Introduction
In recent decades, the quantification of intensive parameters of metamorphism has undergone remarkable evolution with the production of internally consistent thermodynamic databases (Berman, Reference Berman1988; Holland and Powell, Reference Powell, Holland and Worley1998, Reference Holland and Powell2011) and the improvement of our understanding of mixing properties and activity–composition relationships for mineral solid solutions and melts (Green et al., Reference Green, White, Diener, Powell, Holland and Palin2016; White et al., Reference White, Powell and Johnson2014). The development of software using these various databases has had a significant impact on modelling the thermodynamic evolution of metamorphic rocks (Powell et al., Reference Powell, Holland and Worley1998; Connolly, Reference Connolly2005; De Capitani and Petrakakis, Reference de Capitani and Petrakakis2010; Duesterhoeft and Lanari, Reference Duesterhoeft and Lanari2020; Xiang and Connolly, Reference Xiang and Connolly2022). These approaches allow the stability fields, mode and composition of minerals to be estimated for a given rock composition in the pressure‒temperature (P‒T) space (Thinkam and Ghent, Reference Tinkham and Ghent2005; Powell et al., Reference Powell, Guiraud and White2005; Spear et al., Reference Spear, Pattison and Cheney2016; Lanari and Duesterhoeft, Reference Lanari and Duesterhoeft2019; Lanari et al., Reference Lanari, Vho, Bovay, Airaghi and Centrella2019). This, together with decisive progresses in ‘in situ’ geochronology has allowed better understanding of the tectono-metamorphic evolution of high-grade units. (Štípská et al., Reference Štípská, Powell, Hacker, Holder and Kylander‐Clark2016; Plunder et al., Reference Plunder, Agard, Dubacq, Chopin and Bellanger2012; Airaghi et al., Reference Airaghi, Lanari, de Sigoyer and Guillot2017; Luoni et al., Reference Luoni, Zanoni, Rebay and Spalla2019; Lotout et al., Reference Lotout, Poujol, Pitra, Anczkiewicz and Van Den Driessche2020; Ghignone et al., Reference Ghignone, Sudo, Balestro, Borghi, Gattiglio, Ferrero and Van Schijndel2021; Bonnet et al., Reference Bonnet, Chopin, Locatelli, Kylander‐Clark and Hacker2022; Collett et al., Reference Collet, Regnault, Ozhogin, Imantayeva and Garnier2022). In contrast, poor thermomechanical constraints on the tectonic evolution of low- and very low-grade metamorphic units involved in external collision zones were implemented (Muirhead et al., Reference Muirhead, Bond, Watkins, Butler, Schito, Crawford and Marpino2020 and quoted references; Willner, Reference Willner2021). Currently, electron probe microanalysis (EPMA) provides high-quality reliable and reproducible compositional analysis, even on fine-grained minerals that are generally present in low-grade metamorphic rocks (see Lanari et al., Reference Lanari, Guillot, Schwartz, Vidal, Tricart, Riel and Beyssac2012; Airaghi et al., Reference Airaghi, Lanari, de Sigoyer and Guillot2017; Di Rosa et al., Reference Di Rosa, Frassi, Meneghini, Marroni, Pandolfi and De Giorgi2019; Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b, Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022c; Barbero et al., Reference Barbero, Di Rosa, Pandolfi, Delavari, Dolati, Zaccarini and Marroni2023). This has led metamorphic petrologists to investigate a larger dataset of metasedimentary rocks, opening new scenarios for understanding low-grade metamorphism based on robust P‒T estimations (i.e. Massonne and Schreyer, Reference Massone and Schreyer1987; Vidal et al., Reference Vidal, Goffé, Bousquet and Parra1999; Vidal and Parra, Reference Vidal and Parra2000; Parra et al., Reference Parra, Vidal and Agard2002; Agard et al., Reference Agard, Vidal and Goffé2001; Vidal et al., Reference Vidal, Parra and Vieillard2005; Dubacq et al., Reference Dubacq, Vidal and De Andrade2010; Pourteau et al., Reference Pourteau, Bousquet, Vidal, Plunder, Duesterhoeft, Candan and Oberhänsli2014; Lanari et al., Reference Lanari, Rolland, Schwartz, Vidal, Guillot, Tricart and Dumont2014a; Bourdelle and Cathelineau, Reference Bourdelle and Cathelineau2015). Among metasedimentary rocks, metapelites are widely represented lithotypes in mountain belts. The difficulty in studying this kind of metamorphic rock is related to the occurrence of detrital minerals, which remain as relict phases without reacting during metamorphism, thus making the reactive bulk composition difficult to investigate (Tracy, Reference Tracy and Ferry1982; Marmo et al., Reference Marmo, Clarke and Powell2002; Tinkham and Ghent, Reference Tinkham and Ghent2005; Lanari and Engi, Reference Lanari, Engi, Kohn, Engi and Lanari2017). Therefore, a combination of careful microstructural investigations and various thermodynamic methods are required to constrain the P–T paths recorded by these rocks.
In this study, we present an attempt to reconstruct a P–T path by combining crystallochemical indexes and thermodynamic modelling including multi-equilibrium thermobarometry and phase equilibria relationships (i.e. pseudosections). This approach is tested in metapelites sampled in a low-grade unit, i.e. the Cabanaira Unit, located in the High Valle Roja in the southwestern sector of the Marguareis Massif along the Italian–French border (Western Ligurian Alps, Fig. 1). The reconstructed P–T path is then discussed to evaluate its consistency with respect to the available regional tectonic framework of the investigated area.
Figure 1. (a) Geological scheme of the Western Alps-Northern Apennines (modified and re-drawn from Molli et al., Reference Molli, Crispini, Malusà, Mosca, Piana and Federico2010 and Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a). (b) Close-up of the Western Ligurian Alps (black box of a) with the related geological cross-section (c) modified by Bonini et al. (Reference Bonini, Dallagiovanna and Seno2010). (d) Simplified geological sketch (blue box of b) of the boundary between Maritime and Ligurian Alps.
Geology of the sampling area
The samples investigated were collected in the Cabanaira area within the Marguareis Massif. This massif is located along the Western Ligurian Alps which are part of the southwestern prolongation of the Alpine collisional belt (Fig. 1a,b). The latter is characterised by a long-lived convergence history between the Europe and Adria Plates (Handy et al., Reference Handy, Schmid, Bousquet, Kissling and Bernoulli2010; Schmid et al., Reference Schmid, Kissling, Diehl, van Hinsbergen and Molli2017; Manatschal et al., Reference Manatschal, Chenin, Haupert, Masini, Frasca and Decarlis2022). This history commenced in the Late Cretaceous with a subduction system leading to progressive consumption of the Ligure–Piemontese Ocean (ocean subduction stage), which was interposed between the two plates (Stampfli et al., Reference Stampfli, Borel, Cavazza, Mosar and Ziegler2001; Beltrando et al.; Reference Beltrando, Compagnoni and Lombardo2010; Lardeaux, Reference Lardeaux2014; Lagabrielle et al., Reference Lagabrielle, Brovarone and Ildefonse2015; Marotta et al., Reference Marotta, Roda, Conte and Spalla2018; Roda et al., Reference Roda, Regorda, Spalla and Marotta2019), and part of the Adria Plate (Polino et al., Reference Polino, Gosso and Dal Piaz1990; Lardeaux and Spalla, Reference Lardeaux and Spalla1991; Stöckhert and Gerya, Reference Stöckhert and Gerya2005). Subsequently, after the closure of the Ligure–Piemontese Ocean, the European continental crust started to be involved in a subduction system (continental subduction stage, Rosenbaum et al., Reference Rosenbaum, Lister and Duboz2002; Handy et al., Reference Handy, Schmid, Bousquet, Kissling and Bernoulli2010; Schmid et al., Reference Schmid, Kissling, Diehl, van Hinsbergen and Molli2017) in the Eocene until continental collision occurred during the early Oligocene (collision stage, Ford et al., Reference Ford, Duchêne, Gasquet and Vanderhaeghe2006; Simon-Labric et al., Reference Simon‐Labric, Rolland, Dumont, Heymes, Authemayou, Corsini and Fornari2009). Finally, tectonic denudation started during the Miocene (Tricart et al., Reference Tricart, Schwartz, Sue, Poupeau and Lardeaux2001).
According to the classic tectonic framework proposed for the southwestern Alps (Vanossi et al., Reference Vanossi, Cortesogno, Galbiati, Messiga, Piccardo and Vannucci1984; Seno et al., Reference Seno, Dallagiovanna and Vanossi2005; Seno et al., Reference Seno, Dallagiovanna and Vanossi2003; Bonini et al., Reference Bonini, Dallagiovanna and Seno2010; Mueller et al., Reference Mueller, Maino and Seno2020), their westernmost part (outer sectors) is characterised by a SW-verging fold-and-thrust belt consisting of the Dauphinois/Provençal Units, which, together with the External Crystalline Massifs, represent the former European continental margin, poorly involved in the Alpine collisional stage. These units are separated by the southern prolongation of the Penninic Front, whose activity is constrained at 33 Ma in this sector of the belt (Maino et al., Reference Maino, Casini, Ceriani, Decarlis, Di Giulio, Seno and Stuart2015), from the Briançonnais Units (Fig. 1a, see Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021, Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2023). The latter, together with the Internal Crystalline Massifs (blue domains of Fig. 1a), are classically regarded as the thinned portion of the European continental margin that was deeply involved in continental subduction, which is documented for the whole Western Alps (Michard et al., Reference Michard, Avigad, Goffé and Chopin2004; Bousquet et al., Reference Bousquet, Oberhänsli, Goffé, Wiederkehr, Koller, Schmid and Martinotti2008; Lanari et al., Reference Lanari, Guillot, Schwartz, Vidal, Tricart, Riel and Beyssac2012; Strzerzynski et al., Reference Strzerzynski, Guillot, Leloup, Arnaud, Vidal, Ledru, Corrioux and Darmendrail2011; Groppo et al., Reference Groppo, Ferrando, Gilio, Botta, Nosenzo, Balestro and Rolfo2019; Manzotti et al., Reference Michard, Schmid, Lahfid, Ballèvre, Manzotti, Chopin and Dana2022; Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b). They are thrust in turn by oceanic-derived units representative of the Ligure–Piemontese Ocean (i.e. the Moglio-Testico and the Borghetto d’Arroscia Units, Vanossi et al., Reference Vanossi, Cortesogno, Galbiati, Messiga, Piccardo and Vannucci1984; Di Giulio, Reference Di Giulio1988; Di Giulio, Reference Di Giulio1992; Mueller et al., Reference Mueller, Maino and Seno2020; Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022c; and the Voltri Group, Chiesa et al., Reference Chiesa, Cortesogno, Forcella, Galli, Messiga, Pasquar, Pedemonte, Piccardo and Rossi1975; Capponi and Crispini, Reference Capponi and Crispini2002; Capponi et al., Reference Capponi, Crispini, Federico, Piazza and Fabbri2009). The San Remo-Monte Saccarello Unit, whose paleogeographic origin is a matter of debate (Sagri, Reference Sagri1984; Mueller et al., Reference Mueller, Langone, Patacci and Di Giulio2018; Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a; Sanità, Reference Sanità2023), forms the uppermost portion of the tectonic pile (Fig. 1a–c, see also Maino and Seno, Reference Maino and Seno2016 and Mueller et al., Reference Mueller, Maino and Seno2020). Moving towards the Ligurian Sea, the San Remo-Monte Saccarello Unit tectonically overlays the Dauphinois/Provençal Units (Fig. 1a–c see also Perotti et al., Reference Perotti, Bertok, d’Atri, Martire, Piana and Catanzariti2012; Decarlis et al., Reference Decarlis, Dallagiovanna, Lualdi, Maino and Seno2013; Mueller et al., Reference Mueller, Maino and Seno2020).
In the sampling area, the Cabanaira Unit (cf. Roja Unit of Piana et al., Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014 and Rocca-Borbone Unit of Maino and Seno, Reference Maino and Seno2016) is separated, to the east, by the Penninic Front from a SW-verging stack of units (according to Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a, Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021b; Figs 1b,d). The latter, from the upper to the lower structural levels, consists of: (1) the topmost Marguareis Unit (Briançonnais Domain), regarded as a fragment of the European continental margin involved in the subduction zone (Carminati, Reference Carminati2001; Decarlis et al., Reference Decarlis, Dallagiovanna, Lualdi, Maino and Seno2013; Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a; Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021b) that experienced high pressure–low temperature (HP–LT) metamorphism (1.0–0.9 GPa and 280–330°C, Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b); (2) the Helminthoid Flysch Unit (cf. San Remo-Monte Saccarello Unit of Sagri, Reference Sagri1984), which is here considered as being derived from the External Ligurian Domain (according to Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a, Reference Sanità, Lardeaux, Marroni and Pandolfi2022a; Sanità, Reference Sanità2023) and has a diagenetic/anchizone-type metamorphic imprint (T never exceeding 220°C, Piana et al., Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014; Maino et al., Reference Maino, Casini, Boschi, Di Giulio, Setti and Seno2020 and P in the range of 0.5–0.2 GPa, Sanità, Reference Sanità2023). To the southeast, the oceanic-derived Moglio-Testico Unit crops out (Fig. 1d), which is tectonically located between the Marguareis Unit and the Helminthoid Flysch and has a HP metamorphic imprint achieved during the oceanic subduction stage (1.2–1.0 GPa and 260–330°C, Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022c).
According to Sanità et al. (Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a), the Cabanaira Unit (Fig. 2a) consists of a Jurassic platform meta-carbonate (Forte Pepin Limestone, FPL) unconformably underlying the middle Eocene (Gidon, Reference Gidon1972) Nummulites-rich meta-limestones (Nummulitic Limestone, NUM) and foredeep meta-turbidites (Cima Aurusi Formation, CAF). The finite strain pattern of the Cabanaira Unit, which records the superposition of folding and thrusting events (Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a, Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021b), is thought to be the result of its involvement in alpine convergent-related processes. The oldest deformation event (here D1 phase, cf. D1CU ‘pre-stacking structures’ of Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a) produced a SW-verging folding system that was confined within the unit and developed from micro- to map-scale (Figs 2b, 3a). The associated axial planes were cut by the subsequent deformation event (‘syn-stacking event’ of Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a), which was responsible for the thrusting of the Helminthoid Flysch Unit onto the already deformed Cabanaira Unit (see geological cross-section of Fig. 2b). The last deformation events are represented by a fold system (here called the D2 phase, cf. ‘post-stacking folds’ of Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a) with roughly flat-lying axial planes and high-angle faults. D1-related folding produced isoclinal to tight F1 folds that are associated with S1 tectonic foliation (Figs 2b, 3a). At the outcrop scale, this foliation is a continuous surface that is well developed in fine-grained rocks (i.e. metapelites), whereas in the more competent layers (e.g. limestones), it appears as a spaced cleavage or, as in other cases, it is scarcely preserved.
Figure 2. (a) Geological map (modified from Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a and Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021b) of the Marguareis Massif (red box of Fig. 1d) with the related geological cross-section (b). In the map the locations of samples are reported (red dots).
Materials and methods
Materials
To perform this study, we collected samples where the S1 foliation is well preserved, i.e. the metapelites from the topmost deposits (Cima Aurusi Formation) of the Cabanaira Unit. In these specimens, the S1 foliation is a slaty cleavage developed by pressure solution processes (i.e. pressure solution creep) that are diagnostic for fluid-assisted deformation under low-temperature conditions (Rutter, Reference Rutter1983; Gratier et al., Reference Gratier, Dysthe and Renard2013 with references therein). Parallel to this dissolution schistosity plane white mica (Wm), chlorite (Chl) and quartz (Qz) (mineral abbreviations after Warr, Reference Warr2021 with the exception of white mica) with marked shape-preferred orientation are typically well developed. Moreover, there are ultra-local micro-domains where grains of Wm + Chl + Qz ± calcite (Cc) and potassic feldspar (K-Fsp) can be observed (Fig. 3b–d) showing clear textural equilibrium relationships suggesting that metamorphic crystallization of new grains occurred along the S1 foliation. Along the S1 foliation, syn-metamorphic elongated Wm grains have lengths not exceeding 15–25 µm (Fig. 3b–e), whereas Chl grains are stubby, sometimes with recrystallized tails, reaching up to 30 µm in size (Fig. 3b-e). Both have undulose extinction and well defined edges with apparently no evidence of compositional zoning (Fig. 3b,c). Large and altered Wm and Chl crystals (both greater than 60 µm in length) with frayed edges and slight compositional zoning are also present (Fig. 3e) in which the micro-textures suggest a detrital origin. The latter are mostly set along the primary bedding (S0 in Fig. 3d,e) and show clear cross-cutting relationships with the syn-metamorphic mineral growth along the S1 foliation (Fig. 3d,e). Plagioclase (albite?, Fig. 3c) are also present and are characterised by stubby grains with frayed edges wrapped by very fine-grained crystals of Wm and Chl and indicating a detrital nature. In accord with the above microstructural observations the main deformation mechanisms include re-crystallization, pressure solution and passive rotation of detrital phyllosilicates (according to Groshong, Reference Groshong1988; Hirt and Tullis, Reference Hirth and Tullis1992; Passchier and Trouw, Reference Passchier and Trouw2005).
Figure 3. Macro- to micro-scale D1 phase-related structural features. (a) F1 fold system in the Cabanaira Unit (white box of Fig. 2b). (b, c) S1 (yellow line) slaty cleavage (bottom left of b: plane polarized light, top right of b and c: crossed polarized light) marked by syn metamorphic Wm and Chl and Qz in textural equilibrium. Detrital grains of plagioclase and chlorite (c) can be observed. (d) Microphotograph of F1 hinge-zone where clear cross-cutting relationships between S1 (yellow line) and bedding (S0, white dashed line) are shown. (e) BSE image of the micro-area investigated to perform the P–T estimates (orange box of d). The white box represents the area where the local bulk composition was extracted to perform forward and inverse thermodynamic modelling. Along the S0 (white dashed line), detrital Wm and Chl (blue arrows) are indicates. Along the S1 (yellow line) neo-formed Wm and Chl crystals (red arrows in the white boxes) are present.
Methodological approach
Microstructures and micro-textures were the main criteria used to select the microdomains characterised by dynamic recrystallization of Wm and Chl with which the P and T conditions were estimated (white box of Fig. 3e, see also the next section). Figure 3e shows the micro-area (orange box in Fig. 3d) chosen to perform the P–T investigation. The microdomain of Fig. 3d (detail in X-ray maps in Fig. 4a,b) was chosen in order to include neo-formed grains of Chl and Wm and Qz, which grew along the S1 foliation, avoiding detrital grains (Fig. 3d,e). In the selected micro-area, syn-metamorphic Chl and Wm (white box of Fig. 3e) are slightly zoned with rims slightly brighter than cores. Wm and Chl grains occur either as well-defined grains or very fine-grained crystals.
Figure 4. Composition of chlorites in ED208a sample. MgO wt.% (a) and FeO wt.% variation in the microdomain signalled in Fig. 3e. Differences exist between detrital (white dotted line) and syn-metamorphic (yellow line) Chl grains. (c) Enlarged area of the white box of (a) and (b) of the syn-metamorphic Chl (red arrows) grew along the S1 foliation (yellow line). The Al2O3 wt.% variation sheds light the complex mineral chemistry of the neo-formed grains (blue, red and green areas) corresponding to first, second and third groups (see the text). (d) AlVI/Si and (e) XMg/Si plots referred to the average compositions of the three groups of Chl detected along the S1 foliation are reported with their error bars. (f) Octahedral Fe2+ + Mg2+ (apfu, red dashed lines) content vs. Si (grey dashed lines) diagram (according to Bourdelle and Cathelineau, Reference Bourdelle and Cathelineau2015). The blue dotted lines show the octahedral vacancy (blue empty squares) values. The four Chl end-members are also reported.
As the detrital minerals cannot be excluded as relict phases in the micro-domain of interest, and because of the high-variance Chl–Wm–Qz mineral assemblage, different analytical techniques, including inverse multi-equilibrium thermobarometry and forward modelling (i.e. pseudosections), a Wm ‘crystallinity’ index (WmIC) and b0 cell parameters were used. Eleven samples of metapelites (Fig. 2a, the geographic coordinates of each sample are available in table T1 in Supplementary materials S1) were selected for measurements. Thermobarometric estimates and forward modelling were performed using X-ray compositional maps on one sample of metapelite collected in the F1 hinge zone (sample ED208a), whereas the WmIC values were measured on 10 samples of carbonate-free metapelite powders (ED203-208 and ED2017-220). The WmIC (Kubler, Reference Kubler1967a, Reference Kubler1967b) and the b0 cell parameter (Guidotti and Sassi, Reference Guidotti and Sassi1986; Guidotti et al., Reference Guidotti, Sassi and Blencoe1989; Franceschelli et al., Reference Guidotti, Sassi and Blencoe1989) are primarily known as qualitative methods and are thought to be functions of T and P, respectively. In this work, the T and P ranges obtained from the WmIC and b0 cell parameters were compared with the results of the other independent methods applied (Guidotti and Sassi, Reference Guidotti and Sassi1998; Kubler and Goy-Eggenberger, Reference Kubler and Goy-Eggenberger2001; Abad and Nieto, Reference Abad and Nieto2007; Warr and Ferreiro Mählmann, Reference Warr and R.F2015).
X-ray map processing and local bulk composition
The X-ray map was obtained using a JEOL 8800 electron microprobe at the Dipartimento di Scienze della Terra “A. Desio” (Milano, Italy), equipped with five wavelength-dispersive spectrometers and calibrated with the following standards: wollastonite (Ca, Si); orthoclase (K); albite (Al); periclase (Mg); rhodonite (Mn); TiO2 (Ti); Al2O3 (Al); Fe2O3 (Fe); and Cr2O3 (Cr) (for the acquisition settings see supplementary material S2, for the position of each spot analysis within the investigated micro-area see tables T2 and T3 and fig. F1). The X-ray intensities were processed with XMapTools 3.2 software (Lanari et al., Reference Lanari, Vidal, De Andrade, Dubacq, Lewin, Grosch and Schwartz2014b) to produce an oxide composition (wt.%) quantitative map using the spot analyses (see supplementary tables T2 and T3 in S2 for criteria used to evaluate the quality of standard analysis) as an internal standard following the procedure described by De Andrade et al. (Reference De Andrade, Vidal, Lewin, O’brien and Agard2006). The Local Bulk Composition (LBC) of the selected micro-area (white box of Fig. 3e) was obtained using the density-corrected oxide map function implemented in the XMapTools program (see the supplementary material S2 for details about the adopted procedure).
Forward modelling
Forward phase equilibria modelling was performed in order to evaluate the equilibrium of the metamorphic minerals that have grown along the S1 foliation using the LBC extrapolated from the wt.% map (Lanari and Engi, Reference Lanari, Engi, Kohn, Engi and Lanari2017, see table T4 in supplementary material S3). The Gibbs free energy minimization algorithm of Theriak-Domino (de Capitani and Petrakakis, Reference de Capitani and Petrakakis2010), was used to compute isochemical phase diagrams and mineral compositional isopleths of Wm and Chl with or without the ferric iron content. The internally consistent thermodynamic database JUN92.bs (Berman, Reference Berman1988; Pourteau et al., Reference Pourteau, Bousquet, Vidal, Plunder, Duesterhoeft, Candan and Oberhänsli2014) implemented with the solution models of Parra et al. (Reference Parra, Vidal and Agard2002) with the revised Margules parameters for Wm (Dubacq et al., Reference Dubacq, Vidal and De Andrade2010), Chl (Vidal et al., Reference Vidal, Parra and Vieillard2005), carpholite (solution models of Vidal et al., Reference Vidal, Goffé and Theye1992; Dubacq, Reference Dubacq2008) and chloritoid (solution models of Vidal et al., Reference Vidal, Parra and Trotet2001) were used to perform all of the computations (see the supplementary material S3 for details about the solution models included in the database) with an excess of H2O-rich fluid (aH2O = 0.8).
Thermobarometry
The multi-equilibrium modelling approach for partially re-equilibrated minerals (this study) allows us to detect different segments of the P–T path (i.e. peak P, peak T; in accord with Lanari and Duesterofth, Reference Lanari and Duesterhoeft2019). Therefore, to estimate the P and T conditions recorded by the Cabanaira Unit, chemical analyses of Chl and Wm grains grown along the S1 foliation were processed with three different methods using ChlMicaEqui software (Lanari, Reference Lanari2012), which operates assuming the presence of water (H2O) and Qz within the assemblage. In the calibrated compositional map (Fig. 4a), we investigated the micro-area where the relationships between the identified S1 foliation and the primary bedding (white box in Fig. 3e) were unambiguously clear. In this microdomain, we considered all chemical heterogeneity of Chl and Wm using the XMapTools software (Lanari et al., Reference Lanari, Vidal, De Andrade, Dubacq, Lewin, Grosch and Schwartz2014b). This operation allowed us to investigate analytical results for each neo-formed mineral phase (Chl and Wm) avoiding the detrital crystals (Tables 1 and 2). The methods used to retrieve the P–T equilibrium conditions are the Chl-Qz-H2O, the Phg-Qz-H2O and the Chl-Phg-Qz-H2O methods.
Table 1. Average compositions (wt.%) related to the three groups of Chl. Standard deviation is indicated
* average and standard deviation were obtained from 50 analysis for group from XMapTools.
n.d. – not detected
The Chl-Qz-H2O method (Vidal et al., Reference Vidal, de Andrade, Lewin, Munoz, Parra and Pascarelli2006) has been used to calculate the temperature range for chlorite formation (see supplementary material S4 for more details about this geothermometer), with an equilibrium tolerance of 30°C for estimating the percentage of Fe3+ for each Chl analysis with fixed pressure value (see F2 in supplementary S4) and specific water activity (0.8 for the presence of Cc, according to Frassi et al., Reference Frassi, Di Rosa, Farina, Pandolfi and Marroni2022).
The Phg-Qz-H2O method (Dubacq et al., Reference Dubacq, Vidal and De Andrade2010) relies on the de-hydration of Wm to refine the thermodynamic status of water, and is based on temperature-sensitive reactions (see supplementary material S4 for more details about this method). However, P also affects the reaction and, therefore, for a fixed T, the percentage of Fe3+ content and P can be estimated simultaneously (see Lanari, Reference Lanari2012; Scheffer et al., Reference Scheffer, Vanderhaeghe, Lanari, Tarantola, Ponthus, Photiades and France2016; Lanari et al., Reference Lanari, Vho, Bovay, Airaghi and Centrella2019; Di Rosa et al., Reference Di Rosa, Frassi, Malasoma, Marroni, Meneghini and Pandolfi2020; Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b). The Wm composition depends on the relative proportions of the end-members celadonite (Cel), muscovite (Ms) and pyrophyllite (Prl), which are mostly controlled by the Tschermak and Pyrophyllite substitutions (see Guidotti and Sassi, Reference Guidotti and Sassi1998 for a review).
By combining the T and P ranges obtained from the Chl-Qz-H2O and Phg-Qz-H2O methods respectively, the P–T estimates were calculated using Chl-Phg-Qz-H2O multi-equilibrium thermobarometry (Vidal and Parra, Reference Vidal and Parra2000, see supplementary material S4 for more details). Only the Chl-Wm couples whose P–T equilibrium was within the P and T ranges calculated with the two other methods were considered. For this selected group of P–T values, a further equilibrium tolerance was set to consider only the P–T values to which the minimum Gibbs free energy (<5000 J in this case) is related. Finally, to verify the robustness of our results, a comparison between the classic geo-thermobarometers available in the literature was performed (i.e. Massonne and Schreyer, Reference Massone and Schreyer1987; Bourdelle and Cathelineau, 2013; Lanari et al., Reference Lanari, Wagner and Vidal2014c). According to Vidal and Parra (Reference Vidal and Parra2000), the absolute uncertainty on T and P estimates are of ±30°C and ±0.2 GPa, respectively.
White mica ‘crystallinity’ index and b0 cell parameters
X-ray diffraction was used to obtain the WmIC and b0 cell parameters (see supplementary material S5 for more details about the sample preparation). The WmIC was measured on the <2 μm grain-size glycolated powder fraction that was placed on glass slides following the procedure described by Lezzerini et al. (Reference Lezzerini, Sartori and Tamponi1995). The WmIC measures the changes in the shape (Full Width at Half Maximum, FWHM) of the first basal reflection of Wm grains by X-ray diffraction, which is considered sensitive to temperature variation (Weaver, Reference Weaver, Flaun, Goldstein, King and Weaver1961; Kübler, Reference Kubler1967a, Reference Kubler1967b; Sassi and Scolari, Reference Sassi and Scolari1974; Kisch, Reference Kisch1980a, Reference Kisch1980b). The FWHM values expressed in Δ°2θ CuKα units of the Wm 10 Å peaks have been reported in the crystallinity index scale (CIS, Warr and Ferreiro Mählmann, Reference Warr and R.F2015). The spectra of each investigated sample are reported in fig. F3 in the supplementary material S5.
The b0 parameter was calculated (Sassi and Scolari, Reference Sassi and Scolari1974; Franceschelli et al., Reference Franceschelli, Leoni, Memmi and Puxeddu1989; Guidotti and Sassi, Reference Guidotti, Sassi and Blencoe1989) by measuring the d060 spacing in white mica grains using the (211) quartz reflection as an internal standard (i.e. Kisch et al., Reference Kisch, Sassi and Sassi2006). The positions of Wm and quartz reflections were defined on randomly orientated whole-rock powders for each metapelite sample. To verify the reliability of the measurement of b0 in this work, we performed a further calculation of this parameter using the compositions of Wm, which were observed to be in textural equilibrium with Chl using the equations proposed by Guidotti and Sassi (Reference Guidotti, Sassi and Blencoe1989) included in the spreadsheet of Verdecchia et al. (Reference Verdecchia, Collo, Zandomeni, Wunderlin and Fehrmann2019). The authors take into account different cation occupancies of Wm, where the condition Na/(Na+K) < 0.15 is obeyed to calculate the parameters: b0 = 8.9931 + 0.0440(Mg2+ + Fe2+ + Fe3+); b0 = 9.1490 – 0.0258(AlIV + AlVI); and b0 = 8.5966 + 0.0666(Si). Finally, to verify the robustness of our results for WmIC and b0, a comparison with data published in the literature by different authors in the same tectonic unit was performed.
Results
Mineral compositions
Chlorite
All the compositional heterogeneity of Chl grains grown along the S1 foliation (D1 chlorites) were considered. Note, that the quality of the compositions of Chl described in this section were evaluated following the recommendations proposed by Vidal and Parra (Reference Vidal and Parra2000). Although the chlorite investigated is located in the same microdomain (white box of Fig. 3e), compositional differences exist (Fig. 4a,b). At the microscale (Figs 3e,4a), detrital Chl along the S0 appears as large stocky crystals with frayed edges and has a MgO-rich composition (up to 15 wt.%) compared to the neo-formed S1 Chl (which tends to have ca 10 wt.% of MgO, yellow line in Fig 4a). The latter is characterized by crystals with well defined edges never exceeding 20–30 µm in size (Fig. 3e). Overall, a core-to-rim decrease of the MgO content can be observed. FeO wt.% tends to be higher in the core of both detrital and syn-metamorphic Chl (Fig. 4b). The Al2O3 wt.% map of Fig. 4c shows an enlarged area of the Chl grains that grew along the S1 foliation (white boxes of Fig. 4a and b). At least three groups of Chl can be detected (blue, red and green areas in Fig. 4c), whose average compositions with the related standard deviations are reported in Table 1. The first group is characterised by Al2O3 of ca. 23.00 wt.%; the second group tends to have slightly higher Al2O3 content (with a mean of 23.27 wt.%) whereas the third group has a more scattered Al2O3 content with a mean of 21.87 wt.%. XMapTools allows fine-scale selection of pixels located in different areas of the X-ray compositional maps. This tool allowed us to investigate small areas (a few µm-size) of Chl crystals in which the compositions cannot be acquired by EPMA. For each Chl group, the structural formulae were calculated on 14 anhydrous oxygens. The first Chl group is characterised by an XMg mean of 0.42 ± 0.08 (blue dot in Fig. 4d), Si contents of 2.75 ± 0.06 atoms per formula unit (apfu, blue dot in Fig. 4d,e), and AlVI content of 1.63 ± 0.06 apfu (Fig. 4d). The second group of Chl is characterised by higher XMg with a mean of 0.45 ± 0.05, Si of 2.72 ± 0.04 apfu, and AlVI similar to the first group chlorites (Fig. 4d,e) with 1.60 ± 0.04 apfu,. The third group of D1 chlorites is characterised by XMg of 0.34 ± 0.05, mean Si content of 2.77 ± 0.07 apfu and more variable AlVI content (1.52 ± 0.16 apfu) (Fig. 4d,e). All D1 chlorite groups have Si contents that never exceed 3.00 apfu; the total contents of Mg2+ and Fe2+ present in the M octahedral site (M-site) are greater than 3.50 apfu and have vacancies of lower than 0.20 apfu (Fig. 4f). Overall, the D1 chlorite groups have clinochlore (Clc) + daphnite (Dph)-rich compositions (Fig. 4f, see also Table 1). The sudoite (Sud) content tends to progressively decrease from the first to the third group of D1 chlorites, whereas the amesite (Am) content tends to be higher in the second group of chlorites and, generally, never exceeds 30% (Fig. 4f and Table 1).
White mica
All chemical heterogeneities of the D1 Wm were considered, and the quality of these is based on the recommendations proposed by Vidal and Parra (Reference Vidal and Parra2000). In common with Chl, detrital Wm also occurs as large crystals (up to 100 µm in size) characterised by frayed edges, whereas the neo-formed Wm never exceed 20–25 µm (Figs 3e,5). The SiO2 content ranges between 45 and 55 wt.% (Fig. 5a) and the Al2O3 tends to be higher (up to 40 wt.%) in the detrital Wm compared to the neo-formed Wm (20–35 wt.%, Fig. 5b). However, the Al2O3 wt.% variation at a grain-scale (Fig. 5b) reveals that differences exist for both detrital and syn-metamorphic Wm crystals with a general decreasing core-to-rim trend. Figure 5c shows an enlarged area of the X-ray compositional map (white box of Fig. 5a, b) where the variation of Si apfu for all the Wm grains along the S1 foliation is represented. The same XMapTools function used for the Chl grains was applied again. At least three groups of Wm have grown along the S1 foliation, and have textural equilibrium with those detected for Chl (purple box in Fig. 5c indicates the location of the Chl crystal outlined in Fig. 4c). The representative compositions for each group of Wm and the structural formulae calculated on 11 anhydrous oxygens are reported in Table 2. The first group has Si contents with an average of 3.31 ± 0.02 apfu (Fig. 5c,d), and variable AlVI content (average: 1.56 ± 0.1 apfu, Fig. 5f,g). The XMg and K content are more variable (average: 0.46 ± 0.10 and 0.79 ± 0.30 apfu, respectively – Fig. 5d,e). The second group generally has lower Si contents (average: 3.14 ± 0.08 apfu) and higher AlVI contents (1.64 ± 0.09 apfu Fig. 5d–g). XMg is variable with an average of 0.52 ± 0.22, whereas the K contents tend to be less scattered (mean 0.93 ± 0.03 apfu). The third group has the lowest Si content (average: 3.11 ± 0.1 apfu) and the highest K content (average: 0.98 ± 0.01 apfu, Fig. 5d,f). The AlVI content has an average of 1.66 ± 0.07 apfu, while XMg is scattered with an average of 0.48 ± 0.18 (Fig. 5d,g). All Wm groups analysed can be represented by a Ms–Cel solid solution and tend to have Ms-rich compositions, with Cel and Prl contents generally not exceeding 30%, while a few Wm grains have trioctahedral (Tri) and Prl components that never exceed 25% (Table 2). According to Bousquet et al. (Reference Bousquet, Goffé, Vidal, Oberhänsli and Patriat2002) the trioctahedral component could be due to the Fe3+ content of the Wm.
Figure 5. Compositions of Wm in sample ED208a. (a) SiO2 wt.% and (b) Al2O3 wt.% variation in the microdomain of Fig. 3e. Differences exist between detrital (white dotted line) and syn-metamorphic (yellow line) Wm grains. (c) Enlarged area of the white box of (a) and (b) of the syn-metamorphic Wm (light blue arrow) grew along the S1 foliation (yellow line). The Si content variation outlines a complex mineral chemistry of the syn-metamorphic Wm (blue, red and green areas) corresponding to first, second and third groups (see the text). The purple box indicates the Chl the location of chlorite crystal shown in Fig. 4c. (d) XMg/Si, (e) K/Si, and (f) AlVI/Si plots refer to the average composition of the three groups of Wm detected along the S1 foliation and include the related error bars. The orange stars indicate the theoretical compositional of muscovite. (g) AlIV/AlVI plot (according to Bousquet et al., Reference Bousquet, Goffé, Vidal, Oberhänsli and Patriat2002). The grey arrows indicate the solid solution trend of Wm, while the black arrows indicate the main substitution (TK: Tschermak; DT: Di-trioctahedral; PL: Pyrophyllitic).
Table 2. Average compositions related to the three groups of Wm. Standard deviation is indicated
* average and standard deviation were obtained from 50 analysis for group from XMapTools.
n.d. – not detected
Forward thermodynamic modelling
Phase diagrams were computed on a P–T space by setting a pressure and temperature range of 0.2–1.1 GPa and 200–600°C. The KFMASH and KFMASHO systems were used, however Ca, Mn, Ti and Na were removed from the input bulk composition (see the table T4 in the supplementary material S3) because of their minor contents in the observed mineral assemblage found to be in textural equilibrium within the micro-domain of interest (white box of Fig. 3e). The model was computed considering FeOtot = Fe2+ (Fig. 6a), as (Fe-rich) oxides in the investigated domains are scarce or absent. However, a model where FeOtot = Fe2+ + Fe3+ is performed (a 10% of FeOtot as Fe3+ as suggested by Forshaw and Pattison, Reference Forshaw and Pattison2021, see table T4 in S3) to check if the predicted mineral compositions is still comparable with the observed values (Fig. 6b). The lack of carpholite, chloritoid and biotite in the observed mineral assemblage constrains the P/T field (marked in red in Fig. 6a and b) corresponding to a P range of 1.1–0.2 GPa and a T range of 200–550°C (Fig. 6a), and to 1.0–0.2 GPa and 200–375°C in which the predicted mineral assemblage Chl, Wm and Qz matches the observed one. When Fe3+ is considered in the input bulk composition (Fig. 6b), the field where the observed mineral assemblage is predicted to be stable is smaller and hematite (Hem) is modelled (ca. 1%, see table T4 in S3). As the phase proportion of Fe-oxides in the investigated domains is lacking, and additionally, the modal proportion predicted by the model is low (ca. 1%), their occurrence in the P–T fields of interest can be ignored. This test suggests that the observed mineral assemblage, phase proportion and compositions, can be predicted keeping FeOtot = Fe2+ or FeOtot = Fe2+ + Fe3+ (with a 10% of FeOtot as Fe3+). Moreover, the amount of Fe3+ used in this work, given the absence of detailed information regarding mineral oxides, could be considered as the maximum value to be used for the observed mineral assemblage to be predicted (according to the suggestions proposed by Forshaw and Pattison, Reference Forshaw and Pattison2021). Ms-rich and Clc+Daph-rich compositions for Wm and Chl are predicted by both models. As the P–T field for this mineral assemblage is large (Fig. 6a and b) we also computed the isopleths contoured for Si content and XMg content of Wm and Chl, respectively (Fig. 6c,d).
Figure 6. Isochemical phase diagrams related to the micro-domain within the white box of Fig. 3e. (a) Phase diagram with FeOtot=Fe2+. (b) Phase diagram where 10% of FeOtot=Fe3+. Black lines mark the boundary between different stability field, whereas the red ones indicate the field where the observed mineral assemblage is found to be stable. (c, d) Isochemical phase diagrams contoured for Si (Wm) and XMg (Chl) isopleths. The LBC is reported in moles of elements at the top of the diagrams. In the grey boxes, the stable mineral assemblages relate to the different fields are indicated. The yellow area indicates the P/T space of (d) in which the measured value of Si and XMg (for Wm and Chl, respectively) match with that predicted by the model. At the bottom of each diagram, the contour legend is reported and the numbers inside the ellipses represents values of the related observed isopleth. Car: Carpholite; Ctd: Chloritoid; Grt: Garnet; Bt: Biotite; St: Staurolite; Crd: Cordierite; Hem: Hematite; Mt: Magnetite.
All computed isopleths (for Wm and Chl) intersect within the stability field of the mineral assemblage investigated. The steep trend of the XMg contour for Chl allowed us to use it as good T-sensitive parameter. For Wm, the Si isopleth shows a P- and T-dependent trend (Fig. 6b). In the model where no Fe3+ is considered, the predicted Si content and XMg value (for Wm and Chl, respectively) range from 3.26 and 3.36 and 0.28–0.44, respectively. Similarly, when Fe3+ is considered in the input bulk composition, Si content and XMg values (for Wm and Ch, respectively) are 3.26–3.34 and 0.41–0.44. Note that only the observed Si contents and XMg values of Wm (3.31 ± 0.02 apfu) and Chl (0.42 ± 0.08) of the first groups (see the previous section) are coherently predicted by the models. By contrast, the Si and XMg measured for the second and the third groups of Wm and Chl are not predicted by the models. Therefore, taking in mind the observed average values of Si content (for first group Wm) and XMg (for first group Chl) which can be regarded as the maximum values, it is possible to make a comparison with the modelled isopleths. In this approach, we consider that the intersection between the modelled isopleths is reliable for estimates of P and T. When FeOtot = Fe2+ is considered, it is not possible to discriminate an area in the P–T space due to a lack of clear cross-cutting relationships for the modelled isopleths corresponding to the observed average values of XMg and Si (Fig. 6c). By contrast, in the model where Fe3+ is considered, there is a good cross-cutting relationship between the Si and XMg isopleths for Wm and Chl, respectively, which reflects those observed and provides P and T estimates of 275–280°C and 0.85–0.80 GPa (Fig. 6d).
Inverse modelling for P–T estimates
The results obtained with the Chl-Qz-H2O (Vidal et al., Reference Vidal, Parra and Vieillard2005) and Phg-Qz-H2O methods allowed us to estimate the P–T conditions using the local composition of each Chl and Wm group that were found to be in textural equilibrium along the S1 foliation (see the previous section). The Chl-Qz-H2O method that was applied on the three groups of Chl present in the microdomain of Fig. 4c indicates three different ranges for chlorite temperature formation (Fig. 7a). The results are reported in the distribution histograms (Fig. 7a) where along the X-axis the T is reported while the height of each bar (Y-axis) represents the number of chlorites formed in a specific temperature range. For each computation related to each Chl group, at least 10% of Fe3+ was estimated (using the ChlMicaEqui minimization function) at fixed pressure values (0.7, 0.5 and 0.4 GPa for the first, second and third group of Chl, respectively, see the S4 for details). The temperature ranges (Fig. 7a) correspond to the peaks of 170–240°C, 250–285°C and 300–335°C.
Figure 7. P–T estimate results. (a) Histograms of temperature range of Chl formation belonging to the three groups obtained with Chl-Qz-H2O. Each bar represents the counts of each chlorite whose temperature formation falls in a specific range reported along the X-axis. (b) Results of the Phg-Qz-H2O barometer. Within the P/T space the equilibrium lines Wm+Qz+H2O, together with only the hydration state (empty circles along each line) of white mica changes, are reported for each Wm groups. The XH2O values used in the model for each Wm group are also indicated. (c) P/T diagram showing the results of P–T estimates from the Chl-Phg-Qz-H2O method. Each cross indicates the pressure and temperature equilibrium condition for a single Chl-Wm couple belonging to the first, the second and the third groups. The blue, red and green boxes mark the temperature and pressure ranges of Chl and Wm formations obtained with the methods reported in a and b. Along the Y-axis, the histogram shows the results of the Massone and Schreyer (Reference Massone and Schreyer1987) barometer. Each bar represents the counting of pressure values for each pixel inside the micro-domain of Fig. 3e. (e) P/T space showing examples of all the independent (black lines) and dependent (grey lines) equilibria obtained with Chl-Phg-Qz-H2O for each Chl-Wm couple belonging to the first, the second and the third group. (e) Histograms showing the results of the Lanari et al. (Reference Lanari, Wagner and Vidal2014c) geothermometer. Each bar statistically represents temperature range values of chlorite formation. (f) Diagram showing the results of the Bourdelle and Cathelineau (Reference Bourdelle and Cathelineau2015) geothermometer, where the dashed lines indicate different T ranges. Here, the plot of each average values referred to the different groups of chlorite are reported with their error bars.
The P conditions were estimated using the Phg-Qz-H2O method (Dubacq et al., Reference Dubacq, Vidal and De Andrade2010) considering only the Wm analyses contained in the T range previously defined with Chl-Qz-H2O. The P range for the Wm grains grown along the S1 foliation is thus calculated at a fixed T (275, 250 and 320°C, respectively the average T values estimated for different Chl groups) and Fe3+ (10–20% of FeOtot). The results are represented by a line along which the interlayer water content (in the A-site) varies by increasing or decreasing P and T. Three main P ranges can be observed, i.e. 0.38–0.21 GPa, 0.61–0.42 GPa and 0.87–0.62 GPa (Fig. 7b), supporting the proposition that different Wm groups occur along the S1 foliation according with the described mineral compositions.
The P–T estimates obtained with the Chl-Phg-Qz-H2O multi-equilibrium approach (Vidal and Parra, Reference Vidal and Parra2000) were compared with the P and T ranges constrained with the Chl-Qz-H2O and Phg-Qz-H2O methods (Fig. 7c). The comparison sheds light on the good fit between the three different methods that mark the occurrence of three different clusters of Chl-Wm couples along the S1 foliation (Fig. 7c): (1) the first cluster is related to the peak pressure conditions (HP–LT, 0.85–0.68 GPa and 250–285°C – Fig. 7c,d); (2) the second is related to the peak temperature conditions (low pressure–high temperature – LP–HT, 0.6–0.4 GPa and 300–335°C – Fig. 7c,d); whereas (3) the third cluster is related to the Chl-Wm couple, which is stable at the lowest P and T conditions (low pressure–low temperature – LP–LT, 0.35–0.2 GPa and 235–250°C – Fig. 7c,d).
The results obtained were compared with the pressures and temperatures estimated via classic geothermobarometers (Fig. 7c–f). We used the calibrations proposed for low-temperature conditions by Lanari et al. (Reference Lanari, Wagner and Vidal2014c), which can be used even if the amount of Fe3+ in chlorite is unknown, and by Bourdelle and Cathelineau (Reference Bourdelle and Cathelineau2015). The latter calibration was proposed by the authors as an efficient tool to estimate the temperature of chlorite formation (although the non-ideal contribution of site mixing is not considered) by means a graphical representation in which Fe2+ + Mg2+ (M-site) occupancy vs Si content is reported. The first geothermometer (Fig. 7e) for the different groups of chlorite grains indicates the following T ranges: 250–270°C for the first group (HP–LT chlorites); 310–320°C for the second group (LP–HT chlorites); and 180–240°C for the third group (LP–LT chlorites). The Bourdelle and Cathelineau (Reference Bourdelle and Cathelineau2015) geothermometer (Fig. 7f) yields three different T ranges: the first range is 225–275°C; the second is 275–325/ >325°C; and the third again 225–275°C. Altogether, these comparisons show a good fit with the three temperature ranges of chlorite formations modelled with multi-equilibrium thermobarometry. The geobarometer of Massonne and Schreyer (Reference Massone and Schreyer1987) (Fig. 7c) was used to test our P estimates. This geobarometer is based on a calibration where the (Fe, Mg) celadonite component of Wm, which is in equilibrium with K-Fsp + Qz + Phlogopite (Phl), is a function of pressure. The pressure range calculated on the Wm grains (values >200 pixel/map) grown along the S1 foliation (histogram reported along the Y-axis of Fig. 7c) is 1.2–0.2 GPa, with a peak at 0.5–0.2 GPa.
WmIC and b0
The WmIC values with the relative standard deviation, measured along the investigated transect, are shown in the histogram of Fig. 8a. The red dashed lines indicate the lower and upper limits of very low-grade metamorphic conditions proposed by Warr and Ferreiro Mählmann (Reference Warr and R.F2015). Overall, specimens have fewer variable WmIC distributions, ranging from 0.13 to 0.22 Δ°2θ (Fig. 8a and table T5 in the supplementary material S5). Mixed layers are not identified (see fig. F2 in supplementary material S5). These data strongly indicate that the Cabanaira Unit reached typical epizone T conditions (more than 300°C) during its deformation history. Overall, the distribution of the WmIC values into the Cabanaira Unit is in good agreement with the microstructural observations and the observed paragenesis.
Figure 8. Crystallochemical indices distribution. (a) WmIC distribution. (b) b0 cell parameters distribution from XRD patterns. (c) b0 values calculated using the Verdecchia et al. (Reference Verdecchia, Collo, Zandomeni, Wunderlin and Fehrmann2019) spreadsheet. The grey rectangular boxes represent the three T ranges independently obtained with thermobaric methods. Along Y- and X-axes the crystallochemical index distribution (coloured rectangular bars) obtained using the XRD pattern analysis is reported. The T ranges calculated by Piana et al. (Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014) are reported (yellow rectangular bar along the X-axis). The shaded effect indicates that the upper limit of the range is poorly constrained with this method. The triangles, circles and squares contoured in black are referred to the calculations in which 10% of ferric iron is considered.
The b0 value distribution is represented in the histograms of Fig. 8b, with the barometric limit values (blue dashed lines) suggested by Franceschelli et al. (Reference Guidotti, Sassi and Blencoe1989). Samples again show less variability in the distribution values of the b0 cell parameter, with values ranging between 9.006 and 9.018 Å and a peak at 9.011–9.013 Å (Fig. 8b). The b0 values measured on the Cabanaira Unit are within the field related to high-pressure metamorphic field (0.5–0.2 GPa based on Franceschelli et al., Reference Guidotti, Sassi and Blencoe1989). The calculation of the b0 parameter obtained by taking into account the average compositions of white mica groups (according to Verdecchia et al., Reference Verdecchia, Collo, Zandomeni, Wunderlin and Fehrmann2019) found to be in equilibrium with chlorite grains along the S1 foliation was performed (Fig. 8c, see table T6 in supplementary materials S5). In the P/T space of Fig. 8c, the b0 isopleths (after Guidotti and Sassi, Reference Guidotti and Sassi1986) are indicated. Along the X-axes, the equivalent T range related to the measured WmIC values of powders < 2 µm, is reported. Each point in the P/T space represents the b0 calculated for Wm average compositions related to each group at a T corresponding to that estimated for chlorite formation with the Chl-Qz-H2O thermometer of Vidal et al. (Reference Vidal, Parra and Vieillard2005). Overall, the P ranges obtained from the calculated b0 values fall within the range estimated with that extrapolated from powder X-ray diffraction. However, these P ranges associated with the b0 values tend to be related to lower pressure values rather than those calculated by the barometer of Dubacq et al. (Reference Dubacq, Vidal and De Andrade2010) with the HP–LT Wm group showing the highest discrepancy.
Discussion
P–T path of the Cabanaira Unit reconstructed by an integrated approach
The integrated approach applied in this work allows us to reconstruct, for the first time, the P–T path for the low-grade Cabanaira Unit (Fig. 9) which is exposed along the southwestern sector of the Marguareis Massif. Microstructural analyses clearly indicate that mineral the assemblage Chl-Wm-Qz are in textural equilibrium, along the S1 foliation micro-site of interest, with no evidence of minerals like carpholite or chloritoid. The thermobarometric estimates show that the prograde path is not recorded by the unit and that the described mineral assemblages are related to the retrograde path only (Fig. 9). In the following the different approaches will be discussed taking into account their ability to capture the metamorphic history of the low-grade rocks of interest.
Figure 9. Diagram showing the P–T path of the Cabanaira Unit constrained in this work. The paths estimated for the Marguareis and Moglio-Testico Unit are also represented (data from Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b; Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022c). The geothermal gradients typical of subduction and collision setting are also shown. Along the X-axis the T ranges extrapolated by means Illite Crystallinity Index for the Marguareis (red rectangular bar data from Piana et al., Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014), Cabanaira (blue rectangular bar from Piana et al., Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014; orange rectangular bar from this work) and the Moglio-Testico (light blue rectangular bar from Bonazzi et al., Reference Bonazzi, Cobianchi and Galbiati1987) Units are reported. The shaded effect indicates that the upper limit of the range is poorly constrained with this method. Along the Y-axis the P range yielded from b0 measurement (this work) are reported. The length of rectangular boxes represents the range of temperature and pressure estimates.
Forward thermodynamic modelling suggests that the observed mineral assemblage is predicted to be stable under a wide range of P–T conditions. However, using the isopleths approach and considering the effect of the ferric iron, the P–T estimates for the first group of Wm and Chl was constrained (275–280°C and 0.85–0.80 GPa). The occurrence of Wm-Chl groups showing slight differences in their compositions and the limited ability to physically select them in order to obtain at least three local bulk compositions make the usage of forward thermodynamic modelling restricted. Notwithstanding, the database used to perform forward thermodynamic modelling yielded realistic results. In fact, the adopted solution models for Wm include pyrophyllite and paragonite as end-members of white-mica solid solution. The stability of paragonite is not expected due to the very low Na-content of Wm, even though it cannot be correctly evaluated, because Na is not included in the input bulk composition. The possible formation of pyrophyllite predicted by modelling results from a miscibility gap with dioctahedral mica (see Parra et al., Reference Parra, Vidal and Agard2002), which is not observed in the simulation performed in this work. For Chl, among the available solid-solution models, those adopted in this work (Vidal et al., Reference Vidal, Parra and Vieillard2005 with the revised Margules parameters from Vidal et al., Reference Vidal, de Andrade, Lewin, Munoz, Parra and Pascarelli2006) have demonstrated more realistic pressure conditions for carpholite-producing reactions (see solution models of Hunziker et al., Reference Hunziker2003 where carpholite is predicted to be stable at low P conditions). It must be underlined that the chlorite model of Vidal et al. (Reference Vidal, Parra and Vieillard2005), where Si>3 apfu is not allowed and a Sud end-member is included, is not fully compatible with the Theriak-Domino algorithm in which Mg-Fe partitioning cannot be forced. However, the usage of the database of Pourteau et al. (Reference Pourteau, Bousquet, Vidal, Plunder, Duesterhoeft, Candan and Oberhänsli2014) with the solution models chosen in this work (S3 supplementary material) during forward modelling allowed us to compute realistic phase equilibria diagrams even if the observed mineral assemblage is predicted to be stable into a field characterised by a substantial range of P–T conditions.
The reconstructed P–T path of the Cabanaira Unit reported in Fig. 9 was mostly constrained with the use of the multi-equilibrium thermobarometry. Inverse thermodynamic modelling applied to the three groups of Wm-Chl couples indicates that different peak P and T conditions were achieved along the tectono-thermal evolution of the Cabanaira Unit (Fig. 9). The peak P at 0.85–0.68 GPa and 250–285°C (which is coherent with the P–T conditions predicted by forward modelling) were followed by decompressional warming (peak T) constrained by the Chl-Wm couples in chemical equilibrium at lower P (0.6–0.4 GPa) and higher T (300–335°C) conditions. Note that decompressional warming was also documented in other sectors of the Ligurian Alps (see Maino et al., Reference Maino, Dallagiovanna, Dobson, Gaggero, Persano, Seno and Stuart2012). The last part of the P–T path is constrained by the Chl-Wm couples in equilibrium at the lowest P and T conditions (0.35–0.2 GPa and 235–250°C). The P and T estimates for the different groups of Chl-Wm couples show a good fit with the classic geothermobarometers that are usually used for low-grade metamorphism, with the exception of the Massonne and Schreyer (Reference Massone and Schreyer1987) barometer. The latter, in fact, gives pressure values (0.5–0.6 GPa) lower than the peak P estimated with multi-equilibrium thermobarometry approach. According to Guidotti and Sassi (Reference Guidotti and Sassi1998), this is due to the effect of the deviation of the investigated mineral assemblage from the limiting mineral assemblage (i.e. Ms + K-Fsp + Qz + Phl) calibrated by Massonne and Schreyer (Reference Massone and Schreyer1987). In fact, these authors underline that any deviation from the assemblage used to perform their calibration can, at best, give a minimum pressure value.
The WmIC index and the b0 cell parameter calculated and measured in this work are roughly related to P and T range which partially fit with the thermobarometric estimates. The measured WmIC distribution roughly indicates temperatures 300°C < T < 350°C, according to Kisch (Reference Kisch and Frey1987), Niedermayr et al. (Reference Niedermayr, Mullis, Niedermayr and Schramm1984) and Weaver and Boekstra (Reference Weaver and Boekstra1984), which are coherent with the maximum T values (peak T) estimated with thermobarometry. The measured b0 value distribution indicates a P range related to 0.4–0.6 GPa for the Wm grains (Guidotti and Sassi, Reference Guidotti, Sassi and Blencoe1989; Guidotti et al., Reference Guidotti, Sassi and Blencoe1989; Franceschelli et al., Reference Guidotti, Sassi and Blencoe1989). The calculated b0 parameters for the Wm crystals used for thermobarometric calculations further provide constraints. The most coherent data are those related to the LP–HT Wm, indicating a range of 0.48–0.2 GPa, showing a good fit with the measured b0 values. Note that the P range estimated with the b0 value tends to be slightly lower than that obtained with the Phg-Qz-H2O barometer (Dubacq et al., Reference Dubacq, Vidal and De Andrade2010), even if they are within the uncertainty associated with the thermobarometric methods. The b0 values calculated for the LP–LT Wm crystals also indicate P conditions comparable to those estimated with thermobarometry (range of 0.32–0.18 GPa, cf. Fig. 7c and Fig. 8c). The P range, extrapolated through the b0 parameter, calculated for the HP–LT Wm chemical analysis shows significant differences compared with those estimated with thermobaric calculation (cf. Figs 7c and 8c). This can be explained considering the remarks underlined by Guidotti and Sassi (Reference Guidotti and Sassi1976) and Sassi and Scolari (Reference Sassi and Scolari1974), which suggest that the b0 value measurement (or calculation) shows only a qualitative relation with P and it is tested only for a range never exceeding 0.6 GPa. The results beyond this value are therefore not reliable. However, crystallochemical indices measurement (according to Guidotti and Sassi, Reference Guidotti and Sassi1998; Kubler and Goy-Eggenberger, Reference Kubler and Goy-Eggenberger2001; Abad and Nieto, Reference Abad and Nieto2007; Warr and Ferreiro Mahalman, Reference Warr and R.F2015) have two limits: (1) they do not provide a quantitative estimate of P and T conditions; and (2) they were typically measured on rock powders where the microtexture is clearly not preserved. These two factors clearly represent the weakness of this method, and mistakes can be made in constraining the metamorphic history of low-grade rocks if the results obtained with these approaches are not verified and/or supported by other methods.
The ranges of T estimates obtained using different approaches, overall indicating no more than 350°C, are coherent with the observed microstructures resulting from dissolution and recrystallization processes, which were the main deformation mechanisms during the metamorphic history. Similar T estimations for the deformed phyllosilicates-rich rocks have been obtained by other authors (Buatier et al., Reference Buatier, Lacroix, Labaume, Moutarlier, Charpentier, Sizun and Travé2012; Lacroix et al., Reference Lacroix, Buatier, Labaume, Trave, Dubois, Charpentier, Ventalon and Convert-Gaubier2011; Lacroix and Venneman, Reference Lacroix and Vennemann2015; Elmola et al., Reference Elmola, Charpentier, Buatier, Lanari and Monie2017).
Assessment of the P–T estimates within the tectonic framework of the Cabanaira Unit
Our data indicate that the Cabanaira Unit recorded a pressure‒temperature evolution where peak P conditions are followed by decompressional warming (peak T). The consistency between this estimate and the tectonic framework proposed for the Cabanaira Unit can be assessed. The first contribution was made by Brizio et al. (Reference Brizio, Deregibus, Eusebio, Gallo, Gosso and Rattalino1983), in which the authors proposed that the thrusting of the San Remo-Monte Saccarello Unit (cf. Helminthoid Flysch Unit of this work) onto the Briançonnais Units (here represented by the Marguareis Unit, including the Cima del Becco Slices and the Cabanaira Unit) has been responsible for their deformation and the very low-grade metamorphic imprint. In this framework, the peak P conditions estimated in this work for the Cabanaira Unit should be due to the effect of the lithostatic loading produced by the thrusting of the San Remo-Monte Saccarello Unit (cf. Helminthoid Flysch Unit, this work). However, the estimated thickness of the Helminthoid Flysch Unit succession exposed in the investigated area is not sufficient to produce lithostatic loading to explain the pressure values estimated in this work for the Cabanaira Unit.
Alternatively, Piana et al. (Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014) and d’Atri et al. (Reference d’Atri, Piana, Barale, Bertok and Martire2016) proposed that this unit has an Anchizone metamorphic imprint, with temperatures never exceeding 240°C during its tectonic evolution, and that it developed at very shallow structural levels. This tectonic scenario must now be reconsidered with regard to the P and T conditions constrained in this work, where temperatures greater than 300°C and pressures up to 0.8 GPa are well established.
Recently, Sanità et al. (Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a) and Sanità (Reference Sanità2023) proposed a tectonic evolution in which the thrusting of the Helminthoid Flysch Unit (cf. San Remo-Monte Saccarello Unit, see syn-stacking tectonics of Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a) developed immediately after the D1 phase recorded by the Cabanaira Unit that is thought to have developed during its involvement into the Alpine wedge (see pre-stacking structures of Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a). The reconstructed P–T pathway (Fig. 9) seems to be coherent with the tectonic setting of Sanità et al. (Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a, Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021b), where the tectonic evolution of the Cabanaira Unit is thought to have developed during its underthrusting, accretion and subsequent exhumation into the Alpine wedge during the middle to late Eocene. The P–T paths of Fig. 9 are related to each tectonic unit exposed at the boundary between the Maritime and Ligurian Alps. In this P/T space, the trajectories followed by each tectonic unit have been reconstructed using multi-equilibrium thermobarometry (see Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b and Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022c for more details) making the comparison with the estimates performed in this paper justified and reliable. Note that the metamorphic peak P conditions of all the units lie along a geothermal gradient coherent with that estimated by different authors for the Western Alps for oceanic (i.e. the Moglio-Testico Unit) and continental (i.e. Margaureis Unit and Cabanaira Unit) subduction settings (Agard et al., Reference Agard, Vidal and Goffé2001; Bousquet et al., Reference Bousquet, Oberhänsli, Goffé, Wiederkehr, Koller, Schmid and Martinotti2008; Sterzynsky et al., Reference Strzerzynski, Guillot, Leloup, Arnaud, Vidal, Ledru, Corrioux and Darmendrail2011; Agard et al., Reference Agard, Plunder, Angiboust, Bonnet and Ruh2018; Herviou et al., Reference Herviou, Agard, Plunder, Mendes, Verlaguet, Deldicque and Cubas2022). Interestingly, the peak P conditions recorded by the Cabanaira Unit overlap the P–T conditions estimated for the exhumation of the Marguareis Unit (see D2 phase of Sanità et al., Reference Sanità, Di Rosa, Lardeaux, Marroni and Pandolfi2022b). However, only the Cabanaira Unit shows warming during decompression with the associated peak T conditions close to the classic geothermal gradient typical of the continental collision (30°C/km) estimated by previous authors for the Alpine belt.
Factors triggering warming during the exhumation of the Cabanaira Unit
The warming event is well constrained both by multi-equilibrium thermobarometry and crystallochemical indexes measurement and it needed to be clarified. There are more factors that could trigger warming, and, in this section, they are taken into account. Below we explore all the possible processes in which warming can be developed.
(1) Thrusting-related warming. The peak T conditions estimated in this work with different methods are evidently in contrast with those calculated by Piana et al. (Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014) (Fig. 9). These authors, through the WmIC measurement performed on the Roja Unit (cf. Cabanaira Unit of this work) estimated Anchizone field-related temperature conditions never exceeding 200–240°C by means of the WmIC parameter. Considering that the same methodology and procedure for the WmIC measurement (see Piana et al., Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014 for more details) have been used, the differences between their data and the results obtained in this work could be due to the location of the sampling. In this work, sampling was performed along a transect on which structural features, such as unit-bounding thrust systems (Fig. 2a) occur. The lower crystallinity values (= higher metamorphic temperature conditions, this work), which are supported by the thermobarometric estimates, could be related to the thrust activity generally described for the whole investigated area (Mueller et al., Reference Mueller, Maino and Seno2020; Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a; Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021b). In contrast, the higher WmIC values (= lower metamorphic temperature conditions) estimated by Piana et al. (Reference Piana, Battaglia, Bertok, d’Atri, Ellero, Leoni and Perotti2014) were obtained on samples collected in areas where tectonic structures, such as thrusts or fault systems, are lacking. Therefore, having witnessed the impact of this method, the metamorphic temperature conditions calculated in this work could be due to the effect of the thrust activity that has been documented by several authors in different tectonic settings (Fernandez-Caliani and Galan, Reference Fernández-Caliani and Galán1992; Ducci et al., Reference Ducci, Leoni, Marroni and Tamponi1995; Giorgetti et al., Reference Giorgetti, Memmi and Peacor2000; Brogi, Reference Brogi2006; Elmola et al., Reference Elmola, Charpentier, Buatier, Lanari and Monie2017). In this framework, the peak T calculated in this work in the investigated transect could thus be the witness of the syn-coupling tectonic events (cf. syn-stacking events of Sanità et al., Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a) documented in the mapped area. In this context, the warming could thus be regarded as related to the thrusting of the Helminthoid Flysch Unit onto the already exhumed Briançonnais Units (i.e. the Cabanaira Unit). Maino et al. (Reference Maino, Dallagiovanna, Dobson, Gaggero, Persano, Seno and Stuart2012), (Reference Maino, Casini, Ceriani, Decarlis, Di Giulio, Seno and Stuart2015) and (Reference Maino, Casini, Boschi, Di Giulio, Setti and Seno2020) performed an investigation about the thrust activity-related thermal budget on the units exposed above and below the Penninic Front in the Monte Fronté zone. They conclude that the HT event is recorded by the rock volumes because of frictional heating due to thrust activity during nappe emplacement processes. Notwithstanding, the temperature proposed by Maino et al. (Reference Maino, Casini, Boschi, Di Giulio, Setti and Seno2020) are not coherent with those estimated in this study being ∼80–90°C lower than our T-peak estimates (Tmax = 245°C vs. 300–335°C). However, comparable T estimations (>300°C) were documented by Balansa et al. (Reference Balansa, Lahfid, Espurt, Hippolyte, Henry, Caritg and Fasentieux2023) for the Briançonnais Units exposed in the Embrunais-Ubaye sector (Western Alps, Fig. 1) and tectonically located under the Parpaillon Unit (cf. Helminthoid Flysch Unit, this work). The authors proposed that the estimated T conditions are due to the thrusting of the Parpaillon Unit onto the underlying Briançonnais Units. According to the authors, this hypothesis implies a volume for the Helminthoid Flysch Unit higher than the one shown currently. In fact, in their palinspastic reconstruction, Balansa et al. (Reference Balansa, Lahfid, Espurt, Hippolyte, Henry, Caritg and Fasentieux2023) proposed that the leading edge of the Parpaillon Unit was located to the west of its current position. Therefore, in the investigated area, the estimated volume of the Helminthoid Flysch Unit probably does not correspond to that during the thrusting event. Therefore, a similar tectonic framework can be taken into account to explain the warming recorded by the Cabanaira Unit in the investigated area. However, this scenario is in contrast with the field constraints available for the tectonic evolution of the area of interest. In fact, according to Sanità et al. (Reference Sanità, Lardeaux, Marroni, Gosso and Pandolfi2021a), the thrusting of the Helminthoid Flysch Unit postdates the folding system affecting the Cabanaira Unit (D1 phase), during which warming has been recorded.
(2) Fluid‒rock interactions. Fluid‒rock interaction cannot a priori be excluded. Hydrothermal circulation was documented in analogue units exposed along the outer sectors of the southwestern Alps by previous authors (cf. Entracque Unit of Barale et al., Reference Barale, Bertok, Salih Talabani, d’Atri, Martire, Piana and Préat2016; Bertok et al., Reference Bertok, Musso, d’Atri, Martire and Piana2018). The authors invoke the occurrence of fault systems that were used as preferred channels for hot fluid circulation. These are interpreted as responsible for the chemical transformation of the host rock and for the variation in the mineral chemistry before the deformation starts to be recorded by the unit. Moreover, fluid circulation into the subduction zone at shallower structural levels cannot be excluded. In fact, dehydration pulses within the subduction zone were outlined as key features contributing to rapid warming during the onset of exhumation (Camacho et al., Reference Camacho, Lee, Hensen and Braun2005; John et al., Reference John, Gussone, Podladchikov, Bebout, Dohmen, Halama and Seitz2012; Dragovic et al., Reference Dragovic, Baxter and Caddick2015), even if their role is still far from being understood. Notwithstanding, these hypotheses are in contrast with the dataset collection presented in this work because: (i) the T estimated (at least 350°C) for the hot fluids (Barale et al., Reference Barale, Bertok, Salih Talabani, d’Atri, Martire, Piana and Préat2016; Bertok et al., Reference Bertok, Musso, d’Atri, Martire and Piana2018) is higher than our temperature estimates (peak T: 300–335°C); and (ii) in the study area, field evidence of an extended early faulting event was not documented, and the peak T is constrained to the deformation of the Cabanaira Unit (i.e. during its exhumation towards the surface, D1 phase). In addition, the role of fluids to the thermal budget during faulting was investigated by Maino et al. (Reference Maino, Casini, Boschi, Di Giulio, Setti and Seno2020), however, the authors concluded that fluid-rock interaction effects on the thermal budget is neglectable.
(3) Perturbation of the geothermal gradient. The warming could be due to a progressive variation in the Alpine wedge thermal state as a result of the beginning of the continental collision, which was documented in other sectors of the belt where its effects are more pronounced (Goffé et al., Reference Goffé, Schwartz, Lardeaux, Bousquet and Oberhansli2004; Bousquet et al., Reference Bousquet, Oberhänsli, Goffé, Wiederkehr, Koller, Schmid and Martinotti2008; Handy et al., Reference Handy, Schmid, Bousquet, Kissling and Bernoulli2010) and in Alpine Corsica (Di Rosa et al., Reference Di Rosa, Frassi, Meneghini, Marroni, Pandolfi and De Giorgi2019; Di Rosa et al., Reference Di Rosa, Frassi, Malasoma, Marroni, Meneghini and Pandolfi2020 and quoted references). Many contributions have stressed the factors influencing the thermal structure of orogenic belts, such as shear and/or radioactive heating, as well as the thermal conditions of subducting and overriding plates (Gerya et al., Reference Gerya, Perchuk and Burg2008; Syracuse et al., Reference Syracuse, van Keken and Abers2010; Faccenna et al., Reference Faccenna, Becker, Auer, Billi, Boschi, Brun and Serpelloni2014; Zheng et al., Reference Zheng and Chen2016; Regorda et al., Reference Regorda, Spalla, Roda, Lardeaux and Marotta2021). Anomalously high-temperature events are thought to have developed during incipient continental collision (Gerya et al., Reference Gerya, Perchuk and Burg2008), even if associated with partial melting (Burg and Gerya, Reference Burg and Gerya2005). Amadori et al. (Reference Amadori, Maino, Marini, Casini, Carrapa, Jepson and Di Giulio2023) proposed a geological scenario in which the mantle upwelling is invoked to account for the hot and shallow thermal signal recorded by the sediments in the late Eocene-late Miocene epi-sutural basin (i.e. the Tertiary Piedmont Basin).
Unfortunately, the current knowledge based on data presented in this work or available from previously published work is not able to provide a clear snapshot, and therefore, the hypotheses discussed in this paper need further investigation to be supported.
Conclusions
This work outlines the paramount importance of a multidisciplinary approach to perform geological surveys aimed at investigating orogenic mechanisms recorded by very low- to low-grade units exposed in the outer sectors of collisional belts. Quantitative and qualitative methods were tested together for reconstruction of the P–T path of low-grade meta-pelites, capturing as many segments of it as possible. Each method allowed us to constrain a different part of the metamorphic history of the Cabanaira Unit. The good agreement between estimates from inverse (multi-equilibrium thermobarometry) and forward modelling (pseudosection) and crystallochemical indexes measurement, taking into consideration the strengths and weaknesses of each method, allows us to regard them as suitable to reconstruct the P–T path of low-grade rocks. Multi-equilibrium thermobarometry together with high-resolution quantitative compositional mapping remain the best method to unveil the metamorphic history of low-grade rocks in details.
The P–T estimates for the Cabanaira Unit indicate peak P conditions (0.85–0.68 GPa and 250–285°C) that are coherent with the geological models that imply its involvement in the subduction zone during its underthrusting and accretion into the Alpine wedge. In addition, the peak T (0.4–0.6 GPa and 300–335°C) conditions (decompressional warming) were achieved during exhumation after the peak P. The causes triggering the warming recorded by the Cabanaira Unit are far from understood, and its tectonic meaning, based on current knowledge, can only be postulated.
Supplementary material
Data are available from the authors on request. The supplementary material for this article can be found at https://doi.org/10.1180/mgm.2024.80.
Acknowledgements
The authors thank Tim Johnson and Arne Willner for their very useful and stimulating comments and suggestions, which improved the final version of the manuscript. We also thank the editor for managing the paper. Andrea Risplendente is thanked for his assistance with the EPMA data collection.
Author contributions
Conceptualization: ES; Methodology: ES, MDR; Software: ES, MDR; Validation: ES, MDR, JML, ML, MT, MM, LP; Formal analysis: ES; Investigation: ES; Resources: ES; Data Curation: ES. MDR; Writing original draft: ES; Writing - Review & Editing: ES, MDR, JML, ML, MT, MM, LP; Visualization: ES; Supervision: JML, MM, LP; Project administration: ES, LP.
Funding statement
This study was funded by PRIN project 2020 (PI Michele Marroni).
Competing interests
The authors declare none.
Open access
