The metamorphic basement of the Corsica-Sardinia microplate is a fragment of the southern European branch of the Variscan belt, considered by most authors a collisional belt generated by Siluro-Devonian deep subduction of the oceanic crust and by Upper Devonian-Carboniferous continental collision (Burg and Matte, 1978; Autran and Cogné, 1980; Bard et al., 1980; Matte and Burg, 1981; Matte, 1983; Burg et al., 1984; Matte, 1986; Pin and Peucat, 1986; Costa and Maluski, 1988; Franke, 1989; Pin, 1990), with ubiquitous emplacement and stacking of plurikilometric nappes made up of Paleozoic cover formations in the External zone (Arthaud and Matte, 1966; Julivert, 1971; Carmignani and Pertusati, 1977) and of the crystalline Precambrian basement in the Internal one (Ries and Shackleton, 1971; Mattauer and Etchecopar, 1976; Burg and Matte, 1978; Matte,1983; Behr et al.,1984).
For the Variscan Belt in Sardinia, Carmignani et al. (1992, 1994, 2001) proposed the following evolution:
Sea-floor spreading between the passive continental margins of Gondwana and Armorica from the Precambrian to Lower Ordovician;
Middle Ordovician convergence between Gondwana and Armorica, with subduction below the Gondwana margin indicated by volcanic arc products on the continental crust (Andean type);
Silurian subduction of the ocean crust below Armorica, accompanied during the entire Devonian by restored passive behaviour of the Gondwana margin;
Lower Carboniferous continental collision between the Andean margin of Gondwana and the Armorica crust, with crustal stacking of several tectonic units;
Gravitative collapse of the thickened orogenic wedge, with ascent of the deepest metamorphic cores;
Crustal extension allowing the emplacements of calc-alkaline granitoids coeval with late Paleozoic volcanism and the opening of continental molasse basins.
This model is based on the existence of an ocean area between the Armorica plate and Gondwana (Massif Central Ocean: Matte, 1986; South Armorican Ocean: Paris and Robardet, 1990).
The suture of this ocean has been identified in the Massif Central, France, in the external Alpine Massifs (Aiguilles Rouges, Pelvoux, Belledonne, Argentera-Mercantour) and inferred in the Maures Massif (southern Hercynian suture by Bodinier et al., 1986; Matte, 1986). According to Cappelli et al. (1992) and Carmignani et al. (1992), the continuation of this suture might be represented in Sardinia by the Posada-Asinara Line, a major shear zone (cataclasites up to phyllonites for a thickness of 10-15 km), with a NW-SE direction from Asinara Island to the Anglona region and a nearly E-W direction along the Posada Valley.
The proposal by Cappelli et al. (1992), who consider the Posada-Asinara Line as a southern continuation of the suture between Armorica and Gondwana, is based principally on the regional importance of this major tectonic line and the occurrence within mylonites of several metabasite bodies interpreted as fragments of the ocean crust. A different hypothesis was put forward by Elter et al. (1999), who defined the Posada-Asinara Line as a major shear zone, probably once connected to the Ramatuelle-Plan de La Tour shear zone described in the Maures Massif by Vaucher and Buffalo (1998). The same interpretation of the Posada-Asinara Line as a major late Variscan shear zone, in contrast to the suture hypothesis, is proposed by Helbing (2003) on the basis of several analogies between the two facing sides along the line, showing that both sides follow the same evolution.
Moreover, Helbing and Tiepolo (2005), taking into account that the coeval orthogneisses of Tanaunella and Lodé are placed on opposite sides of the Posada-Asinara shear zone, emphasized that "the displacement along the Variscan Posada fault did not exceed the dimensions" of the "Mid-Ordovician magmatic Belt" of NE Sardinia and cannot have played a major role as a suture between Gondwana and Armorica, a hypothesis also rejected by Giacomini et al. (2005a).
According to Carmignani et al. (1992, 1994) and Carosi and Palmeri (2002), during the first collisional phase D1, the high grade metamorphic complex (HGMC) on the NE side of the Posada- Asinara Line overthrust the L-MGMC on the SW side of the same tectonic line. This overthrusting produced early inversion of metamorphic zones in underthrust units (Carosi and Palmeri, 2002).
Similarly juxtaposition of midcrustal rocks with supracrustal sequences was reported by Schaltegger (2000) for the Schwarzwald basement. Fluck (1980) and Rey et al. (1992) described an overthrust of a hot granulitic lower crust over an amphibolitic medium one. According to von Raumer (1998, p.419 and references therein), uplift and exhumation of deeper crustal rocks appearing as thrust sheets above less metamorphic levels could explain the ubiquitous occurrence of lower crust remnants within Variscan migmatite outcrops in the Monts du Lyonnais (Massif Central), Schwarzwald, Vosges, Eastern Austroalpine basement, External Alpine and Maures Massifs.
In Sardinia, the D2 phase is considered purely extensional by Carmignani et al. (1992, 1994, 2001), but transpressional by Carosi and Palmeri (2002). For all these authors, the D2 phase was responsible for extensional tectonics with exhumation of deep crustal levels and, according to Carmignani et al. (1992, 1994, 2001), for the genesis of metamorphic core complexes such as the NE part of Sardinia, the Capo Spartivento dome and the Flumendosa Antiform. During this stage, thrust surfaces were reactivated as low-angle listric faults, enhancing the detachment of external Nappes occurring on the peripheral molasse and Culm basins. At the end of the Upper Carboniferous and in the Lower Permian, the sialic root of the collisional orogenic belt was completely destroyed by extensional tectonics and by the coeval intrusion of huge masses of calc-alkaline granitoids.
On the contrary, according to Carosi and Palmeri (2002), transpressional tectonics, due to the indentation of the Cantabrian block within the narrowing Ibero-Armorican arc, lasted during the entire D2 and D3 phases up to the uppermost Carboniferous, and caused not only a slowing of the exhumation rate but also both reversal and west-to-northwestward displacement of the L-MGMC, bringing internal crystalline Nappes onto the migmatites of the HGMC. This means that Nappe transport changed from orthogonal to the belt in D1 to a parallel trend in D2.
Finally, besides the Posada-Asinara line, another discontinuity in southern Sardinia is claimed by Loi and Dabard (1997). They found the Caradoc-Ashgill sediments of Sulcis-Iglesiente (SW Sardinia, External zone) radically differ from coeval sediments of Sarrabus (SE Sardinia, External Nappe zone) as regards source of detrital materials and faunas. Some baltic Taxa are present in Sarrabus, but lacking in Sulcis-Iglesiente. The authors think that Sulcis-Iglesiente was part of north Gondwana margin, Sarrabus was distant and intermediate between Gondwana and Baltica.
The occurrence of pre-Variscan relics in Sardinia has long been debated. Naud (1979 and references therein) considered the eclogite and granulite relics in NE Sardinia, the augen gneisses at Capo Spartivento and the phyllite cover to be of Precambrian age. Carosi et al. (1995) described a pre-Variscan deformation in the Bithia Fm and Conti et al. (1978) proposed a pre-Variscan metamorphism for the Cabitza Fm. Carmignani et al. (1982a) attributed the polymetamorphic complex at Monte Grighini (central Sardinia) to a pre-Variscan basement. Carmignani et al. (1992) proposed a Precambrian (?) age for the protolith of sedimentary-derived migmatites in NE Sardinia, and Franceschelli et al. (2005) have recently hypothesised that the banded amphibolites at Monte Plebi, north of Olbia, were remnants of a pre-Variscan mafic-silicic layered intrusion.
Helbing (2003) and Helbing and Tiepolo (2005) attributed all pre-D2 minerals crystallised during the M1 event to a pre-Variscan basement in NE Sardinia intruded by late orogenic Ordovician granitoids (orthogneisses) and overlain by a cover with interlayered Middle-Ordovician acidic igneous rocks (porphyroids). D1 structural features would be the expression of Cadomian Pan-African tectonics. On the contrary, Franceschelli et al. (1982a,b), Elter et al. (1986) and Carosi and Palmeri (2002) attributed D1 deformations of the low-to medium-grade metamorphic rocks of NE Sardinia to Variscan orogeny.
According to Helbing and Tiepolo (2005), the garnet in the lower garnet zone (i.e. garnet +albite zone of Franceschelli et al., 1982a) is a pre-Variscan relict, whereas albite derives from Variscan metamorphism. Pre-Variscan and Variscan events reflect different metamorphic grades. Helbing and Tiepolo (2005) suggest a metamorphic jump (Figure 8, p. 693) from biotite to the garnet zone marking cover/basement contact, but field and petrological evidence of this jump is not provided.
Cortesogno et al. (2004) suggest that some eclogites associated with gneisses underwent lower Ordovician metamorphism in northern Sardinia. However, in spite of the several attributions of lithotypes and deformation phases to pre-Variscan orogenic events unconfirmed by radiometric or paleontological data, the only clear evidence of pre-Variscan deformation may be observed in the External zone (Iglesiente, Sulcis in SW Sardinia), where the weaker Variscan overprint still allows recognition of angular unconformity between Cambrian-Early Ordovician and Upper Ordovician sediments. This event has been present in the literature, and called the Sardic phase, for considerable time. However, Helbing and Tiepolo (2005) refuse to consider the Sardic phase a distinct orogenic event and attribute the Sardic unconformity to "the final deformational overprint" caused by the detachment of Avalonia from Gondwana. They assert: "the Sardic phase at its type locality of SW Sardinia is not, therefore, an orogenic event, but relates to rifting and evolution of the Rheic Ocean" (see following paragraph).
The existence of pre-Variscan deformation and metamorphism in northern Sardinia has not been proven to date with exhaustive, convincing argumentation. However, further field and petrographic work, such as new geochronological data, could help to finally solve this long-debated problem.