The above presented data allow us to discuss a geodynamic evolution model. Several models involving the Massifs Central and Armoricain have been already proposed (e. g. Matte, 1986; Paris and Robardet, 1990; Robardet et al., 1994; Tait et al., 1997; Cocks, 2000; Robardet, 2003). Presently, two types of scenario are invoked. The first one emphasizes a continuous convergence between Gondwana and Laurussia from Silurian to Early Carboniferous (e. g. Matte, 1991; Lardeaux et al., 2001). The second one points out a polycyclic evolution (Pin, 1990; Faure et al., 1997). According to the later model, an Early Paleozoic cycle, from Cambrian to Early Devonian, is related to the opening and closure of the Medio-European Ocean and correlatively the drifting and rewelding of Armorica to Gondwana. Then, a second orogenic cycle, from Middle Devonian to Carboniferous, accounts for the closure of the Rheic Ocean and the collision of Gondwana and Laurussia. The structural and magmatic data presented in sections 3 and 4 support the polycyclic model developed in the following.
From Early Cambrian to Ordovician, the Massif Central and Massif Armoricain belong to Gondwana. The Central Armorican and Léon Domains are pieces of two microcontinents progressively drifted from the North Gondwana margin. The numerous alkaline granitoids found in the Para-autochthonous Unit, LGU and IGU (e. g. Duthou et al., 1984) and also in Léon (Tréglonou, Plounevez-Lochrist) or in Central Armorica (Brest, Lanvaux or Douanenez) argue for an Early Ordovician rifting. In the Para-autochthonous Unit, alkaline mafic volcanics (locally with pillow lavas), diabase dykes, gabbro intrude the grauwacke-pelite series (Pin and Marini, 1993). In the UGU, the bimodal magmatism of the leptynite-amphibolite complex is interpreted as the consequence of crustal thinning and oceanisation. The oceanic basin, is variously called South Armorican Ocean (Paris and Robardet, 1990), or Massif Central Ocean (e. g. Matte, 1986). Since this domain probably extends farther east up to Bohemia, the name of “Medio-European Ocean” will be used here. The width of this ocean is presently unknown. Since Ordovician faunas appear quite similar between N. Gondwana and Armorica, the existence of the Medio-European Ocean is questionned (e. g. Robardet, 2003). However, a narrow Medio-European Ocean did not act as a significant paleobiogeographic barrier. Indeed, as discussed in the next section, the Medio-European ocean closed in Late Silurian. Thus, the duration of this oceanic domain lasted less than 80 Ma from Early Ordovician to Late Silurian, and its width can be assumed whatever the opening and closing rates, ranging between 500 and 1000 km. The resemblance of Cambrian facies between Montagne Noire-Rouergue and Normandy although located in N. Gondwana and Armorica respectively can be easily understood since at that time, the two areas still belong to the same Gondwana continent. Paleomagnetism could help to solve this question, however, up to know all attempts show remagnetization of those Early Paleozoic series. As a matter of fact, the Ordovician rifting led to the formation of continental stripes such as Avalonia, Armorica, Léon drifted from Gondwana.
On the basis of available dates of the high-pressure metamorphism (Figure 6, section 3), the Medio-European Ocean closed in Late Silurian by northward subduction of the Gondwana margin. However, structural constraints (i. e. kinematics coeval with the development of high pressure assemblages) or geodynamic evidence (i. e. relics of a magmatic arc) are lacking (for a discussion of this problem, see Faure et al., 1997). The 440 +/-12 Ma age of the high pressure metamorphism in the Léon Domain (cf section 2. 3), suggests that continental subduction of this microcontinent might have occurred in Early Silurian, i. e., earlier than the collision of the North Gondwana Margin with the Central Armorican Domain. However, the date of collision between Léon and Armorica is not settled yet.
The existence of the Mauges basement nappe overlying the high-pressure Champtoceaux Complex in the South part of the Massif Armoricain (Figures 3, 4) supports a collision model. However, in the Central Armorican Domain, the lack of any disturbance in the Silurian-Devonian sedimentation and the Famennian unconformity observed near Angers (Lardeux, 1969) shows that this domain did not experienced any deformation before Late Devonian. The deep sea origin of Silurian radiolarite and arc derived magmatic rocks recovered as olistoliths in the St-Georges-sur-Loire olistostrome, led Cartier et al., (2001) to consider that a rift basin (called the Layon Rift) separated Armorica from an island arc installed upon a continental basement : the Mauges microblock. The complete closure of the Layon rift will occur only in Late Devonian-Early Carboniferous during D2 event. Therefore, in the Massif Central-Massif Armoricain area, the Early Paleozoic tectonics between Armorica and Gondwana corresponds to a continental subduction and collision of the N. Gondwana margin below an island arc.
Subduction of oceanic and continental rocks is followed by their exhumation in Early to Middle Devonian, around 390-385 Ma. The D1 event (section 3.1) coeval with pervasive retrogression of the high-pressure rocks of the UGU and migmatisation of the pelitic parts occurred at that time.
The exhumation mechanism is not well understood yet. A buoyancy driven model accommodated by southwestward thrusting and normal faulting (e. g. Chemenda et al., 1995) is a likely mechanism to be tested by further works.
According to faunal distribution, paleomagnetism and geodynamic reconstructions, (e. g. Tait et al., 1997; Paris and Robardet, 1990; Robardet et al., 1994; Matte, 2001; Robardet, 2003) in Late Silurian-Early Devonian, an oceanic basin called «Rheic Ocean», of ca 1500-2500 km width, separated the North Gondwana margin from Laurussia (Figure 12A). The Middle-Late Devonian magmatic arc and back-arc basins observed both in N. Gondwana margin and Central Armorican Domain show that at that time these two paleogeographic and tectonic domains already behaved as a single plate. The Late Devonian geodynamic setting is the southward subduction and subsequent closure of the Rheic Ocean that separated Laurussia (i. e. North America-Baltica-Avalonia already welded by the Early Paleozoic Caledonian Orogeny) and Gondwana-Armorica, also rewelded by the Early Paleozoic Variscan Orogeny, to the North and South respectively (Figure 12).
Figure 12. Reconstruction of the Late Devonian paleogeodynamics
A: Global reconstruction showing the active margin of Gondwana (in Laurussia, Av: Avalonia,
B: Baltica, NA: north America; simplified from Robardet, 2003). B: Close up of the study area.
C: Corresponding lithosphere scale cross section.
The closure of the Rheic Ocean will control the development of the D2 tectonics (see section 3.2, Figure 8). A general understanding of the Massifs Central and Armoricain tectonics requires to take into consideration the tectonics of SW England. On the basis of 40Ar/39Ar dates on metamorphic amphibole in the thrust sole, the Lizard ophiolitic nappe emplaced to the North in Late Devonian (ca 366-360 Ma, e. g. Le Gall and Darboux, 1986; Holder and Leveridge, 1986; Sandeman et al., 1995; Cook et al., 2002). Seismic profiling through the English Channel shows that the Lizard ophiolitic nappe is rooted into the Channel gravimetric and magnetic anomaly that can be considered as the Rheic suture (Figure 8). In turn, the ophiolitic nappe is overthrust by basement rocks belonging to the Léon Block (Le Gall, 1990). The Lizard ophiolite is sometimes rapported to the Rheno-Hercynian basin which, in Germany, opened in Mid Devonian along the trace of an Ordovician precursor : the Rheic Ocean assumed to be closed in Early Devonian (Franke, 2000). Along the SW England-Massif Armoricain transect, the available data do not support the existence of such an Early Devonian suture overprinted by a Late Devonian one. Therefore, in the present state of knowledge, the Late Devonian-Early Carboniferous closure of the Rheic ocean appears as the simplest interpretation. East of the Paris Basin Magnetic Anomaly (AMBP), a similar N-NWward directed stack of nappes is recognized along the “Nord de la France” seismic line south of the Ardenne Carboniferous Variscan thrust front (Cazes et al., 1985; Figure 8). The suture zone recognized east of the Bray fault might be correlated to the Rheic suture. Nevertheless, a comprehensive discussion of the whole system is beyond the scope of this paper.
The tectonic significance of the NW-SE trending D2 stretching lineation with top-to-the-NW shearing is one of the most puzzling problem of the Paleozoic evolution in the French massifs. Models involving a progressive rotation of shearing trends from transverse to longitudinal to the belt have been proposed (e. g. Brun and Burg, 1982; Burg et al., 1987). However, the Middle to Late Devonian distension is not taken into account. Moreover, the 20 Ma gap between the top-to-the-SW (D1) and top-to-the-NW (D2) shear events is hardly compatible with the model of a progressive and continuous rotation of shear trends. An extensional setting has also been invoked (Mattauer et al., 1988) but this interpretation is based on a misunderstanding between the D2 lineation and the Namurian-Westphalian stretching direction associated to the synconvergence extensional tectonics (Faure, 1995; cf section 3.2). Although similar in trend and often in kinematics, these two diachronous lineations differ by their P-T conditions of deformation. Moreover, the extensional model with a top-to-the-NW shearing cannot account for the tectonic superposition of the Thiviers-Payzac Unit above the Upper Gneiss Unit. In the South Massif Armoricain, the overthrust of the Mauges nappe upon the Upper Gneiss Unit precludes a northerly origin of the Thiviers-Payzac Unit from the Central Armorican Domain. Northwestward thrusting, as suggested by Bouchez and Jover, (1986); or Friedrich et al., (1988) complies with the prograde middle pressure/middle temperature metamorphism and the per-aluminous Guéret-type magmatism (section 4. 2).
A lithosphere-scale model linking the top-to-the-NW shearing on flat-lying foliations in the Massif Central and left-lateral wrenching in the Massif Armoricain is presented here (Figure 13). During the closure of the Rheic Ocean and subsequent collision, the lithosphere of the North Gondwana margin is extruded to the Southeast. The Silurian Nort-sur-Erdre suture is reactivated as left-lateral wrench fault, and the St-Georges-sur-Loire Unit is overthust to the NW. This boundary accounts for the lack of widespread flat-lying ductile deformation in the Central Armorican Domain and the development of left-lateral wrenching in the St-Georges olistostrome and Lanvaux Unit. In the Central Armorican Domain, the deformation is not distributed penetratively in the whole crust but instead is localized along belt-parallel strike-slip faults. This rigid behaviour of the Central Armorican Domain may be a consequence of the Neo-Proterozoic (or Cadomian) orogeny responsible for a general strain hardening of the Central Armorican crust. Conversely, the crust of the North Gondwana margin which never experienced the Neo-Proterozoic tectonics, but was already heated by the Devonian magmatism deforms penetratively.
Figure 13. Late Devonian-Tournaisian D3 geodynamics
Interpretative block diagram of the Late Devonian-Tournaisian D3 geodynamics in the Massif Central and Massif Armoricain. The top-to the-NW shearing in the North Gondwana Margin is coeval with the left-lateral wrenching in the Central Armorican Domain while the Rheic Ocean is closing by southward subduction. The top-to-the-NW ductile deformation is interpreted as an effect of a SEward escape of the North Gondwana Margin during the closure of the Rheic Ocean and subsequent Gondwana-Laurussia collision.
After the Tournaisian collision, intracontinental tectonics characterizes the Middle Carboniferous evolution of the Variscan Belt. The time interval from 340Ma to 290 Ma can be divided into a first period during which compression and extension regimes are both active and a second one during which extension controls the tectonic activity.
i) The Late Visean is a key period since the orogenic core, i. e. the central and northern parts of Massif Central or the entire Massif Armoricain, is characterized by the onset of syn-orogenic extension whereas, D3 compression is still going on both in the south Massif Central and in the northern foreland of SE England orArdenne (Figure 10). In the northern foreland, the Middle Carboniferous tectonics are characterized by northward directed thrusts with hundred kilometers displacement as shown by seismic lines (Cazes et al., 1985; Raoult and Melliez, 1987; Le Gall, 1990 and enclosed references). Those nappes are rooted in the eastern equivalent of the Rheic suture which is reworked by the dextral Bray fault.
The Late Visean period corresponds also to a huge magmatism that spreads out from north Limousin to south Vosges (Figures 2, 10). This magmatism is a direct consequence of the Early Carboniferous stage of crustal thickening (e. g. Pin and Duthou, 1990). In order to account for the high temperature magmatism and the mantle contribution, a lithospheric delamination model has been speculated (Faure et al., 2002).
ii) Extensional tectonics develops in Late Carboniferous (Namurian-Stephanian). As in most orogens, the evolution of the Variscan Belt ends with extensional tectonics that accommodates crustal re-equilibration. In the Massif Central, two stages have ben distinguished (e. g. Faure and Becq-Giraudon, 1993; Burg et al., 1994; Faure, 1995 and enclosed references). The syn-orogenic (325-310 Ma) stage is represented by NW-SE stretching as shown by the development of planar and linear fabrics developed in syntectonic plutons. As discussed in the previous section, the NE-SW trending acidic dykes of the Tufs Anthracifères in the North part of the Massif Central show that extensional tectonics initiated earlier in Late Visean.
The late orogenic stage (310-290 Ma) is associated to N-S to NNE-SSW maximum stretching of the crust responsible for the opening and infill of intramontane coal basins either by half-grabens or pull apart (for further details, see Faure, 1995). The Late Carboniferous high temperature granulites scavenged by Tertiary volcanoes that form the layered lower crust observed in the seismic lines are also the result of the syn-extension thermal input (Pin and Vielzeuf, 1983).