. In: Frey M, Robinson D (eds) Low-grade metamor- Alps. Geology 17:404–407 phism. Blackwell Science, Oxford, pp 202–226 Roig JY, Calgano P, Bouchot V, Maluski H, Faure M (1997) Clauer N, Rais N, Schaltegger U, Piqu A (1995) K–Ar systematics Modelisation 3D du palochamp hydrothermal As + Au (330– of clay-to-mica minerals in a multi-stage low-grade metamor- 300 Ma) le long de la faille d’Argentat (Massif Central phic evolution. Chem Geol 124:305–316 franais). Chron Rech Mineral 528, 63–69 Copard Y (1998) Contributions la mise en vidence de Schaltegger U (1997) Magma pulses in the Central Variscan Belt: palocirculations de fluides dans les basins houillers du Massif episodic melt generation and emplacement during lithospheric Central Franais. B.R.G.M. Geofrance 3D Public Report n94/ thinning. Terra Nova 9:242–245 98, Orlans 347 Schaltegger U, Corfu F (1992) The age and source of late Scott AC, Galtier J, Clayton G (1984) Distribution of anatomically- Hercynian magmatism in the Central Alps: evidence from preserved floras in the Lower Carboniferous in Western precise U–Pb ages and initial Hf isotopes. Contrib Mineral Europe. Trans R Soc Edinb 75:311–340 Petrol 111:329–344 Steiger RH, Jger E (1977) Subcommission on geochronology: Schaltegger U, Corfu F (1995) Late Variscan “Basin and Range” convention on the use of decay constants in geo- and magmatism and tectonics in the Central Alps: evidence from cosmochronology. Earth Planet Sci Lett 36:359–362 U–Pb geochronology. Geodin Acta 8, 82–98 Van der Voo R (1982) Pre-Mesozoic paleomagnetism and plate Schaltegger U, Zwingmann H, Clauer N, Larqu P, Stille P (1995) tectonics. Ann Rev Earth Planet Sci 10:191–220 K–Ar dating of a Mesozoic hydrothermal activity in Carbonif- Von Raumer JF (1998) The Palaeozoic evolution in the Alps: from erous to Triassic clay minerals of northern Switzerland. Gondwana to Pangea. Int J Earth Sci 87:407–435 Schweiz Mineral Petrogr Mitt 75:163–176 Wenzel T, Mertz DF, Oberhnsli R, Becker T, Renne PR (1997) Schaltegger U, Schneider JL, Maurin JC, Corfu F (1996) Precise Age, geodynamic setting, and mantle enrichment processes of a U–Pb chronometry of 345–340 Ma old magmatism related to K-rich intrusion from the Meissen massif (northern Bohemian syn-convergence extension in the Southern Vosges (central massif) and implications for related occurrences from the mid- Variscan belt). Earth Planet Sci Lett 144:403–419 European Hercynian. Geol Rundsch 86:556–570 Schneider JL, Maas R, Gall JC, Duringer P (1989) L’vnement Zoback MD, Stephenson RA, Cloething S, Larsen BT, Van Horn B, intra-visen dans la zone Moldanubienne de la chane Varisque Robinson A, Horvath F, Puigdefabregas C, Ben Avraham Z d’Europe: les donnes des formations volcano-sdimentaires- (1993) Stresses in the lithosphere and sedimentary basin dvono-dinantiennes du Massif Central Franais, des Vosges du formation. Tectonophysics 226:1–13 Sud (France) et de la ForÞt Noire (RFA). Geol Rundsch 78:555–570 U–Pb Silurian age for a gabbro of the Platinum-bearing Belt of the Middle Urals (Russia): evidence for beginning of closure of the Uralian Ocean

DELPHINE BOSCH1, OLIVIER BRUGUIER2, ALEXANDER A. EFIMOV3 & ARTHUR A. KRASNOBAYEV3 1Laboratoire de Tectonophysique, Universite´ de Montpellier II, UMR 5568-CNRS/UMII, Place E. Bataillon, 34095 Montpellier Cedex 05, France (e-mail: [email protected]) 2Service ICP-MS, ISTEEM, Universite´ de Montpellier II, Place E. Bataillon, 34095 Montpellier Cedex 05, France 3Institute of Geology and Geochemistry, Pochtovyi Pereuloic 7, Ekaterinburg 620151, Russia

Abstract: The Platinum-bearing Belt of the Urals consists of a series of zoned ultramafic bodies obducted onto the passive continental margin of the East European Craton during the Palaeozoic. Conventional U–Pb and secondary ionization mass spectrometry U–Th–Pb analyses of single zircon grains from a pegmatitic gabbro of the Kumba massif provide Mid-Silurian ages of 425 + 3 Ma and 419 + 10 Ma, respectively, interpreted as dating crystallization of the gabbroic magma. This contrasts with the c. 360 Ma age of the neighbouring Kytlym massif and indicates that the Uralian Platinum-bearing Belt is best interpreted as remnants of an island-arc oceanic lithosphere that started forming in Mid-Silurian times and had a protracted lifetime of about 70 Ma. The Uralian Platinum- bearing Belt is thus coeval with the Sakmara arc of the Southern Urals, and is of similar age to other ultramafic bodies such as the Kempersai and Mindyak massifs. Ophiolitic fragments, now preserved along the Main Uralian Fault, may thus represent the assemblage of contemporaneous oceanic and arc terranes brought together during the final stages of the evolution of the Uralide orogen. The age of the Kumba massif, in addition, indicates that the Uralian Ocean underwent contractional events as early as the Mid-Silurian, which suggests a possible link with the pivotal rotation of Baltica.

The convergence of the East European Craton (EEC) and the gabbro of the Kumba massif, which crops out in the Uralian Siberian–Kazakhstan terrane assemblage that ultimately led to Platinum-bearing Belt. The significance of these results is dis- the formation of the Urals mountain belt at the end of the cussed in relation to the available isotopic data from other Palaeozoic was preceded by closing of the Uralian Ocean that sep- massifs in the Platinum-bearing Belt and from other occurrences arated the various continental masses. Remnants of oceanic litho- along the Main Uralian Fault in an attempt to understand the sphere that became trapped along the suture between the colliding evolution of the Uralian Ocean. continents provide important time markers for the pre-collisional history of the orogen. Ophiolite complexes are numerous in the Urals (Savelieva & Nesbitt 1996) and crop out in the hanging Geological setting wall of the Main Uralian Fault (MUF), the main suture zone that runs the entire length of the mountain belt (Matte 1995). These The Uralian Platinum-bearing Belt (UPB) forms a giant structure massifs appear as lens-shaped bodies mainly of lherzolitic exposed in the Central and Northern Urals (see Fig. 1a). It is and harzburgitic composition elongated parallel to the MUF located east of the MUF where the suture sector widens and is charac- (Savelieva et al. 1997). The continuity of the belt and similar terized by the abundance of zoned mafic–ultramafic complexes. structural position of the massifs (i.e. allochthonous position over- These massifs are similar to Alaskan-type complexes (Irvine 1967) lying either the Precambrian basement or Palaeozoic sediments of and show the typical association dunite–clinopyroxenite–gabbro the EEC margin) suggest a common origin. However, whereas (DCG units of Fershtater et al. 1997). The UPB consists of a series some massifs show a nearly complete petrological association of 13 rounded to elongated massifs forming a discontinuous but typical of ophiolite sequences, others lack the upper gabbro, linear elongated zone, which crops out for a distance of about sheeted dyke and lava complexes, and crop out as zoned plutons 900 km (Fig. 1a). The mafic–ultramafic bodies are associated with of mafic–ultramafic rocks. Thus, it is unclear whether the rocks of the Tagil Zone, which consists of Early Ordovician to c. 2000 km long oceanic terrane exposed in the Urals represents Early Silurian volcanic and volcano-sedimentary units, tectonically fragments of a single extensive dismembered ophiolite (where, associated with shelf-facies and continental-slope deposits. in some cases, the upper lava complexes have been removed by They commonly developed a thermal aureole in the volcano- tectonic processes) or whether some massifs are remnants of sedimentary sequences (Savelieva & Nesbitt 1996). The massifs island-arc terranes derived from the Uralian Ocean and progress- are mainly composed of gabbros (olivine gabbros, gabbro–norites, ively telescoped with the EEC margin. Despite their importance clinopyroxene–hornblende gabbros), and also of clinopyroxenites, in understanding the evolution and destruction of the Uralian wehrlites, dunites, pegmatitic gabbros and plagiogranites (Fig. 1b) Ocean and thus the earliest evolutionary stages of the Uralide with a progressive magma differentiation from dunite through wehr- Orogen, isotopic age information is sparse on these complexes. lite–clinopyroxenite to gabbro. Sharp compositional changes of rock Most of our knowledge is from a few isotopic datings on Alaskan- associations suggest that intrusions were pulsating (Savelieva et al. type zoned mafic–ultramafic complexes located in the Uralian 1999). In the study region, field relationships indicate a complex Platinum-bearing Belt (UPB; Ivanov & Kaleganov 1993; allochthonous association with at least two main nappe systems. Bea et al. 2001; Ronkin et al. 2003), and from typical ophiolitic Structurally lower units consist of Late Ordovician to Early Silurian massifs displaying well-developed plutonic and volcanic units, deposits of palaeo-oceanic affinities (greenstone basalts, metacherts, such as the Kempersai (Southern Urals) and Voykar massif psammitic and pelitic tuffs, along with tectonized relics of possibly (Polar Urals) (Edwards & Wasserburg 1985; Sharma et al. 1995; Ordovician ophiolites), whereas the structurally upper system is Melcher et al. 1999). In this study, we present conventional composed of a stack of tectonic slices. These are composed of U–Pb single-zircon analyses and U–Th–Pb in situ secondary ion- volcano-sedimentary and volcaniclastic terrigenous units of Late ization mass spectrometry (SIMS) analyses for a pegmatitic Ordovician to Early Silurian age in contact with and apparently

From:GEE,D.G.&STEPHENSON, R. A. (eds) 2006. European Lithosphere Dynamics. Geological Society, London, Memoirs, 32, 443–448. 0435-4052/06/$15.00 # The Geological Society of London 2006. 443 444 D. BOSCH ET AL.

Fig. 1. Simplified geological map of the Uralide orogen (a) and a detailed map of part of the Platinum-bearing Belt close to the sample locality (b). Square with cross indicates sample location. The simplified map is modified after Friberg et al. (2000); (b) is from Efimov et al. (1993).

intruded by the mafic–ultramafic bodies of the UPB (Savelieva & isotopic composition of radiogenic Pb was determined by subtract- Nesbitt 1996). ing first the blank and then the remainder, assuming a common Pb composition at the time of initial crystallization as determined from the model of Stacey & Kramers (1975). Isotopic measure- Sample description ments were carried using on a VG Sector mass spectrometer at the University of Montpellier II, with a Daly detector. SIMS The Kumba massif is located in the centre of the UPB, and is one analyses were carried out on the same zircon population as was of the smallest of the 13 massifs of the belt. The general shape of analysed by TIMS. The grains, together with chips of standard the Kumba massif is a 308E dipping oval form. This massif zircon, were mounted in epoxy resin and polished to approxi- consists of the typical association of dunite and clinopyroxenite mately half their thickness to expose internal structure. U– enclosed by gabbros and gabbro–norites (Fig. 1b). In the northern Th–Pb analyses were performed with an elliptical spot size of part, it mainly consists of olivine–anorthite gabbro with rare pyro- about 25 mm Â 35 mm on the CAMECA IMS 1270 ion micro- xenites, the central part contains gabbro–norites, and the western probe at the CRPG Nancy (France), following the technique part comprises dunites, pyroxenites and some metamorphic outlined by Deloule et al. (2001). Isotopic ratios were measured products. with a primary O2 beam of 10 nA at a mass resolution of c. The study sample is an amphibole pegmatitic gabbro enclosed in 5000, at which no significant interferences on the Pb, U and Th gabbro–norites. The mineralogical assemblage consists of plagio- isotopes were detected. Oxygen flooding was used to enhance clase, clinopyroxene, orthopyroxene, hornblende, spinel, ilmenite sensitivity. Under these operating conditions, the sensitivity and magnetite. The sample is fresh without evidence of alteration. for Pb isotopes was c. 20 cps per ppm per nA of primary beam. Zircon appears as a scarce accessory mineral in this rock and more Pb/U ratios were normalized using quadratic working curves, to than 50 kg of rock fragments have been crushed to perform this study. values measured on the G91500 standard zircon (Wiedenbeck et al. 1995). Common Pb was corrected using measured 204Pb and a composition taken from the model of Stacey & Kramers Analytical procedures (1975). Corrected isotopic ratios, regression lines and intercepts were calculated according to Ludwig (1987). Decay constants The minerals were prepared for U–Pb analyses according to stan- are those recommended by the IUGS Subcommission on dard techniques (e.g. Bosch et al. 1996) where only the best Geochronology (Steiger & Jager 1977). quality zircons (non-magnetic and free of visible inclusions and fractures) were selected for analysis. For thermal ionization mass spectromery (TIMS) analyses, single grains were weighed Results on a Cahn electrobalance and then processed according to Bosch & Bruguier (1998). Total Pb and U blanks over the period of the Zircons occur as a uniform population of large (.100 mm) colour- analyses were ,15 pg and ,5pg(+50%), respectively. The less translucent grains. They show euhedral shapes with sharp SILURIAN AGE FOR PT-BEARING BELT, MIDDLE URALS 445 angle terminations, often interrupted by irregular boundaries (see U concentrations ranging from c. 200 to 2000 ppm (see Tables 1 left upper termination of grain B in Fig. 2), which we interpret as and 2). These U contents are surprising for zircons from gabbroic reflecting late magmatic crystallization in a small melt volume rocks, which usually have U concentrations in the range 50– adjacent to already crystallized minerals. SEM images of polished 400 ppm (Scha¨rer et al. 1986; Pedersen & Dunning 1997), but surfaces (Fig. 2a and b) show that the grains exhibit almost struc- concentrations up to 6000 ppm have been reported (Dunning & tureless internal domains, although irregular, broad-banded zoning Pedersen 1988). This is likely to reflect a large degree of differen- can be recognized, which is often observed in magmatic zircons tiation before zircon crystallization, consistent with the pegmatitic from mafic lithologies (e.g. Rubatto 2002). All grains have high nature of the rock studied and with crystallization of the zircon in the latest stages of the magmatic evolution. Four of the six grains analysed by TIMS were air abraded (Krogh 1982) to minimize Pb losses. On a concordia diagram (Fig. 3a), the abraded grains cluster close to concordia and, with the two unabraded grains, yield a discordia line with upper and lower intercept ages of 424.9 + 2.7 Ma and 226 + 8Ma (MSWD ¼ 0.72). The unabraded grains (Zr5 and Zr6) are discordant, which indicates Pb loss. The near-zero lower intercept indicates recent, surface-correlated, Pb loss rather than an ancient, metamorphic disturbance of the U–Pb system of these zircons. The high degree of concordance of the abraded zircons and the magmatic morphology of the grains indicate that the 424.9 + 2.7 Ma intercept age corresponds to the age of zircon growth during crystallization of the pegmatitic magma. Sixteen SIMS analyses were performed on 14 grains, which dis- played Th/U ratios typical of magmatic zircons (Th/U . 0.1; after Williams & Claesson 1987). Overall, the SIMS analyses yielded a pattern consistent with the TIMS results (Fig. 3b). Although all analysed grains are concordant within error margins, spot analyses yield a tendency to scatter along the discordia between 390 and 430 Ma, suggesting that the grains have undergone various degrees of Pb loss. The best age information is thus given by a discordia line where the upper intercept with the concordia is 419 + 10 Ma. This age is slightly younger, but within error, of the more precise TIMS age, which is interpreted as our best estimation of the age of the Kumba gabbro.

Discussion and conclusion

The Wenlock U–Pb zircon age of 424.9 + 2.7 Ma obtained for the Kumba gabbro provides a new time constraint for the formation of the zoned mafic–ultramafic massifs that crop out in the UPB, which has implications for the early evolution of the Uralide orogen. Available geochronological data for mafic and ultramafic bodies of the UPB are sparse and limited to K–Ar ages of minerals from gabbros of the Kachkanar massif (Ivanov & Kaleganov 1993) and to recently published Pb–Pb evaporation and SIMS U–Pb zircon ages from a dunite of the Kytlym massif (Bea et al. 2001). The latter study resulted in the discovery of inherited zircons, but no Fig. 2. Back-scattered SEM images of zircons from the Kumba gabbro firm age was proposed for emplacement of the massif. Excluding (Platinum-bearing Belt, Central Urals). very old grains (Proterozoic and Archaean in age), the data yielded

Table 1. Conventional single-zircon U–Pb isotopic data (Kumba pegmatitic gabbro, Platinum-bearing Belt, Central Urals)

Sample Weight Concentrations Corrected ratios r Apparent ages (Ma) number (mg) U Pb 206Pb/204Pb 208Pb/206Pb 207Pb/206Pb 206Pb/238U 207Pb/235U 206Pb/238U 207Pb/206Pb (ppm) (ppm) measured (+2s) (+2s) (+2s) (+2s) (+2s)

1, ab 0.001 2010 326 2384 0.277 0.0554 + 2 0.0700 + 2 0.534 + 2 0.76 436 + 3 427 + 07 2, ab 0.001 543 84 666 0.215 0.0551 + 6 0.0687 + 3 0.522 + 6 0.48 428 + 4 415 + 23 3, ab 0.002 475 71 1399 0.179 0.0551 + 7 0.0693 + 2 0.527 + 7 0.57 432 + 3 417 + 27 4, ab 0.002 302 43 791 0.130 0.0552 + 3 0.0675 + 2 0.514 + 3 0.60 421 + 2 421 + 11 5 0.005 507 61 7511 0.203 0.0555 + 1 0.0548 + 1 0.419 + 1 0.94 344 + 1 431 + 02 6 0.007 596 60 12240 0.217 0.0556 + 1 0.0458 + 1 0.351 + 1 0.96 288 + 1 436 + 02

Grains were selected from non-magnetic separates at full magnetic ﬁeld in a Frantz magnetic separator. ab, grains that have been air abraded following Krogh (1982). Isotopic ratios have been corrected for fractionation, blank (204:206:207:208 ¼ 1:18.31:15.59:37.88), and initial common Pb after the growth curve of Stacey & Kramers (1975). Errors are 2s and refer to the last digits. 446 D. BOSCH ET AL.

Table 2. SIMS in situ U–Th–Pb isotopic data (Kumba pegmatitic gabbro, Platinum-bearing Belt, Central Urals)

Spot U Th PbÃ Th/U 204Pb/ 208Pb/ 207Pb/ +(1s) 207Pb/ +(1s) 206Pb/ +(1s) r Apparent age (Ma) number (ppm) (ppm) (ppm) 206Pb 206Pb 206Pb 235U 238U 207Pb/ 206Pb/ +(1s) 206Pb 238U

1–1 486 189 29 0.39 0.0002 0.122 0.056 0.009 0.530 0.012 0.0692 0.0015 0.92 433 431 9 1–2 499 194 28 0.39 0.0002 0.124 0.056 0.009 0.508 0.009 0.0659 0.0011 0.87 448 411 7 1–3 525 213 30 0.41 0.0002 0.129 0.055 0.008 0.514 0.011 0.0675 0.0014 0.93 423 421 8 2 507 294 29 0.58 0.0002 0.185 0.055 0.008 0.502 0.011 0.0663 0.0013 0.93 407 414 8 3 469 273 25 0.58 0.0001 0.182 0.056 0.008 0.483 0.009 0.0626 0.0011 0.92 449 392 7 4 626 349 34 0.56 0.0001 0.176 0.055 0.008 0.485 0.010 0.0637 0.0011 0.91 424 398 7 5 650 239 36 0.37 0.0001 0.115 0.055 0.011 0.491 0.011 0.0642 0.0012 0.87 428 401 7 6 897 469 51 0.52 0.0001 0.167 0.055 0.006 0.499 0.010 0.0657 0.0013 0.96 417 410 8 7 1518 1374 86 0.91 0.0001 0.292 0.055 0.003 0.502 0.011 0.0660 0.0014 0.99 418 412 8 8 527 294 30 0.56 0.0005 0.179 0.056 0.015 0.511 0.014 0.0668 0.0015 0.83 434 417 9 9 259 125 14 0.48 0.0003 0.155 0.054 0.025 0.477 0.017 0.0636 0.0016 0.71 387 397 10 10 378 156 21 0.41 0.0003 0.131 0.055 0.013 0.479 0.013 0.0635 0.0015 0.88 400 397 9 11 718 412 40 0.57 0.0003 0.180 0.054 0.012 0.482 0.011 0.0642 0.0012 0.85 391 401 7 12 758 272 42 0.36 0.0001 0.113 0.056 0.006 0.497 0.010 0.0649 0.0012 0.94 434 406 7 13 264 134 15 0.51 0.0003 0.161 0.054 0.014 0.483 0.013 0.0644 0.0015 0.85 384 403 9 14 551 300 31 0.54 0.0002 0.172 0.055 0.009 0.493 0.011 0.0648 0.0013 0.91 418 405 8

ÃRadiogenic. Errors are given at 1s.

a continuum of ages from 435 + 18 Ma to 315 + 13 Ma (1s). On the basis of age clustering at 350–370 Ma, Bea et al. (2001) suggested emplacement and crystallization of the massif at 360 Ma, broadly coeval with the high-pressure metamorphism in the Urals, dated at 370–390 Ma (Matte et al. 1993; Glodny et al. 2002). The inferred age of the Kytlym massif, located only c. 50 km south of the Kumba massif, differs by 60–70 Ma, although some values at c. 400–430 Ma reported by Bea et al. (2001) are in good agreement with our proposed age. The observed discrepancies may be explained in two ways, which have different geodynamical implications. (1) The Kumba and Kytlym massifs are contemporaneous and the continuum of ages observed for the Kytlym dunite zircons by Bea et al. (2001) reflects varying degrees of recrystallization and associated lead loss of a zircon population that crystallized at c. 425 Ma. This is consistent with some of the SEM images provided by Bea et al., which show oscillatory zoned areas interrupted by unzoned patchy domains. Because the zircon lower intercept for the Kumba massif is nearly zero, recrystallization and Pb losses in the Kytlym massif should be related to a local phenomenon. Late gabbroic or granitic intrusions associated with locally developed high-temperature plastic deformation have been observed in the Kytlym massif (Savelieva et al. 1999) and may have driven recrystallization and associated lead loss in the 425 Ma zircons. However, this would result in an episodic disturbance of the zircon U–Pb system instead of the continuum of ages observed by Bea et al. (2001). This continuum of ages, on the contrary, requires multiple resetting of the U–Pb zircon ages in the Kytlym massif during a protracted time interval. The occurrence of metasomatized ultramafic rocks around the massifs (Savelieva et al. 1999) and particularly around the Kytlym massif, where the so-called kytlymite developed at the expense of basement country rocks, indicates fluid flows associated with metasomatic episodes. Leaching of radiogenic lead and/or low-temperature recrystallization or dissolution–reprecipitation processes through fluid flow under experimental (Geisler et al. 2002) or natural (Ho¨gdahl et al. 2001) conditions can, episo- dically, strongly disturb the U–Pb systematics of zircon. However, this requires that the zircons in the Kytlym dunite have responded nonuniformly to a common history, a possibility that could be Fig. 3. U–Pb concordia diagram for zircon grain analyses of the Kumba gabbro. explained by proposing that some zircons may have been armoured (a) TIMS single-grain analyses. Error ellipses are 2s.(b) SIMS U–Th–Pb against Pb loss by incorporation in other grains, such as large analyses. Error ellipses are 1s. olivine crystals. SILURIAN AGE FOR PT-BEARING BELT, MIDDLE URALS 447

(2) The age difference between the two massifs is evidence that particularly from the Late Silurian–Early Devonian westward the UPB contains an association of mafic–ultramafic rocks that obduction events, during which the MUF already acted as a formed over a long time interval ranging from the Mid-Silurian major thrust (Matte 1995; Echtler et al. 1997). to Mid-Devonian (425–360 Ma). The geochemical signatures of Available geochronological information from other ophiolitic mafic and ultramafic rocks of the Kytlym massif show island fragments that crop out along the MUF allows a direct comparison arc-like composition, interpreted as evidence for a with massifs from the UPB. A Sm–Nd age for the Kempersai suprasubduction-zone setting (Bea et al. 2001). The geology of massif in the SW Urals is 397 + 20 Ma (Edwards & Wasserburg the surrounding Late Ordovician to Early Silurian volcano- 1985), consistent with the 394–427 Ma age range yielded by sedimentary and volcaniclastic units of the Tagil Zone is consist- Sm–Nd mineral isochrons from various lithologies in the same ent with this view. Combining the ages of the two massifs would massif and with a 420 + 10 Ma Pb/Pb apparent age on one therefore reflect a long-lived island-arc history of about 60– zircon fraction (Melcher et al. 1999). The Voykar massif in the 70 Ma. Alternatively, the two contrasting ages could also indicate northeastern Urals yielded a less precise younger, but identical a more complex history in which the c. 425 Ma Kumba massif rep- within error, Sm–Nd age of 387 + 34 Ma (Sharma et al. 1995). resents remnants of a Mid-Silurian oceanic lithosphere on top of Lastly, although its age is not precisely known, the Mindyak ophio- which a Late Devonian–early Tournaisian arc, represented by lite in the Southern Urals formed before 414 + 4 Ma, the age of the Kytlym massif, subsequently developed. Although further peak metamorphism affecting this massif (Scarrow et al. 1999). studies are needed, all the massifs in the UPB have similar charac- All these dates are identical to or within the limits of error of the teristics and the latter hypothesis is thus considered unlikely. ages available for the UPB and, given the large errors on most Existing models for the Palaeozoic tectonic development of the dates, no significant age discrepancies can be detected for the Uralides involve break-up of the EEC margin and initiation of various oceanic fragments. In the light of the present dataset, this the opening of the Uralian Ocean during Ordovician times, at would suggest that the fragments of oceanic lithosphere preserved c. 480–500 Ma (Zonenshain et al. 1984; Fershtater et al. 1997). in the hanging wall of the MUF may represent a single ophiolite, The subsequent evolution of this oceanic domain is characterized tectonically dismembered during collision of the continental by the building of several volcanic island arcs within or peripheral masses, rather than an assemblage of various oceanic and arc ter- to the Uralian Ocean. Several independent lines of evidence ranes brought together subsequently during the final stages of the concur with a Silurian age for the UPB. The schematic distribution evolution of the Uralide orogen. As the closure of the Uralian of the Uralian magmatism in space and time given by Fershtater Ocean began as early as the Mid-Silurian, it may thus show simi- et al. (1997), in particular for rocks of the dunite–clinopyroxe- larities to the closure of the Iapetus Ocean in the sense that the ear- nite–gabbro series, is in very good agreement with the Wenlock liest stages of contraction of the Uralian Ocean may be the age of the Kumba massif. Those researchers indicated a Late counterpart of the rotation of Baltica that preceded the main Scan- Ordovician–Early Silurian age for rocks of the suture sector, dian continent collision with Laurentia. east of the MUF, including the UPB. Lastly, K–Ar ages of amphiboles and phlogopites from clinopyroxenites and gabbros of the We thank H. Austrheim, F. Bea and J. Glodny for detailed constructive reviews Kachkanar massif, south of the Kytlym massif, range from 420 that improved the manuscript. We thank the ISTEEM institution for financial + support during the 1996 field trip in the Urals. F. Boudier, P. Matte, to 430 Ma, and a mean amphibole age of 423 3 Ma was A. Nicolas, A. Pertsev and G. Savelieva are thanked for scientific interaction proposed for the gabbroic rocks (Ivanov & Kaleganov 1993). and fruitful discussions on the Urals geology. Further north, the Chistop massif yielded a Sm–Nd isochron age of 419 + 12 Ma (Ronkin et al. 2003). All these ages are similar References to the U–Pb zircon age of the Kumba massif. Given the similarities between all the massifs of the UPB and BEA, F., FERSHTATER, G. B., MONTERO, P., ET AL. 2001. Recycling of the suprasubduction-zone setting of the Kytlym massif, we cur- continental crust into the mantle as revealed by Kytlym dunite rently favour a model in which the UPB represents remnants of zircons, Urals Mts, Russia. Terra Nova, 13, 407–412. an island-arc oceanic lithosphere that existed in Mid-Silurian BOSCH, D., BRUGUIER,O.&PIDGEON, R. T. 1996. The evolution of an times, or earlier, by subduction of the Uralian oceanic domain. Archean metamorphic belt: a conventional and SHRIMP U–Pb In Early Silurian times, anticlockwise rotation of Baltica study on accessory minerals from the Jimperding Belt, Yilgarn (Torsvik et al. 1996; Torsvik & Rehnstrom 2001; Roberts 2003) craton, West Australia. Journal of Geology, 104, 695–711. may have been accommodated by subduction of the Uralian BOSCH,D.&BRUGUIER, O. 1998. An Early Miocene age for a high temp- oceanic domain and development of an island arc. The erature event in gneisses from Zabargad Island (Red Sea, Egypt): Mid-Silurian UPB could have formed in such a geodynamic mantle diapirism? Terra Nova, 10, 274–279. setting, contemporaneously with the Sakmara arc of the Southern DELOULE, E., CHAUSSIDON, M., GLASS,B.P.&KOEBERL, C. 2001. Urals (Zonenshain et al. 1984). The 425 Ma U–Pb zircon age of U–Pb isotopic study of relict zircon inclusions recovered from the Kumba massif would thus represent a minimum age for Muong Nong-type tektites. Geochimica et Cosmochimica Acta, 65, early contraction processes and for onset of the closure of the 1833–1838. Uralian oceanic domain. According to model (1), the UPB DUNNING,G.R.&PEDERSEN, R. B. 1988. U/Pb ages of ophiolites and island arc may have been short lived and would thus represent arc-related plutons of the Norwegian Caledonides: implications for closure of a discrete marginal basin of the Uralian Ocean. Conver- the development of Iapetus. Contributions to Mineralogy and sely, a continuous arc history of the UPB implies the occurrence Petrology, 98, 13–23. of a major arc whose lifetime (from 425 to 360 Ma), broadly ECHTLER,H.P.,IVANOV,K.S.,RONKIN,Y.L.,KARSTEN,L.A.,HETZEL,R. corresponds to the time interval between the opening of the &NOSKOV, A. G. 1997. The Paleozoic tectono-metamorphic evolution of gneiss complexes in the middle Urals: a reappraisal. Tecto- Uralian Ocean (480–500 Ma) and the beginning of its destruction nophysics, 276, 229–251. (425 Ma). The currently available dataset of ages does not exclude EDWARDS,R.L.&WASSERBURG, G. J. 1985. The age and emplacement either of these models. Nevertheless, the occurrence of gabbroic + of obducted oceanic crust in the Urals from Sm–Nd and Rb–Sr rocks (known as tilaites) dated at 340 22 Ma (Pushkarev et al. systematics. Earth and Planetary Science Letters, 72, 389–404. 87 86 2003) within the Kytlym massif, with a low Sr/ Sr initial EFIMOV, A. A., EFIMOVA,A.A.&MAEGOV, V. I. 1993. Tectonics of the ratio (,0.705), suggests that rocks of the UPB were still within Platinum-bearing Belt of the Urals: correlation of complexes and an oceanic environment at that time. We thus consider model mechanism of substantial-formation. Geotectonika, 3, 34–46. (2) more likely, and suggest that the UPB represents a long-lived FERSHTATER, G. B., MONTERO, P., BORODINA, N. S., PUSHKAREV, E. V., island arc (and its associated marginal basin) that has been SMIRNOV,V.N.&BEA, F. 1997. Uralian magmatism: an overview. preserved from the sequential subduction–obduction cycles and Tectonophysics, 276, 87–102. 448 D. BOSCH ET AL.

FRIBERG, M., LARIONOV, A., PETROV,G.A.&GEE, D. G. 2000. Paleo- Isotope Geology, Salvador–Bahia, Brazil. Journal of Conference zoic amphibolite–granulite facies magmatic complexes in the Abstracts, 762–765. hinterland of the Uralide Orogen. International Journal of Earth RUBATTO, D. 2002. Zircon trace element geochemistry: distribution Sciences, 89, 21–39. coefficients and the link between U–Pb ages and metamorphism. GEISLER, T., PIDGEON, R. T., VAN BRONSWIJK,W.&KURTZ, R. 2002. Chemical Geology, 184, 123–138. Transport of uranium, thorium and lead in metamict zircon under SAVELIEVA,G.N.&NESBITT, R. W. 1996. A synthesis of the strati- low-temperature hydrothermal conditions. Chemical Geology, 191, graphic and tectonic setting of the Uralian ophiolites. Journal of 141–154. the Geological Society, London, 153, 525–537. GLODNY, J., BINGEN, B., AUSTRHEIM, H., MOLINA,J.F.&RUSIN,A. SAVELIEVA, G. N., SHARASKIN, A. Y., SAVELIEV, A. A., SPADEA,P.& 2002. Precise eclogitization ages deduced from Rb/Sr mineral sys- GAGGERO, L. 1997. Ophiolites of the southern Uralides adjacent tematics: the Maksyutov complex, Southern Urals, Russia. to the East European continental margin. Tectonophysics, 276, Geochimica et Cosmochimica Acta, 66, 1221–1235. 117–137. HO¨ GDAHL, K., GROMET,L.P.&BROMAN, C. 2001. Low P–T Caledonian SAVELIEVA, G. N., PERTSEV, A. N., ASTRAKHANTSEV, O. V., DENISOVA, resetting of U-rich Paleoproterozoic zircons, central Sweden. E. A., BOUDIER, F., BOSCH,D.&PUCHKOVA, A. V. 1999. Structure American Mineralogist, 86, 534–546. and dynamic history of the Kytlym Pluton, North Urals. Geotec- IRVINE, T. N. 1967. The Duke Island ultramafic complex, southeastern tonics, 33, 119–141. Alaska. In:WYLLIE, P. J. (ed.) Ultramafic and Related Rocks. SCARROW, J. H., SAVELIEVA, G. N., GLODNY, J., MONTERO, P., PERTSEV, Wiley, Chichester, 97–121. A. N., CORTESOGNO,L.&GAGGERO, L. 1999. The Mindyak IVANOV,O.K.&KALEGANOV, B. A. 1993. New data on the age of Palaeozoic lherzolite ophiolite, Southern Urals: geochemistry and concentrically zoned dunite–pyroxenite intrusions in the Ural geochronology. Ofioliti, 24, 239–246. Platiniferous Belt. Doklady Earth Sciences, 328, 720–724. SCHA¨ RER, U., KROGH,T.E.&GOWER, C. F. 1986. Age and evolution of KROGH, T. E. 1982. Improved accuracy of U–Pb zircon ages by the cre- the Grenville Province in eastern Labrador from U–Pb systematics in ation of more concordant systems using an air abrasion technique. accessory minerals. Contributions to Mineralogy and Petrology, 94, Geochimica et Cosmochimica Acta, 46, 637–649. 438–451. LUDWIG, K. R. 1987. Isoplot200, a plotting and regression program for SHARMA, M., WASSERBURG, G. J., PAPANASTASSIOU, D. A., QUICK, J. E., 143 144 isotope geochemists, for use with HP series 200 computers.US SHARKOV,E.V.&LAZ’KO, E. E. 1995. High Nd/ Nd in Geological Survey Open-File Report, 85–513. extremely depleted mantle rocks. Earth and Planetary Science MATTE, P. 1995. Southern Uralide and Variscides: compared anatomy Letters, 135, 101–114. and evolution. Geologie en Mijnbouw, 74, 151–166. STACEY,J.S.&KRAMERS, J. D. 1975. Approximation of terrestrial lead MATTE, P., MALUSKI, H., CABY, R., NICOLAS, A., KEPEZHINSKAS,P.& isotope evolution by a two-stage model. Earth and Planetary Science 39 40 SOBOLEV, S. 1993. Geodynamic model and Ar/ Ar dating for Letters, 6, 15–25. the generation and emplacement of the High Pressure (HP) meta- STEIGER,R.H.&JA¨ GER, E. 1977. Subcommission on geochronology: morphic rocks in SW Urals. Comptes Rendus de l’Acade´mie des convention on the use of decay constants in geo- and cosmochrono- Sciences, 317, 1667–1674. logy. Earth and Planetary Science Letters, 36, 359–362. MELCHER, F., GRUM, W., THALHAMMER,T.V.&THALHAMMER,O.A.R. TORSVIK,T.H.&REHNSTROM, E. F. 2001. Cambrian palaeomagnetic 1999. The giant chromite deposits at Kempirsai, Urals: constraints data from Baltica: implications for true polar wander and Cambrian from trace element (PGE, REE) and isotope data. Mineralium palaeogeography. Journal of the Geological Society, London, 158, Deposita, 34, 250–272. 321–329. PEDERSEN,R.B.&DUNNING, G. R. 1997. Evolution of arc crust and TORSVIK, T. H., SMETHURST, M., MEERT, J., ET AL. 1996. Continental relations between contrasting sources: U–Pb (age), Nd and Sr break-up and collision in the Neo-proterozoic and Palaeozoic—a isotope systematics of the ophiolitic terrain of SW Norway. tale of Baltica and Laurentia. Earth-Science Reviews, 40, 229–258. Contributions to Mineralogy and Petrology, 128, 1–15. WIEDENBECK, M., ALLE´ , P., CORFU, F., ET AL. 1995. Three natural zircon PUSHKAREV, E. V., FERSHTATER, G. B., BEA, F., MONTERO,P.& standards for U–Th–Pb, Lu–Hf, trace element and REE analyses. SCARROW, J. 2003. Isotopic Rb–Sr age of the pseudoleucite Geostandards Newsletter, 19, 1–23. tilaite in the Uralian Platinum Belt. Doklady Earth Sciences, 388, WILLIAMS,I.S.&CLAESSON, S. 1987. Isotopic evidence for the Precam- 373–377. brian provenance and Caledonian metamorphism of high grade ROBERTS, D. 2003. The Scandinavian Caledonides: event chronology, paragneiss from the Seve Nappes, Scandinavian Caledonides. paleogeographic settings and likely modern analogues. Tectonophy- Contributions to Mineralogy and Petrology, 97, 205–217. sics, 365, 283–299. ZONENSHAIN, L. P., KORINEVSKY, V. G., KAZMIN, V. G., PECHERSKY, RONKIN, Y. L., IVANOV, K. S., SHMELEV,V.R.&LEPIKHINA, O. P. 1997. D. M., KHAIN,V.V.&MATEVEENKOV, V. V. 1984. Plate tectonic The Platinum-bearing Belt of the Urals, Russia: Rb–Sr, Sm–Nd and model of the south Urals development. Tectonophysics, 109, geochemical constraints. In: IV South American Symposium on 95–135.