doi: 10.1111/ter.12001 A non-collisional, accretionary Sveconorwegian orogen Trond Slagstad,1 Nick M. W. Roberts,2,3 Mogens Marker,1 Torkil S. Røhr1 and Henrik Schiellerup1 1Geological Survey of Norway, Postboks 6315 Sluppen, 7491 Trondheim, Norway; 2Department of Geology, University of Leicester, Leicester, LE1 7RH, UK; 3NERC Isotope Geosciences Laboratory, Keyworth, Nottingham, NG12 5GG, UK ABSTRACT The late Mesoproterozoic Sveconorwegian orogen in southwest in the period 990–920 Ma. This magmatic and metamorphic Baltica is traditionally interpreted as the eastward continuation evolution may be better understood as reflecting a long-lived of the Grenville orogen in Canada, resulting from collision with accretionary margin, undergoing periodic compression and Amazonia, forming a central part in the assembly of the Rodinia extension, than continent–continent collision. This study has supercontinent. We challenge this conventional view based on implications for Grenville–Sveconorwegian correlations, com- results from recent work in southwest Norway demonstrating parisons with modern continental margins, Rodinia reconstruc- voluminous subduction-related magmatism in the period tions and how we recognize geodynamic settings in ancient 1050–1020 Ma, followed by geographically restricted high- orogens. T ⁄ medium-P metamorphism between 1035 and 970 Ma, suc- ceeded by ferroan magmatism over large parts of south Norway Terra Nova, 00, 1–8, 2012 Introduction Accretionary orogenesis, involving years of the onset of collision (Beau- gian orogenic belt is widely regarded oceanic subduction and terrane accre- mont et al., 2010; Jamieson et al., as a Himalayan-type and -scale oro- tion along a convergent margin, and 2010; Rivers, 2012). Evidence of this gen (e.g. Bingen et al., 2008b; Hynes continent–continent collision between is seen in collisional orogens, such as and Rivers, 2010) resulting from col- two major land masses, represent two the Himalayas and the Greenland lision with an unknown continent, very different geodynamic regimes Caledonides (e.g. Kalsbeek et al., possibly Amazonia, to the south. This (Cawood et al., 2009). Nevertheless, 2001). In contrast, syn-orogenic mag- orogen is typically regarded as form- orogen-scale cross-sections of the col- matism in accretionary orogens is ing an integral part in the assembly of lisional Tibetan Plateau (Yin and typically calc-alkaline with mixed the Rodinia supercontinent (refs. in Harrison, 2000; Searle et al., 2009) crust-mantle sources (e.g. Ort et al., Fig. 1). The Li et al. (2008) Rodinia and the accretionary Altiplano–Puna 1996; Davidson and Arculus, 2006), reconstruction is the most recent advo- Plateau (Elger et al., 2005; McQuarrie and may vary periodically in volume cating this ÔclassicÕ Baltica–Laurentia– et al., 2005) show a number of simi- and composition if the orogen alter- Amazonia configuration (Fig. 1A); larities, including major crustal short- nates between compression and exten- however, other reconstructions exist ening, crustal-scale shear zones and sion (Kemp et al., 2009). The style of that are incompatible with this classic high-grade metamorphism resulting in high-grade metamorphism and P–T interpretation. The most radical of mid-crustal partial melting. Whereas, conditions will also differ, with colli- these alternative reconstructions is the geodynamic settings are relatively sional orogens characterized by that of Evans (2009), who suggested easy to determine in these modern mid-crustal temperatures typically that the Baltica–Laurentia margin was orogens, distinguishing between them <800 °C (e.g. Jamieson et al., 2004), external, facing a large ocean to the in ancient orogens, where causal rela- whereas accretionary orogens under- southwest, with Amazonia located tionships are generally obscured or going periodic extension ⁄compression north in Rodinia (Fig. 1B). removed by later geological processes, may reach temperatures up to 900 °C Modern orogenic systems com- is far from straightforward. Conti- at similar crustal levels, over compar- monly display significant along-strike nent–continent collision involves atively geographically restricted areas variations in tectonic style. For exam- thrusting of a continent and its lead- (Collins, 2002). Thus, the timing and ple, the collisional Himalayan orogen ing, passive margin beneath the over- composition of pre-, syn- and post- continues south-eastwards to become riding continental plate. Burial of orogenic magmatism and the style of the accretionary Indonesian arc, and the passive-margin sediments to mid- high-grade metamorphism may be westwards into the active arc in Mak- crustal levels results in radioactive two of the most powerful ways of ran, with a combined length of over self-heating and extensive dehydration determining the geodynamic settings 5000 km. The Grenville–Sveconorwe- melting typically within c. 20 million of ancient orogens. gian orogenic belt is similar in scale The Sveconorwegian Province in and identifying variations of this type Correspondence: Dr. Trond Slagstad, Geo- southwest Baltica is commonly inter- along the length of the orogen is logical Survey of Norway, Leiv Eirikssons- preted as the eastward continuation of essential for constraining the presently vei 39, Trondheim 7491, Norway. Tel.: the Grenville Province in Canada (e.g. incompatible reconstructions of Rodi- +47 73 90 42 29; fax: +47 73 92 16 20; Gower et al., 1990; Karlstrom et al., nia. Here, we focus on the magmatic e-mail: [email protected] 2001), and the Grenville–Sveconorwe- evolution of the Sveconorwegian Ó 2012 Blackwell Publishing Ltd 1 A non-collisional, accretionary Sveconorwegian orogen • T. Slagstad et al. Terra Nova, Vol 00, No. 0, 1–8 ............................................................................................................................................................. they interpreted to reflect short-lived imparted on the SMB. The main (A) subduction of the southwest margin of crust-forming event in south Norway Baltica prior to continent–continent was at c. 1500 Ma (Bingen et al., collision. Figure 2 shows a new gen- 2005), and the available isotopic data Laurentia Baltica eralized regional map of south Nor- (Fig. 4C) show that a simple remelting way based on several years of of this crust, as would be expected mapping, and indicates locations of during a continent–continent colli- Study area new geochronological and geochemi- sion, cannot account for the isotopic Amazonia cal data from pre- to syn-Sveconor- composition of SMB rocks. Nd–Sr wegian granitoid rocks. The prefixes mixing calculations by Andersen et al. 1000 km pre-, syn-, late- and post- used in the (2001) suggest that the Feda augen Sveconorwegian orogen (Baltica) text are used relative to the timing of gneiss has a 65–80% mantle-derived (B) { Grenville orogen (Laurentia) high-grade Sveconorwegian metamor- component. Moreover, the onset of Sunsas orogen (Amazonia) phism in southwest Baltica, defined by SMB magmatism 15 Ma prior to Bingen et al. (2008a) as between c. high-grade metamorphism is inconsis- 1035 and 970 Ma (Fig. 3), with conti- tent with crustal reworking. We there- Amazonia nent–continent collision believed to fore interpret the available geochemical have initiated at c. 1050 Ma. Methods and isotopic data to suggest formation and data are presented in Electronic of the SMB in a continental magmatic Supplements 1–4. These undeformed arc between c. 1050 and 1020 Ma. Laurentia to weakly deformed pre- ⁄syn-Sveco- Contemporaneous magmatism with Baltica norwegian granitoids form a NNW- transitional calc-alkaline–anorogenic trending belt several tens of kilometres affinity in Vest-Agder (1035 ± 2 Ma Study area wide, i.e. significantly wider than the Fennefoss augen gneiss, Bingen and Feda augen gneiss (Fig. 2). We coin van Breemen, 1998), and ferroan mag- Fig. 1 Rodinia maps vary significantly, the term Sirdal Magmatic Belt (SMB) matism in Aust-Agder (1036 ± 23 Ma reflecting the difficulties in reconstruct- for this granitoid belt, of which the Rosskreppfjord granite, Andersen ing Precambrian supercontinents where Feda augen gneiss is a constituent. et al., 2002) and Telemark (1024 ± 24 the palaeomagnetic record is sparse, Although granulite-facies conditions Ma Otternes granite, Andersen et al., palaeontological data are absent and have been reached locally in the study 2007), requires a heat source for geological information is commonly area, the investigated granitoids gen- crustal melting, and a degree of rela- obscured by later geological events. (A) erally reached only amphibolite facies. tively juvenile mantle input to produce ÔClassicÕ configuration with the Lauren- We therefore interpret our zircon U– their isotopic signatures (Andersen tia–Baltica margin facing Amazonia, Pb dates to reflect igneous crystalliza- et al., 2009). This is compatible with roughly along the lines proposed by tion rather than later metamorphic lithospheric thinning and astheno- Cawood et al. (2007), Dalziel (1997), overprinting, in line with the textural spheric uprise, indicating an exten- et al. et al. Pisarevsky (2003), Li (2008). and compositional data from the zir- sional setting at this time, i.e. intra- or (B) Alternative reconstruction with con grains. The age data show that back-arc, rather than compressional, Amazonia located north of Laurentia, magmatism in the SMB was continu- and show that this magmatic phase with the Laurentia–Baltica margin fac- ing a Pacific-scale ocean, proposed by ous from c. 1050 to 1020 Ma, affected