Kinematics and Timing Constraints in a Transpressive Tectonic Regime: The Example of the Posada-Asinara Shear Zone (NE Sardinia, Italy)

Detailed geological field mapping, integrated with meso- and microstructural investigations, kinematic of the flow and finite strain analyses, combined with geochronology, are fundamental tools to obtain information on the temperature–deformation–timing path of crystalline rocks and shear zone. The Posada-Asinara shear zone (PASZ) in northern Sardinia (Italy) is a steeply dipping km-thick transpressive shear zone. In the study area, located in the Baronie region (NE Sardinia), the presence of mylonites within the PASZ, affecting high- and medium-grade metamorphic rocks, provides an opportunity to quantify finite strain and kinematic vorticity. The main structures of the study area are controlled by a D2 deformation phase, linked to the PASZ activity, in which the strain is partitioned into folds and shear zone domains. Applying two independent vorticity methods, we detected an important variation in the percentage of pure shear and simple shear along the deformation gradient, that increases from south to north. We constrained, for the first time in this sector, the timing of the transpressive deformation by U–(Th)–Pb analysis on monazite. Results indicate that the shear zone has been active at ~325–300 Ma in a transpressive setting, in agreement with the ages of the other dextral transpressive shear zones in the southern Variscan belt.


Introduction
Shear zones are characterized by strain localization at different scales and the complex behavior of the involved rocks [1][2][3][4]. The activity of shear zones controls the exhumation of crystalline complexes during the collisional and post-collisional stages of orogenic belts [5][6][7][8][9]. Their geometry and kinematics strongly affect the way by which large crustal blocks are exhumed [7,10,11]. Structural geologists have long been interested in the kinematic reconstruction of the crustal-scale shear zones, which is thought to be one of the most effective ways to constrain key aspects of the tectonic evolution of collision orogens. Deformation in shear zones is often approximated to simple shear, especially in high-strain zones [7]. However, an important component of pure shear has been detected in many regional-scale shear zones developed in different tectono-metamorphic conditions [5][6][7][12][13][14][15][16][17][18]. Several vorticity gauges have been developed and used to quantitatively assess kinematic vorticity in natural shear zones [19]. During progressive deformation, the common pattern of vorticity path along transpressive shear zones suggests a strain partitioning deformation, characterized by a steep, narrow domain of simple-shear  [9] and Simonetti et al. 2018 [18]). The geographic location of the Sardinian island is shown; (b) the simplified geological map of the Variscan basement in Sardinia showing the distribution of the different tectono-metamorphic zones (modified after Carmignani et al. 1994 [36]). The PASZ high-strain zone is showed by the red shadow.
The lack of the Alpine age overprint makes Sardinia island an important locality to investigate a sector of the Southern Variscan belt and the transpressive deformation associated with the EVSZ (Figure 1a).
The Cambrian to Lower Carboniferous rocks, affected by a prograde Barrovian metamorphism, from anchizone (SW) to amphibolite facies metamorphism (NE), are characterized by folds and thrusts indicating a SW tectonic transport [29,[36][37][38][39]. Carmignani et al. (1994Carmignani et al. ( , 2001 [36,40] divided the Sardinian Variscan chain into three main tectono-metamorphic zones (Figure 1b). From SW to NE they are: (i) the External Zone (foreland area, southwestern Sardinia), a thrust and fold belt foreland consisting of a sedimentary sequence from pre-Cambrian(?)-Cambrian to Lower Carboniferous in age [36,40], showing very low-grade metamorphism; (ii) the Nappe Zone, subdivided into External (central to southern Sardinia) and Internal Nappes (northern to central Sardinia), results from southward thrusting with km-scale isoclinal folding and syn-tectonic metamorphism [36]. The External Nappe is made of a Paleozoic metasedimentary sequence bearing a thick continental Middle Ordovician arc-related volcanic suite [36]. This zone was deformed mainly under low-grade metamorphic conditions. The Internal Nappe is constituted by two main metamorphic complexes: the Low-Grade Metamorphic Complex (LGMC) and the Medium-Grade Metamorphic Complex (MGMC). The LGMC consists of a Cambrian meta-sedimentary sequence, Ordovician arc-related volcanic rocks, Silurian black-shale and Devonian marble [36,40]. This complex equilibrated mainly under greenschist facies conditions, except for the Monte Grighini Unit (Figure 1b) where  [9] and Simonetti et al. 2018 [18]). The geographic location of the Sardinian island is shown; (b) the simplified geological map of the Variscan basement in Sardinia showing the distribution of the different tectono-metamorphic zones (modified after Carmignani et al. 1994 [36]). The PASZ high-strain zone is showed by the red shadow.
The boundary zone between the Low-to Medium Grade Metamorphic Complex (L-MGMC) and the HGMC is marked by the Posada-Asinara shear zone (PASZ) [8,17,23,47], a nearly 150 km-long transpressive Variscan shear zone affecting both complexes. The metamorphic basement is widely intruded by Late Carboniferous-Early Permian granitoids constituting the Sardinian batholith, and/or unconformably overlain by the Late Carboniferous-Early Permian sedimentary basins [48]. During the syn-collisional shortening and thickening stage, both L-MGMC and HGMC underwent high-pressure (HP) metamorphism, related to the D 1 deformation phase, with metapelites reaching pressures of nearly 1.7 GPa [49]. The HP metamorphic signature [49], allows linking the prograde metamorphism to underthrusting and nappe stacking. The different portions of the metamorphic basement in northern Sardinia show different P-T trajectories, in response to the diachronous metamorphic evolution of the different areas [50][51][52], but the general trend is a clockwise P-T path, typical of subduction/collisional belts [53]. A northward increase in the Barrovian metamorphism has been recognized [39,50]. Barrovian minerals began to grow at the end of the collisional stage (D 1 phase). Sometimes they grow syn-kinematically at the beginning of transpressive stage (D 2 phase) [17]. In the L-MGMC, the Barrovian metamorphism has been dated at~350-330 Ma, based on Ar-Ar on mica in meta-pelitic rocks and U-Th-Pb monazite datings on para-and ortho-derivate metamorphic rocks [9,28,54]. The D 1 collision-related deformation is well recorded in the L-MGMC, where it is associated with a penetrative S 1 axial plane foliation of SW-facing folds. Towards the north, the S 1 foliation is progressively transposed by the D 2 phase developed in an orogen-parallel dextral transpression regime [17]. This was responsible for most of the exhumation of high-grade metamorphic rocks with initial increasing temperature followed by decompression and cooling [8,23,27,29]. The onset of the transpression, giving rise to the PASZ [8,15,17,55], developed in amphibolite-facies metamorphism and it has been constrained at~320-315 Ma (Ar-Ar on white mica and U-Th-Pb on monazite [9,28]). The transpressive activity of the PASZ continued during the D 3 phase under shallower crustal conditions. D 3 deformation phase is characterized mainly by metric chevron and/or asymmetric folds, associated with an S 3 axial plane crenulation cleavage [56]. The post-transpressional evolution was characterized by the inversion from contractional to extensional tectonics. The end of the orogenic activity is characterized by the collapse of the belt (D 4 phase), associated with open folds, brittle-ductile shear zones and locally by the emplacement of syn-tectonic granites [45,57].
The study area, located in the north-eastern Sardinia between the MGMC and the HGMC, extends for nearly 200 km 2 between the Posada lake and the Montalbo Massif ( Figure 2a) and comprises the south-eastern sector of the PASZ (Figure 1b). An updated geological map of the area, derived from a new detailed geological survey at the 1:10.000 scale, is shown in Figure 2. The Paleozoic metasedimentary sequence of this sector, consisting of micaschist and paragneiss, is characterized by a metamorphic zonation showing an increase in metamorphic grade from south to north. Four metamorphic zones occur, marked by the growth of: (i) garnet; (ii) staurolite + biotite; (iii) kyanite + biotite; (vi) sillimanite + white mica. The garnet zone is further subdivided into garnet + albite and garnet + oligoclase sub-zones [39,51,58]. Granodioritic and granitic augen orthogneiss and amphibolite lenses within the kyanite-bearing micaschist are also present. From the southern (low- Figure 2. (a) Geological sketch map of the study area (Baronie area, see Figure 1). The two sampling transects (1 and 2) and the low-and high-strain zone are indicated. The Posada-Asinara shear zone (PASZ) is also displayed. A geological cross-section, not in the same scale of the map, is present below; (b) the details of the two studied areas. Labels of the selected study samples and the corresponding types of analysis are reported; (c) 3D simplified block diagram showing the final structural architecture related to the superposition of different deformation phases. Stereoplot (equal angle, lower hemisphere projections) of the main structural elements are given.
The Paleozoic metasedimentary sequence of this sector, consisting of micaschist and paragneiss, is characterized by a metamorphic zonation showing an increase in metamorphic grade from south to north. Four metamorphic zones occur, marked by the growth of: (i) garnet; (ii) staurolite + biotite; (iii) kyanite + biotite; (vi) sillimanite + white mica. The garnet zone is further subdivided into garnet + albite and garnet + oligoclase sub-zones [39,51,58]. Granodioritic and granitic augen orthogneiss and amphibolite lenses within the kyanite-bearing micaschist are also present. From the southern (low-strain zones) to the northern sector (high-strain zones), the rocks become progressively more sheared (Figure 2a).

Field Data, Deformation History and Mesoscale Observations
In the study area, four deformation phases were recognized (Figure 2c). The oldest deformation phase (D 1 ) is testified by an older penetrative relict foliation (S 1 ) mainly recognizable at the microscale, and poorly developed at the mesoscale, observed only in the hinges of D 2 structures. F 2 folds are tight, isoclinal to locally ptygmatic. They are both cylindrical ( Figure 3a) and non-cylindrical (sheath fold; Figure 3b). They show rounded and thickened hinges with stretched limbs (class 2 [59]), that can be classified as B5 according to Hudleston (1973) [60]. F 2 km-scale folds (Lodè Antiform and Mamone Synform; Figure 2a) are developed at the map scale. A 2 fold axes show a main E-W to NNE-SSW trend gently plunging with quite scattered values ( Figure 2c). F 2 axial planes, parallel to the S 2 foliation, dip moderately toward S-SSE, with a local variation toward N-NNW ( Figure 2c). A well developed and penetrative S 2 foliation is the main planar structural element recognized in the study area. The S 2 foliation shows a general E-W to WNW-ESE strike and dips at high angles both to the N-NE and to the S-SW (Figure 2c). S 2 changes from an F 2 axial plane foliation, classified as a disjunctive cleavage with sub-parallel cleavage domains, in the southern part of the area, to a mylonitic continuous foliation, toward the northern sector. In migmatitic gneiss, the S 2 is defined by biotite-and sillimanite-rich layers. The L 2 object lineation [61] is represented by both grain and aggregate lineation. L 2 becomes prominent as the shearing increases. It is defined by elongate biotite and muscovite crystals, by millimetre to centimetre-scale quartz ribbons, elongate quartz-feldspar aggregates and elongated and boudinated staurolite (Figure 3c). The L 2 shows a N80-N120 main trend and gently plunges toward both E and W (Figure 2c). The lithotypes located in the southern sector of the study area show few kinematic indicators, whereas, approaching the high-strain zone of the PASZ in the northern sector, they become more frequent. Shear sense indicators have been observed at the mesoscale on a section parallel to the XZ plane of the finite strain ellipsoid (i.e., perpendicular to the S 2 foliation and parallel to the L 2 lineation). The main kinematic indicators are represented by C-S and C'-S fabrics (Figure 3d), σand δ-type porphyroclasts (Figure 3e), asymmetric strain fringes around staurolite and garnet porphyroclasts and shear bands boudins. All kinematic indicators indicate a top-to-the-NNW sense of shear. F 3 folds, affecting the S 2 foliation, are the main evidence of D 3 phase in the study area. The F 3 folds show kink (Figure 3f), chevron and/or asymmetric geometry (Figure 3g), with rounded hinges and locally asymmetric profiles. The D 3 deformation phase increases in intensity moving toward the high-strain zone. F 3 fold axes, A 3 , generally trend parallel both to the F 2 fold axes and to the L 2 object lineation, with higher plunging values (Figure 2c). In some areas, an intersection lineation, between F 3 axial planes and S 2 foliation occurs. An axial plane foliation is generally not well developed, but a D 3 -related crenulation cleavage (S 3 ) is locally present. A syn-kinematic growth of chlorite in the hinges of F 3 folds was observed. The F 2 -F 3 fold interference pattern shows parallel axes and sub-orthogonal axial planes (Type 3 [59]). During the D 4 deformation phase, metric-and decimetre-scale F 4 open folds developed with both sub-horizontal axes and axial planes (Figure 3h). A 4 axes plunge at low angles toward E or W with very high dispersion ( Figure 2c). F 4 folds re-orient and deform the original attitude of both the D 2 and D 3 previous structural elements, causing a variation in the orientation of the S 2 mylonitic foliation. D 4 phase is not associated with the development of foliations and object lineations. In some sectors, folds often coupled with thin, millimetre-scale, brittle/brittle-ductile shear planes, developed along the axial plane surfaces. These observations suggest that the D 4 phase developed at a shallower structural level.

Field Data, Deformation History and Mesoscale Observations
In the study area, four deformation phases were recognized (Figure 2c). The oldest deformation phase (D1) is testified by an older penetrative relict foliation (S1) mainly recognizable at the microscale, and poorly developed at the mesoscale, observed only in the hinges of D2 structures. F2 folds are tight, isoclinal to locally ptygmatic. They are both cylindrical ( Figure 3a) and non-cylindrical (sheath fold; Figure 3b). They show rounded and thickened hinges with stretched limbs (class 2 [59]), that can be classified as B5 according to Hudleston (1973) [60]. F2 km-scale folds (Lodè Antiform and Mamone Synform; Figure 2a) are developed at the map scale. A2 fold axes show a main E-W to NNE-SSW trend gently plunging with quite scattered values (Figure 2c).

Microstructures
Microstructural analysis was performed on samples collected along two parallel transects ( Figure 2b) perpendicular to the shear zone boundaries and the deformation gradient. The summary of the microstructural and petrographic investigation is reported in Figure 4. Microstructures are described according to the different lithologies. Mineral abbreviations are after Whitney and Evans (2010) [62]. Geosciences 2020, 10, x FOR PEER REVIEW 8 of 29  Micaschist and paragneiss are characterized by both a disjunctive cleavage, with sub-parallel cleavage domains, and a continuous schistosity mainly made of biotite + white mica and quartz ( Figure 5a).
Geosciences 2020, 10, x FOR PEER REVIEW 9 of 29 Micaschist and paragneiss are characterized by both a disjunctive cleavage, with sub-parallel cleavage domains, and a continuous schistosity mainly made of biotite + white mica and quartz ( Figure 5a). . Subgrain rotation recrystallization (SRR) overprinting is highlighted by the presence of "core and mantle" structure. An oblique foliation, Sb, defined by elongated quartz, is present (XPL); (b) oligoclase porphyroclast, in oligoclase-bearing micaschist, showing an internal foliation S1 (green line), defined by ilmenite + graphite. Oligoclase is wrapped by the main foliation S2 (blue line) defined by quartz + white mica + biotite (XPL); (c) garnet porphyroclast, in garnet-bearing mylonitic micaschist, showing an internal foliation S1 (green line) defined by ilmenite + elongated quartz. Garnet is wrapped by the main foliation S2 (blue line) defined by white mica + biotite (PPL); (d) staurolite in staurolite-bearing micaschist, showing an internal foliation S1 (green line), discordant with respect to the foliation in the rim which is in continuity with the external foliation S2 (blue line). This observation point to an interto early-syn tectonic growth of staurolite (PPL); (e) dynamically recrystallized quartz in staurolite- These observations indicate an inter-kinematic growth of Barrovian minerals (i.e., between the D 1 phase and the D 2 mylonitic shearing). In some staurolite crystals, two internal foliations are recognizable. In the core of the grains, an internal foliation S 1 is present. The core is characterized by the presence of an internal foliation S 1, discordant with respect to the external foliation. In the rims, another internal foliation, concordant with the external one, is present ( Figure 5d). These observations indicate an inter-to syn-tectonic (syn-S 2 ) growth of the staurolite (i.e., between the D 1 and D 2 phase and during the early D 2 phase). Quartz in mylonite displays undulose extinction and tilt walls. Quartz lobate and ameboid grain boundaries suggest dynamic recrystallization by grain boundary migration (GBM, [63,64] (Figure 5e). In some samples, quartz shows new grains of smaller size surrounding larger crystals forming a "core and mantle structure" [65], with weakly bimodal grain size. This indicates an overprinting of subgrain rotation recrystallization (SRR) on GBM [66]. A top-to-the-NW sense of shear is highlighted by oblique foliation in recrystallized quartz domains, by mica-fish and S-C-C' fabric ( Figure 5f) and asymmetric strain fringes around porphyroclasts. In granodioritic and the granitic augen orthogneiss, the S 2 disjunctive cleavage is marked by biotite ± white mica. The occurrence of asymmetric myrmekites and asymmetric K-feldspar porphyroclasts points to a top-to-the-NW sense of shear. In the migmatite gneiss, the mylonitic foliation is defined by biotite, white mica, ± fibrolitic sillimanite and wraps around sheared leucosomes consisting of quartz and feldspar. Sillimanite, growing parallel to the S 2 foliation, is syn-kinematic with the D 2 phase. However, sillimanite is frequently retrogressed to fine-grained white mica. In migmatite gneiss, the chessboard extinction of quartz is also recognizable, indicating deformation temperatures above 650 • C [64]. Feldspar shows mechanical twinning, undulose extinction and sometimes fractures. At the microscale, amphibolite lenses hosted in kyanite-bearing rocks are constituted by the alternations of plagioclase-rich and green amphibole-rich levels intercalated with quartz + plagioclase and garnet bands.

Methodology
A total of 34 samples (micaschist, para-and orthogneiss) from the PASZ were analyzed (Figure 2b, Figure 4). Vorticity analyses were performed on the samples poorly affected by post-D 2 deformation, on sections parallel to the L 2 lineation and perpendicular to the S 2 mylonitic foliation (XZ plane of the finite strain ellipsoid). The complete list of the samples selected for vorticity and strain analyses and the corresponding results are summarized in Table 1.
The flow regime of the PASZ was investigated, applying two independent kinematic vorticity gauges: the C' shear band method [22] for 26 samples, and the stable porphyroclasts method [12,66] on eight samples. The first method is based on the observation of whether C' plane nucleates are parallel to the bisector of the angle between the two apophyses of the flow [67][68][69][70]. If so, the orientation of C' planes concerning the shear zone boundaries, approximated by the C planes, is related to the kinematic vorticity (Wk) of the flow. It is, therefore, possible to measure the angle ν between the C and C' planes and to estimate the Wk, according to the relation: The maximum value of ν is preferable as it is representative, or close to, the original angle of nucleation of C' planes [22]. Table 1. Results of the kinematic vorticity and the finite strain calculations collected along the two study transects (see Figure 2 for samples location).

C' Shear Bands Method
Porphyroclasts Method Finite Strain Analysis Convergence Angle The second vorticity gauge is based on the rotational behavior of rigid objects within a flowing matrix [8,12,14,71]. The stable porphyroclasts method is based on three main assumptions ( [15] and references therein): (i) porphyroclasts are considered to be rigid objects in a Newtonian viscous matrix constantly deformed; (ii) clasts and matrix should be perfectly coupled together and slips do not occur between them; and (iii) there is no interference or interaction between clasts. To prove that the first condition has been assessed, it must be verified that the porphyroclasts do not show internal deformation and they are harder to deform compared to the matrix [15]. The second condition was tested by plotting porphyroclast distribution against the theoretical curves of Mulchrone [72,73], calculated for different Wk values for a system in which the slip between porphyroclasts and matrix occurs, as suggested by Iacopini et al. (2011) [15]. The third condition is satisfied if the porphyroclasts are positioned far from each other and are free to rotate. This method consists in the identification of a critical aspect ratio (Rc) that separates the porphyroclasts that reached a stable position from the ones that experienced a continuous rotation. According to Passchier (1987) [12] and Wallis et al. (1993) [66], Rc is related to Wm according to the relation: Measurements were performed using the software EllipseFit 3.2 [74] on the samples with an adequate amount of garnet and K-feldspar porphyroclasts. Examples of the applications of the two vorticity gauges are given in Figure 6a,b.
The full dataset, related to the vorticity calculations and also containing the Mulchrone tests [72,73] is given in Figures S1-S5. To check the type of deformation of the shear zone, according to Fossen and Tikoff, (1993) [75] and Fossen et al. (1994) [76], we calculated the θ angle formed between the maximum instantaneous stretching axis (ISAmax) in the horizontal plane and the shear zone boundary. The θ angle calculation was performed using the formula [19]: Finite strain analysis was performed with the centre-to-centre method [77,78] on both the XZ and YZ sections of the finite strain ellipsoid. The complete finite strain results are reported in Figure S6. Selected examples are reported in Figure 6c. To quantify the amount of shortening perpendicular to the mylonitic zone, finite strain and vorticity data from the same sample were combined [7,66]. Convergence angles were estimated according to Fossen and Tikoff (1998) [79]. The proposed relation is: where α is the convergence angle, kx and ky are the X and the Z axis of the finite strain, respectively, and γ is the shear strain. Geosciences 2020, 10, x FOR PEER REVIEW 13 of 29 The full dataset, related to the vorticity calculations and also containing the Mulchrone tests [72,73] is given in Figures S1-S5. To check the type of deformation of the shear zone, according to Fossen and Tikoff, (1993) [75] and Fossen et al. (1994) [76], we calculated the θ angle formed between the maximum instantaneous stretching axis (ISAmax) in the horizontal plane and the shear zone boundary. The θ angle calculation was performed using the formula [19]: Finite strain analysis was performed with the centre-to-centre method [77,78] on both the XZ and YZ sections of the finite strain ellipsoid. The complete finite strain results are reported in Figure  S6. Selected examples are reported in Figure 6c. To quantify the amount of shortening perpendicular to the mylonitic zone, finite strain and vorticity data from the same sample were combined [7,66]. Convergence angles were estimated according to Fossen and Tikoff (1998) [79]. The proposed relation is:

Vorticity and Finite Strain Results and Implications
Vorticity values obtained with the C' shear band method range between 0.40 and 0.74 (mean value of 0.58 and a modal value of 0.53). These results indicate a component of pure shear between 73% and 48%. The stable porphyroclasts method gives values ranging from 0.27 to 0.74, indicative of pure shear between 82% and 48%, in good agreement with the results obtained with the previous vorticity gauge (Figure 4). Comparison between the flow regime and the structural position of the studied samples (see Figure 2b for sample location) reveal a variation of the component of simple shear along the deformation gradient from SW to NE (from 27% up to 52%; Figure 7a) approaching the high-strain zone. In the Flinn diagram, most of the strain ellipsoids have an oblate shape, except one sample (MIC5E) falling near the plane-strain conditions (Figure 7b). value of 0.58 and a modal value of 0.53). These results indicate a component of pure shear between 73% and 48%. The stable porphyroclasts method gives values ranging from 0.27 to 0.74, indicative of pure shear between 82% and 48%, in good agreement with the results obtained with the previous vorticity gauge (Figure 4). Comparison between the flow regime and the structural position of the studied samples (see Figure 2b for sample location) reveal a variation of the component of simple shear along the deformation gradient from SW to NE (from 27% up to 52%; Figure 7a) approaching the high-strain zone. In the Flinn diagram, most of the strain ellipsoids have an oblate shape, except one sample (MIC5E) falling near the plane-strain conditions (Figure 7b).  Finite strain data suggest a general flattening, in agreement with pure shear-dominated transpression [75,76,81]. The scatter in strain geometry indicates that deformation in the shear zone was slightly heterogeneous. Combining Wk values and finite strain estimates from the same sample, following Wallis et al. (1993) [66] and Law et al. (2004) [7], we calculated 18-30% shortening, respectively. Since the studied samples are not deformed in the plane strain conditions, it is necessary to correct the shortening and stretching values [19,82,83]. Applying the aforementioned correction (Figure 7c), the stretching parallel to the transport direction is between 21% and 41%. According to the calculated θ angles, most of the studied samples (Figure 7d) plot in the field of pure shear-dominated transpression, and only two samples fall in the simple shear-dominated transpression area. Combining kinematic vorticity and finite strain, as stated before, a convergence angle, ranging between 51 • and 60 • , was estimated.

Methods and Analytical Techniques
To obtain time constraints on the evolution of the studied sector, monazite geochronology was performed [84,85]. Grain locations, internal features (inclusions, fractures, etc.) and Backscattered-Electron (BSE) images were obtained with the aid of a scanning electron microscope (JEOL JSM IT300LV; JEOL Ltd. Japan Electron Optics Laboratory Co. Ltd, Akishima, Tokyo, Japan) at the University of Torino (Italy). Mineral chemistry and X-ray compositional maps, to highlight the compositional zoning required for the correct interpretation of monazite grains and ages, were investigated with a JEOL 8200 Super Probe electronic microprobe hosted at the University of Milano (Italy). We followed the working conditions of Montomoli et al. (2013) [25]. Isotopic dating was performed at the CNR-Istituto di Geoscienze e Georisorse U.O. Pavia (Italy), using the laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) directly on 30 µm-thick sections. The instrument consists of a Ar-F 193-nm excimer laser (GeolLas 102 from Micro-Las) coupled with a magnetic sector ICP-MS (Element I from Thermo-Finnigan). The full description of the analytical procedures is reported in [86,87]. Single analyses were performed by a one-minute procurement of the background signal followed by recording, for at least 30 s, the ablation signal of the masses: 238 U, 232 Th, 206 Pb, 207 Pb, 208 Pb, 202 Hg and 204 (Hg + Pb). The analysis conditions were a spot size of 10 µm in diameter, an energy density of 8 J cm −2 and with a repetition frequency of 3 Hz. Matrix-matched external monazite standard (Moacir monazite) [88,89] was used for correcting the laser-induced elemental fractionation and mass bias. Monazite EMP analysis and isotopic results for standards and unknown are reported in Tables S1-S3. IsoplotR software [90,91] was used for the data processing and plotting.  Monazite grains are mostly subhedral to anhedral with a size between 80100 µm (ST2A), 70-150 µm (ST2B), and 40-80 µm (26-16-15). In sample ST2A, five grains (Mnz5, Mnz11, Mnz13, Mnz21 and Mnz24; Figure 8b) lie along the S 2 foliation; one grain (Mnz9; Figure 8c) is included in staurolite and is oriented parallel the S 2 foliation. In sample ST2B, three grains (Mnz2, Mnz10 and Mnz21) are along the S 2 foliation and two grains (Mnz11 and Mnz12; Figure 8d) are included in staurolite with Mnz11 parallel to the S 1 internal foliation, and Mnz12 located near the intragranular fractures. In sample 26-16-15, five grains (Mnz10, Mnz20, Mnz21, Mnz22 and Mnz24; Figure 8e) lie along the S 2 foliation or are included within syn-kinematic biotite and white mica; one grain (Mnz9) is included in staurolite, but near fractures connecting it with the matrix. Locally, monazite grains from sample ST2A and 26-16-15 show a partial replacement by a fine-grained allanite ± apatite aggregate. Thirteen elements were analyzed for 17 monazite grains within the three samples (Table S1). In order to get information about correlation within and among samples we applied a Principal Component Analyses (PCA) to the multivariate dataset ( Figure S7). The two main components PC1 and PC2 account for 63.8% of total data variance. Eigenvectors for Ca, Ce, La and Pr drive PC1 whereas Gd, Sm and U drive PC2. Combining the information from the two main components, monazite grains within ST2A are characterized by a higher Ca and Si content and lower P, according to PC1, and higher U and Pb concentrations, according to PC2, concerning the monazite from other samples. In sample ST2A, the CaO ranges between 0.6 and 1.4 wt.%, whereas in the other two samples, it is generally lower than 0.2 wt.% with a few values ranging from 1.0 to 1.4 wt.% (Table S1). Analogously, the UO 2 and PbO concentrations of monazite from ST2A are significantly higher (UO 2 > 0.4 wt.%; PbO > 0.06 wt.%) than those reported for monazite within the other samples (UO 2 < 0.1 wt.%; PbO < 0.01 wt.%). Monazite data from sample 26-16-15 show a major correlation according to PC2, positively correlated to light-rare-earth elements (i.e., La, Ce, Pr). Grains from the sample ST2B have an average higher Y absolute content with respect to the other two samples (Table S1). The X-ray maps of Ca, P, Y, Ce, La, Pr, Th, and U revealed a complex zoning consistent within each sample. We noted that, in both the ST2A and ST2B samples, monazite is characterized by the apparent zoning of Y and Ca that are generally inversely correlated to Th. Monazite from 26-16-15 is characterized by the marked zoning of Y, Ca, Ce, Th and U, where Y and Ca are generally inversely correlated with the other elements. In order to unravel a possible correlation between monazite and garnet, we focused on Y-Th-Ca zoning and chemistry. Combining the zoning features with quantitative chemical analyses, we observed that monazite shows concentric Y-Th-Ca zoning (Mnz5, Mnz9, Mnz21 and Mnz24 in sample ST2A; Mnz2, Mnz11 and Mnz21 in sample ST2B; Mnz21 and Mnz24 in sample 26-16-25) Figure 8b; Mnz9, sample ST2A). Most monazite along the main foliation or included in syn-kinematic minerals generally present either all the compositional domains described above or a homogeneous high-Y and high-Ca composition (chemically corresponding to the inner rim domain). The 206 Pb/ 238 U and 207 Pb/ 235 U isotopic ratios resulted as mainly discordant up to 10% (Figure 9a,b), whereas the discordance between 206 Pb/ 238 U and 208 Pb/ 232 Th data was generally better (<6%). The 206 Pb/ 238 U ages range from 334 ± 5 Ma to 295 ± 6 Ma, whereas the 208 Pb/ 232 Th ages range from 334 ± 6 Ma to 297 ± 5 Ma. These ages were taken from the ST2A and 26-16-15 samples, whereas the sample ST2B records a smaller age range between 332 ± 5 and 322 ± 5 Ma ( 206 Pb/ 238 U ages) and 332 ± 4 to 317 ± 4 Ma ( 208 Pb/ 232 Th). Taking into account the microstructural position of the grains and the chemical composition of the dated domains, we distinguished two main 206 Pb/ 238 U and 208 Pb/ 232 Th data clusters from~325-315 Ma and at~330 Ma. The most common age cluster mainly comprises high-Y rims/domains along the S 2 mylonitic foliation and within staurolite rims (syn-S 2 ) from ST2A and 26-16-15 samples. The oldest 206 Pb/ 238 U and 208 Pb/ 232 Th data cluster is generally defined by monazite included both in the staurolite crystals (parallel to the S 1 ) and within the matrix. The youngest ages (~310-300 Ma), were mainly found in discontinuous low and very low-Y outer rims, generally observed in grains where a retrograde allanite ± apatite formation is documented (several grains in the matrix and few in staurolite, where monazites are not completely shielded due to late fractures).

Geometry and Strain Partitioning of the PASZ
Fieldwork and microstructural analysis allowed us to recognize four deformation phases in the study area. The only evidence of D1 structures was the S1 relict foliation both in the hinge of F2 folds or as an internal foliation in the Barrovian index minerals. A progressive increase in metamorphic

Geometry and Strain Partitioning of the PASZ
Fieldwork and microstructural analysis allowed us to recognize four deformation phases in the study area. The only evidence of D 1 structures was the S 1 relict foliation both in the hinge of F 2 folds or as an internal foliation in the Barrovian index minerals. A progressive increase in metamorphic grade, as suggested by the progressive occurrence of Barrovian metamorphism (Figure 4), can be recognized from SW to NE in the studied sector, in agreement with Carmignani et al. (1994) [36] and Franceschelli et al. (1989) [51]. Barrovian index minerals mainly grew between the collisional (D 1 ) [17,92,93] and the transpressional D 2 event. However, we recognized that part of the staurolite growth is syn-kinematic with the D 2 shearing (Figure 4). The syn-kinematic mineral assemblage (sillimanite + biotite, biotite + white mica) parallel to the S 2 mylonitic foliation, is indicative of amphibolite-facies. This metamorphic condition is also in agreement with the occurrence of grain boundary migration as the main dynamic recrystallization mechanism of quartz, indicative of temperatures ≥500 • C ( [61]; Figure 4) and of feldspars [64]. The syn-kinematic chlorite in the hinges of F 3 folds, grown at the expense of biotite, testifies a decreasing temperature during deformation. Furthermore, local incipient subgrain rotation recrystallization, overprinting GBM in quartz, has been recognized, supporting a decrease in the temperature down to the greenschist-facies condition, as also noted by Graziani et al. (2020) [56]. The PASZ developed under conditions of decreasing temperature, starting from the amphibolite-facies (D 2 ) down to greenschist-facies conditions (D 3 ).
The principal structures of the study area are controlled by the D 2 progressive deformation phase, linked to the PASZ activity, in which the strain is partitioned into F 2 tight and isoclinal folds (low-strain zone) and shear zone domains (high-strain zone). We highlight the development of shearing domains that gradually increase in number and dimension moving toward the PASZ. In the same direction, D 2 folds become smaller and less frequent, while the mylonitic foliation becomes gradually more penetrative and continuous (strain partitioning, [94]; Figure 10).
The presence of a sub-vertical S 2 foliation, parallel to the boundaries of the shear zone, and F 2 fold axes, parallel to the sub-horizontal L 2 object lineation, led us to assume the simultaneous development of D 2 shear and F 2 fold domains. Kinematic indicators, both at the meso and microscale, point to a top-to-the-NW sense of shear, in agreement with Carosi and Palmeri (2002) [17]. The geometry and kinematics of the PASZ are compatible with a transpressive regime, as a consequence of dextral shear deformation with a shortening component perpendicular to the shear zone boundaries [8]. F 2 and F 3 fold axes parallelism suggests that the transpressive regime was active until the end of the D 3 deformation phase. Subsequent post-collision and post-transpression gravitational instability was characterized by the development of open folds (F 4 ) and low-angle brittle-ductile shear zones.

Kinematics of the PASZ
The kinematic of the flow allowed us to characterize the deformation, in terms of the percentage of pure and simple shear components of the D 2 phase, linked to the PASZ activity. Two independent methods were used. The mean kinematic vorticity number ranges from 0.27 to 0.74 (stable porphyroclasts method) and between 0.40 and 0.74 (C' shear bands method). Taking into account uncertainties in the estimates of the vorticity number, the two methods return consistent results (Figure 4). Despite the analyzed lithotypes along the studied transects being different, i.e., gneiss and micaschist, and thus have different rheology and possibly strain memory, a progressive northward increase in the amount of simple shear component (from~27% up to~52%) approaching both the high-strain zone and the HGMC was observed and highlighted by both the vorticity gauges. Finite strain data suggest a general flattening, in agreement with pure shear-dominated transpression [75,76,81]. Kinematic vorticity and finite strain analysis confirm that the deformation along the PASZ occurred under a general shear condition with a major component of pure shear acting together with simple shear in a transpressive regime. Geosciences 2020, 10, x FOR PEER REVIEW 20 of 29 The presence of a sub-vertical S2 foliation, parallel to the boundaries of the shear zone, and F2 fold axes, parallel to the sub-horizontal L2 object lineation, led us to assume the simultaneous development of D2 shear and F2 fold domains. Kinematic indicators, both at the meso and microscale, The transition from pure to simple shear dominated transpression, approaching the high-strain zone, is in agreement with the transpressional theoretical model proposed by Fossen and Tikoff (1993;[75]) and by Fossen et al. (1994;[76]). Fossen and Tikoff (1998) [79] defined a spectrum of transpressive and transtensive deformations which have in common a simple shear component and a perpendicular shortening component. In all the models of transpression [79] (types A-E) with any wrench component, there is an angular difference between the plate motion (oblique flow apophysis) and the fastest horizontal shortening direction. In type A, the strain ellipsoid has a prolate geometry, while in type B no extension or shortening along the strike of the zone was involved. Evidence of exhumation or the tectonic uplift of rocks within transpressive zones results from type C-E transpression. Despite the complex evolution, a series of observations collocate the PASZ in the type C shear zones: (i) a vertical foliation parallel to the shear zone boundary; (ii) a sub-horizontal lineation; (iii) a general flattening strain regime where the stretching along the X axis parallel to the extensional lineation matches or is greater than the vertical exhumation; (iv) the progressive increase in simple shear approaching the PASZ; (v) the convergence angle is not constant, varying from about 50 to 60 • . However, heterogeneous and non-steady-state transpressive deformation and strain partitioning observed along the PASZ led us to support also a change from type B (pure shear dominated transpression) to type C (simple shear dominated transpression). This is supported by the strain partitioning observed in the field, by the progressive increase in the Wk value and by a change in the convergence angle.

Monazite Chemistry and Dating: Timing of Deformation in the PASZ
Chemical analyses and X-ray maps of monazite revealed that significant chemical differences are present among the samples and within each grain. Sample chemical variations can be related to the bulk chemistry, whereas single grain zoning depends on mineral assemblages and/or modal abundances. According to the X-ray map, monazite shows apparent zoning that is consistent within each sample. Monazite in ST2A is characterized by a higher absolute concentration of Ca, U, and Pb, whereas monazite in sample ST2B showed higher Y average concentrations. To explain the composition of monazite from ST2A richer in Ca, U and Pb, we suggest a possible bulk control or the lack or low modal abundance in the sample of Ca-, U-, and Pb-bearing mineral phases (e.g., apatite). It is noteworthy that sample ST2B is also characterized by a lower modal abundance of garnet, which is also localized in some layers, with respect to ST2A and 26-16-15 samples. This implies a reduced reactivity of monazite at the thin-section/microdomain scale during garnet-involving reactions. The absolute higher Y concentrations observed in the ST2B could be related to the lower modal abundance of garnet with respect to the other samples. The low reactivity of monazite in this sample is also documented by the lack of retrograde replacement features (i.e., allanite ± apatite aggregates).
Microstructural position, texture, zoning and chemical analysis (Figure 11a,b), integrated with the P-T path obtained by Franceschelli et al. (1989) [51] and Carosi and Palmeri (2002) [17] (Figure 11), highlight a complex history of monazite growth during the tectono-metamorphic history. The rare oldest ages (around c. 330 Ma) were detected both along the S 1 foliation in staurolite and in the low-Y inner core and medium-high-Y outer core of the matrix grains (Figure 11a,b). According to the evolution of a garnet-monazite system [95], low-Y (and high-to medium-Th) inner core indicates that monazite grows coevally or just after the garnet growth [96,97]. Medium-high-Y (with a medium-low-Th) outer core can be explained by a progressive garnet breakdown [95] during staurolite growth, both together and subsequently with respect to the garnet. We did not detect any significant difference, within the resolution of the method, in the absolute ages of the two cores domains. The Barrovian index minerals grew during prograde metamorphic conditions under a near-peak amphibolite-facies condition (Figure 11c) [51]. The timing of these monazites domains, recording the D 1 phase history, are in agreement with the timing of the prograde metamorphism in Northern Sardinia (~350 Ma, [55]; 330-340 Ma, [28];~350-320 Ma, [98];~350-320 Ma [9]). The most common ages (~325-315 Ma) have been detected in high-Y and low-Th inner rims. This supported the growth of monazites during retro-metamorphism and garnet breakdown. Monazite grows as tiny rims around the previous domains, or as a new grain along the main S 2 foliation with a homogeneous high-Y composition. High-Y monazites are also present in parallel to the S 2 in the rim of syn-kinematic D 2 phase staurolite crystals (Figure 11a,b). The presence of syn-kinematic biotite, white mica and staurolite, with the presence of GBM quartz microstructures, indicate temperatures of amphibolite-facies condition. The prograde metamorphism is associated with the collisional stage while the retro-metamorphism is related to the transpressional deformation linked to the PASZ activity (D 2 ). In several monazite grains, along the mylonitic foliation or within staurolite crystals, but not completely shielded, we noted the presence of discontinuous outer rims with a low to very low-Y concentration with the youngest ages (~310-300 Ma; Figure 11a,b). These youngest domains document the latest stages of monazite re-crystallization slightly before (or even during) the progressive destabilization and monazite replacement by allanite [85,97] during the upper greenschist facies evolution of the transpressional deformation (D 3 ;~450-500 • C; Figure 11c). The subsequent cooling is also confirmed by the local occurrence of syn-kinematic chlorite and the overprinting of SRR on GBM (Figure 11a,c) [65]. All the collected data, coupled with previous P-T estimates, constrain the activity of the PASZ starting from~325 up to~300 Ma (Figure 11c), indicating a decrease in temperature and pressure from amphibolite-facies conditions down to the greenschist-facies condition in the time span of~25 Ma. The upper limit of the deformation age is also constrained by the emplacement at~300-280 Ma (U-Pb zircon dating) [99] of post-transpressional granitoids. The ages of transpression obtained in this work are in good agreement with the data collected in other areas of the PASZ (~320-300 Ma; [9,28]). Previous data of Carosi et al. (2012) [9] lack the chemical characterization of used monazite grains, because no X-ray compositional maps were produced. This makes it more difficult to link the obtained ages with the tectono-metamorphic evolution of sheared rocks. Our new data allow to better constrain the complete evolution (P-T-D-t path) of this sector of the Sardinian Variscan Basement by precisely linking age, metamorphic conditions and deformation. Geosciences 2020, 10, x FOR PEER REVIEW 23 of 29

PASZ in the Framework of Late Variscan Shearing
In the late Variscan framework, similar ages of transpressive tectonic were collected in the Maures-Tanneron Massif [35,100] and in the External Crystalline Massif [18,101] by applying the same methodologies. This reinforces the correlation of the different sectors of the Southern European Variscan Belt along a regional-scale shear zone, known as East Variscan Shear Zone (EVSZ; [9,34]). It is important to highlight that the EVSZ is not a unique shear zone, but it is a network of shear zones

PASZ in the Framework of Late Variscan Shearing
In the late Variscan framework, similar ages of transpressive tectonic were collected in the Maures-Tanneron Massif [35,100] and in the External Crystalline Massif [18,101] by applying the same methodologies. This reinforces the correlation of the different sectors of the Southern European Variscan Belt along a regional-scale shear zone, known as East Variscan Shear Zone (EVSZ; [9,34]). It is important to highlight that the EVSZ is not a unique shear zone, but it is a network of shear zones made by several branches with comparable tectono-metamorphic evolution [101], affecting the Variscan Belt during the late Carboniferous age (Figure 1a). Our new data enhance the potential correlations of the PASZ with other shear zones in the European Variscan Belt [34]. Texturally-controlled geochronological data indicate the similar ages of the transpression onset (~320 Ma) detected in different sectors of the Southern European Variscan Belt [9,18,35,100,101]. In several studied shear zones, a general flow regime was observed with the coexistence of simple and pure shear [9,18,100,101]. Moreover, the increase in the simple shear component recognized along the PASZ confirms the progressive change of orientation of the belt with respect to the regional stress field.

Conclusions
Our study, based on a multidisciplinary approach, highlights the following points:

I.
New meso-and microstructural data, combined with the kinematics of the flow and finite strain analysis, point to an increase of the simple shear component approaching the high-strain sector of the PASZ. The principal structures of the D 2 phase are controlled by a deformation in which the strain is partitioned into folds and shear zone domains. II. The transpressive activity and non-coaxial flow, coupled with an oblate finite strain ellipsoid and a variable convergence angle, suggests that the PASZ evolved from a type B (pure shear-dominated transpression) to a type C (simple shear-dominated transpression) shear zone models (Fossen and Tikoff, 1998 [79]). III. Texturally-controlled U-(Th)-Pb geochronology on monazite reveals that the onset and the progressive evolution of the PASZ lasted from~325 up to~300 Ma, in a time span of~25 Ma, and culminated with copious granitoid magmatism at~300 Ma. The PASZ evolved from the amphibolite-facies condition down to the greenschist-facies condition. The collisional stage, and prograde metamorphism, before the transpressive deformation, occurred at~330 Ma. IV. The PASZ shows striking structural, kinematics and chronological similarities with nearly vertical dextral ductile shear zones occurring in other fragments of the Variscan basements in Southern Europe, such as in the Maures-Tanneron Massif, Corsica and the Western Alps [9,18,35,[100][101][102].

Computer Code and Software
Structural data were projected with the software StereoNet©. The sketch-map and all figures were drawn using Adobe Illustrator®CC 2018. The kinematic vorticity analysis with the stable porphyroclasts method and the finite strain analysis were performed using the software EllipseFit 3.2 by Vollmer (2015) [74]. Calculation of the convergence angle was performed using the software Strain Calculator 3.2 [103]. Geochronological data were treated with Isoplot 3.70 [90].
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3263/10/8/288/s1, Figure S1- Figure S3: Polar histograms used for estimating the kinematic vorticity of transect A and B. ν = angle between the C' planes and the shear zone boundary; A1= flow apophysis 1; A2 = flow apophysis 2; n = total number of data; Wk = kinematic vorticity number. The dashed line represents the bisector of the angle between A1 and A2 apophyses. Figure S4: Mulchrone test [71,72] on the samples selected for the stable porphyroclast method. Curves represent the expected distribution of porphyroclasts at different Wk values if the distribution of porphyroclasts does not fit these curves, revealing that no significant slip between the porphyroclasts and matrix occurred. Figure S5: Diagrams showing the porphyroclasts distribution and the related mean vorticity number Wm. RCmin = minimum critical axial ration; RCmax = maximum critical axial ratio; Wmmin = vorticity number calculated with RCmin; Wmmax = vorticity number calculated with RCmax. Figure S6: Finite strain diagrams based on the centre-to-centre method of Fry (1979) [75]. Figure S7: Principal Component Analyses (PCA) of three analyzed samples (ST2A, ST2B and 26-16-15). Table S1: Chemical analysis of the monazite selected for geochronology. Structural formula calculated on the basis of 4 oxygens. Table S2: Standards analyzed for the geochronological study. Table S3: Isotopic results and geochronological data from the monazites of selected samples (ST2A, ST2B and 26-16-15).