Rodinia without Baltica? Constraints from Sveconorwegian orogenic style and palaeomagnetic data Open Access

Supplementary material: Method and data from zircon trace element analyses are available at https://doi.org/10.6084/m9.figshare.c.6627988

In the absence of intact oceanic lithosphere, pre-dating the pre-Late Jurassic. and with only sparse faunal evidence, pre-Pangaean, Precambrian supercontinent reconstructions are largely based on palaeomagnetic data and matching orogenic and/or magmatic events on different continents. However, available palaeomagnetic data are unevenly distributed in time and space and geological evidence is variably preserved, typically allowing significant leeway in Precambrian plate tectonic reconstructions. Notwithstanding these limitations, however, many of the currently accepted reconstructions of the Precambrian supercontinent Rodinia are nearly indistinguishable from one another and from some of the original reconstructions dating back more than three decades (e.g. Dalziel 1991; Hoffman 1991; Moores 1991), particularly with regard to the Rodinia core blocks Laurentia, Amazonia and Baltica (Fig. 1). While this does not invalidate either the original or the more recent work, it does warrant a closer look at the current evidence for these reconstructions.

Since the work of Hoffman (1991) and Gower (1985), the SW Baltican and SE Laurentian margins (present-day coordinates) have been viewed as contiguous prior to and during collision with Amazonia at around 1 Ga, forming a triple junction in nearly all Rodinia supercontinent reconstructions, along the Sveconorwegian, Grenville and Sunsas orogens (Fig. 1). The continents are generally believed to have remained together until separation in the Ediacaran, during opening of the Iapetus Ocean (Wilson 1966). However, the paper by Slagstad et al. (2013a), that argues for a non-collisional setting for the Sveconorwegian orogeny, triggered both debate (Möller et al. 2013; Slagstad et al. 2013b) and a series of publications offering a variety of tectonic models (Möller et al. 2015; Tual et al. 2017; Bingen and Viola 2018; Slagstad et al. 2018, 2020; Bingen et al. 2021).

Cawood and Pisarevsky (2017) proposed a model, which suggests that Baltica and Laurentia formed a unity before c. 1265 Ma. Sometime between 1265 and 990 Ma Baltica and Laurentia became separate plates. This separation was marked by some 95° clockwise rotation and 1000 km southward motion of Baltica during the opening the Asgard sea. This model is in many ways very similar to that proposed earlier by Cawood et al. (2010).

In their most recent study, Kulakov et al. (2022), based on new palaeomagnetic and ⁴⁰Ar–³⁹Ar geochronological data, argued that simultaneously with or shortly after the 95° rotation, Baltica did not stay in its secondary position adjacent to Laurentia but instead drifted away (see Kulakov et al. 2022 for more details) latitudinally and probably longitudinally too. In this contribution, we briefly review the recently proposed tectonic models, and show that despite the long-held view that the Sveconorwegian orogen (SNO) is collisional and integral in assembly of Rodinia, this is not supported by the current geological and palaeomagnetic data. Instead, these data suggest that Baltica may have remained in solitude until early Silurian collision with the eastern Laurentian margin (Greenland).

The oldest rocks in the Sveconorwegian Province formed between 1.7 and 1.5 Ga during the Gothian and Telemarkian orogenies, which represent juvenile arc magmatism at or outboard of the SW Fennoscandian margin (Andersen et al. 2002; Bingen et al. 2005; Åhäll and Connelly 2008; Roberts et al. 2013). Following a tectonically and magmatically quiescent period between 1.5 and 1.3 Ga, juvenile bimodal magmatism dominated between 1.3 and 1.1 Ga and extended geographically well into central Sweden. Although the tectonic setting of this period is unclear, most workers favour a wide continental back-arc setting (Brewer et al. 2004; Söderlund et al. 2006; Spencer et al. 2014; Slagstad et al. 2020). The late Paleo- and Mesoproterozoic western Baltican and eastern to southern Laurentian margins share many similarities and the two are commonly inferred to have been contiguous throughout this time period (Gower 1985; Karlstrom et al. 2001; Condie 2013).

The Sveconorwegian Province consists of five main lithotectonic units (Fig. 2) that record tectonically and temporally distinct metamorphic, magmatic and depositional histories. Following a prolonged period of extension between 1.34 and 1.14 Ga across entire southwestern Baltica, the Sveconorwegian orogeny commenced at 1140 Ma with high-grade, granulite-facies metamorphism (between 7 kbar/800°C and 11.5 kbar/>850°C) in the Kongsberg and Bamble lithotectonic units in central parts of the orogen (Harlov 2000; Bingen et al. 2008; Engvik et al. 2016; Slagstad et al. 2020). At the same time, continued extension and sedimentation is recorded in the Telemark lithotectonic unit until c. 1100 Ma (Spencer et al. 2014). Cessation of sedimentation in the Telemark unit coincided with thrusting of the Bamble unit onto Telemark at 1100 Ma, indicated by rapid cooling of the former (Slagstad et al. 2020).

Between c. 1080 and 930 Ma, widespread, voluminous granitic and minor mafic magmatism in the Telemark unit, west in the orogen, coincided with high- to ultrahigh-temperature (HT/UHT) metamorphism (>900–1000°C) under low- to moderate pressure (6–8 kbar) (Drüppel et al. 2013; Blereau et al. 2017; Laurent et al. 2018; Slagstad et al. 2018). The Sveconorwegian magmatism is typically subdivided into the c. 1080–10 Ma Sirdal Magmatic Belt and the younger, c. 990–930 Ma hornblende–biotite suite (Granseth et al. 2020). A large anorthosite complex, the Rogaland Anorthosite Province (RAP) and the geographically associated Bjerkreim–Sokndal (BKSK) layered intrusion are spatially associated with the HT/UHT rocks (Marker et al. 2003), and earlier studies have argued that HT/UHT metamorphism was a contact metamorphic effect of short-lived emplacement of the RAP and BKSK at 930 Ma, with the development of contact-metamorphic isograds (Tobi et al. 1985; Schärer et al. 1996; Westphal et al. 2003), superimposed on an earlier regional metamorphism. More recent work, however, has shown that these contact-metamorphic ‘isograds’ reflect HT/UHT conditions over an extended period of time (Blereau et al. 2019), similar to the >150 myr range of metamorphic and magmatic ages obtained from the western and central SNO, and that the distribution of HT/UHT rocks is tectonically controlled rather than related to heat from the RAP (Slagstad et al. 2022). The short-lived emplacement of the RAP has also been challenged based on both geochronological and structural arguments (Bybee et al. 2014; Slagstad et al. 2022), which point to formation and emplacement of the RAP on time scales of >100 myr, between 1040 and 930 Ma.

Recent structural and geochronological data show that the HT/UHT metamorphic rocks constitute the footwall of a wide extensional detachment that formed in the period 980–930 Ma (Slagstad et al. 2022). This finding suggests that similar metamorphic conditions existed beneath much of the western Sveconorwegian orogen in the period 1080–930 Ma, consistent with the magmatic record showing extensive lower-crustal melting throughout this period and an indirectly observed increase in lower-crustal temperatures coinciding with the onset of orogen-wide extension at c. 990–980 Ma (Granseth et al. 2020, 2021).

To the east, the tectonic style was very different. High-pressure metamorphism (10–15 kbar/700–740°C) is recorded in the Idefjorden lithotectonic unit at 1050 and 1030 Ma (Söderlund et al. 2008) and eclogite-facies metamorphism (16–19 kbar/850–900°C) in the Eastern Segment at c. 990 Ma was followed by migmatization until c. 975 Ma (Möller et al. 2015; Tual et al. 2017). Extension-related magmatism in the eastern part of the SNO, including the foreland, between 978 and 946 Ma (Blekinge–Dalarne dolerite dyke swarm; Söderlund et al. 2005) coincides with extension and exhumation of HT/UHT crust in the western part of the orogen (Slagstad et al. 2022).

Widespread granite-dominated bimodal magmatism between c. 1300 and 1140 Ma is generally regarded to be related to back-arc extension in SW Fennoscandia (Bingen et al. 2002; Brewer et al. 2002, 2004; Roberts et al. 2011), the effects of which reached all the way to central Sweden (Söderlund et al. 2006). Studies of Sveconorwegian granitic magmatism have largely focused on the c. 990–930 Ma hornblende–biotite granite suite and its relationship with the RAP and BKSK, arguing that the dry nature of the latter two was related to derivation from and emplacement into granulite-facies crust, in contrast to the more hydrous hornblende–biotite granite suite (Vander Auwera et al. 2003, 2011). More recently, however, the so-called orthopyroxene-in isograd, separating the two regions of variably hydrous crust, was shown to be intrusive in nature and therefore not an isograd at all (Coint et al. 2015), with the granulite-facies UHT rocks hosting the RAP and BKSK resulting from late-orogenic exhumation rather than representing a different crustal domain (Slagstad et al. 2022).

The Sirdal Magmatic Belt was largely unrecognized until 2013 and limited to the c. 1050 Ma Feda suite (Bingen 1989; Bingen and van Breemen 1998; Slagstad et al. 2013a). Granseth et al. (2020, 2021) presented a large geochemical and isotopic dataset from both the Sirdal Magmatic Belt and the hornblende–biotite granite suite, and used the data to model the petrogenesis and mantle-crust source contributions to the magmatic rocks. They found that the Sirdal Magmatic Belt could have formed by c. 50% partial melting of a lower-crustal source, similar in composition (geochemically and isotopically) to the surrounding 1.5 Ga gneisses. The hornblende–biotite granite suite appears to constitute two subgroups; one subgroup that formed east of the Sirdal Magmatic Belt with a similar chemical composition and petrogenesis, and another subgroup intruding the Sirdal Magmatic Belt that is more strongly enriched in incompatible elements and may have formed by c. 5–10% melting of the lower-crustal residue following extraction of the Sirdal Belt magmas. This was interpreted to reflect partial melting over a wider region as a result of increased mafic magma underplating at the onset of extension, including an increase in lower-crustal temperatures allowing partial melting of previously melted, more refractory lower crust.

The relationship between Baltica and Laurentia during Sveconorwegian–Grenvillian orogeny has long been considered well understood. However, palaeogeographical models for this period vary considerably (Fig. 1). For instance, some models assume that the two continents became juxtaposed sometime around 1050 Ma and formed a single entity until at least 850 Ma (Pisarevsky et al. 2003; Li et al. 2008). Alternative views posit that northern Baltica became contiguous with northeastern Laurentia (eastern Greenland, present-day coordinates) sometime before c. 1250 Ma and then committed a c. 90° clockwise rotation after 1250 Ma but before 1000 Ma (Cawood et al. 2010). Following rotation, Baltica remained in Rodinia until the eventual breakup of the supercontinent in the late Neoproterozoic. A more recent study by Kulakov et al. (2022), however, proposed a different scenario, in which Baltica did not remain adjacent to Laurentia after the c. 90° clockwise rotation, but instead rifted and drifted latitudinally (and possibly longitudinally) away from Laurentia prior to the onset of Grenvillian–Sveconorwegian orogeny. The rift-to-drift scenario is consistent with the geological record from the Sveconorwegian orogen, as discussed further below.

At present, two competing hypotheses for the tectonic setting of the SNO exist, one advocating collision with another continent at c. 1050 Ma during assembly of Rodinia, and another proposing orogeny at or behind an active continental margin (Fig. 3) with no continent–continent collision at an unconstrained distance from Laurentia. Here, we briefly review the two models before discussing their validity in a later section.

The continent–continent collisional model for the Sveconorwegian orogeny has been around since the 1990s (e.g. Romer 1996) and, based on a general temporal similarity to the Grenvillian orogeny, this model has been the main argument for incorporation of Baltica in Rodinia (Li et al. 2008). Recently, significant modifications to the collisional model have been made in response to a number of new findings concerning the metamorphic and magmatic evolution of the orogen and is summarized below (time brackets and keywords from Bingen et al. 2021).

Upwelling asthenosphere and development of a (non-Tibetan, i.e. no crustal thickening) orogenic plateau during widespread extension (Fig. 3a). Partial melting of the lower crust initiated a decoupling between the crust and lithospheric mantle and progressive delamination or convective removal of the lithospheric mantle between 1280 and 1145 Ma. The lithospheric mantle was not entirely removed though, because the onset of Sveconorwegian orogeny, marked by high-grade metamorphism between 1150 and 1120 Ma in the Bamble and Kongsberg units, is ascribed to foundering of the continental lithospheric mantle pulling the overlying crust down with it (Fig. 3b). After 1120 Ma, break-off of the mantle slab resulted in exhumation of the Bamble and Kongsberg high-grade rocks (Fig. 3c).

Orogenic activity widened significantly to include the Idefjorden and Telemark units, with the development of a Tibetan-style orogenic plateau. Mantle upwelling under Telemark was balanced by mantle downwelling and continental lithospheric delamination and foundering under the Idefjorden unit, leading to the latter being pulled down by a lithospheric mantle slab (Fig. 3d). Eventual partial melting and weakening led to slab break-off and exhumation of the high-grade Idefjorden crust. Continued protracted mantle upwelling under the Telemark unit provided heat for lower-crustal melting, magmatism and high-grade metamorphism. Crustal thickness in Telemark is constrained to c. 40 km at 1040 Ma by high-alumina orthopyroxene megacrysts in the RAP (Charlier et al. 2010; Bybee et al. 2014).

At c. 1000 Ma, the Tibetan-style orogenic plateau propagated to the east to also include the Eastern Segment (Fig. 3e, f). Continued high-temperature medium-pressure metamorphism in the Telemark unit suggests continued mantle upwelling there, in contrast to high-pressure, eclogite-facies metamorphism in the Eastern Segment. The eclogite-facies metamorphism is ascribed to foundering of continental lithospheric mantle pulling the overlying crust down to great depth, followed by break-off and exhumation. The 978–946 Ma Blekinge–Dalarna dolerite dyke swarm is interpreted to reflect a ‘dynamic response’ of break-off in the asthenosphere. The RAP is interpreted to have formed towards the end of this period, at c. 930 Ma, by re-melting of a c. 1040 Ma mafic underplate, triggered by a c. 930 Ma mafic underplate. The Tibetan-style orogenic plateau was sustained until 930 Ma, when it underwent gravitational collapse.

A non-collisional interpretation for the Sveconorwegian orogeny was first advocated by Torske (1977) and Falkum and Petersen (1980), although a general lack of geochronological data hindered interpretations about timing and duration. These early interpretations were largely ignored until Slagstad et al. (2013a) revived the non-collisional model based on geochronological, chemical and isotopic data constraining the timing and style of Sveconorwegian metamorphism and magmatism. Since then, the non-collisional model has been further developed and corroborated, mainly through providing new geochronological, geochemical, isotopic and structural data from magmatic and high-grade metamorphic rocks and shear zones. The following account of the non-collisional model is largely based on Slagstad et al. (2020, 2022) and corresponds roughly to the time brackets of the collisional model.

During this period, large parts of southwestern Baltica underwent extension, suggested by widespread bimodal magmatism, sedimentation and relatively juvenile isotopic compositions (Brewer et al. 2004; Granseth et al. 2021). The extension most likely took place in a wide continental back-arc setting (Fig. 3i), and the sediments were fed by detritus from the active margin to the west as well as the craton to the east (e.g. Spencer et al. 2014). Extension and rifting are interpreted to have resulted in fragmentation of southwestern Baltica and separation of crustal domains to account for later diverse Sveconorwegian metamorphic, magmatic and depositional histories. There are no indications that rifting evolved into ocean-floor spreading and a scenario involving crustal blocks of near normal thickness separated by thinned or even hyperextended crust is more likely.

The first high-pressure metamorphism in the Sveconorwegian orogen is recorded in the Bamble and Kongsberg units in the central parts of the orogen, marking the transition from widespread extension to compression (Fig. 3j). The cause of this high-pressure event is unclear, partly because the eastern margins of the two units have been overprinted by the Permian Oslo Rift or lies out at sea. However, sparse evidence of coeval metamorphism in the Idefjorden unit could suggest that the two units (or at least parts of them) collided at this time. In the active-margin model, this event marks the onset of re-amalgamation of the fragmented Baltican margin (Slagstad et al. 2020).

In the Telemark unit, extension continued until c. 1100 Ma, indicating significant separation from the Bamble–Kongsberg units and showing that extension in parts of the system continued while other parts were in compression. In this model, the observed differences in stress regime are interpreted to reflect differential lower-crustal drag from convection cells developed in the wide continental back-arc (e.g. Tikoff et al. 2004; Macey et al. 2022).

Extension and sedimentation in Telemark ended at c. 1100 Ma, coincident with westward thrusting of Bamble and Kongsberg onto Telemark (Fig. 3k). HT/UHT metamorphism in the Telemark unit started at about the same time and appears to have been more or less continuous until c. 930 Ma, a period of more than 150 myr. Voluminous granitic magmatism coincides closely in time with high-grade metamorphism, consistent with isotopic data suggesting widespread lower-crustal melting. Although structural data suggest compression during much of the 1100–1000 Ma period (Henderson and Ihlen 2004; Scheiber et al. 2015; Stormoen 2015), the lack of high-pressure metamorphism at this time in western parts of the orogen and petrological/geochronological evidence indicating a Moho depth of c. 40 km, suggest a lack of significant crustal thickening. The most likely interpretation is that the HT/UHT lower crust was too weak to thicken significantly.

The first Sveconorwegian metamorphism to affect the Eastern Segment took place at c. 990 Ma and involved continental subduction beneath the crustal block comprising the Idefjorden, Bamble–Kongsberg and Telemark units (Fig. 3l; Möller 1998; Möller et al. 2015). Shortly thereafter, extension affected the entire orogen, from foreland to hinterland, and lasted until termination of orogenic activity at c. 930 Ma, or perhaps until as late as c. 880–850 Ma (Walderhaug et al. 1999; Mulch et al. 2005; Viola et al. 2011). Lower-crustal temperatures in the western Sveconorwegian orogen remained high during this entire (990–930 Ma) period, as demonstrated by HT/UHT lower-crustal rocks exhumed along a major, 980–930 Ma extensional detachment in the southwesternmost parts of the orogen (Slagstad et al. 2022).

The evolution from active-margin oceanic subduction to continent–continent collisional orogeny is typically marked by a change in magma source and/or melting conditions. One of the best examples of this change is the Himalayan–Tibetan orogen, where subduction-related arc magmatism prior to India–Asia collision was succeeded by low-temperature partial melting of the subducted Indian passive margin, forming orogen-wide peraluminous leucogranites (Searle et al. 2009). In cases where no passive margin is subducted on collision (i.e. a doubly vergent subduction system), the change in melt composition may be less marked, but would still be expected to reflect a shift in melting conditions and sources. Spencer et al. (2019), for example, argued that an isotopic pull-down marked the onset of Grenvillian continent–continent collision. However, similar pull downs are also observed along active margins, such as the western American Cordilleras, where periods of convergence and thickening alternate with periods of extension and gravitational foundering of arc roots, reflected in a c. 25–50 myr cyclicity of isotopic pull-downs and pull-ups, respectively (DeCelles et al. 2009). Thus, care must be taken when interpreting tectonic setting based on such data.

A relatively large database of chemical and isotopic data now exists from the Sveconorwegian Province that, along with metamorphic and structural data, can help constrain tectonic setting. Figure 4 presents plots of chemical and isotopic data v. age from 1300 to 900 Ma in the Telemark unit, which displays nearly continuous magmatic activity throughout this period. Figure 4a shows zircon saturation temperatures based on whole-rock geochemical data (Watson and Harrison 1983) and Ti-in-zircon temperatures calculated from zircon trace element compositions (Watson et al. 2006). The zircon saturation temperatures range between 750 and 950°C and there is no variation with time (R² = 0.002). Ti-in-zircon temperatures tend towards somewhat lower values, mainly between 650 and 850°C, with a slight trend toward higher temperatures with decreasing age (R² = 0.39). Although there are petrogenetic arguments for the youngest Sveconorwegian granites being hotter than earlier granites (Granseth et al. 2020), we note that there are comparatively few analyses at either end of the age range and that the apparent trend should be interpreted with caution. Caution should also be exercised when interpreting the zircon saturation temperatures as the calculation assumes that the analysed plutonic rocks represent melt composition; an assumption that is probably invalid to some extent for most of the samples.

In either case, the data suggest that the onset of Sveconorwegian orogeny was not associated with a significant change in melt temperature. Figure 4b shows alumina saturation (A/CNK; Maniar and Piccoli 1989) plotted against age. Most analyses straddle the boundary between peraluminous and metaluminous compositions (0.8–1.1) and there is a weak trend towards more metaluminous compositions with time but, as for temperature, no major change following the onset of orogeny. Other geochemical indicators that can be useful for discerning changes in tectonic setting, such as La_N/Yb_N and Sr/Y (e.g. Moghadam et al. 2022), are difficult to use with any degree of confidence due to the highly evolved compositions of the Sveconorwegian granites. Figure 4c shows zircon Hf data from rocks dated between c. 1290 and 920 Ma. The data plot along a trend consistent with reworking of c. 1.5 Ga continental crust, but modelling by Granseth et al. (2021) suggest some addition of mantle-derived material as well. An isotopic pull-down associated with the youngest period of magmatism (c. 980–930 Ma) could reflect subduction of the Eastern Segment under the orogen at c. 990 Ma (Brueckner 2009). Similar to the chemical data, the Sveconorwegian orogeny did not have a strong impact on the isotopic evolution of SW Fennoscandia. Sm–Nd data yield results similar to the Hf data (Granseth et al. 2020).

Persistently high lower-crustal temperatures between 1340 and 930 Ma in the Telemark unit, with no major breaks in the chemical or isotopic composition of magmatism, suggest that more or less similar tectonic processes operated throughout most of this period. Although outside the scope of this paper, the return to an extensional regime at c. 990 Ma suggests a certain long-term cyclicity between compression- and extension-dominated tectonic stresses on time scales of >100 myr, starting at least as early as 1.3 Ga, but possibly well before (e.g. Åhäll et al. 2000). In this perspective, the Sveconorwegian orogeny appears more as part of a continuum rather than a well-defined orogenic event, at least from the perspective of the Telemark unit.

Metamorphic and magmatic activity and deformation in continent–continent collisional orogens are typically interpreted to result from crustal underthrusting and thickening, followed by internal radiogenic heating (e.g. Clark et al. 2011). The thermal regime is to some extent dependent on the pre-collisional tectonic evolution (e.g. development of a continental back arc will lead to higher-temperature metamorphism; Hyndman 2019a), and duration of orogeny is a direct function of the driving force of convergence. When the convergent force ceases, the gravitational potential of the overthickened and thermally weakened crust typically results in gravitationally driven extension and thinning (collapse) (Dewey 1988; Rey et al. 2001).

During long-lived convergence and thickening, the lower and middle crust may undergo widespread partial melting and weakening, making it unable to sustain differential loading (i.e. topography) (Clark and Royden 2000). The result of this weakening is the formation of orogenic plateaus, of which the Tibetan plateau is the best known with a maximum crustal thickness of c. 90 km. However, similar orogenic plateaus can also develop in other tectonic settings where convergence causes long-term compression and thickening resulting in horizontal lower- and middle-crustal flow, as seen in the c. 70 km-thick Altiplano–Puna plateau in the Andes (McQuarrie et al. 2005). Thus, even if the presence of orogenic plateaus can be identified in the geological record, relating them to specific tectonic settings is not straightforward.

Interpretations of the Sveconorwegian orogeny as a continent–continent collision make frequent references to the Himalayan–Tibetan system, arguing for the development of an orogenic plateau that propagated eastwards during long-lived convergence (Möller et al. 2015; Bingen et al. 2021). However, unlike most models of the Himalayan–Tibetan system, where the mantle plays a comparatively minor role in providing heat and material, recent models of Sveconorwegian collision invoke a persistently thin crust (at most c. 40 km thick) throughout much of the orogen, with HT/UHT metamorphism on time scales of 150 myr driven by continuous mantle upwelling. More localized high-pressure metamorphism to the east is interpreted to be caused by crust being pulled down to great depth by gravitationally unstable, foundering lithospheric mantle. While we recognize that such instabilities are a likely result of orogeny, in the case of the Sveconorwegian there is no preceding processes causing the instability, and rather than resulting from orogeny the instabilities appear to cause it.

Thus, in our view there are very few similarities between this scenario and that proposed for the Himalayan–Tibetan orogenic system. It is also noteworthy that the processes of mantle upwelling and crustal pull are argued to have started around 1280 Ma and continued unabated following collision at c. 1065 Ma. In other words, the collision itself has no impact on the tectonic process.

The current collisional model contains a number of other inconsistencies and unresolved issues as well, such as apparently spontaneous lithospheric mantle foundering on time scales of several tens of millions of years, which is much longer than suggested by numerical models (e.g. Li et al. 2016) and without an obvious process causing the mantle instability. While underplating may eventually result in delamination, it is noteworthy that most of the underplating is deemed to have taken place in western part of the orogen, whereas most of the foundering took place in the east. The assertion that the Mesoproterozoic lithospheric mantle was generally denser than the underlying asthenosphere (Bingen et al. 2018) is not only unfounded and contrary to literature on the subject (e.g. Poudjom Djomani et al. 2001; Griffin et al. 2009), it also begs the question of how the Earth would have looked given such an unstable configuration.

It is unclear to us why the ‘collisional’ model requires a continent to collide with the southwestern margin of Baltica at all, since no significant crustal thickening is suggested and with foundering lithospheric mantle being the only driver for high-pressure metamorphism, both before and after collision. It is possible to envisage the other continent sidling up to southwestern Baltica in a soft collision, as has been proposed for Mid- to Late Ordovician Baltica–Avalonia (e.g. Torsvik and Rehnström 2003). However, in this case, the driving force for continuous mantle upwelling (subduction of oceanic crust) would have been lost as soon as the intervening oceanic lithosphere had subducted. Thus, not only is the presence of a continent SW of Baltica unnecessary in the current collisional model, but it would also appear to be tectonically complicated if long-term upwelling is accepted.

The most recent attempts at describing the Sveconorwegian orogeny in terms of an active margin advocates long-lived asthenospheric upwelling or convection in a wide continental back arc (Currie and Hyndman 2006; Hyndman 2019a, b; Macey et al. 2022). In this model, mantle convection driven by oceanic subduction provides heat to the lower crust and may also have been capable of causing compression/extension through viscous drag (‘clutch tectonics’ of the thermally weakened crust. Although, until very recently, no actual evidence for the existence of Sveconorwegian-age arc(s) was available, recent work in the tectonostratigraphically lowermost Caledonian nappes, widely inferred to represent the pre-Caledonian southwestern margin of Baltica, indicate that c. 960–950 Ma arc-related rocks exist there (Corfu 2019).

Accounting for the high-pressure events at c. 1140 Ma in Bamble–Kongsberg, 1050 and 1020 Ma in Idefjorden, and 985 Ma in the Eastern Segment is, admittedly, not straightforward in either the collisional or active-margin model. One challenge, in particular, is explaining the first high-pressure event at c. 1140 Ma in central parts of the orogen. Stress transmitted from the plate margin cannot account for this event as the intervening region in Telemark was still undergoing extension, thus some other mechanism is required. While the collisional model suggests that crust was being pulled down by foundering lithospheric mantle, the active-margin model proposes that the crust in southwestern Baltica had undergone hyperextension and thinning, at least locally, prior to 1140 Ma and that the negative buoyancy of this thin crust, perhaps in combination with drag at the base of the lithosphere, facilitated subduction and amagmatic basin closure (see also McCarthy et al. 2018, 2020). The latter model is, in principle, testable, for example by identifying pre-Sveconorwegian extensional structures, associated sedimentary basins and perhaps even exhumed lithospheric mantle. An alternative explanation is that margin-ward movement of the thick Baltican craton caused the high-pressure event at 1140 Ma; however, this idea appears to be contradicted by the fact that the eastern parts of the orogen did not undergo metamorphism until much later. It is possible that progressive heating and weakening towards the east allowed new parts of the Baltican crust to be involved in the orogeny; however, the eastward propagation of metamorphism does not appear to be progressive but rather stepwise and separated by many tens of of millions of years, which would make this propagation orders of magnitude slower than ordinary plate tectonic rates. Another orogenic driving force that may bear some similarity to the delamination-and-foundering model proposed by Bingen et al. (2021) is decoupling and subduction of negatively buoyant lower crust and mantle lithosphere (e.g. Capitanio et al. 2010; Chowdhury et al. 2020). This process has not previously been considered but is, in theory, capable of causing convergence on time scales of several tens of millions of years; it does not, however, require collision with another continent.

The above discussion illustrates that there are a several similarities between the recently proposed collisional and active-margin models: high-temperature metamorphism was driven by upwelling (or convecting) asthenosphere and high-pressure metamorphism is related to gravitational instabilities. However, whereas the gravitational instabilities in the active-margin model are ascribed to the preceding well-documented extensional history of southwestern Baltica, the spontaneous and long-lived foundering of lithospheric mantle does not appear feasible given the lack of processes that could have caused the instabilities, their duration of several tens of millions of years and their undocumented ability to pull hot and weak crust down. It is also unclear why such a process requires, or even permits, a continent–continent collision given that it is purported to have commenced c. 200 myr prior to collision. The upwelling asthenosphere in the collisional model is also required to keep on welling up with no obvious driving force following the cessation of oceanic subduction. Thus, despite some similarities in the processes alluded to by the different orogenic models, borne out of large amounts of recently acquired data, it would appear that placing a continent at the southwestern margin of Baltica would render these processes untenable, at least during the period of Rodinia assembly between <1100 and c. 930 Ma.

The most commonly proposed Rodinia reconstructions, placing southwestern Baltica against Amazonia and Laurentia, are inconsistent with the metamorphic, geochemical and isotopic data from the Sveconorwegian Province and require modifications. One possible solution is to argue for a long, tectonically varied margin, similar to the present-day Himalayan–Indonesian system, where the Tibetan part of the margin is undergoing continent–continent collision, whereas oceanic subduction is ongoing under Indonesia (Slagstad et al. 2019). Another solution is to separate Baltica from Rodinia. These two alternatives cannot easily be tested based on geological data alone and appear to require palaeomagnetic data that can constrain the continents’ absolute latitudinal position at the time of orogenesis; however, some insight may be gleaned from geological data from the younger Neoproterozoic period.

Copious amounts of zircon was produced along one or more of the Baltican margins during the Neoproterozoic (Andresen et al. 2014; Sláma and Pedersen 2015; Zhang et al. 2015; Andresen 2021), with most authors arguing for a Timanian source along the northern margin of Baltica. Slagstad et al. (2023), however, showed that the Timanian signal is much stronger farther away from the proposed source region and that production of zircon was coeval with thermal events of unknown nature along parts of the western Baltican margin. They argued that the western margin, typically but not exclusively inferred to have faced Laurentia until Iapetus opening around 600 Ma (Cawood et al. 2001; Hartz and Torsvik 2002; McCausland et al. 2007), could have been active during most of the Neoproterozoic, which is consistent with separation of Baltica from Laurentia (and Rodinia) prior to the Sveconorwegian orogeny.

The Precambrian palaeomagnetic database contains approximately 4000 palaeomagnetic poles, (https://paleomagia.it.helsinki.fi; Veikkolainen et al. 2014) and, in principle, one can reconstruct Baltica and Rodinia either juxtaposed or separated at any given time during Rodinia's tenure depending on the palaeomagnetic data selection. However, most of the available palaeomagnetic poles are of questionable suitability for constructing robust apparent polar wander paths and palaeogeographical reconstructions because most of these data lack robust temporal constraints and were very tentatively dated by fitting to assumed, and, in most cases, ill-defined apparent polar wander tracks. A detailed overview of Baltican and Laurentian Meso-to-Neoproterozoic data is beyond the scope of this study, and we refer readers to the relevant section in Kulakov et al. (2022). A recently published compilation of palaeomagnetic data for the Precambrian represents a series of palaeomagnetic community-approved and agreed-upon sets of reliable palaeomagnetic poles for all Precambrian cratons (Evans et al. 2021). Reliable palaeomagnetic data for Baltica and Laurentia from the compilation of Evans et al. (2021) and more recently published 1090–900 Ma data for Baltica (Kulakov et al. 2022) do not support the universally used juxtaposition of the two blocks during Rodinia's tenure, but instead suggest latitudinal separation between the respective margins of Baltica and Laurentia between c. 1090 and at least 1050 Ma (Fig. 5). This observation further questions the classical models arguing for a continuous Sveconorwegian–Grenvillian orogenic belt, including a possible Himalayan–Indonesian scenario as discussed above. While the palaeomagnetic data do not rule out contiguity of Baltica and Laurentia sometime after 950 Ma (but, importantly, do not support it either), there is no geological evidence supporting such a scenario.

Meert (2014) discussed the apparently fortuitous coincidence that many reconstructions of pre-Pangaean supercontinents (Columbia or Nuna, Rodinia and Gondwana) place the various continents in configurations that strongly resemble that of Pangaea. Meert (2014) suggested that while there may be some geological control on these continents’ tendency towards small horizontal movements and rotations, for example a partial stagnant lid regime, we also need to consider the possibility of a perception bias towards Pangaean-like reconstructions – arguably the only well-constrained supercontinent to date.

The seminal publication edited by Gower et al. (1990) contained a series of papers pointing out similarities in late Paleoproterozoic evolution along the southwestern Baltican and eastern and southern Laurentian margins, culminating in orogenic events around 1 Ga. Following the work of Gower et al. (1990), new data from the Sveconorwegian Province have largely been interpreted to support the inferred Grenville–Sveconorwegian connection. In most cases, however, such new data have been either non-unique with respect to tectonic setting or interpreted somewhat isolated from other data. Thus, an argument can probably be made that many interpretations of the Sveconorwegian orogeny are biased towards models of Grenvillian orogeny in eastern Canada.

For example, eclogite-facies metamorphism in the Eastern Segment at c. 990 Ma is typically correlated with the coeval Rigolet event along the Grenville Front (e.g. Möller et al. 2015), whereas the significant differences between the two events are largely ignored: the Rigolet marked the final orogenic event in the Grenville and succeeded the main Ottawan orogenic phase by several tens of millions of years, suggesting that it represents a final, separate compressional event following widespread post-orogenic extension (Rivers 2008). In contrast, the Eastern Segment eclogites formed at the end of a long period of compressional tectonics in the rest of the orogen and mark the transition from regional compression to extensional tectonics with HT/UHT metamorphism farther west in the orogen that lasted for another c. 60 myr. Also, no previous Sveconorwegian activity is recorded in the Eastern Segment unlike the Rigolet event that overprints Ottawan metamorphic fabrics. Thus, while the age of eclogite-facies metamorphism is broadly similar, there are also several significant differences that need to be considered.

Furthermore, the presence of a Tibetan-style orogenic plateau appears to be an important feature in Grenville geology (Jamieson et al. 2007; Rivers and Schwerdtner 2015) and also figures prominently in current continent–continent collisional models for the Sveconorwegian (Möller et al. 2015; Bingen et al. 2021). The Sveconorwegian plateau is proposed despite the seemingly general agreement that the crust throughout much of the orogen remained at c. 40 km thickness, probably because it was too hot to thicken more, and nowhere near the double thickness that characterizes the Grenville and Tibetan plateaus. There is also general agreement that heat was largely derived from the mantle rather than from radiogenic self-heating, also very different from the Grenville and Tibetan situations.

Biases also exist among advocates of an active-margin tectonic setting of course. For example, Granseth et al. (2021) argued that zircon Hf isotopes showed that Sveconorwegian orogeny involved near-continuous crustal growth (i.e. addition of juvenile mantle material). However, as pointed out by Wang et al. (2021) it is also possible to interpret the isotopic data to suggest reworking of older crust with few or no juvenile additions. On the other hand, it does appear significant that no change in isotopic evolution of granitic magmatism is recorded between c. 1.3 and 0.9 Ga (Fig. 4), which means that collision surprisingly did not instigate a change in the proportion of lower-crustal/mantle sources melting processes (Spencer et al. 2019; Granseth et al. 2020, 2021).

While some unresolved questions still exist in Sveconorwegian geology, several features appear to be more or less agreed upon, although the underlying tectonic driver is not. We highlight the following as a basis for further discussions on the tectonic evolution of the Svenonorwegian orogen:

1. The crust in western parts of the Sveconorwegian orogen did not undergo significant crustal thickening and reached a thickness of at most c. 40 km.

2. The main heat source in the western part of the orogen was upwelling/convecting mantle, operating on time scales well in excess of 150 myr (and probably closer 350 myr) and unimpeded by onset of the main phase of convergence around 1060 Ma.

3. Geochemical and isotopic data from granitic rocks ranging in age between c. 1.3 and 0.9 Ga do not signify changes in melting conditions and/or sources.

4. High-pressure metamorphic events are restricted to central and eastern parts of the orogen and appear to be discrete events rather than part of a continuum.

5. High-pressure metamorphism in central and eastern parts of the orogen are difficult to explain by plate-margin forces and appear to require some sort of internal driver.

The lack of significant crustal thickening and the necessity of oceanic subduction to drive upwelling or convection of asthenosphere unperturbed on time scales of several hundred million years are inconsistent with the presence of a continental landmass SW of Baltica during Sveconorwegian orogeny. This geological constraint invalidates the most common Rodinia reconstructions but does not require that Baltica was completely detached from the supercontinent as such. That said, there is no a priori evidence that the two continents remained juxtaposed after c. 1.2 Ga. Recently obtained palaeomagnetic data show that Baltica and Laurentia were widely separated latitudinally during the early and late stages of Sveconorwegian orogeny, providing a minimum separation since longitude is unconstrained.

We thank Toby Rivers and an anonymous reviewer for insightful and constructive comments to an earlier version of the manuscript and a large number of co-workers for discussions on the tectonic significance of the Sveconorwegian and Baltica's place in Rodinia. Analytical work was carried out under the auspices of MiMaC (Norwegian Laboratory for Mineral and Materials Characterization, NGU node).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

TS: conceptualization (lead), writing – original draft (lead); EVK: conceptualization (supporting), writing – original draft (supporting).

Analytical work was supported by the Research Council of Norway project number 269842/F50. EVK thanks the Research Council of Norway (RCN) for support through its Centres of Excellence funding scheme (project 223272: CEED).

All data supporting this study are provided as supplementary information accompanying this paper.